Professional Documents
Culture Documents
THOMPSON BOWCHER FONTAINE SCHONTHAL 2019 The Cambridge Handbook of Systemic Functional Linguistics PDF
THOMPSON BOWCHER FONTAINE SCHONTHAL 2019 The Cambridge Handbook of Systemic Functional Linguistics PDF
Introduction 1
Wendy L. Bowcher, Lise Fontaine and David Schoenthal
Index 793
Preface
This volume has been several years in the making. It was first conceived of
in 2013, when Cambridge University Press approached Lise Fontaine with
the possibility of including a Handbook on Systemic Functional Linguistics as
part of its series of Handbooks on Language and Linguistics. Recognizing
this as a wonderful opportunity, Lise, Geoff Thompson, and Wendy
Bowcher discussed the possibility of co-editing the volume. It was decided
that Geoff would take the lead, and in consultation with various scholars,
including Michael Halliday, he developed the conceptual framework for the
book – a volume with a comprehensive, somewhat historical but also
forward-looking overview of Systemic Functional Linguistics. Later, after
Geoff’s untimely death, David Schönthal was invited to join the editorial
team. As editors, we encouraged contributors to include both theoretical
and practical details where possible – the latter noted by Halliday in
personal correspondence as being an important part of the character of a
‘handbook’. Finding contributors to this volume was difficult, but in a
positive way, as there are so many scholars around the world with expertise
in the various areas covered who could have been approached. The final
line-up, we feel, offers a wide scope of perspectives from a range of estab-
lished and emerging scholars, some expert in more than the field of
research which they have written about in this volume. We would like to
take a moment here to thank all our contributors for the effort and
expertise they have brought to this collection. Readers will notice that at
the beginning of some of the chapters there is a note of tribute to several
scholars who have passed away since the volume’s inception: Chapter 4
pays tribute to Geoff Thompson (see also the tribute to Geoff at the
beginning of this volume), Chapter 7 to Bill Greaves, Chapter 23 to Johna-
than Fine, and Chapter 26 to Ruqaiya Hasan. We felt it was important to
include these tributes – to Geoff himself as the person who really got this
project off the ground, and to Geoff and all the other scholars who have
been such an important influence not only in the development of the
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:30:31, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.001
xxx PREFACE
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:30:31, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.001
Introduction
Wendy L. Bowcher, Lise Fontaine, and
David Schönthal
tackles problems and tries to answer questions – but questions that are
asked, and problems that are raised, not by professional linguists so much
as by other people who are in some way concerned with language, whether
professionally or otherwise. There are large numbers of such people: educa-
tors, translators, legal and medical specialists, computer scientists, students
of literature and drama, . . .; and it is their ‘take’ on language that is being
addressed, at least to the point of clarifying what sorts of questions can
usefully expect to be asked, and whether or not there is any hope of coming
up with an answer.
(Halliday 2013:128)
Unlike theories of language that separate ‘langue’ from ‘parole’ and which
consider parole as a somehow ‘flawed’ version of language, SFL recognizes
the ‘symbiotic relation’ between langue and parole:
Downloaded from https://www.cambridge.org/core. Stockholm University Library, on 30 Oct 2019 at 11:18:10, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.002
2 WENDY L. BOWCHER, LISE FONTAINE, AND DAVID SCHÖNTHAL
This symbiotic relation is also seen in the relation between language and
society, and is explained through the social semiotic perspective on lan-
guage that SFL holds:
The universal characteristics of parole – its orderly variation, its flexible regu-
larities – are functional (Halliday 1970): they have their origin in the relations of
parole to the community’s living of life, while at the same time, the various
dimensions of a community’s social contexts of living depend on parole for
their creation, maintenance, and evolution. Language as a social semiotic is
predicated on this mutual relation between parole and social contexts.
(Hasan 2009:309–10)
Downloaded from https://www.cambridge.org/core. Stockholm University Library, on 30 Oct 2019 at 11:18:10, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.002
Introduction 3
pursue some of the challenges and issues raised within the volume – be
they theoretical or practical in nature.
The volume is divided into three parts: Part I, ‘SFL: The Model’; Part II,
‘Discourse Analysis within SFL’; and Part III, ‘SFL in Application’. The
following paragraphs summarize the contents of the chapters in each of
these sections.
As the introductory part of the volume, Part I, ‘SFL: The Model’, has been
designed in such a way as to cover the core features and terminology of the
SFL framework. The organization of this part is designed in part to reflect
the key perspectives on the theory. Opening the part, in Chapter 1, David
G. Butt lays out in considerable detail the origins and history of how the SFL
approach evolved. Halliday’s interest in a language-based linguistics is
shown to derive directly from J. R. Firth, and, by better understanding
Firth’s concerns, the reader gains valuable insight into Halliday’s develop-
ment of SFL. A description of key terms in the SFL model is then given in
Chapter 2, by Jonathan J. Webster. These two chapters form a necessary
background for the more specific chapters that follow.
In Chapter 3, Miriam Taverniers takes up the central concept of seman-
tics and explores how it is conceptualized and modelled in SFL theory. In
particular, she teases out the different conceptions of semantics within SFL.
Importantly, she relates the key concepts of abstraction, patterning, and
actualization to stratification and metaredundancy. Chapters 4 through 7,
then, combine to provide a detailed discussion of four key approaches or
perspectives on language in the SFL framework: the clause, units of the
clause, context, and sound patterns. The multifunctional view of the clause
is detailed by Margaret Berry in Chapter 4 on the lexicogrammar. Berry
presents a concise analytical overview of the clause from the experiential,
interpersonal, and textual metafunction. In Chapter 5, Lise Fontaine and
David Schönthal present a critically engaged description of the units of
‘group’ and ‘phrase’. After reviewing the different units below the clause,
they go on to challenge the distinction between the units of ‘group’ and
‘clause’. Context and its relation to text type is examined by Wendy
L. Bowcher in Chapter 6, as she details the concepts of context and register
within the model. Specifically, Bowcher discusses the history of these
two concepts and their relation between one another, and reviews
seminal SFL research on context and register. Chapter 7, by Wendy
L. Bowcher and Meena Debashish, details how intonation and English tone
groups are situated within the SFL framework. Not only do the authors
offer a usable description of English intonation, they also raise important
issues related to topics currently under debate and possible areas of future
research.
Downloaded from https://www.cambridge.org/core. Stockholm University Library, on 30 Oct 2019 at 11:18:10, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.002
4 WENDY L. BOWCHER, LISE FONTAINE, AND DAVID SCHÖNTHAL
The final three chapters of Part I each step outside the standard SFL
model in different ways. In Chapter 8, Mick O’Donnell explores aspects of
the model which are under debate within the field. He outlines the main
points of interest, showing where there is scope for significant contribu-
tions from researchers. Chapter 9, written by Anke Schulz and Lise Fon-
taine, presents the model of functional syntax as developed within the
Cardiff Grammar. As a model with its roots in SFL theory, the Cardiff
Grammar shares many of the same principles as outlined in the other
chapters in this part. However, there are some important differences illus-
trated in this chapter. In the final chapter of Part I, Chapter 10, Christopher
S. Butler situates SFL in its theoretical context in relation to other func-
tional approaches, or what he refers to as ‘functional-cognitive space’. This
chapter, based on a detailed comparison of sixteen different models, shows
that while some differences are highlighted, there are also some interesting
points of shared concerns.
The second part of the volume contains eight chapters which present
various discourse analytical tools developed within the framework of SFL
theory. This part begins with Chapter 11, by Tom Bartlett, who first
describes SFL from a discourse analytical perspective and then discusses
some of the main approaches to discourse analysis within SFL. This chapter
includes critical comments on some of the analytical approaches and effect-
ively sets the background for the chapters which follow.
Chapter 12, by Maite Taboada, focuses on cohesion and conjunction. This
chapter describes Halliday and Hasan’s (1976) early work on cohesion, and
includes a brief description of cohesive harmony (Hasan 1985; Khoo 2016), as
well as work on rhetorical structure theory (RST) (Mann et al. 1992). In her
discussion of conjunction, Taboada briefly points out some of the differences
between Martin’s (1992) and Hasan’s (1985) descriptions. The chapter also
includes a brief description of some areas in which Halliday and Hasan’s
model of cohesion and coherence has been applied, such as in computational
studies and foreign language teaching and learning.
Chapter 13, by Andy Fung and Francis Robert Low, focuses on semantic
networks as developed by Hasan (1996), but the chapter first situates
this discourse analytic framework in relation to other semantic-level
approaches within SFL. The chapter describes the basic unit of analysis,
the ‘message’, and demonstrates the utility of this framework through an
analysis of a mother–child interaction. The latter part of the chapter illus-
trates how semantic networks have been applied, with a focus on pedagogic
and journalistic discourse.
Chapters 14 (J. R. Martin) and 15 (Susan Hood) describe two related analyt-
ical frameworks: discourse semantics and the system of appraisal. Martin
Downloaded from https://www.cambridge.org/core. Stockholm University Library, on 30 Oct 2019 at 11:18:10, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.002
Introduction 5
briefly explains how his ‘discourse semantics’ differs from Halliday and
Hasan’s (1976) concept of coherence and cohesion before elaborating on
various systems within his framework. The second part of the chapter
presents a text analysis demonstrating the application of the different dis-
course semantic systems and the value of this analytical approach in high-
lighting threads of ideational, interpersonal, and textual meanings
throughout a text, and how these threads relate to each other. Towards the
end of the chapter, Martin discusses the interpersonal system of appraisal,
which provides a natural segue to Hood’s chapter on Appraisal. Hood first
situates the appraisal system within SFL theory and then introduces the
three main appraisal sub-systems: attitude, graduation, and engage-
ment. The latter part of her chapter discusses some of the ways that the
system of appraisal has been applied, ending with a discussion of the
current trends in Appraisal research, including multimodal studies,
research into legal language, and studies of identity and affiliation.
Chapter 16, Diachronic Studies by David Banks, while not technically
about a discourse analytic approach, is about a discourse analytic perspec-
tive – diachrony – and Banks illustrates some of the discoursal features that
are focused on by SFL researchers whose data and analytical perspectives
would fall within the domain of diachronic research.
Chapter 17, by Kay L. O’Halloran, Sabine Tan, and Peter Wignell, covers
multimodal discourse analysis, a particularly fruitful and growing area
within the SFL theoretical framework, abbreviated as SF-MDA (Systemic
Functional Multimodal Discourse Analysis). The chapter highlights the
features of SFL theory which inform SF-MDA before presenting an exemplar
analysis of a multimodal text, an internet webpage, using the analytical
tools of SF-MDA.
The last chapter in this part, Chapter 18, by Gerard O’Grady, outlines the
relationship between SFL theory and Critical Discourse Analysis (CDA),
particularly that developed by Norman Fairclough (1989, 2015). After
tracing the development of CDA, O’Grady describes Fairclough’s method-
ology, pointing out its compatibility with SFL theoretical perspectives and
analytical concepts. Towards the end of the chapter, O’Grady presents a
critique of the criticisms of SFL-inspired CDA, and in the last part of the
chapter he outlines some of the particularly productive areas of CDA
research which make use of SFL theoretical tools.
This third part of the volume presents several fields of research in which
the theory of SFL has been applied, with the first three chapters
(Chapters 19, 20, and 21) on different areas related to language develop-
ment and learning. Geoff Williams’ chapter describes Halliday’s ground-
breaking work on child language development (Chapter 19), and is
Downloaded from https://www.cambridge.org/core. Stockholm University Library, on 30 Oct 2019 at 11:18:10, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.002
6 WENDY L. BOWCHER, LISE FONTAINE, AND DAVID SCHÖNTHAL
organized around two main thrusts of the SFL approach: how research into
language development informs the SFL theory of language, and how learn-
ing language goes hand in hand with learning culture and in developing
one’s position and identity within society. Williams relates how this latter
perspective, in particular, is demonstrated through research into
mother–child interaction conducted by Ruqaiya Hasan and analyzed using
her semantic networks (e.g. Hasan 1996).
Chapter 20, by Heidi Byrnes, moves on to the application of SFL in the
field of second language development. She first locates SFL within the
general domain of second language acquisition research, and then discusses
the application of several key features of SFL theory and L2 teaching and
learning, noting, in particular, the value brought to the L2 teaching/learn-
ing environment of the SFL approach to the description of language and
language in use, as well as its description of a lexicogrammar rather than
the typical and ‘unsustainable’ separation of lexicon and grammar. She also
raises the significance of SFL research into and description of grammatical
metaphor in second language learning and teaching.
In Chapter 21, Peter Mickan outlines how SFL theory has been applied
within the field of general education. Mickan’s chapter covers work focus-
ing on early childhood and primary school education, secondary school,
tertiary education, and finally teacher-training and educational research.
Underlying all these chapters is the principle that learning language and
learning through language is a process of ‘learning how to mean’.
In Chapter 22, John Bateman, Daniel McDonald, Tuomo Hiippala, Daniel
Couto-Vale, and Eugeniu Costetchi note the long history of connection
between SFL and computational linguistics, mentioning Halliday’s involve-
ment in ‘some of the earliest attempts to achieve automatic translation
systems in the 1950s’ and his key role in some of the ‘most well-known
language-oriented systems to emerge in computational linguistics and Arti-
ficial Intelligence in the 1970s and 1980s’. The chapter discusses recent SFL-
related research, noting some of the challenges that a meaning-based
theory of language poses for computational models, but also the distinct
and far-reaching possibilities that SFL can offer the field.
The next three chapters (Chapters 23, 24, and 25) are connected in terms
of their focus on SFL in relation to science and medical research. Chapter 23,
by Elissa Asp and Jessica de Villiers, concerns clinical linguistics. After
briefly describing the field of clinical linguistics, Asp and de Villiers present
an overview of research that falls within the SFL theoretical approach,
including work on schizophrenia, neurodevelopmental and neurocognitive
disorders, Alzheimer’s disease, and aphasia. They also discuss the signifi-
cance of SFL-informed research and some future directions.
The focus in Chapter 24, by Michael Halliday and David G. Butt, is
‘science’ and scientific language. They describe the part that language has
played in the development of science and how scientific language, as a
register, has come to construe knowledge and experience in a specific and
Downloaded from https://www.cambridge.org/core. Stockholm University Library, on 30 Oct 2019 at 11:18:10, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.002
Introduction 7
‘uncommonsense’ way, which they argue is in line with the social purposes
of science. Another argument that the authors make is that science lan-
guage, or ‘verbal science’, operates in a similar way to verbal art (see Miller,
this volume) in that there is a ‘symbolic articulation’ – that is, conventional
language choices realize metaphorical constructions of meanings to create
novel understandings.
Alison Rotha Moore, in Chapter 25, comprehensively reviews SFL-
informed research into language and medicine. She describes the kind of
health problems and medical contexts in which research has been con-
ducted, such as HIV, emergency services, surgery contexts, and health
curricula. She discusses the analytical tools used and then outlines the
achievements SFL researchers have made in this field. The chapter also
suggests some of the directions this kind of research can take and the
possible knock-on improvements that could emerge in healthcare and in
the healthcare system.
The next chapter (Chapter 26), by Donna R. Miller, entitled Language and
Literature, presents some of the most innovative work on the analysis of
‘verbal art’ available. Miller presents a historical recount of the field of
stylistics and the place of Halliday’s and Hasan’s work in relation to this,
noting the possible reasons why Halliday’s work has been acknowledged
outside the circle of SFL scholars, whereas, surprisingly, Hasan’s has
received little recognition. Several key influential figures emerge in her
chapter, such as Jakobson and Mukařovský. The chapter describes Hasan’s
systemic socio-semiotic stylistics (SSS) model and its value and insights for
understanding the ‘art’ in verbal art, demonstrating that literature is not
like other varieties of language, but is a special variety.
In Chapter 27, Michele Zappavigna describes current work in the appli-
cation of SFL tools to analyzing social media platforms such as Twitter,
Facebook, Instagram, Weibo, etc. Specifically, she focuses on the way SFL
tools can unlock the construction of ‘identity’ and ‘affiliation’ in such
platforms, pointing out that this kind of investigation derives from Firth’s
(1950) work on the language of persons and personalities and his concepts
of ‘communion of feeling’ and ‘the user in uses’ (see also Martin 2009). The
chapter also discusses issues involved in collecting social media data.
Chapter 28, by Erich Steiner, focuses on translation studies and the
usefulness of SFL ideas and concepts for theorizing and modelling the
process of translation. His chapter compares translation with multilingual
text production and interpreting, and discusses the relationship between
translation, text variation, and paraphrase. It also includes a discussion on
the SFL perspectives on equivalence, the translation of registers or text
types, and the role of the translator in the process of translation. Steiner
describes some of the SFL tools for text analysis that are relevant to transla-
tion and possible future directions of SFL-informed translation studies.
The last chapter in the volume is Abhishek Kumar Kashyap’s chapter on
language typology (Chapter 29). Kashyap briefly traces the development of
Downloaded from https://www.cambridge.org/core. Stockholm University Library, on 30 Oct 2019 at 11:18:10, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.002
8 WENDY L. BOWCHER, LISE FONTAINE, AND DAVID SCHÖNTHAL
the field of language typology, the kinds of questions that typologists ask,
and some of the different theoretical approaches taken. The chapter covers
the major contributions to language typology from an SFL perspective and
applications of language typology to other fields such as translation, inter-
cultural communication, and language teaching and learning.
References
Downloaded from https://www.cambridge.org/core. Stockholm University Library, on 30 Oct 2019 at 11:18:10, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.002
1
Firth and the Origins
of Systemic Functional
Linguistics
Process, Pragma, and Polysystem
David G. Butt
The ideas and principles of Professor John Rupert Firth are an essential
source of what is important and distinctive about the development, after
Firth’s death, of Systemic Functional Linguistics (SFL). Firth was the first
Professor of General Linguistics in England (1944) at the School of Oriental
and African Studies at London University (SOAS). He also headed the
Department of Phonetics in London. Though first trained in history,
between the outset of WWI and 1928, Firth had worked as a teacher of
English and as a professor in British Imperial India (at Lahore, now Paki-
stan) as well as in Afghanistan and East Africa. He later returned to these
communities to conduct further descriptions of languages. His students
and colleagues are notable for the extent and depth to which they
developed the study of languages of these regions, as well as languages
of East and South East Asia. Firth emphasized the importance of
de-Anglicization, and of looking back at one’s own language from the
perspective of another culture. This is a form of ‘de-familiarization’ quite
remarkable in a person who seemed a conservative Yorkshireman. Yet, as
emphasized by Roman Jakobson after their two meetings, Firth shared with
the pioneer of British linguistics, Henry Sweet (1845–1912), an ‘unusual
courage to see the world’ with his own eyes ‘irrespective of the environ-
mental usage, habit and predilection’ of conventional thinking (Jakobson
1966:242). For others, especially to phonemicists and morphologists in
America, this ‘unusual courage’ appeared to be eccentricity, and a lack of
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
12 DAVID G. BUTT
We might note at this point that there is little difficulty in citing what Firth
recommended for linguistics, and also in establishing what he believed a
linguist should abjure. Firth’s view is at the polar extreme from Chomsky’s
claim that linguistics is a ‘branch of cognitive psychology’ (Chomsky
1972:1). Similarly, the idea that linguists might concede the domain of
semantics to a philosopher of ‘intentions’, like J. R. Searle (1969), or that
we pass pragmatics over to the maxims of a logician like H. P. Grice (1989),
would have seemed a total abrogation of the roles and tool power of
linguistics. For a start, such an approach assumes a diminution of linguis-
tics, namely, that each of the strata of language could be contracted out and
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 13
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
14 DAVID G. BUTT
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 15
expression? Those at greater distance, in time and space, raise other issues –
for example, the claim that it is inconsistent of Firth to state the anomalies
of phonological practice using a phonemic alphabet.
As an initial response to such reactions and interpretations of Firth’s
ideas, it is advisable to consider two points: first of all, the problem of
anachronism in seeing the past according to the fashion of speaking in
succeeding decades; and secondly, to ask how a genuinely different way of
seeing problems can achieve a rational evaluation in any science with a
strong prevailing set of assumptions.
Certain ways of seeing scientific and linguistic issues in particular have
made it difficult for many linguists to construe what Firth was delivering
with the brevity of ‘gnomic utterances’. Strevens emphasized, on the other
hand, in his editorial comments (Firth 1964), that Firth wrote his papers for
expert audiences of linguistic specialists, to whom the assumption of
expertise was naturally extended. The problems arise with the beliefs a
reader brings. If reading or hearing Firth in the late 1940s, the prevailing
views were positivist and Bloomfieldean. ‘Language’ (Bloomfield 1933)
espoused a behaviourist psychology and eschewed any ‘science’ in seman-
tics – any science of meaning. Perhaps more problematic was the ideal of
what could be a science at all. The discovery procedures of the era meant
treating levels like phonology and grammar as if they needed to be man-
aged autonomously: quite the inverse of Firth’s ‘meaning is made at all
levels’. Bloomfield’s statement may not do justice even to Bloomfield’s own
linguistic methods (e.g. his work on Menomini, including categories
like ‘verbs of being’ (Bloomfield 1962:274); and ‘verbs of undergoing’
(Bloomfield 1962:298)). Firth’s reactions were patently clear, but bewilder-
ing to the adherents of Bloomfield: the phoneme is dead, and Nida’s work
on morphology is ‘“nonsense” . . . added to “nonsense”’ (Firth 1957:170).
A decade and more later, with Chomsky’s work, syntax is presented as
formal and autonomous; meaning is deferred and passed over to philoso-
phy; the individual is the domain of study; the assumption of a genetically
based universal grammar is used against any evidence of the ‘typical actual’
of language behaviour; a language is the collection of sentences generated
by the formal rules of the language; and intuitions of grammaticality
become the least assailable form of linguistic testimony.
As the adequacy of Chomsky’s assumptions in linguistics was variously
contested, semantics returned but via formalisms from logic (entailment;
presupposition), on the one hand, and from ‘speech act’ theory in philoso-
phy, on the other. The study of context or pragmatics was also re-invented,
but through a kind of ‘Alice in Wonderland’ idiom of maxims of conversa-
tional logic. These supposed reinvigorations of semantics and cultural
analysis were two doses of North Atlantic ‘conventional wisdom’ which
were, from their dominant spokespersons at least (i.e. Searle and Grice,
respectively), without ethnographic evidence or any historical basis or
cross-cultural complexity. Grice himself, in his idiom of gentle whimsy,
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
16 DAVID G. BUTT
is purported to have claimed that his maxims were just special versions of
what a ‘decent chap ought to do’ (cited by Thomas 2011:219). Had not
Malinowski and Firth already died, their reactions to these developments
might well have ended them. But my point is that by the time Firth and
Malinowski are relinquished to the status of ‘proto-pragmatists’ (Levinson
1983:xii), they have been occluded each decade for quite different reasons –
firstly, due to assumptions that most have relinquished in linguistics; and
secondly, because we now draw our semantic inspiration from speech act
theory, Gricean maxims, the psychological load of relevance, or other
sources extrinsic to the recording of actual linguistic exchanges. If we
reflect on the decades of the dominant American linguistic theories after
WWII, the idea of paradox can be invoked most strongly against each of the
prevailing assumptions of each decade. Consider, for instance: the rejection
by both Bloomfield and Chomsky of the study of meaning; the promotion of
intuitive competence over observable behaviour; the dramatic genetic
speculations about how recent language might have evolved; the idea that
the basis of language might not have been for communication between
people; and the final reduction of UG to recursion (and Merge) only
(Chomsky and McGilvray 2012:16–20, 245).
There were, then, at least on the face of things, both paradoxical state-
ments and what might be called serious revisions. Yet, all these statements
have been influential principles for a period in modern linguistics. One
might be reminded then that dogma and rhetorical forcefulness character-
ize much of what achieves a high degree of visibility in science (see Brooks
2011: Chapter 6). Firth’s approach, with its emphasis on actual languages
and variation, deserves fresh evaluation in the light of Halliday’s develop-
ment of Firth’s polysystemic approach to linguistic description. Much as
with the work of Sapir, important work can be brought back to a more
rational evaluation.
The efficacy and consistency of Firth’s linguistics can be seen clearly in his
emphasis on ‘restricted languages’: the variation in relation to social pur-
poses and varying social contexts. By setting the scale of analysis close to
the ‘typical actual’ of social events, Firth’s approach helps resolve (or
‘dissolve’) many of the practical and theoretical conundrums of ‘doing
linguistics’. These include, for example:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 17
Firth’s principles appear to be all directed to bringing out the way mean-
ingful practices in a culture ‘hang together’ or cohere through an ensemble
of organized behaviours. From the access we have to the ‘aggregate of
experience’ (Firth 1962:1), we can follow the social relations and their role
in events down to the posture of body and voice in the reciprocation of a
social situation. It is the expectancies that hold the social show together.
These mutual expectancies can be regarded in probabilistic terms: when we
get ‘A’, to what degree does that predict (or prehend) the presence of ‘B’,
and of ‘C’, or of ‘B: C’, etc.? Clearly, this suggests we will be investigating
aspects of text beyond the domains of normative grammar and lexicology,
and beyond phonemic ideas of sound. Such an open sense of co-occurrence
encompasses more too than the ideas that Firth passed down to us
in collocation (the co-presence of word-like units) and colligation (the clus-
tering of grammatical categories). Mutual expectancy opens up for
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
18 DAVID G. BUTT
consideration any regularity that carries a value in the text; and this
breadth of conception leads Firth to step away from assuming the efficacy
of conventional ‘units’, preferring to use the term ‘piece’. With this neutral
‘place-holder’ for sequence and order, Firth can propose statements that
are not limited by the procrustean assumptions of classical, Western
‘syntagma’. One might ‘net in’ co-occurrences that are significant in char-
acterizing the text, or the persona of an author, or the typical personae of a
group with their restricted language, their trade language, their signs of
solidarity or ‘insider’ talk, their total sociality, their semantic variety.
Yet, it still needs to be asked, how then do we proceed under the guidance
of these principles? To produce ‘statements of meaning’, the language
needs the separation of its various polyphonic strands, of its ensemble of
congruent levels. For Firth, the serial contextualization of a language piece
is like the diffraction of white light through another medium: one sees that
the totality of the white beam is actually a spectrum of waves all contrib-
uting at different scales, but with each essential to the patterning across all
scales. Hence, ‘meaning is made at all levels’. And that cross-level relation
must be handled according to ‘the differences that make a difference’
(Bateson 1982).
The alliterative terms – patterns, pragma, and polysystem – supply three
of the motifs that guide Firth’s approach to structure and function.
Structure is that syntagmatic order of ‘mutual expectancy’ within a given
social event – not the simple sequence of the structuralist’s fixed units.
Function is the profile of relations that pertain to paradigmatic aspects of
that order – essentially, where a ‘piece’ fits in all up and down the relations
on one level and then, where crucial to the issue under investigation, across
those relations from other levels that determine the value of the ‘piece’. For
example, the expression He kept popping in and out all afternoon challenges
segmental descriptions at the ranks in the grammar of verb/verbal group;
but statements of meaning relating to this wording would also need to
include that it only fits into specific social situations, between certain
participants, face to face, and somewhere contextually between personal,
confidential exchange and gossip. It is part of a network of relations that
narrow down the potential of wording to what might be thought of as its
place in a ‘restricted language’. Put quite simply, a polysystemic profile
suggests that the piece could not appear just anywhere. It carries with it a
penumbra of collocational and colligational patterns that reflect the habits
of a social membership and of a personality: an instance of the ‘typical
actual’ for a personality in a social moment.
The linguistic importance and cross-disciplinary resonance of this obser-
vation can be brought out by again reflecting on ‘mutual expectancy’, and
the more unusual, technical word from Whitehead’s ‘process’ interpret-
ation of nature: namely, prehension (Whitehead 1979:379). Considered
from a phonetician’s point of view, mutual expectancy encompasses much
that linguistics up to today has overlooked about patterning in language.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 19
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
20 DAVID G. BUTT
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 21
The human body is that region of the world which is the primary field of
human experience but it is continuous with the rest of the world. We are in
the world and the world is in us.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
22 DAVID G. BUTT
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 23
Mitchell uses three ‘attested’ texts of sales in markets and shops in Cyrena-
ica. From these, he exemplifies the most significant contextual distinctions
made by Firth. In particular, the final discussion of negotiations over a fast
horse – a prestige purchase, involving long-standing traditions and commu-
nity expertise – illustrates the practical necessity of beginning at context
rather than treating context as a final form of pragmatic ‘tweaking’. Mitch-
ell also raises a number of problems, issues to which I would also attest
from my own collaborations in surgical care and psychiatric contexts.
These problems include the following: the absence of talk on the activity;
the presence of oblique talk; and the difference between talk on the job and
talk explaining the job; and (I would add) talk that runs a bulletin and guide
which, as a commentary, brings all participants to a shared understanding
of how the process is progressing. Mitchell certainly emphasizes ‘mutual
expectancy’ and even cites prehension (Mitchell 1957:39, 49, 54, 55), noting
the cumulative effects of unfolding connections between relevant words
and actions. In SFL, following Halliday, the process of choices unfolding has
been referred to as logogenesis. The term encompasses the changing values
in the text as new choices accord with what has gone before, and as these
choices direct the changing expectancies as to what is to come in the light
of the most recent choices in the potential. Consider how generic elements
or stages are related to both order and succession (Mitchell 1957:43, 47).
The link between ‘habitual’ patterns, collocation, and expectation is
expressed by Mitchell (1957:55). An interesting distinction is also made
plain between personalities and persons (Mitchell 1957:36–59): the differ-
ence is between the role you are playing and the physical being involved.
Firth emphatically rejected the idea of individuals in language; but he
invoked ‘personality’ (as did Sapir). These distinctions lead to Mitchell’s
four-column tabulation of Text, Translation, Personality, and Stage. At
other points there are signs of some emergence of the more recent style
of Pragmatics: ‘essential conditions’ (Mitchell 1957:36) suggests J. L.
Austin’s influence (although Austin’s 1955 lectures were not in print until
1962); and mutual expectancy appears to encompass overt connections like
those that were later referred to by ‘adjacency pairing’ (Mitchell 1957:59).
Mitchell (1957:34) is working out the ‘complex pattern of activity’ per-
taining to a central pattern (the way ‘a line needs to be distinguished as to
its role in either a rectangle, or as the radius of a circle’). In relation to
the sale of the horse in the market, the exchange takes semantic directions
that the outsider is unlikely to predict or construe. The quoting of a
quatrain of traditional poetry is not unique to the Bedouin stages and styles
of negotiation – many negotiations across cultures invoke gnomic sayings
or apothegms. But the complex relationship between the horse’s speed and
the rider’s good fortune, on the one hand, and the whorls on the horse’s
flanks, on the other, are opaque for anyone not enculturated to Bedouin
horsemanship. These issues are detailed (Mitchell 1957:70–1) and are strik-
ingly parallel to Malinowski’s example of canoe racing, relying as they do
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
24 DAVID G. BUTT
1
See articles by Allen (1953); Halliday (1959; 1963); Henderson (1987); Mitchell (1975); Palmer (1970); Robins
(1957).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 25
1.4.2 Sound
While Firth and Halliday are both cited for systemicization in phonology
and the semantics of intonation, the consistency of their work (along with
the work of Abercrombie at Edinburgh) is notable when viewed against the
current trends in phonological and phonetic debate. In an attempt to set
the record right, Palmer emphasizes that Firth’s distinctive approach can be
tracked in publications from 1934 to 1937 (see Palmer 1970:x–xi). The
polysystemic principle was argued in Firth’s (1935:51) discussion of Marathi,
where he points out that eight /n/ sounds needed to be recognized as
linguistically and functionally distinct, and therefore not the same unit even
if seemingly identical phonetically. This was part of the broader contextual
principle: namely, that one had to recognize not only the phonetic contexts,
but also ‘lexical and grammatical functions’ (Firth cited in Palmer 1970:xi).
The value of an item was dependent on the systemic character of ‘recurrent
contexts’. So too, there were the issues of y and w prosodies; the urging of
the importance of the syllable and of extended phenomena in phonological
analysis; and the notion of ‘articulation types’. Palmer recommends the
completeness of Henderson’s (1949) analysis of Siamese in that she shows
a full hierarchy of prosodies: in sentence; sentence parts; polysyllables and
sentence pieces; in syllables and syllable parts; and in consonant and vowel
units. Firth’s arguments against rigid phonemic methods concerned the
unreality of segmentation and of discovery procedures. These arguments
were bewildering to linguists of his day, despite the vigorous critique of
the phoneme concept by Twaddell in the USA (see Anderson 1985) and the
questions raised by other linguists in the USA (Palmer (1970:x) cites Harris
and Hockett in this regard). Today, views homogeneous with those of Firth
are part of the ‘natural’ background in the study of phonetics. Since Firth
emphasized that the /d/ that was word initial was not the same as /d/ word
final since the two acoustic elements operated in distinct systems (i.e. with
differing ‘valeurs’), then we might regard this as a rigorous application of
relational thinking (suggesting some affinity with the extreme relationalism
of Hjelmslev, a linguistic alter ego with whom Firth enjoyed frequent
exchanges and debates). We can see that Firth’s declarations against Saus-
sure’s work – namely, that it creates a system ‘in rebus’ – need to be seen
alongside what may be the strictest application of Saussure’s relational
universe (see the comment on valeur by Palmer 1968:xx).
The problems urged by Firth in relation to phonetics and phonology are
usefully explained and criticized in an evaluation of Linguistic Thought in
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
26 DAVID G. BUTT
Firth consistently refers to the body going out to the world – of the ears
being active in exploring sounds; of the brain as a guide to acting and
moving in ‘situations’; of the human as engaged in the pursuit of a ‘joy’
at structure. Firth sets out in linguistics with the explicit assumption of this
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 27
active engagement with the world, quite the opposite of the passive, tabula
rasa notion often foisted upon empiricists in human sciences. Firth’s citing
of sources, however, needs to be regarded for its own cultural context and
the ways in which Firth’s approach is oriented to current thinking.
Firth cites Charles Sherrington’s work for its emphasis on movements
and proprioception as ways into understanding emotions – the ‘felt ME’ –
and as the way of building relations with other humans. Firth responds
positively to Sherrington’s view that the brain is a ‘manager of muscle’, a
view in accord with the fact that current analysis suggests that the cerebel-
lum, although only 10 per cent of the brain’s volume, encompasses more
than fifty per cent of its neurons.
More centrally, Firth argues for the continuities between human inner
and outer worlds, with the brain creating a form of ‘mutual grip’ between
the world and us. This side of Firth’s thinking is congruent with the
position of neuropsychologists like Trevarthen (1998; see also Panksepp
and Trevarthen 2009). Trevarthen, having worked with Sperry, Bruner,
and Halliday, has argued for a related going out to the world in neonates,
a bridge-building through intersubjectivity to person-ness. Trevarthen
(in Stensæth and Trondalen 2012) summarizes his own thinking thus:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
28 DAVID G. BUTT
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 29
modifies the value of what has preceded, namely, the choices that have
brought the logogenetic unfolding to this point. The expansion of
computer-based corpus studies has not only activated Firth’s concepts of
collocation and colligation, but has also given a practical, probabilistic
character to register studies and for ‘restricted languages’ (Halliday and
James 1993; Teich 2003; Bartsch et al. 2005; Matthiessen 2015b).
1.7 Conclusion
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
30 DAVID G. BUTT
the robustness of the SFL model. Furthermore, the ideas around context
have insulated SFL from the autonomous formalisms that Duranti states
had taken over from the anthropological linguistics in America (see Duranti
2009:3).
Firth was first of all a naturalist of languages – the languages of Eastern
and African cultures. Firth was focused on instances of language rather
than on the essence of what language ‘is’. Like Henry Sweet before him,
Firth’s efforts were directed to shifting the axis of linguistic discussion
away from the focus on a Eurocentred, reconstructed classical tradition
towards the ‘typical actual’ of current speech. His interests were data
driven and problem oriented: from apprehending the polyphony of human
articulation to improving the system of ‘shorthand’ for rapid recording of
speech. In his writings he tended to reprise his speeches before learned
societies, a spoken context in which much could be taken for granted.
On the other hand, he wrote two clear and prescient books on language
matters for a wider public that had, between the Wars, a demonstrated
keenness to absorb knowledge from experts. Between linguists, it might be
said that he insisted on consistency of theory, part of which meant being
vitriolic at any reifying of theoretical abstractions. Firth was politically
conservative, yet found common ground with the Russian critique of Gene-
van structuralism as well as with Halliday’s Marxist perspective on linguis-
tics. Most remarkable, perhaps, is that he demanded of his students and
colleagues that they defamiliarize their worlds through the semantics of a
non-European language, adopting the lens of linguistic evidence and only
looking back at English with a renewed, extended grammatical
imagination.
References
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 31
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
32 DAVID G. BUTT
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
Firth and the Origins of SFL 33
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
34 DAVID G. BUTT
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:12:20, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.003
2
Key Terms in the
SFL Model
Jonathan J. Webster
1
The selections for further reading included in the footnotes of this chapter are adapted from Bloomsbury’s
The Essential Halliday (Halliday 2009) edited by J. J. Webster, which includes selected extracts and additional readings
from the ten volumes of Halliday’s Collected Works for twenty key concepts in Systemic Functional Linguistics.
2
For further reading on ‘functions and use of language’, see Halliday 2003a:298–322; Halliday 2003b:47, 51–6,
68–74, 81–7, 270–80; Halliday 2007a:41–2, 50–3, 56–7; Halliday 2007b:88–92, 120–2.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
36 J O N AT H A N J . W E B S T E R
[t]he usual way we talk about language is by saying that language ‘expresses’
meaning, as if the meanings were already there – already existing, in some
formation or other, and waiting for language to transpose them into sound,
or into some kind of visual symbols. But meaning is brought about by
language; and the energy by which this is achieved, the source of its semo-
genic power, is grammar.
(Halliday 2013:194–5)
In his book Reinventing the Sacred, Kauffman (2008:193–4) writes that ‘while
the human mind, central to our human embodied agency, is sometimes
algorithmic and sometimes computes, it does some things we do not yet
understand; it makes meanings’.
Halliday (personal communication) describes language as ‘a basic human
resource with potentially immense power, which is hidden, partly because
people are genuinely not aware of how much they are, in fact, depending
on it’. We depend on language to construe the world around us and
describe our feelings within, and exchange this meaning with others.
Meaning, in SFL theory, is not limited to referential meaning, i.e. word
meanings. We use language not only to construe experience but also to
enact social relationships, and create the discourse.
A text is an instance of meaning, a construct of meaning that is formed
out of a continuous process of choice among the innumerable interrelated
sets of semantic options organized into three main functional components
or metafunctions: ideational, interpersonal, and textual.5 Operating in par-
allel, the three metafunctions comprise the total meaning potential of a
language.
3
For further reading on ‘act(s) of meaning’, see Halliday 2002a:201, 206, 354–6; Halliday 2002b:50, 52; Halliday
2003a:171, 174, 355–74, 375–89; Halliday 2003b:11–12, 14–15, 18–20, 113–43, 212–18, 239, 245–6,
249–50, 327–52; Halliday 2005a:198–202; Halliday 2013:253, 264.
4
For further reading on ‘semiotics’, see Halliday 2002a:196–218, 384–418; Halliday 2002b:23–84, 150–2; Halliday
2003a:2–7, 93, 113–15, 116–24, 131–7, 147–51, 171, 192–8, 199–212, 213–31, 275–7, 355–74, 375–89,
390–432; Halliday 2003b:6–27, 90–112, 140–3, 157–95, 212–26, 250–66, 281–307, 327–52; Halliday
2004:43–4, 53–5, 102–34, 216–25, 198–202; Halliday 2007a:81–96; Halliday 2007b:179–86, 193–6, 259–63.
5
For further reading on ‘metafunction’, see Halliday 2002a:21–36, 390–2; Halliday 2003a:15–18, 248–50, 277–8;
Halliday 2003b:209–25, 332–3, 335–6, 338–41, 343–4, 346, 348–9; Halliday 2005a:200–2, 215–22; Halliday
2007b:183–4.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
Key Terms in the SFL Model 37
SFL is first and foremost a theory about how language works at the level of
grammar.7 ‘[T]hinking about meaning means thinking grammatically’,
writes Halliday (2013:207). A grammar is a theory of experience of everyday
life. It is that abstract stratum of coding between meaning and expression; it
is a resource for making meaning. Just as linguistics is language about
language – or ‘language turned back on itself’ (Firth 1957:181) – grammatics
is a theory of grammar; it is a theory for explaining how the grammar
constructs a theory of experience. Grammatics is theorizing about a theory;
it is a theory of a second order, a part of a more general theory of meaning.8
In categorizing the grammar, the grammarian comes at the task from
three perspectives, each of which corresponds to a different stratum. First,
there is the higher stratum of semantics. Here, the grammarian’s perspec-
tive is from above. Second is the stratum of lexicogrammar, where the
perspective is from around. The third perspective is from below and looks
at the morphological and phonological realization of meaning. In a func-
tional grammar, priority is given to the perspective from above, as form
follows function, and the meaning of an expression will decide its phono-
logical and morphological realization.9
6
For further reading on ‘semantic systems’, see Halliday 2002a:196–218, 310–11; Halliday 2002b:23–8, 45–52;
Halliday 2003a:323–54; Halliday 2003b:90–8, 109–12, 115–25, 281–94; Halliday 2007a:345–6; Halliday
2007b:131–3, 143–4, 153, 158, 164–6, 183–4, 186–95, 256–7.
7
For further reading on ‘theory and description’, see Halliday 2002a:37–42, 58–61, 72–7, 86, 98–9, 106–17, 158–72,
396, 403–6, 414–15; Halliday 2003a:7–15, 37–47, 199–212, 327–30; Halliday 2004:53–8; Halliday
2005a:227–38; Halliday 2005b:156–63; Halliday 2006:294–322; Halliday 2007a:136–9, 149.
8
For further reading on ‘grammatics’, see Halliday 2002a:296–8, 365–6, 369–73, 384–6, 416–17; Halliday
2003a:264–5, 274–6, 286, 362, 373, 385; Halliday 2005a:213–38.
9
For further reading on ‘trinocular vision’, see Halliday 2002a:398, 402, 408–9; Halliday 2003a:202–5, 254–5, 266;
Halliday 2005a:231.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
38 J O N AT H A N J . W E B S T E R
the dimensions that define the overall semiotic space of lexicogrammar, the
relationships that inhere in these dimensions – and its relationship to other
sub-systems of language – to semantics and to phonology (or graphology).
Thus, according to systemic functional theory, lexicogrammar is diversified
into a metafunctional spectrum, extended in delicacy from grammar to
lexis, and ordered into a series of ranked units.
The fundamental categories for the theory of grammar are unit, structure,
class, and system. Rank is the scale on which the units are ranged.
Rankshift occurs when a given unit is transferred to a lower rank.
A structure is made up of ordered elements. Sequence is one formal expo-
nent of the more abstract notion of order. Delicacy has to do with the depth
of detail, ranging along a cline from least delicate at one end, i.e. primary,
to those small infinities at the opposite ‘where distinctions are so fine that
they cease to be distinctions at all’ (Halliday 2002a:48), i.e. secondary.
Structural types include configurational, prosodic, and periodic. Experi-
ential meaning can be accounted for in terms of the configuration of
process, participant, and circumstance. Interpersonal meaning tends to be
prosodic, involving intonation. Textual meaning is more periodic, with the
flow of discourse understood less in terms of discrete constituents than
wave-like movements.10
Based on the findings from his study of the English language, Halliday
observed how each functional component or sub-component produces its
own distinct dimension of structure. For example, experiential meaning,
i.e. the ‘construing experience’ function (a sub-component within the Idea-
tional Metafunction), is realized as the structural configuration of process
as an integrated phenomenon involving participant(s) and circumstance(s).
In Systemic Functional Grammar (SFG), the structural configuration of
process, participant, and circumstance is referred to as a clause’s transitiv-
ity structure.
We talk about our experience of the world in terms of processes, plus the
participants and circumstances that enter into them. Processes are typically
realized as verbs which may describe an action, or a feeling, or a state of
being, or a way of behaving, either happening in the world around us or
within our own consciousness. Processes are often accompanied by
10
For further reading on ‘structure and rank’, see Halliday 2002a:40–9, 75–81, 95–105, 106–17, 118–26, 196–218;
Halliday 2002b:24–5, 27, 79–80; Halliday 2005a:29–36; Halliday 2005b:xii–xxix, 154–63, 249–51.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
Key Terms in the SFL Model 39
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
40 J O N AT H A N J . W E B S T E R
11
For further reading on ‘intonation’, see Halliday 2002a:55, 78, 90–1, 192–3, 205–7, 262–4, 269–70; Halliday
2002b:27–9, 32–6, 204–5, 232–3, 255; Halliday 2003b:50–1, 106–7, 162, 177, 184–9, 233, 317–19; Halliday
2004:69–71; Halliday 2005a:77–8; Halliday 2005b:57–70, 106–7, 139–40, 155–6, 161, 192–5, 213–15, 218,
237–86, 287–92; Halliday 2007a:71–3, 101, 158–9.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
Key Terms in the SFL Model 41
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
42 J O N AT H A N J . W E B S T E R
<p 1> <c 1a> Over the past 40 years, species have been extending their
ranges toward the poles <c 1b> and populations have been migrating,
developing, or reproducing earlier in the spring than previously (1–4).
<c 2a> These range expansions and changes in the timing of seasonal events
have generally been attributed to ‘phenotypic plasticity’ – that is, the ability
of individuals [<c 2b> to modify their behavior, morphology, or physiology
in response to altered environmental conditions (5, 6).] <c 3> Phenotypic
plasticity is not the whole story. <c 4a> However, recent studies show
<c 4b> that over the recent decades, climate change has led to heritable,
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
Key Terms in the SFL Model 43
Dear Friends, on the 9th of October 2012, the Taliban shot me on the left
side of my forehead. They shot my friends too. They thought that the bullets
would silence us. But they failed. And then, out of that silence, came
thousands of voices. The terrorists thought that they would change our
aims and stop our ambitions but nothing changed in my life except this:
Weakness, fear and hopelessness died. Strength, power and courage was
born. I am the same Malala. My ambitions are the same. My hopes are the
same. My dreams are the same.
(Yousafzai 2013)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
44 J O N AT H A N J . W E B S T E R
It freezes the whole thing, and then you have to introduce the dynamic in
the form of paths through the system. Your problem then is to show how the
actual process of making paths through the system changes the system.
(Halliday in Martin 2013:110)
12
For further reading on ‘quantifying language’, see Halliday 2002a:70–2, 92–4, 166, 168–9, 352–68; Halliday
2003a:23–6, 122, 253, 404–13, 425–6, 430; Halliday 2005a:8–9, 13–19, 42–62, 63–75, 76–92, 93–129,
130–56, 157–90, 235–8; Halliday 2006:209–48; Halliday 2007a:310–16.
13
For further reading on ‘markedness’, see Halliday 2002a:305, 320–1, 326, 376–7; Halliday 2002b:28–38, 199–200,
205; Halliday 2003b:342–3; Halliday 2005a:22–3, 81, 88, 91, 96–7, 101–2, 131–2; Halliday 2005b: 5–54,
55–109, 110–53, 154–63, 193–5, 203, 220–31, 249–61, 264–86, 288–9; Halliday 2006:5–174, 209–48,
330–2.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
Key Terms in the SFL Model 45
To illustrate his point, Halliday gives the analogy of climate and weather.
The weather is what we experience on a daily basis. Over time, we general-
ize this daily occurrence of the weather into our sense of what the climate
is. So the climate becomes a theory about the weather. Similarly, like the
notion of climate, the system of meaning – meaning potential – is a theory
of the text – the instantiation of meaning potential (see Halliday
2008:79–80). Just as the weather and climate are not two distinct and
separate phenomena, so too parole (instances of language) and langue
(system of language) are one and the same phenomenon.
14
For further reading on ‘indeterminacy in language’, see Halliday 2002a:399–402, 409–10; Halliday 2002b:33, 51,
139–40, 145–6; Halliday 2003a:54–5, 254–5, 266–7; Halliday 2005a:204–7, 211, 226–30; Halliday
2007b:193, 200.
15
Halliday defines a register as ‘a syndrome, or cluster of associated variants; and again only a small fraction of the
theoretically possible combinations will actually be found to occur’ (2002b(1990):168) rather than the obligatory
incidence of particular features. Dialects are identified by their users. Codes are patterns or speech habits of speakers
of the same language. For further reading on ‘varieties and variation in language: dialect, register, code’, see Halliday
2002b:17, 168–70, 231–4; Halliday 2003a:255–6, 268, 360, 362–3, 382–3, 416–17; Halliday 2005a:225–6,
248, 263–4; Halliday 2005b:214–16; Halliday 2007a:29–31, 240–3, 296–300; Halliday 2007b:5–40, 85–8,
93–7, 103–7, 115–16, 129–30, 138–9, 140–2, 147, 174–5, 181–3, 196–7, 205–9, 235, 242–3, 252–5,
259–61.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
46 J O N AT H A N J . W E B S T E R
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
Key Terms in the SFL Model 47
16
For further reading on ‘grammatical metaphor’, see Halliday 2002a:219–60, 346–8, 358–60, 397; Halliday
2002b:23–84, 160, 164, 219–23, 226; Halliday 2003a:130–4, 139–76, 248–70, 282, 284–5, 384, 388, 415,
419–23; Halliday 2003b:339–40, 347–9, 367–9; Halliday 2004:7–23, 32–43, 49–101, 102–34, 143, 147–52,
156–7, 162, 171–9, 190–7, 214–16, 220–5; Halliday 2005a:42–62, 63–75, 196–212, 213–38; Halliday
2006:325–33, 339–44; Halliday 2007a:63–80, 105–10, 117, 123, 126–8, 301–3, 354–67, 379, 381; Halliday
2007b:239, 243–4, 278.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
48 J O N AT H A N J . W E B S T E R
17
For further reading on ‘context of culture and context of situation’, see Halliday 2002a:29, 35, 201, 211, 217, 221,
225–31, 243, 246, 263, 283–5, 311, 357, 359, 405; Halliday 2002b:38, 44, 51–64, 150–2, 229–34, 243–4, 251,
254; Halliday 2003a:154–6, 185, 195–7, 210, 273, 279, 298–9, 358, 362, 382, 420; Halliday 2003b:81, 87, 95,
101, 111, 121, 134, 204, 207, 286–95, 302–4; Halliday 2005a:207–8, 217, 225, 238, 249, 256, 260, 266;
Halliday 2005b:199, 306–37; Halliday 2006:10, 13, 16, 20, 64, 355–7; Halliday 2007a:85–7, 94, 96, 271–90, 298,
300, 307, 311, 349, 354–67, 368–82; Halliday 2007b:59, 62, 77, 82, 90–7, 105, 110–20, 127, 130, 133–7, 140,
142, 172, 180–2, 184, 187, 192–9, 203, 209, 235, 242, 258–9, 262.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
Key Terms in the SFL Model 49
18
In a letter, Professor M. A. K. Halliday identified the following seven distinctive features of SFL: ‘(1) Quantitative studies
and probability: This has been a constant thread since “Linguistics and machine translation” (it was even raised in my
PhD thesis on the Secret History), with references also to Shannon Weaver’s theory of information (now at last
recognized for its importance, e.g. in recent work by Terrence Deacon). And as far as I know nobody has followed up
the work I did with collaboration from Zoe James on the probabilistic nature of grammatical systems. (2) Metafunction:
again a recurrent theme, with the point that metafunction determines the way that languages have evolved, and the
insistence that the interpersonal and textual metafunctions are equally fundamental, along with the ideational, to the
functioning of language as a semiotic system, the form taken by grammatical structures maximizing the possibilities for
different meanings to combine freely with one another. (3) Historical contexts of linguistics: Wang Li’s and Firth’s
departments were unusual (even unique?) in building the history of linguistics into their teaching of the subject; Wang
Li the Chinese tradition, Firth the European (and also the Indian, though I never learnt so much about that); more
recently, linguistics in the context of the history of ideas, relation of the human to the natural sciences (linguistics itself
being also physical, biological and social). (4) Marxism: early with Jeff Ellis, Dennis Berg, Jean Ure and others, mostly
existing only in typescript; not made explicit in the McCarthy era (I had lost too many openings because of it in those
years!) but always part of my own thinking: values of language and language varieties, language as political process
and political tool; linguistics as (critical component of ) a science of meaning <more recent of course>. (5) Unity of
lexis and grammar. (6) Continuity of protolanguage. (7) Thinking in terms of patterns of analogies, complementarities,
compatibility with other views and theories (where others only see contractions)’ (Halliday, personal communication).
19
This interview was conducted by Annabelle Lukin, David Butt, and myself with Professors Halliday and Hasan in their
home in 2012.
20
Jeffrey Ellis, Jean Ure, Dennis Berg, Trevor Hill, and Peter Wexler.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
50 J O N AT H A N J . W E B S T E R
Human beings are incapable of living, surviving alone. In this lies their
humanity. Everything in them is created through being part of a society.
Either a reaction to it, or a following of it, whichever form you take.
As he states in his interview with Kress, Hasan, and Martin in 1986 (Martin
2013:118), Halliday saw Firth’s approach to the study of language as being
‘perfectly compatible’ with a Marxist linguistics:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
Key Terms in the SFL Model 51
is the way they have always developed in conversation with other people,
often as a by-product of my activities as a teacher; I have tried to acknow-
ledge those who have been part of this enterprise, though I am conscious of
having done so only very inadequately. It is impossible to track the proven-
ance of scientific ideas, and with our modernist ways of thinking we attach
too much importance to the individual anyway. It was my privilege to
encounter so many congenial and thoughtful colleagues.
I hoped that what I was trying to achieve as a linguist might make some
contribution to improving the human condition, however minuscule and
oblique. This is what I meant by calling the theory ‘appliable’. The term is
less specific than ‘applicable’, which denotes applicable to some specific task,
and therefore less immediate, and more indirect; its relevance is less obvi-
ous, but more long term. But other than this feature of being appliable, what
other aspect of the theory might be considered as marxist?
Practising appliable linguistics has indeed been a driving force behind the
development of Systemic Functional Linguistic Theory. Describing what he
means by ‘an appliable linguistics’, Halliday (2013:128) writes:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
52 J O N AT H A N J . W E B S T E R
References
21
For further reading on ‘language teaching, learning and development’, see Halliday 2002a:323–4, 349–51; Halliday
2003a:228–30, 273–4, 378–9, 384, 397–404, 429–30; Halliday 2003b; Halliday 2005b:297–305; Halliday
2007a; Halliday 2007b:63–4, 75–81, 118, 128–9, 175–6, 193–5, 212–13, 223–30.
22
For further reading on ‘linguistic computing’, see Halliday 2003b.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
Key Terms in the SFL Model 53
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
54 J O N AT H A N J . W E B S T E R
Shaviro, S. 2008. Reinventing the Sacred (Stuart Kauffman). Available online at:
www.shaviro.com/Blog/?p=636. (Last accessed 16/05/17.)
Yousafzai, M. 2013. The Full Text: Malala Yousafzai Delivers Defiant Riposte
to Taliban Militants with Speech to the UN General Assembly. Independent,
12 July 2013. Available online at: www.independent.co.uk/news/world/
asia/the-full-text-malala-yousafzai-delivers-defiant-riposte-to-taliban-mili
tants-with-speech-to-the-un-8706606.html. (Last accessed 05/06/17.)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:07:52, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.004
3
Semantics
Miriam Taverniers
3.1 Introduction
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
56 MIRIAM TAVERNIERS
In order to explain what ‘semantics’ can mean in SFL, we will consider the
following questions:
1
Note that the duality of the initial two conceptions of meaning that are pointed out here is widespread in linguistics:
the former, as the type of meaning that is interfacing with context, has also been called ‘contextual meaning’, or
‘extra-linguistic meaning’ (and more specific sub-types of this are ‘reference’, ‘ontological meaning’, ‘speech act
meaning’, etc.), whereas the latter has been called ‘intra-linguistic’ or ‘formal meaning’ (with ‘sense’ as a sub-type), or
has been defined in relation to grammar as the ‘semantics of grammar’.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
xt
te
n
co
ics
nt
a
sem
ram
-g
lex
ph
gra
realization statement
phon/
doings meanings wordings systemic option realization statement content expression language as a whole
context meaning form value token plane plane as ‘meaning-making’
as a ‘meaning potential’
context semantics lexicogrammar paradigmatic valeur syntagmatic structure
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
[1] [2] [3] [4]
Figure 3.1 Four types of conceptions of ‘meaning’ in SFL
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
58 MIRIAM TAVERNIERS
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 59
Asking the question of what a (separate) semantic stratum can be, more
specifically, correlates with asking what this stratum can contain, or how this
stratum can be modelled in relation to other – especially surrounding –
strata. In the visual metaphor of stratification, with the strata represented
as cotangential circles (since Martin and Matthiessen 1991; see Figure 3.1), the
higher levels of the language system are the more abstract ones, and the lower
ones the more concrete realizations. In keeping with this orientation, seman-
tics in general can be conceived of as ‘higher order valeur’: a ‘valeur’ which can
be recognized above the lexicogrammar, as a more abstract meaning.
Three complementary ways of modelling this ‘higher order valeur’ can be
distinguished. I will initially characterize them here, and we will then
return to the different types of modelling in more detail when looking at
specific semantic analyses (Section 3.3) after we have addressed the ‘why’
question (Section 3.2.3).
2
In this respect the topological modelling of (higher-order) meanings is similar to the method of semantic maps which is
used in functional typology (e.g. Haspelmath 2003).
3
Note that the ‘regrouping’ of distinct lexicogrammatical phenomena into topological ‘domains’ does not necessarily
imply that those domains are conceived as pertaining to a different (i.e. usually ‘higher’) stratum. Martin and
Matthiessen (1991) talked about regroupings within a stratum, with a focus on the lexicogrammar. This has been
interpreted by Halliday (1996:15) as suggesting a view of lexicogrammar as typologically organized and semantics as
topologically organized strata. This is also the view highlighted here, although it should be borne in mind that the
topological/typological distinction does not correlate with semantics/lexicogrammar per se (indeed, it perturbates
throughout the system as a fractal motif – see Section 3.4.2 below on fractality and 3.4.3 on the related concept of an
extravagant theory of language).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
60 MIRIAM TAVERNIERS
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 61
context
semantics
lexicogrammar
Figure 3.2 Three design alternatives in modelling a semantic stratum: topological [1],
systemic/typological [2], and discourse-structural [3]
other hand, is not enough. For some systemicists or from one perspective, this
is always the case, for others it is the case for specific contexts (as we will see
below). In order to come to an understanding of how the recognition of a
stratum of semantics is motivated in relation to this concept of variability, it
is necessary to first clarify what exactly this variability is, and how different
research purposes can put this variability into different types of perspectives.
4
Lamb (1962) (whose stratificational theory has been a source of inspiration in SFL) refers to those different situations
of variability between strata as ‘composite realization’ (=[2] here), ‘portmanteau realization’ (=[3]), and ‘interlocking
diversification’ (=[4]).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
62 MIRIAM TAVERNIERS
meanings
[1] [2] [3] [4]
f
forms
Figure 3.3 One-to-many [2], many-to-one [3], and many-to-many [4] conceptions of
variability between meanings and forms in relation to the default view of one-to-one
meaning-to-form couplings within lexicogrammar [1]
5
In Hjelmslev’s (1963) terms, a ‘connotative semiotic’ is then recognized (i.e. a semiotic that has a higher-order content
plane) (see Taverniers 2008).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 63
6
Other areas in relatively early SFL studies which point to a semantic stratum in addition to a lexicogrammatical one
include the following: stylistics and socio-semantic variation (esp. with the notion of a de-automatization of a grammar à
la Mukařovský, e.g. Halliday 1982); language development (ontogenesis) (with the view that the adult language system
contains more strata – an extra content stratum? – than the proto-language of the child); and mood (with variability
between mood choices and socio-semantic roles, on the one hand, and between mood types and lexicogrammatical
realizations, on the other hand).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
64 MIRIAM TAVERNIERS
‘command’ ‘probability’
projecting
modalized relational clause
interrogative It is obvious that …
Could you …
modalized projecting
imperative* declarative modal elements mental clause
You should … will, probably* I think, I suppose …
Figure 3.4 Interpersonal grammatical metaphors of mood (left) and modality (right)
theorized within a view of variation of one-to-many
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 65
‘part of a
piece of furniture’ ‘event’
‘part of an
animal’*
leg expressed
expressed as nominal
as clause*
Figure 3.5 The traditional view of metaphor (left), and an alternative view which is
equally inspiring in the study of grammatical metaphor (right)
and expression: there is one form, and this form has multiple meanings,
among which are at least one literal and at least one metaphorical interpret-
ation. This is the conception that prevails in thinking about metaphor in
general, and especially in traditional views of lexical metaphor (e.g. legs can
be ‘parts of animals’, i.e. have a literal meaning, or ‘parts of furniture’, as in
the lexical metaphor table legs). However, Halliday argues, the complemen-
tary view, i.e. of ‘one-to-many’, is equally important in analyzing grammat-
ical metaphor (see Figure 3.5, which shows the relation between those two
views). Thus in introducing ‘ideational grammatical metaphor’, Halliday
(1985) gives an example of a clause which realizes a process with its partici-
pants (i.e. a processual or ‘event’ meaning), and then shows that this pro-
cessual meaning could be realized by a nominal group, which functions as a
participant in another process configuration. The first example offered by
Halliday concerns the congruent wording Mary saw something wonderful,
compared to the incongruent A wonderful sight met Mary’s eyes or Mary came
upon a wonderful sight, where the process of ‘Mary seeing something wonder-
ful’ is construed as a nominal group, a wonderful sight.
In explaining grammatical metaphor as ‘one-to-many’ in addition to
‘many-to-one’, Halliday both connects grammatical metaphor with the
traditional concept of metaphoricity (‘many-to-one’) and brings the new
concept of grammatical metaphor more in line with what had earlier been
recognized in relation to typicality in the interpersonal metafunction
(where the ‘one-to-many’ view was the initial source of inspiration:
one speech function such as ‘command’ can have several realizations such
as ‘imperative’, ‘interrogative’, etc.). In terms of the stratal type of perspec-
tive, the approach to grammatical metaphor that is initially taken is that
from below: the starting point is a form, and the question is what this
form ‘means’, and then this ‘meaning’ is analyzed as having other
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
66 MIRIAM TAVERNIERS
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 67
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
68 MIRIAM TAVERNIERS
xt
te
on
c
s
tic
an
m
xt
se
te
on
c
s
tic
an
m
se
s
tic
an
m
se
s
tic ar
an m
m am
gr
se
mar
ico
am
lex
gr
ico
lex
m ar
am ar
gr m
am
ico
gr
lex
ico
ph
ra
/g
lex
phon
ph
ra
/g
phon
Figure 3.6 Metaredundancy relationships between strata, starting from the top focusing
on the content side of language (left), and starting from the bottom focusing on the
expression side of language (right)
on the relation (i.e. it forces us to choose one thing that realizes the other,
as with the terms ‘signifier’ and ‘signified’), metaredundancy is neutral in
this respect. Accordingly, the opposite view with metaredundancy cycles
starting from the lowest stratum becomes a complementary frame of
interpretation. Here, phonology redounds with the metaredundancy
between lexicogrammar and semantics, and if context is also taken into
account, phonology redounds with the metaredundancy between lexico-
grammar and the metaredundancy between semantics and context. This
view, as Halliday (1992) argues, is the most relevant in research focusing on
phonology. Figure 3.6 is an attempt to capture accumulative metaredun-
dancy cycles in a three-dimensional way.
The conception of language as a dynamic open system puts the motiv-
ation for distinguishing between lexicogrammar and semantics into a
wider perspective. It is because of the requirement of adaptability of the
system to different contexts that the distinction between lexicogrammar
and semantics is needed, because it is exactly through the flexible relation
between lexicogrammar and semantics that language can be open, and that
gaps and contradictions, as room for innovation, can be allowed. In other
words, it is through variability between meanings and forms that language
can continue to function. It is through the specific type of relationship of
semantics, not just to lexicogrammar, but to the rest of the lower layers,
that language has semogenic power, that it is a ‘meaning potential’.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 69
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
70 MIRIAM TAVERNIERS
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 71
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
72 MIRIAM TAVERNIERS
3.4.1 Multiperspectivism
An inherent feature of SFL is that it is ‘multiperspectival’. Its theory of
language is based on a number of dimensions along which differentiations
can be made (see Taverniers 2002 for a further exploration of ‘differentiat-
ing dimensions’), the most important of which are the distinction between
metafunctions, between different strata (stratification, as well as the rela-
tionship of realization or metaredundancy), between ranks, between
system and instance (instantiation), and between syntagmatic and paradig-
matic modelling (see Halliday and Webster 2009). Within each of those
dimensions, a linguistic phenomenon can be looked upon in different ways,
and different perspectives are possible, depending on the viewpoint and the
focal depth:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 73
context
semantics
‘sequence
‘event’ ‘entity’
of events’
* * *
lexicogrammar
y/graphology
phonolog
Figure 3.7 Shunting perspectives in theorizing grammatical metaphor as a tension
between lexicogrammar and semantics
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
74 MIRIAM TAVERNIERS
context
semantics
ideational interpersonal textual
meaning meaning meaning
combined syntagm
lexicogrammar
y/graphology
phonolog
Figure 3.8 Shunting perspectives in theorizing the metafunctions as semantic
components of the system of language
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 75
‘semantic components of the system’ (while at the same time (i.e. meta-
redundancy at play!) depending critically on the bundling of systems at the
lexicogrammatical stratum).
3.4.2 Fractality
By combining complementarities in order to come to a richer theory of the
complex phenomenon that language is, sometimes recurrent patterns are
discovered, and this feature is referred to in SFL as ‘fractality’ or ‘fractal
resonance’ (Martin 1995). Thus, fractality is an additional design feature
which is exploited at the meta-level, the level of the linguistic practice itself.
Put simply, a pattern that appears in one view, in one focus, at one level –
i.e. a type of modelling that is useful in accounting for a phenomenon
within that view – may also be recognized elsewhere, in another view that
appears by shunting the perspective. The newly recognized pattern is then
said to ‘resonate’ with the first one. Focusing on the meta-level, resonance
means that a model, a type of distinction, which works well for one area,
one view, or one dimension, is used as an inspiration to explore a different
area, view, or dimension of language.
The resonance may be between different ranks or units. This is the case
with the semantic motifs of expansion and projection (see Section 3.2.2).
A meaning such as causality (as a subtype of expansion: enhancement) can
be realized between clauses within a clause complex (This happened because
that happened.), within clauses through relational processes (This caused
that.), and within groups (The cause of that was . . .) (see Halliday and Mat-
thiessen 1999:222–6 on this and related types of fractal resonance). Such
resonating lexicogrammatical phenomena can be grouped together into a
topological domain, or a ‘(transcategorical) semantic domain’ (see the
notion of semantic domain in Halliday and Matthiessen 2004:593–4).
The resonance may also be inter-stratal. This kind of parallelism is
explored when the model of lexicogrammar, which is relatively familiar,
is used as an inspiration to tackle the stratum of semantics. This type of
inter-stratal fractality, which ties in well with the notion of metaredun-
dancy between levels of the system, will become clearer in Section 3.5,
which focuses on recent semantic models in SFL.
3.4.3 Extravagance
We have seen how different conceptions of semantics are not just possible
alternative views (see Sections 3.2.1 and 3.2.2), but exist side by side and
moreover, all appear, naturally, in a view of language as a dynamic
open system based on metaredundancy relations between its layers (see
Section 3.3.1). What has now been added, in Section 3.4, is the idea that this
varied view of semantics is ‘fostered’ in SFL. A theory which is based on
complementarities and fractal patterns resonating across its components, is
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
76 MIRIAM TAVERNIERS
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 77
the theory as a whole (see Butler and Taverniers 2008). This is in vein with a
key aim in SFL since its inception, viz. to make a theory that prioritizes
meaning, to set up a system that is as semantic as possible (a ‘semanticky’
grammar, see Halliday and Matthiessen 2014:31), or a grammar that is
‘pushed’ as far as possible towards the semantics (see Halliday 1985:xix),
but the ‘semanticization’ of the model lies more specifically in the fact that
the modelling of the ‘semantic stratum’ has become an important research
aim across different areas and metafunctions.
In the later, more fully-fledged semantic endeavours in SFL, there is
continuity and also a major new research focus. There is continuity from
the initial motivations for separating semantics and lexicogrammar in that
central themes have remained important to the present day: grammatical
metaphor, and the semantics of speech function (as seen in Hasan’s model
of message semantics, see Section 3.5.3). Socio-semantic variation and a
focus on text also remain important. At the same time, it is possible to
recognize an oscillation in a specific approach to semantics, which is inher-
ently in accordance with the overall architecture of SFL, viz. the modelling
of (separate) semantic ‘networks’ which then interact with the lower net-
works in the lexicogrammar. This had already started in earlier work on the
interpersonal metafunction focusing on the relation between speech func-
tion and mood. In more recent models this approach is generalized to other
interpersonal domains (e.g. stance and sourcing of stances), to the idea-
tional metafunction and to the analysis of discourse. Hence, what we see in
the design of the stratum of semantics is fractality, with the method of
system networks being fractally extended from the lexicogrammar to
the semantics (see Matthiessen 2009:14–15).
Beyond this strong ‘systemic’ motif, the other two types of modelling are
also important sources of inspiration, viz. ‘topological’ and ‘discourse-
structural’ designs. As indicated above, those three types of designs are
complementary, and although the ultimate aim may be a systemic model
(as in recent semantic models in SFL), the topological and discourse-
structural views of semantics may be used as a primary source of inspir-
ation to disentangle ‘meanings’ which can then be ‘networked’ in a more
full-blown, systemic semantic model. It turns out that these different types
of designs, topological and discourse-structural, can be seen as very primary
lines of thought for conceptualizing ‘semantics’ (in addition to a third one
that will be introduced below). In other words, they are not only alternative
types of design, but they are also two complementary pre-systemic
approaches to ‘semantics’, and each of them is based in a specific vision
of what a semantics can be.
In all approaches to semantics, the question is, What ‘higher’ meaning
can be recognized, above the lexicogrammatical networks, that forms an
interface between lexicogrammar and context? In a ‘topological’ approach,
this question is explored in terms of different levels of ‘abstraction’. The
‘higher meaning’ is of a different level of coding in two ways: it is not exactly
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
78 MIRIAM TAVERNIERS
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 79
antics
antics
grammar
grammar
s em
s em
co
co
i i
lex lex
abstraction-approach pattern-approach
to semantics to semantics
visually in Figure 3.9. This figure shows how the approaches can be under-
stood as interpretations of the variability relation between a higher mean-
ing and a dispersed realization in the lexicogrammar.
As indicated above, the abstraction-based and pattern-based approaches
have been complementary sources of inspiration in modelling a semantic
stratum, and in setting up networks for this stratum. This has been the case
across the metafunctions, but there are differences in the roles and relations
between these two approaches for the different metafunctions. We will now
turn to specific semantic models in SFL. We will first look at each metafunc-
tion, paying attention to what exactly the abstraction and patterning
approaches have resulted in for those metafunctions, i.e. in setting
up semantic models that flesh out the specific nature of each metafunction.
After that we will turn to a third approach to theorizing a semantics, comple-
mentary to the abstraction-based and pattern-based conceptions, viz. a
register/probability-approach, which has been applied more globally across
the metafunctions.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
80 MIRIAM TAVERNIERS
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 81
semantics. However, a general speech functional model has been part of the
theory for a long time (see an early overview in Halliday 1984) and has been
incorporated into textbooks and presentations of the theory.7 In analyzing
the mapping between speech functions and their realizations in lexico-
grammar, some meaning–form couplings can be seen as the ‘default’ ones
and this leads to the concept of interpersonal grammatical metaphors of
mood (see Section 3.2.3.3). However, in another view, such ‘typicality’ is not
relevant or is at least problematic (in relation to mood as well as modality)
(see Hasan 2010:287).
Speech functions also feature in a pattern-approach to semantics, which
highlights relations between them in sequences, such as pairs consisting of an
initiating and a responding move (as in the traditional concept of adjacency
pairs) and in larger stretches of discourse, especially dialogue. This interper-
sonal dimension which focuses on the scaffolding of interaction is usually
referred to as negotiation or exchange, as in the framework of discourse
semantics (Martin 1992; Eggins and Slade 2005; Martin and Rose 2003).
In addition to speech function and exchange structure, a further inter-
personal area for which a conception of a semantic stratum is essential is
the expression of evaluation, which is modelled in the sub-theory of
appraisal. Appraisal deals with how attitudes and values are conveyed,
how those values are sourced, and how interactants are aligned in relation
to those values (White 2015). One source of inspiration (see White 1999) for
setting up appraisal theory was the concept of interpersonal grammatical
metaphors of modality, which revealed that one interpersonal meaning
such as a modal value of ‘probability’ (see examples in Section 3.2.3.3) can
be realized in various ways in the lexicogrammar (see Figure 3.4). In
appraisal theory, this conception is extended to a range of other types of
interpersonal values in addition to modality, for instance, affective mean-
ings and value judgements, which are realized in dispersed ways, through
interpersonal resources such as forms of modality, through experiential
resources – explicitly or through connotation, and/or through logical
resources (e.g. a conjunction but signalling a concessive meaning of
counter-expectation). Hence appraisal theory is a good example of an
abstraction-based approach to semantics which (re-)organizes various types
of lexicogrammatical means at a higher level in topological areas.
Appraisal meanings can also be looked at from a pattern-approach to
semantics, which then focuses on how attitudes are negotiated in a text and
how different alignments are set up across the text (e.g. Martin and Rose
2003; Martin and White 2005). A specific patterning, i.e. sequential, aspect of
appraisal that is useful to mention is that of ‘semantic prosody’. The combin-
ation of various appraisal resources in a text can form a pattern which sets a
tone or an attitudinal mood which is spread across a stretch of discourse, and
7
For overviews in textbooks, see, for instance, Thompson 1996: Chapter 4; Martin et al. 1997: Chapter 3; Halliday and
Matthiessen 2014: Section 4.1; Matthiessen 1993: Section 5.1.2.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
82 MIRIAM TAVERNIERS
which unfolds with varying degrees of strength, just like musical prosody
(see Martin and White 2005:59). For instance, certain words can trigger
positive or negative connotations in other words that occur in their neigh-
bourhood (see Louw 1993). Note, too, that the featuring of ‘connotation’,
both as a dimension of appraisal (the evocation, rather than explicit con-
strual, of evaluative meanings) and in the concept of semantic prosodies, as
an intrinsic aspect of interpersonal semantics, resonates with a general
conception of the stratum of semantics as a ‘connotative’ layer added to
the ‘denotative’ system of language (i.e. semantics interpreted in terms of
Hjelmslev’s model of a connotative semiotic, see Section 3.3.1 above).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 83
8
It will be noted that in this way, semantics as register bears a fundamental similarity to that component of
interpersonal semantics which models ‘speech functional’ meanings through which social semiotic ‘roles’ are
enacted. This similarity is not surprising, since a register is a procedure for ‘functioning’ in a specific context, i.e. taking
a specific ‘role’ in an institutional setting. The similarity has been noted before in SFL, but it has not been studied
systematically (but rather has been seen as an ‘inconsistency’ or an unresolved issue). See Butler (2003), who also
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
84 MIRIAM TAVERNIERS
antics
grammar
sem
co
i
lex
actualization-approach
to semantics
design probabilistic
refers to Gregory’s (1967) theory of register in which both dimensions pointed out are seen as two different aspects
of the interpersonal component, viz. one which relates to speech functions through which the relation between the
interactants are enacted (‘personal tenor’), and another which deals more broadly with the purpose of the text
(‘functional tenor’). In later work, the ‘purpose of the text’ is not tied to the interpersonal metafunction, but is realized
across the different metafunctions.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 85
9
One problem, pointed out by Hasan (2009), is that when linguists determine what it is, in the context, that activates
particular choices in a text, they are already reasoning from the text (or with an imaginary text in mind), i.e. they always
reason from language. Hence context thus perceived is what Hasan calls ‘relevant context’, and this is different from the
more general non-linguistic ‘eco-social context’ which features in other interpretations of semantics. It will be noted that
this conception of ‘relevant context’ is itself a consequence and an inherent feature of language as a semiotic dynamic
open system with relations of metaredundancy between its strata, hence also between semantics and context-as-seen-
from-semantics.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
86 MIRIAM TAVERNIERS
approach to semantics
t
t
contex
contex
contex
a n t ic s
a n t ic s
a n t ic s
g rammar
g rammar
g rammar
se m
se m
se m
co
co
co
i i i
lex lex lex
te n o
identification e r
od
m
appraisal reference chains
tracking participants
interpersonal metaphor [earlier work on cohesion ideational metaphor
l
na
field
reinterpreted across
so
rper
l
ua
the metafunctions]
al
inte
text
tion
pattern negotiation conjunction
idea
sequencing of
t
exchange structure
contex
periodicity
a n t ic s
+ negotiation of projection & expansion
g rammar
attitudes (hyper-Theme, relations between
se m
macro-Theme) larger units of text
Rhetorical Structure Theory
co
i
lex
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 87
3.6 Summary
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
88 MIRIAM TAVERNIERS
about semantics in SFL, and, (2) on the other hand, to look at various specific
proposals for recognizing a semantics in SFL and to place them against the
theoretical background of possible conceptions of semantics.
We started off in Section 3.2 by looking at where ‘meaning’ is situated in
the model of the system network and stratification in SFL, and by distin-
guishing three related ways in which semantics is designed in SFL. The ‘why’
question led us to the notion of variability relations between semantics and
lexicogrammar, and we saw how grammatical metaphor was an important
initial motivation for recognizing a (separate) stratum of semantics in SFL.
In Section 3.3, conceptions of meaning and semantics in SFL were placed
against a wider theoretical background, in two steps. First the concept of
variability was reconsidered against the background of viewing language as
a semiotic dynamic open system which is characterized by metastability
and by metaredundancy relations between its coding levels. In a second
step, the apparent indeterminate or multi-faceted view of meaning and
semantics in SFL was explained in relation to a design rationale that is
based on multiperspectivism and fractality.
In Section 3.4 we turned to specific semantic models in SFL. In connecting
how exactly semantics is fleshed out in those models to the different
conceptions of semantics which had been disentangled in the previous
sections, three basic approaches to semantics were distinguished, viz.
abstraction, patterning, and actualization. Those were linked to the ways
in which a semantics can be designed (topological, systemic/typological, and
discourse-structural), to the role of variability, and to stratification and
metaredundancy.
References
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 89
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
90 MIRIAM TAVERNIERS
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
Semantics 91
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 15:31:15, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.005
4
The Clause
An Overview of the Lexicogrammar
Margaret Berry
This chapter was to have been written by Geoff Thompson, but sadly he died before he
could write it. I would like what I have to say to be regarded as my tribute to Geoff.
Certainly I shall be drawing deeply on his work, particularly on the third edition of
his Introducing Functional Grammar (Thompson 2014).
4.1 Introduction
I am grateful to Chris Butler, Jeff Wilkinson, and the editors of this volume for comments on the first draft of this chapter.
Of course I alone am responsible for any errors or misrepresentations.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 93
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
94 MARGARET BERRY
material process
behavioural process
mental process
clause
verbal process
relational process
existential process
Figure 4.1 Types of process represented as a system network
mental processes third. There will be room in this chapter only to discuss
the three main types.1
The types of process are presented as options in a system relevant to the
clause (see Figure 4.1). The options in the system are given the names of the
types of process. However, it is important to note that what is really being
chosen is a kind of package deal, a configuration of a process together with
relevant participants. Hopefully this will become clear as I discuss the main
types of process in detail.
1
Halliday and Matthiessen (2014:215) comment that ‘[t]he minor process types appear to vary more across languages
than the major ones’.
2
|| indicates a clause boundary. I am using standard SFL notation, as set out in Halliday and Matthiessen (2014:ix–xi).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 95
(a) Can the clause be rephrased in the form What X did was (to) . . .?
(b) If the process is taking place at the present moment, is the most
natural form in which to refer to it that of the continuous present
(i.e. the be + ing form)?
The processes in the two clauses of (3) pass both these tests. What I did was to
feed my hamster and What I did was to trek downstairs are both perfectly
acceptable. And if we switch the processes into the present, the most natural
way to refer to them would be I am feeding my hamster and I am trekking
downstairs. Other forms would have different implications. For instance,
I feed my hamster and I trek downstairs would seem to refer to habitual pro-
cesses rather than single processes taking place at the present moment.3
On the other hand, the process of ‘knowing’ in (4), from Charlotte’s
health and nutrition essay, fails both tests.
What I did was to know mango juice had a thick consistency sounds very strange;
‘knowing’ is not really a form of ‘doing’. And if we switch the process into
the present, I am knowing mango juice has a thick consistency sounds equally
strange. Much more likely would be I know mango juice has a thick consistency.
The two processes in (3) then are material processes. The ‘knowing’ process
in (4) is not. (Example (4) will be discussed later in Section 4.2.2 on mental
processes.)
3
The point about the continuous present being the most natural way of referring to a single present process is perhaps
truer of some registers than of others. If a register has something of the character of a commentary, the simple present
form can be used to refer to a present process. (2c) and (2d) are from a passage in Charlotte’s horror story that reads
like a commentary on what is going on. In this context, Cold clammy hands grasp my own and Cold clammy hands are
grasping my own would seem to be equally possible ways of referring to the present process, as are A pearly aura
emits from their body and A pearly aura is emitting from their body.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
96 MARGARET BERRY
The first of the probes for material process can also be used as a probe for
the participant ‘Actor’. If the clause can be rephrased in the form What
X did was (to) . . ., where a constituent of the clause naturally replaces X, then
that constituent is representing the Actor. Applying this to (2a) and (2b),
repeated here for ease of reference,
we get What the English did was to rush boldly forward and What the Normans did
was to turn their horses. The English and The Normans are the Actors of their
respective processes.
The probe for ‘Goal’ is an extension of the probe for material process and
Actor. If the clause can be rephrased in the form What X did to Y was (to) . . .,
where a constituent of the clause naturally replaces Y, then that constituent
is representing the Goal of the process. (2b) can be rephrased as What the
Normans did to their horses was turn them. Their horses is the Goal of the process
of ‘turning’. There is no equivalent to Y in (2a). (As we have already seen, in
traditional grammar terms, the process in (2b) is transitive, the process in
(2a) is intransitive.)
Applying this to examples (1a) to (1c), again repeated here for ease of
reference,
we get What I did to the snooze button was press it and What I did to my hamster
was feed it. The snooze button and my hamster are the Goals of their respective
processes. However, there is a problem with (1c): i.e. What I’ll do to the house
is leave it. Here, I is not really doing anything to the house. The snooze button
is presumably in a different position as a result of the pressing. And a fed
hamster is presumably different from an unfed hamster. But there is no
change in the house as a result of the leaving. SFL regards the house as not a
Goal but an instance of another kind of participant which it calls ‘Scope’.
The processes in (1a) and (1b) have Goals and so are transitive. The process
in (1c) is regarded as intransitive, but having Scope. A Scope resembles
a Goal in that typically it occurs immediately after the process and in
that it does not include a preposition. But it represents the domain to
which the process relates, rather than something the process is done
to. Another example of a Scope (a made-up example this time) would be
the hill in (5a).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 97
Both (5a) and (5b) are regarded as intransitive. (5a) is intransitive with
Scope. (5b) is intransitive without Scope – it just has a circumstance.4
I have discussed the probes for material process, Actor, and Goal in detail,
in order to show how probes work. Unfortunately there will not be room in
this chapter to discuss in the same way the probes for the other categories
I am about to cover. For more on probes, see Fontaine (2013:85–91); Bartlett
(2014:48–82).
Moving on to discuss other distinctions within material process, I will quote
from Thompson (2014:96), who says ‘there are many different suggestions for
ways in which [material processes] can be subcategorized at more delicate
levels’. In fact Thompson himself (2014:95–7, 111–14), Berry (1975:151–2,
154–61), and Bartlett (2014:48–58) offer different selections from among the
possible distinctions. The most comprehensive account is that of Halliday and
Matthiessen (2014:224–44, see especially their Figure 5–10, on page 229), but
even that does not cover all the distinctions that have been suggested.
The distinctions I have chosen to present in this chapter are those which
in the past I have found most helpful in distinguishing different registers.
Figure 4.2 shows the distinctions I have discussed here so far and the
distinctions I am going on to discuss.
I have already discussed the transitive (having both Actor and Goal) and
intransitive (having Actor but no Goal) system and the further + scope and -
scope system. The transformative/creative system, the next system down in
Figure 4.2, distinguishes between processes that bring Goals into existence
(creative) and those which do something to existing Goals (transformative).
Thompson’s (2014:96) examples of this are as in (6a) and (6b).
(6a) is creative as the process of making actually brings the puddings into
existence. (6b) is transformative as the pudding exists before the eating of it
(or in this case the not eating of it!). Examples (7a) and (7b) are both from
Charlotte’s health and nutrition essay.
4
For discussion of different types of Scope, see Halliday and Matthiessen (2014:239–42).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
98 MARGARET BERRY
5
As well as discussing the creating or transforming of Goals in transitive clauses, as I just have, Halliday and Matthiessen
(2014:231) discuss the creation of Actors in intransitive clauses. This is why I have followed them in presenting the
transformative/creative system as simultaneous with the transitive/intransitive system, instead of presenting the
transformative/creative system as dependent on the selection of transitive.
6
Halliday and Matthiessen may want to regard example (9b) as a relational clause – see what they say, for example,
(2014:265) about These plates went from the head to the tail. However, for me (9b) passes the tests for material
processes. We can say What this point does is link to my next point and This point is linking to my next point, though
these forms are perhaps unlikely in the register of Charlotte’s history essay.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 99
The first paragraph of Charlotte’s horror story includes five material pro-
cesses. These all have inanimate Actors. And material process clauses with
inanimate Actors continue to feature in later paragraphs.
We cannot really say either What I did was to know mango juice had a thick
consistency or I am knowing mango juice has a thick consistency. ‘Knowing’ is not
a form of ‘doing’. And, if we switch the process into the present, the most
likely form for it would be the simple present – I know mango juice has a thick
consistency – not the continuous present which is most usual for material
processes.
The tests for material processes can thus be negatively used for mental
processes. Material processes pass the tests, mental processes fail them. (For
more on probes for mental processes, see Bartlett 2014:64–5; Fontaine
2013:87.)8
SFL recognizes four different types of mental process, as shown in
Figure 4.3.9
7
I am here disagreeing with Halliday and Matthiessen (2014:250), who say that in material clauses ‘the distinction
between conscious and non-conscious beings simply plays no part’. It is true that, in the class of material processes
viewed as a whole, Actors may be either conscious (animate) or non-conscious (inanimate), but this is not true of all
individual members of the class. Some material processes will normally have an animate Actor (e.g. trek in example
(3)), while others will normally have an inanimate Actor (e.g. flicker in example (10d)). There seems to be a cline of
material processes from this point of view. Examples with untypical animacy are usually viewed as metaphorical. For
discussion, see Berry (1975:151–2, 155).
8
Also see below. A mental process can take a clause as phenomenon. A material process cannot take a clause as goal.
9
For a more detailed system network showing more delicate choices within mental processes, see Halliday and
Matthiessen (2014:258).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
100 MARGARET BERRY
(13a) I thought the mango juice would counteract the tart flavour of the
raspberries
(13b) I thought it would make an attractive red
(13c) I have concluded that the Norman tactics are the most important
reason for the Norman victory10
10
It could be said that conclude passes the tests for material process. However, although there has not been room to
discuss it fully here, a further difference between mental and material processes is that mental processes can take a
clause as complement, while material processes cannot (see footnote 9). Conclude can take a clause, as in example
(13c).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 101
Mental processes will have a participant whose function is the sensing. SFL
calls this the ‘Senser’. Typically the Senser will be animate, indeed human.
All the examples I have so far given of mental processes have human
Sensers – the I’s and you’s. When a mental process occurs with an inani-
mate Senser, this is seen as an example of personification. Example (15) is
from Thompson (2014:98).
The thing that is sensed is called the ‘Phenomenon’. This may be a simple
Phenomenon, realized by a nominal group, such as any trouble in (14b), a
cigarette in (14c), cold weather in (15). Or it may be more complex, involving
another clause which itself contains another process. (11c), repeated here,
has the Phenomenon the parents spectating on the sidelines.
The Phenomenon of see includes another process; what is seen is the parents
spectating on the sidelines. Halliday and Matthiessen (2014:256) call this
type of complex phenomenon a ‘macrophenomenon’. In a macrophenome-
non the second process will be realized by a non-finite verb, usually the
‘-ing’ form.
Another type of complex phenomenon, called by Halliday and Matthiessen
a ‘metaphenomenon’, is found in (13a) to (13c) and (14a), again repeated here.
(13a) I thought the mango juice would counteract the tart flavour of the
raspberries
(13b) I thought it would make an attractive red
(13c) I have concluded that the Norman tactics are the most important
reason for the Norman victory
(14a) I hope my dream will come true
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
102 MARGARET BERRY
All four main types of mental process – perceptive, emotive, cognitive, and
desiderative – can take simple Phenomena, but they vary in the types of
complex Phenomena with which they are associated. Macrophenomena
occur with perceptive and emotive processes, but not with cognitive or
desiderative processes. All four types can take metaphenomena, but per-
ceptive and emotive processes typically take ‘fact’ type, while cognitive and
desiderative processes typically take ‘idea’ type. For more on the possibil-
ities for the various types of mental process, see Halliday and Matthiessen
(2014:256); Bartlett (2014:64–5).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 103
Everything, I, The snooze button, and We. The Attributes are hazy and unclear, a
ghost, a saviour, and big breakfast people. The Attribute may be adjectival, as in
(17a), or a nominal group, usually indefinite, as in (17b) to (17d).
Examples of identifying relational processes are (18a), from Charlotte’s
day in her life, (18b) and (18c), from her horror story, and (18d) and (18e),
from her history essay.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
104 MARGARET BERRY
process clauses are not. It would be perfectly possible to say The most
important meal of the day is breakfast or My eyes are the only reminder of my
original features, instead of the versions in (18a) and (18b), and still sound
naturally idiomatic. While it is just about possible to reverse some attribu-
tive relational processes – e.g. Big breakfast people are we – the result would
sound marked and highly rhetorical.
There is a great deal more that could, and should, be said about relational
processes, but space does not permit. I will close this subsection simply by
saying that I think Thompson (2014:101–5, 122–7) is particularly helpful on
relational processes. For more detailed discussion, see Halliday and Mat-
thiessen (2014:259–300). Figure 4.4 shows the one choice within relational
processes that I have had room to discuss. There are many more. See Halli-
day and Matthiessen (2014:264).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 105
which have to do with the way in which we interact with other people – and
that these interpersonal choices also have an effect on the structure of the
clause.
Halliday (2002:203) says of the experiential options that they will tend to
be realized particulately – that is, a structure that represents experiential
meanings ‘will tend to have this form: it will be a configuration, or constel-
lation, of discrete elements’. We have already seen that the experiential
choices lead to the basic constituents of a clause, their number, and their
nature.
Halliday (2002:205) says of interpersonal meanings that they will tend to
be realized prosodically – that is, the interpersonal meaning is ‘strung
throughout the clause as a continuous motif or colouring’. This section will
attempt to show what Halliday means by this.
4.3.1 MOOD
SFL recognizes two main kinds of clause rank interpersonal choice, which
are usually discussed under the headings of MOOD and MODALITY. MOOD
choices will be discussed in this subsection, MODALITY choices in the
following subsection.
MOOD choices have to do with the forms we use when indicating the
kind of interaction in which we are engaged; e.g. whether we are making
statements or asking questions or giving commands. The main choices are
shown in Figure 4.5.11 Only independent clauses have access to these
choices, not subordinate clauses.
Concepts such as ‘declarative’, ‘interrogative’, and ‘imperative’ are of
course familiar in most approaches to the study of language. SFL’s particu-
lar take on them is to present them as options from which choices can be
made, and to relate them both to the semantics in the stratum above and to
the lexicogrammatical structure of the clause.
Relating them to the semantics is not a simple matter. While it is true
that in the unmarked cases questions are realized by interrogatives,
11
For a more detailed version of this network, see Thompson (2014:60). For an even more detailed version, see
Halliday and Matthiessen (2014:162).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
106 MARGARET BERRY
(20) I pressed the snooze button twice more, but eventually clambered
out of the warmth of my bed. A typical school morning.
The first clause has the Predicator pressed and the second clause the Predi-
cator clambered. But in the third clause, a minor clause, there is no explicit
process. Charlotte uses minor clauses in her horror story and in the day in
her life, but in her other writings minor clauses appear only as headings of
sections and subsections.
The choice between ‘indicative’ and ‘imperative’ is realized by the pres-
ence or absence of a Subject. The first clause in (20) has the Subject I and is
indicative. The second clause would also be regarded as indicative.
Although the Subject does not actually appear in that clause, it is recover-
able from the previous clause. On the other hand, the main clause in (21a),
again from Charlotte’s day in her life, is imperative. There is no Subject of
Imagine. Similarly (21b), from Thompson (2014:58), is imperative. There is
no Subject of Answer.12
12
It is possible for imperatives to be accompanied by things that look like Subjects. Thompson (2014:59) gives the
example You listen to me, young man. However, Thompson argues that these are not ‘normal’ Subjects. The question
of what is meant by ‘Subject’ in SFL is a complex one. For varying views, see Halliday and Matthiessen
(2014:147–50); Fawcett (1999); Fontaine (2013:109–15).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 107
All three clauses are major clauses; they all have Predicators. And all three
are indicative; they all have Subjects. But in (22b) and (22c) the Predicators
are split between a Finite part and a main Lexical part: shall . . . tell and
will . . . cry. In (22a) the Predicator is not split – sparkle is itself a finite verb –
the Finite part and the main Lexical part have been conflated.
To return to the realization of the choice between ‘declarative’ and
‘interrogative’, in a declarative clause the Subject precedes the Finite part
of the Predicator, but in an interrogative clause the Subject follows
the Finite part. (22a) and (22b) are both declarative while (22c) is
interrogative.
This is in fact my working definition of the Subject, that it is the constitu-
ent in indicative clauses that inverts with the Finite verb to show whether
the clause is declarative or interrogative.13 Where there is already an
auxiliary verb to carry the finiteness, it is simply a matter of inversion.
We can turn the declarative (22b) into an interrogative simply by inverting
the I and the shall – Shall I now tell you my story. We can turn the interrogative
(22c) into a declarative simply by inverting the Will and my mother – My
mother will cry at the sight of me. Where there is not already an auxiliary
to turn a declarative into an interrogative, it is necessary to import one.
To turn (22a) into an interrogative, we would need to import do – Do my
eyes sparkle with tears. (For more on Subject and Finite, see Fontaine
2013:110–20.)
What I have just said is truer of yes/no interrogatives than of
wh-interrogatives. While it is the case that wh-interrogatives usually
involve inversion of Subject and Finite verb, this is not always so. What
really realizes a wh-interrogative is of course the presence of a wh-word –
who, what, where, when, why, how. When the wh-word is itself the Subject,
there is no inversion.
As well as the main MOOD options already discussed, there are other
possibilities. For instance, tag phrases may be used. A clause with declara-
tive form may be given an interrogative tag – e.g. My eyes sparkle with tears,
don’t they or I’ll now tell you my story, shall I. An imperative may be ‘softened’
by a tag – Imagine, will you, that you are standing on a race track or Imagine you
are standing on a race track, will you.
As usual there is much more that could be said. But hopefully I have said
enough to show what Halliday means by saying that the realizations of
13
I am here taking a more syntactic, less semantic view of the Subject than is probably usual in SFL these days. In my
view, we still need a syntactic perspective if we are to account for the ways in which the MOOD options are realized.
But see the references in footnote 12.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
108 MARGARET BERRY
4.3.2 MODALITY
‘Modality’ is the name which SFL gives to the whole area of clause rank
speaker/writer assessment of what is being said. If a clause is communi-
cating information, the assessment may be of the probability that the
information is true. In the made-up examples (23a) to (23c) different assess-
ments are made of the degree of probability. (The assumption is that all of
them are spoken on hearing someone arriving.)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 109
However, she does include a few assessments of probability and one assess-
ment of usuality. (28a) and (28b) are from the horror story, (28c) to (28e)
from the day in her life.
There is a further modality choice which cuts across the ones discussed so
far (see Figure 4.6).14
Cutting across the choice between modalization and modulation is a
choice among low, median, and high. This has to do with the strength of
the assessment. In examples (23a) to (23c), repeated below, (23a) represents
14
For a more detailed version of this network, see Thompson (2014:77). For an even more detailed version, see
Halliday and Matthiessen (2014:185).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
110 MARGARET BERRY
Where there is a modal Adjunct, such as possibly or perhaps, this may be at the
beginning of the clause, as in (29b), or in the middle as in (29a) and (29c), or at
the end as in (29d). Hopefully this can be regarded as another illustration of
what Halliday means by saying that the realizations of interpersonal mean-
ings are ‘strung throughout the clause’.15 Bartlett (2014:113) provides a
helpful table summarizing modality options and showing their realizations.
15
Halliday and Matthiessen (2014:183–4) say ‘when we move around the languages of the world, we find a great deal
of variation in the grammaticalization of modality and other types of interpersonal judgement’. They go on to discuss
some of the differences.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 111
16
Textual meaning is sometimes regarded as ‘second-order’ meaning, since it is about the arrangement and relative
prominence of bits of experiential meaning and interpersonal meaning. For discussion, see Matthiessen (1992).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
112 MARGARET BERRY
were the main point, and the new point in that paragraph, of what I was
saying, and I wanted them to have end focus.
(30a) This means that, like Child C, Writer 2 in both his passages has clear
methods of development
(30b) This means that, like Child C, Writer 2 has clear methods of devel-
opment in both his passages
17
For different definitions of ‘Given’ and ‘New’, see Berry (forthcoming).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 113
Halliday and Matthiessen (2014:89) say that the Theme is ‘that which
locates and orients the clause within its context’. The orienting may take
a number of different forms. For instance, the Theme, the first experiential
constituent in the clause, may be indicating that the text is staying with a
topic entity already mentioned, as in (31b). Or it may indicate a moving on
in time, as in (32a). Or it may simply be setting the scene, as in (32b).18
We have already seen that first position in the clause plays an important
part in English in realizing some of the interpersonal options. For a yes/no
interrogative, the finite verb will usually be in first position; for a wh-
interrogative, it will usually be the wh-word; and for an imperative, the
lexical verb may be in first place.
In a declarative clause, the most usual thing to happen will be that the
Subject will be at the beginning. When this happens, SFL calls it ‘Unmarked
Theme’. The first three clauses of the second paragraph of Charlotte’s
horror story all have Unmarked Themes.
(33a) I am a ghost
(33b) I wear a long white dress
(33c) My hair flows down my back in silvery strands
We then get the clauses about her eyes, quoted above, and then in the rest
of the paragraph the main clauses all have I Subjects, these being in first
experiential place and so being Unmarked Theme. This is the paragraph in
which Charlotte establishes the persona she has adopted for the story. The
repeated I’s as Unmarked Theme keep this persona firmly in view.19
Subject in first position then is the unmarked order for a declarative
clause. But other elements of the clause may precede the Subject for special
effects, in which case they are regarded as ‘Marked Themes’. Perhaps the
most common kind of Marked Theme is where an Adjunct, a constituent
representing a transitivity Circumstance, precedes the Subject. Examples
(34a) to (34d) are all from the day in Charlotte’s life.
18
Opinions differ in SFL as to how much of a clause should be regarded as the beginning. Is the Theme just the very first
experiential constituent, or should it include any other experiential constituent near the beginning that has an orienting
function? I have said that They in (31b) has an orienting function in that it shows the text is staying with a topic entity
already mentioned. But I in (32a) and it in (32b) also have this function. Should they not also be included in the
Themes of their respective clauses? If one takes the notion of orienting function seriously, there are grounds for saying
that the Theme extends right up to the main lexical verb. Indeed in a language such as Spanish the Theme may even
include the lexical verb as orienting functions carried by early parts of clauses in English are carried by verbal inflections
in Spanish. For discussion, see Berry (1996).
19
Fries (1981) links Theme with what he calls the ‘method of development’ of a text. The pattern of the experiential
constituents selected to be Theme shows how the subject matter of the text is being organized and developed.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
114 MARGARET BERRY
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 115
(38) I don’t know about you but, controversially in the evening when I’m going
to bed, I don’t feel tired
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
116 MARGARET BERRY
For a more detailed network for Theme, see Thompson (2014:170). For a
much more detailed network, see Halliday and Matthiessen (2014:106).
References
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
The Clause 117
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 23 Aug 2019 at 07:38:48, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.006
5
The Rooms of the House
Grammar at Group Rank
5.1 Introduction
We would like to thank Margaret Berry for her very useful comments on drafts of this chapter.
1
From the film Animal Crackers (1930).
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 119
the following terms: groups form around a particular type of word, for
example, a noun forms a group called the nominal group, such as the
beautiful scenery, where the other words in the group somehow modify the
noun, in this case scenery. A phrase, in contrast, is less like a relationship
among words and rather combines two components through complemen-
tation, as in the prepositional phrase by the lake, where the lake is seen as a
complement to the preposition by.
In this chapter we explore these two types of intermediary unit in SFL. We
take as given that there are units larger than the word and smaller than the
clause. There is generally considerable agreement in terms of the descrip-
tions of these intermediary units (see McDonald 2017 for an excellent over-
view of the historical development of groups). However, there is also room
for debate. The aim of this chapter is to critically examine the theoretical
reasons for including two fundamentally different types of grammatical unit
between clause and word. The main question we ask is whether these
reasons hold in all cases, i.e. is it theoretically justified to maintain two types.
The way we will approach this is as follows. In the next section, we provide
an overview of the existing description of the units below the clause as
currently represented in the theory. Following this, Section 5.3 will present
and evaluate three main criteria for classifying a unit as either a group or a
phrase. These distinctions, drawn primarily from Matthiessen (1995) and
Halliday and Matthiessen (2014), include (i) the concept of (primary) class
and the relation between the functional potential of the unit and the ‘head
word’; (ii) univariate versus multivariate structures; and (iii) the role of
rankshifted units. In considering these perspectives on the units at the
intermediate rank between words and clauses, we conclude that there is no
significant theoretical or practical value in maintaining two different types of
unit at this level. We argue that it is important to ask questions such as those
we propose here in order to evaluate the strength of the position of the
theory and its usefulness in an appliable theory of language.
The category set up to account for the stretches that carry grammatical
patterns is the ‘unit’. The units of grammar form a hierarchy that is a
taxonomy. . . . The relation among the units, then, is that, going from top
(largest) to bottom (smallest), each ‘consists of’ one, or of more than one, of
the unit next below (next smaller). The scale on which the units are in fact
ranged in the theory needs a name, and may be called ‘rank’.
(Halliday 1961:251)
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
120 LISE FONTAINE AND DAVID SCHÖNTHAL
Clause unit The brown foxes were jumping over the laziest dog
Group unit [The brown foxes] [were jumping] [over the laziest dog]
Word unit [The / brown / foxes] [were / jumping] [over / the / laziest / dog]
(1) the county plans to spread the money among a mixture of govern-
ment securities and bank investments (EnTenTen13, SketchEngine)
(2) I once worked in an investment bank (EnTenTen13, SketchEngine)
2
As introduced later in this chapter, we use the term ‘group unit’ to refer to the units at group rank, i.e. both groups and
phrases alike. Further, for a full account of how rank relates to the key SFL concepts of stratum, delicacy, and realization,
see Berry (2017).
3
For details of SketchEngine, see Kilgarriff et al. (2014) or online: www.sketchengine.co.uk.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 121
For English, for the two units between sentence and word the terms ‘clause’
and ‘phrase’ are generally used. It is at the rank of the phrase that there is most
confusion – because there are here the greatest difficulties – in the description
of English; one reason is that in English this unit carries a fundamental ‘class’
division, so fundamental that it is useful to have two names for this unit in
order to be able to talk about it: I propose to call it the ‘group’, but to make a
class distinction within it between ‘group’ and ‘phrase’.
(Halliday 1961:252–3)
Thus, Halliday proposes two types of unit, group and phrase, at the same
rank on the scale, effectively having two different types of unit with no
difference in rank. We will use the term ‘group unit’, as in Table 5.1, as an
umbrella term for units at this rank. This avoids having to say ‘groups and
phrases’ which is not only lengthy, but there is only one phrase and it
seems reasonable to have a single term for each rank along the scale. In this
paper, we have opted for the term ‘group’ rather than ‘phrase’ simply
because it is by far the more common term in Systemic Functional Linguis-
tics and as we will argue below can be applied to all intermediary units. The
difference between groups and phrases is described by Halliday and Mat-
thiessen (2014:262–3) as follows:
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
122 LISE FONTAINE AND DAVID SCHÖNTHAL
Figure 5.1 Word classes in SFL (adapted from Halliday and Matthiessen (2014:75))
these intermediary units, although this is from a slightly different SFL theor-
etical approach, the Cardiff Grammar (see Schulz and Fontaine, this volume).
There is another difference between group and phrase that must be covered
briefly before moving on to the description of each class of unit at the group
unit rank. This concerns the metafunctional nature of group units. In SFL, the
approach to lexicogrammatical description at the group unit rank parallels
the description of the clause in terms of the three main metafunctions. There
is an assumption that all group units express, at least to some extent, experi-
ential, interpersonal, and textual meaning. Halliday and Matthiessen
(2014:361) explain that ‘[a]lthough we can still recognize the same three
components, they are not represented in the form of separate whole struc-
tures, but rather as partial contributions to a single structural line’.4 When it
comes to logical meanings, however, we find that phrases, unlike groups, are
said to not express logical meanings. This metafunction explains one of the
key differences between groups and phrases. While groups have both experi-
ential and logical structure (Halliday and Matthiessen 2014:361–2), phrases
do not express logical structure (Halliday and Matthiessen 2014:425). This is a
distinction we will return to in Section 5.3 so that we can critically examine
some of the implicit assumptions that underlie this claim.
One final distinction must be discussed in relation to the description of units
in SFL. We mentioned above that within SFL, units are defined in terms of the
function they serve to express in the unit above, rather than by structural
similarities that groups might serve. This is what motivates Halliday’s classifi-
cation of units, and it explains why, for example, adverbials are classed differ-
ently than adjectives, but also why nominal groups and prepositional phrases
are seen as different classes in English, whereas in other languages they might
not be. The three primary classes of word are shown in Figure 5.1, where each
primary class corresponds to one of the ‘three main classes of group: nominal
group, verbal group and adverbial group’ (Halliday and Matthiessen 2014:362,
emphases in original). These classes are defined by the functions served in the
clause. For example, there is a tendency for nominal groups to function as
Subject (or Complement) and/or Actor (or Goal), whereas adverbial groups serve
4
A detailed account of the metafunctions at clause rank is given in Berry (this volume).
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 123
A class is always defined with reference to the structure of the unit next
above, and structure with reference to classes of the unit next below. A class
is not a grouping of members of a given unit which are alike in their own
structure. In other words, by reference to the rank scale, classes are derived
‘from above’ (or ‘downwards’) and not ‘from below’ (or ‘upwards’).
In what follows, each group and phrase will be described in turn. Following
this, in Section 5.3, we will examine the criteria for the distinction made
here between the unit of group and phrase.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
124 LISE FONTAINE AND DAVID SCHÖNTHAL
5
The original expression included only a as a Deictic: in order to include a Numerative, two, the Deictic was changed to a
plural one, those, and stop was made plural as well. Finding a naturally occurring example with all elements included
was very difficult.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 125
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
126 LISE FONTAINE AND DAVID SCHÖNTHAL
Table 5.5 Analysis of the children in blue hats (adapted from Halliday and
Matthiessen 2014:383)
Nominal group
the children in blue hats
Logical structure Premodifier Head Postmodifier
Experiential structure Deictic Thing Qualifier
Prepositional Phrase
Process Range
Nominal group
Logical structure Premodifier Head
Experiential structure Epithet Thing
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 127
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
128 LISE FONTAINE AND DAVID SCHÖNTHAL
6
Although this depends on the lexical representation assumed. If these are single lexemes, i.e. multi-word expressions
constituting effectively a single ‘word’, then their morphological composition may be more complex in terms of the
formation of the item, but nevertheless only a single item in the lexicon.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 129
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
130 LISE FONTAINE AND DAVID SCHÖNTHAL
Table 5.9 Epithet headed nominal group (adapted from Halliday and
Matthiessen 2014:391)
You ‘re very lucky
Carrier Process: relational Attribute
Nominal Group Verbal Group Nominal Group
Premodifier Head
Epithet
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 131
While these two analyses illustrate the group unit structure of two basic
examples, other more complex examples would give rise to further com-
plex issues in the description of these units. Unfortunately, it is beyond the
scope of this chapter to deal with these here, but see Fontaine (2013) for a
more detailed discussion of the structure of group units.
7
These criteria do not address the criterion related to the function the unit serves in the unit above, but see Halliday and
Matthiessen (2014:362–4). We thank Margaret Berry for this point.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
132 LISE FONTAINE AND DAVID SCHÖNTHAL
If we recall the word classification discussed above in Figure 5.1, the Pgp
(e.g. right behind, not without, all along) should, following the principle of class
membership, be made up of a group of words of the same primary class, but
if prepositions are in the verbal primary class and adverbs are in the
adverbial primary class, then the Pgp cannot be said to form a group.
Indeed, there is very little evidence for either a prepositional group or a
conjunction group, since they could be treated as word complexes at the
word rank.
There is a potential problem with these ‘compositional hierarchies’ (Hal-
liday and Matthiessen 2014:22). As McDonald (2017:252) explains, ‘Because
the conceptualisation of classes of word is dependent on their function in
classes of group, which are in turn dependent on their function in the
clause, only those word classes that commonly function as the Heads of
groups are “assigned” to a discrete group class.’ This is precisely why the
adjective class is a sub-class of nominal (see above). It is not entirely unrea-
sonable to class the adjective as a type of nominal, since its function is
always in relation to some noun, and its function is far more directly
related to expressing participant meaning in the clause than process mean-
ing or circumstance meaning. The apparent similarities between adjectives
and adverbs are considered as structural (i.e. morphological) rather than
functional, although see Tucker (1998). Recall that unit is not defined by
similarity of structure: see Berry (1975); Fawcett (2010) for a discussion of
the differences between class and type descriptions.
It is difficult to consider prepositions as a sub-class of verb other than in
the sense that they profile a relation (see Langacker 2016). The reason for
this is that in terms of class, its function in the unit above, i.e. in the PP, and
thereby in the clause, is not to express a process but rather to express a
circumstance. In fact the preposition rarely has a function in the clause (see
example (9) below and also Fontaine 2017b). In terms of class, the prepos-
itional units could belong to the sub-class of adverbials rather than verbals.
Berry (1975:76–7) maintains a useful distinction between class (defined by
function) and type (defined by structure), which allows, for example, clarity
when discussing the class of a unit (e.g. Advgp) vs. the type of formal item
(e.g. p, preposition, or c, completive). See also Fawcett (1980, 2010), who also
maintains such a distinction in terms of functional elements and structural
units and items (also see Schulz and Fontaine, this volume).
For the primary class of nominals, we do find within the Ngp word classes
identified as nominal (e.g. determiner, numeral, adjective, and noun as
shown in Figure 5.1), but also adverbial as in example (9), where the Ngp
is in bold.8
8
Items such as ‘outside’ that have no complement are not treated as prepositions in SFL but rather as adverbials. This is
somewhat problematic, but see Fontaine (2017b) for a discussion of such prepositional items in SFL.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 133
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
134 LISE FONTAINE AND DAVID SCHÖNTHAL
structure, although, as we shall see, this split is often blurred, which brings
into question the distinction between group and phrase units.
We can compare univariate and multivariate structures following Mat-
thiessen et al. (2010:145, emphasis in original), who describe univariate
structures as structures which are ‘generated as an iteration of the same
functional relationship’. Indeed, Matthiessen (1995:627) states that ‘[groups]
have a univariate structure with a Head, which is the only obligatory elem-
ent, and optional Modifiers’. This is a principle that is generally accepted
throughout the SFL literature but has never really been examined critically,
except perhaps in Fawcett (2010). Two examples of univariate structures are
given with the nominal word complex in (10) and the Vgp in (11).
(10) investment trust cash management account
(11) will have been eaten
In example (10), we find a series of five nouns, and in example (11) a series
of four verbs. This type of structure is therefore seen as serial and recursive,
involving an interdependency which can be accounted for through the
logical metafunction in terms of logical relations or modification (see
Martin 1996). However, the example in (10) is not a complete Ngp: it is
missing something. Even if we consider this as a word complex, as the Head
element of the Ngp, it is not the only obligatory element. It must be
grounded as an instance (see Langacker 2016).
Actual instances of Ngps with similar word complexes are given in (12)
and (13). In both cases, it should be clear that the rankshifting of the person’s
and with a 3 per cent annual rate of return prevents a description of these
expressions as word complexes.
(12) the person’s income management account (EnTenTen13, SketchEngine)
(13) a cash management account with a 3 per cent annual rate of return (EnTen-
Ten13, SketchEngine)
As mentioned above, multivariate structures are said to differ from univari-
ate ones in that the structure is ‘a configuration of elements each having a
distinct function with respect to the whole’ (Matthiessen et al. 2010:145,
emphasis in original). This configuration relates to constituency as discussed
above in the sense that there is a kind of generic structure involved which is
configurational and non-recursive. This type of structure is accounted for by
the experiential metafunction. In this sense, we can describe the Ngp as a
configuration of the following elements: Deictic + Numerative + Epithet +
Classifier + Thing. Halliday and Matthiessen (2014:437, emphasis in original)
describe multivariate structure as follows:
Groups have developed their own multivariate constituent structures with
functional configurations . . . Here the elements are (i) distinct in function,
(ii) realized by distinct classes, and (iii) more or less fixed in sequence.
A configuration of such a kind has to be represented as a multivariate
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 135
structure. Treating the group simply as a ‘word complex’ does not account
for all these various aspects of its meaning. It is for this reason that we
recognize the group as a distinct rank in the grammar.
It is often argued that the PP is the only unit to have only a multivariate
structure, i.e. no univariate structure. However, this is also true of the clause;
it is a multivariate structure. In this section, we will consider the criteria of
variate structure in terms of how it applies to the group units described above.
Halliday and Matthiessen (2014:390) maintain that the Ngp is a ‘multi-
variate structure: a configuration of elements each having a distinct func-
tion with respect to the whole’, which is illustrated in Table 5.12. When
compared to the Vgp, as shown in Table 5.13, it seems that the experiential
configuration includes elements with very little distinction in terms of
class, and that the arrangement appears more similar to the iterative
structure of example (10). Quiroz (2017:304–5) points out that ‘a univariate
structural interpretation does not map onto the multivariate organisation
of the verbal group in any self-evident way’. Fontaine (2017a) has argued
against a univariate analysis of the Ngp, although according to different
criteria. If the Ngp is more multivariate in nature, the Vgp seems more
univariate. Halliday and Matthiessen (2014:398) explain that ‘the verbal
group is also structured logically, but in a way that is quite different from,
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
136 LISE FONTAINE AND DAVID SCHÖNTHAL
and has no parallel in, the nominal group. The logical structure of the
verbal group realizes the system of tense.’ The nature of this unit is not
clear and open to debate.
Furthermore, with the exception of the Vgp, all group units have the
potential to include elements that are rankshifted. These rankshifted elem-
ents are said to have no logical relationship with the Head element of the
unit, e.g. the Qualifier without sugar in the Ngp a cake without sugar does not
have a functional relationship to the Head cake. It is this feature which casts
some doubt on the univariate structure of groups. Halliday and Matthiessen
(2014:425) take the position that ‘prepositional phrases are phrases, not
groups; they have no logical structure as Head and Modifier, and cannot be
reduced to a single element’. We can conclude from this that phrases do not
have univariate structure (although there is only one such unit at group
rank). Given that most groups also have this as potential (except, as already
stated, the Vgp), we would have to consider whether we need a nominal
group for nominal units without rankshifted elements and a phrase for one
that do, i.e. nominal group and nominal phrase, as is the case for prepos-
itions (i.e. prepositional group and prepositional phrase). A similar case
could be made for adverbial units at group rank, for example, comparative
adverbs such as more quickly than he can, which is more phrase-like than
group-like. However, it is much easier to take the position that any unit
with the potential for rankshifted elements has a more configurational
nature to it and therefore has no univariate structure. Doing so, however,
has important theoretical consequences, since it would suggest that, by
definition, the Vgp is a group, and that all other units are in fact phrases.
9
The role of rankshifting is debatable here but without a clear status of lexical representation within SFL theory, along
with a robust debate about the rank scale, e.g. ‘be willing/keen/eager to do; be afraid/scared to do’ (Halliday and
Matthiessen 2014:427 note 17), but see Tucker (1998). Further discussion of this is beyond the scope of this chapter.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 137
In theory, all group units adhere to the first principle, which situates the
primary classes of unit in terms of constituency. Principles (b) and (d) will
be discussed below, whereas (c) and (e) will not, since the potential to form a
unit complex (e.g. through coordination) and the potential to interrupt a
unit have no significant bearing on the nature of the class of unit. As we
will show, however, the requirement for a group or phrase to consist of the
rank below (principle (b)) and the potential for rankshifting (principle (d))
are relevant to the classification of intermediary units.
In terms of constituency, principle (b) states that a given unit must
contain at least one unit of the rank below it. This principle holds for all
group units, and it is difficult to ignore the role that word class (as per
Figure 5.1) plays in determining the nature of the unit at this rank, even
though originally class is meant to be related to the function served in the
unit above. However, as we have seen for the Ngp, there are some difficul-
ties as concerns Ngps that have an adjective as Head. If, for example, we
were to eliminate the logical structure at the group unit rank (Fontaine
2017a) and to uncouple adjectives from the nominal word class, then we
would find that the Ngp adheres in a more parallel way to the other
primary classes. In other words, the current anomaly in the verbal primary
class, which includes both verbs and prepositions, connected respectively to
a Vgp and a PP, would be addressed in the nominal primary class by
including a noun (nominal) group, an adjective group, a determiner group,
and a numeral group. This is perhaps unnecessarily complex, and it is not
suggested here as a position that should be adopted. However, this raises
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
138 LISE FONTAINE AND DAVID SCHÖNTHAL
important questions. The point being made here is that this is an area that
requires attention and development in the theory.
Principle (d) states that rankshifted units can serve as postmodifiers in
Ngps and Advgps (although not in verbal ones), but that they are not an
obligatory part of the structure as they are in PPs. However, as we have
seen above, in the case of Advgps with comparative adverbs, the rankshifted
postmodifier is obligatory, as in the PP. If we accept an adjective group as
Tucker (1998, 2017) does, this also applies to comparative adjectives (e.g.
kinder than most people).
This distinguishes Vgps from all other units at group rank, since they
cannot have a postmodifier and therefore the potential is not there.
Another distinguishing feature relates to the fact that Vgps cannot be
rankshifted. As Matthiessen (1995:715) points out, ‘Unlike nominal groups,
adverbial groups and prepositional phrases, verbal groups serve a single set
of functions in the clause and they cannot be rankshifted.’
Hence, these two principles of the rank scale lead us to ask whether there
is any need to maintain a distinction between phrase and group at group
rank, and whether the Vgp belongs as a unit at this rank. If we adhere to
the rank scale principles, we find that many of the assumptions about
groups and phrases are challenged, and the picture that emerges is quite
different from what might have been expected.
Having completed the analysis of the three principle differences between
groups and phrases, we will now consider their function in the unit above
and the unit below (or same rank) in order to get a full sense of groups and
phrases as a unit of rank.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 139
Ngp No, not all are No, only Yes No, if you assume phoricity
nominal multivariate must be indicated (e.g.
*boy is nice vs. boys are
nice, the boy is nice), i.e.
it must be a grounded
instance of a type (see
Langacker 2016).
Vgp Yes, all are verbal Yes, both, but No rankshifting Can only be reduced to one
this is highly potential element in non-finite
debatable clauses, otherwise both
the Finite and Event
elements are necessary.
Advgp Yes, all are Both for simple No, for comparative Yes
adverbial adverbials, adverbials,
but only rankshifting is
multivariate obligatory
for
comparatives
PP Yes, all are ‘verbal’ No, only This depends on No, cannot be reduced to a
multivariate assumptions about single element. In SFL
prepositions, but intransitive prepositions
generally are considered adverbial.
rankshifting is
obligatory
Adjgp (if we No, if adjectives No, only No, for comparative Yes
accept are classed as multivariate adjectives,
that there nominals, but rankshifting is
is a need yes, if classed obligatory
for this as adverbials
unit)
the unit can be reduced to a single element, i.e. a single word. As the tables
show, it is very difficult to clearly distinguish between phrases and groups.
Many of the features that are meant to account for PPs also apply to groups,
with the exception of the Vgp. In fact, when all units and criteria are con-
sidered, the Vgp is the only unit to stand out from the others so distinctly.
This chapter set out to consider the nature of units at the rank between
clause and word (i.e. the group rank) and to examine the theoretical reasons
for distinguishing two different units at this rank. What this has revealed is
that depending on the criteria used, comparing units can reveal a variety of
outcomes. There are very clearly different types of unit at the group rank,
but whether the nature of the units is substantially different remains a
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
140 LISE FONTAINE AND DAVID SCHÖNTHAL
References
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
The Rooms of the House 141
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:33:15, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.007
6
Context and Register
Wendy L. Bowcher
6.1 Introduction
6.2.1 Context
In the early stages of the development of SFL theory, ‘context’ referred to
the semantic level of language; there was phonology, grammar and lexis,
and context (see Halliday 1961; Ellis 1966; Gregory 1967). At that time, the
extra-linguistic environment was known as ‘situation’ (Halliday 1961; Halli-
day et al. 1964). Gregory makes the distinction between situation and
context in this way:
I am grateful to Edward McDonald for comments on an earlier draft of this chapter. I alone am responsible for any
shortcomings.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 143
During the 1960s, as the ideas of what would become SFL theory began to
take shape, context, in the sense of semantics, came to be called ‘meaning’
or simply ‘semantics’, and situation came to be referred to as ‘context of
situation’. The changes, however, were not merely terminological, as we
will see, but resonated with the changing shape of Halliday’s thinking on
the architecture of language.
The coinage of the term ‘context of situation’ is attributed to Malinowski,
whose ethnographic work in the Trobriand Island communities led him to
realize that the concept of the context of an utterance needed to encompass
the linguistic, situational, and cultural context of a language (Malinowski
1923:306). Malinowski’s research complemented other work at the time,
such as Ogden and Richards’ (1923) discussion of the relation between signs
and meaning, and their suggestion that a description of ‘sign situations’
should come from observation of a corpus of instances rather than individ-
ual or ‘exceptional’ cases (Ogden and Richards 1923:19) – a scientific, not an
‘intuitive’ approach (Ogden and Richards 1923:20). Within this scholarly
climate, Malinowski made important claims regarding the relation between
meaning, language, and situation: ‘Language is essentially rooted in the
reality of the culture, the tribal life and customs of a people, and . . . it
cannot be explained without constant reference to these broader contexts
of verbal utterance’ (Malinowski 1923:305). However, being an anthropolo-
gist and not a linguistic theoretician, he did not develop the concept of
‘context of situation’ within an explicit theory of language.
Firth, a linguist who worked with Malinowski in the 1930s when Mal-
inowski ‘was especially interested in discussing problems of languages’
(Firth 1950:43) stated clearly that, while the term context of situation was
‘first widely used in English by Malinowski’, it became a ‘key concept in the
technique of the London group’1 for the study of language (Firth 1950:42).
Firth differentiated Malinowski’s concept of context of situation from his
own; Malinowski’s concept was essentially a material idea: ‘an ordered
series of events’ (Firth 1950:43). For Firth, however, the concept of context
of situation was ‘best used as a suitable schematic construct to apply to
language events . . . a group of related categories at a different level from
grammatical categories but rather of the same abstract nature’ (Firth
1950:43) through which the meanings of language in use could be inter-
preted. From the start, Firth could see the connection between this abstract
concept and language use and how the concept might become part of the
description and analysis of language. For example, he presents the
following utterance:
1
The London Group or London School refers to those linguistic scholars in Britain taught and influenced by J. R. Firth.
These scholars were also often referred to as ‘neo-Firthians’.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
144 WENDY L. BOWCHER
and asks:
[a]s we know so little about mind and as our study is essentially social, I shall
cease to respect the duality of mind and body, thought and word, and be
satisfied with the whole man, thinking and acting as a whole, in association
with his fellows.
(Firth 1957:2)
With this view of language, context, and ‘the whole man’, the kind of lan-
guage that was to be studied from Firth’s point of view was ‘actual language
text’. Further, the primary sets of relations for enquiry were ‘the interior
relations connected with the text itself’, which include the syntagmatic and
paradigmatic relations within language elements; the situational relations,
which include the ‘interior relations within the context of situation, the focal
constituent for the linguist being the text’; the ‘analytic relations set up
between parts of the text’; and ‘special constituents, items, objects, persons
or events within the situation’ (Firth 1957:5). As for the interior relations of
the context of situation, Firth (1957:9) proposed the following:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 145
6.2.2 Register
The recognition that language varies according to how it is used played a
significant role in forming the concept of ‘register’ – the term ‘register’ as
used in linguistics being attributed to Reid (1956). It goes without saying
that one does not speak the same with one’s mother as with one’s friends or
husband, nor does a sports commentator speak the same way when com-
mentating a game as when ordering a meal in a restaurant. Firth high-
lighted this in his advice to the Air Ministry in suggesting the type of
Japanese that would be most appropriate to learn in order to take up an
effective position against them during World War II:
When I was consulted by the Air Ministry on the outbreak of war with
Japan, I welcomed the opportunity of service for the Royal Air Force
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
146 WENDY L. BOWCHER
because I saw at once that the operating reconnaissance and fighter air-
craft by the Japanese could be studied by applying the concept of the
limited situational contexts of war, the operative language of which we
needed to know urgently and quickly. We were not going to meet the
Japanese socially, but only in such contexts of fighting as required some
form of spoken Japanese.
(Firth 1950:43–4)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 147
These ‘careful and continuous contrasts’ are indeed important for a scien-
tific approach to understanding the relationship between language and
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
148 WENDY L. BOWCHER
context, but Halliday took things a step further than this. Halliday’s under-
standing of language as social activity led to important theoretical tenets
associated with the concepts of context of situation and register in SFL
theory. One of these is the stratal nature of language and the inclusion of
context as a stratum. This is a legacy of Firth’s description of language into
three strata (Firth 1957; although see also Lamb (1966), whose stratifica-
tional grammar was influential in Halliday’s thinking): graphology/phon-
ology, grammar, semantics, and his proposal for the extra-linguistic level of
context of situation. Because of his view that language is social activity,
Halliday has not considered context to be an isolated concept, but one that
is fully integrated into the description of language:
tenor and mode of the situation. The principle is that each of these elements in
the semiotic structure of the situation activates the corresponding compon-
ent in the semantic system, creating in the process a semantic configuration,
a grouping of favoured and foregrounded options from the total meaning
potential that is typically associated with the situation types in question. This
semantic configuration is what we understand by the ‘register’: it defines the
variety . . . that the particular text is an instance of. The concept of register is
the necessary mediating concept that enables us to establish the continuity
between a text and its sociosemiotic environment.
(Halliday 1977:203, emphasis added)
at the semantic level, not above it. Shifting in register means re-ordering the
probabilities at the semantic level . . . whereas the categories of field, mode
and tenor belong one level up. These are the features of the context of
situation; and this is an interface. But the register itself I would see as being
linguistic; it is a setting of probabilities in the semantics.
(Thibault 1987:610, emphases in original)
With regard to text and its relation to context of situation and to register,
Halliday makes the following claim:
SYSTEM INSTANCE
contexts of context of
REALIZATION:
Figure 6.1 Register and context in SFL theory (adapted from Halliday 1999:8)
In Figure 6.1 we can see two basic features of SFL theory: the hierarchy of
strata (Context – Language) and the cline of instantiation (System –
Instance). As already noted, the strata are related through the relation of
realization/construal: context is realized in language and language con-
strues context, and this is indicated by the vertical lines. The cline of
instantiation (the horizontal axes) depicts varying perspectives on the same
phenomenon: a specific language event represents a selection from the
language system; a specific context of situation represents a selection from
the sociocultural context. Between the system and instance ends of the
cline are ‘types’ or categories of contexts of situation correlating with types
of texts or registers. Instantiation and realization are fundamental prin-
ciples explaining how language changes and is also maintained, that is, how
it is an ‘open dynamic system’ in that each instantiation (whether that be
of context or language) ‘resets’ the ‘overall probabilities’ of the respective
systems (Halliday 1992:27).
Within SFL theory, there have been several developments and variations of
Halliday’s conception of context and register. This section briefly describes
the most influential of these.
(Hasan 2005:68, emphasis in original). She has bemoaned the idea that
context is often used merely to
illuminate the nature of parole, [but] not of langue . . . [or] . . . a good device
for ‘mopping up’ some of the problems that inhere in an essentially
extra-social view of language. . .where context becomes a mundane ‘real-
ity’ to be taken for granted, while language becomes a mysterious mental
organ, and grammar a body of knowledge encrypted in the human brain
at birth.
(Hasan 2001:3, italics in the original)
Rather, Hasan has argued that context is central to the ontogenesis, phylo-
genesis, and logogenesis of language (see Hasan 2001), and is part of ‘a
productive principle, reflection on which [has] enabled SFL to offer a scien-
tific description of “how language works”’ (Hasan 2005:55–6). In sum,
Hasan’s work speaks to her view that ‘there can be no language without
context’ (Hasan 2001:8).
Overall, Hasan’s contribution has added detail, precision, and depth
to Halliday’s conception of context and register, not least because she
has questioned many of the assumptions and terminology associated
with these two concepts (see Lukin 2016 for a lengthy discussion of
Hasan’s contribution to the development and description of context).
An example of her questioning includes an early discussion on the
concept of relevancy. Although Halliday theorized the probabilistic
correlation between features of language and features of context, it
was in probing the idea of relevancy and its relation to context and
register that Hasan has proposed that relevant context, or the context
of situation, is embedded in a material situational setting (MSS): ‘Situ-
ation-type is an abstraction from the totality of material situational
setting’ (Hasan 1973:275; also see Hasan 1981:110, Hasan 1995:219).
Her later work has honed this idea: those features of the social situation
which are ‘illuminated’ in the text (Hasan 1995:219, 2005:61) realize
the relevant contextual parameters of field, tenor, and mode, the fea-
tures of the context of situation. The MSS, on the other hand, is not part
of the context of situation, but acts as ‘a dormant source for affecting
the verbal goings on’ (Hasan 1981:110). Features of the MSS enter the
description of the context of situation when they are directly refer-
enced, no matter how minor that reference may be. Hasan has proposed
that both the MSS and the context of situation could be encapsulated in
the overarching terms Action, Reflection, and Contact (ARC): ‘I think of
ARC as relevant to any form of joint social practice, whether this
involves language or not. By contrast, field of discourse, tenor of dis-
course and mode of discourse are . . . specifically discourse related’
(Hasan 2001:7; also see Hasan 2014:11–13).
The concept of Material Situational Setting has been linked with various
other concepts in the theory, including that of institutionalization, which
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 153
In ordinary life, the word has many meanings and each seems clear in its
‘context’, but what exactly did it mean in the description of field of dis-
course? Here it seemed to have multiple values: it was not clear if the word
referred to precisely the same phenomenon in its various appearances, such
as ‘social activity’, ‘relevant activity’, ‘language activity’. There were also
‘descriptive references’ such as ‘what is going on’, and ‘the area of the
operation of the language activity’. Sometimes ‘the whole activity’ was said
to consist of two kinds of ‘activities’, a ‘language activity’ which ‘assisted’
‘the whole event’, in which case it would seem that the ‘whole event’ was to
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
154 WENDY L. BOWCHER
consist of both ‘language activity’ and some other kind of ‘activity’ which
was not linguistic. At other times, the ‘whole of the relevant activity’ could
be accounted for ‘practically’ by ‘language activity’.
(Hasan 2014:6)
Such querying has led to more rigorous accounts of, for instance, the
concept of action in relation to field and mode (see Hasan 1999, 2014;
Bowcher 2013, 2014), but there is still more work to be done in this regard
as evidenced by Hasan’s comment: ‘I have long felt the need to explore the
differences between “act”, “action” and “activity”: this would bring greater
order in understanding field as the parameter concerned with “doing” of
some kind’ (Hasan 2014:51 note k).
Hasan’s conception of context and register as ‘configurations’ of features
has been an important influence on how researchers now research and
model the relationship between context and text. With reference to context
of situation, Hasan prefers the term ‘configuration’ to ‘combination’
because the features of context do not simply ‘combine’; ‘rather, contextual
configuration is like a chemical solution, where each factor affects the
meanings of the others’, thus claiming an ‘interdependence between the
three parameters’ (Hasan 1995:231).
The concept of contextual configuration (CC) being ‘an account of the
significant attributes of [a given] social activity’ (Hasan 1985:56) is central to
understanding two key features of the semantic nature of text: its structure
and texture. Texture refers to the relations among the meanings of a text
and is determined through an analysis of a text’s ‘cohesive harmony’,
which relates to the cohesive ties among elements of a text (see Taboada,
this volume). Text structure is termed ‘generic structure potential’ (GSP),2
as it is used
to describe the structure of not only a specific text type but also a range of
other related text types. Each member of that range of text types will have some
structural properties in common with other members: no individual text
type will have the same structural shape as any other, and none will be
entirely different. The entire range of such text types will constitute a single
register family.
(Hasan 2014:9–10, emphases in original)
A GSP has both obligatory and optional elements. The obligatory elements
are defining in terms of the text’s register, while the optional elements
indicate variation among texts belonging to a given register. Hasan has
demonstrated how the CC can be used to ascertain obligatory and optional
elements of a text as well as the sequence of these elements and their
iteration (Hasan 1985:56). While both GSP and texture reside at the seman-
tic level of language, GSP is said to act as a link between the texture of a text
2
GSP was originally called ‘generalised structure potential’ (see Hasan 1978, 2014:51 note i).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 155
and the relevant context of a text (Hasan 1985:99). With regard to the
relation between texture and context, Hasan comments that ‘situation type,
at a high degree of specificity, is relevant to texture; you could see it as the
motivating force of texture. But by the same token, the facts of texture
construe the very detailed aspects of the situation in which the text came to
life’ (Hasan 1985:115). Hasan also observes that cohesive chains may ‘dis-
play a close relationship to the structural movement of the text’ (Hasan
1985:115). This has been demonstrated in work such as Butt et al. (2010),
Cloran (1999) (in relation to Rhetorical Units and Material Situation), and
Lukin (2010) (see Khoo 2016 for a survey of work on Cohesive Harmony).
Hasan’s GSP was innovative, and perhaps a concept before its time, as it
has not been fully understood within SFL. Some critics have suggested it is a
static or ‘synoptic’ representation of text structure (see Hasan 1995), but the
concept of GSP offers a means of accounting for structure in terms of both
variation and stability across registers, or text types, and in relation to
features of context of situation. Hasan’s argument has always been that
‘the elements of text structure cannot be defined by reference to the rank
status or sequential ordering of the lexicogrammatical units which have
the function of realizing these elements’ (Hasan 1978:229). Rather,
the controls upon the structural make-up of a text are not linguistic in
origin . . . instead, the control is contextual: the nearest non-linguistic ana-
logue of a text is not a logico-mathematical formula, but a non-verbal social
event. A text is a social event whose primary mode of unfolding is linguistic.
(Hasan 1978:229)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
156 WENDY L. BOWCHER
the semantic systems (see Hasan 1995). She points out that the ‘predicting’
power of the contextual parameters with the metafunctions is a reflection
of the nature of the relationship between social variables and linguistic
choices, a mode of explanation wholly acceptable in sociolinguistics, where
‘predictions about linguistic features correlating with situational ones such
as age, gender, geographical and/or social provenance, have typically been
stated in probabilistic terms’ (Hasan 2014:7; see also Hasan 1995, 1999).
With regard to the concept of ‘permeability’, Hasan posits that this has to
do with the ‘conditioned environments’ and the ‘regularity of a set of
relations’ that show how meanings are ‘expanded’ in the sense that the
meaning ‘space’ of certain linguistic categories is permeated by the mean-
ings of another category, and complementary to this, certain linguistic
categories ‘may abandon’ their meaning space and permeate that of other
categories (Hasan 2016:passim). Permeability exists between categories at
the same stratum of language, with ‘the diagnosis of permeability [being]
assisted by inter-stratal relation’ (Hasan 2016:374). Realization is thus
‘needed to recognize the relation of permeability’ (Hasan 2016:374), but
realization is of a different kind of relation. The concept of permeability is
thus different from both realization and interdependency.3
Another important contribution made by Hasan to the study of context
and register is her work on the paradigmatic representation of context.
For this she uses system networks, the modus operandi for representing
the paradigmatic nature of different features at different language strata
(see Webster, this volume). Hasan has argued that ‘the design of the
system network is well suited to contextual parameters’ (Hasan
2014:14) as much as it is to representing choices at other strata. This
does not mean this form of representation is necessarily straightforward
(see Bowcher 2014 for a discussion of some of the issues): networks can
get highly complex, and the different levels of abstraction require a
different ‘value’ assigned to the descriptive choices. Essentially, however,
system networks are ‘a form of argument’ and ‘are a consistent means of
checking what is the better motivated proposal in linguistic description
[in that] [t]he network either accounts for the linguistic variation and its
consequences, choice by choice, or it does not’ (Butt 2001:1825, emphasis
in original). While there has been considerable progress in the develop-
ment of system networks for each of the contextual variables of field,
tenor, and mode, including work by Berry (2016), Bowcher (2007, 2013,
2014), Butt (2004), and Hasan (1999, 2009, 2014), there is still much work
that needs to be done in specifying their features and applying them to
the study of context and register. An example of a system network for
contextual field is shown in Figure 6.2.
3
See Hasan (2016) for a discussion and demonstration of her concept of permeability, and Miller and Bayley (2016) for
work on the concept of hybridity.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 157
praccal
ACTION
conceptual
natural (Sensible)
irrealis (Intelligible)
specialized
FIELD SPHERE OF ACTION
quodian
instuonal
individualized
bounded
spao-temporal
locaon connuing
constant
variable
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
158 WENDY L. BOWCHER
This work on mapping registers derives directly from Jean Ure’s unpub-
lished work on register classification, which focused primarily on field and
mode (Matthiessen 2015:3–5). The starting point for classifying registers is
context, since registers are categorized by reference to their contexts of
situation (Matthiessen 2015). Thus, features from all parameters of context
of situation are relevant to categorizing a register, although so far in this
project, it is the contextual variable of field that has been given the most
attention.
Since field concerns the ‘nature of the activity’, Matthiessen uses the
concept of activity, or social process, as his starting point. He distinguishes
three broad types of processes which he defines in the following way:
(a) Semiotic processes (i.e. ‘meaning’ processes – semiotic processes consti-
tutive of context, manifested through social processes)
(b) Semiotic processes potentially leading to social processes (i.e. ‘meaning’
leading to ‘doing’)
(c) Social processes (i.e. ‘doing’ processes – social processes constitutive of
context, semiotic processes facilitating (i.e. ‘meaning’ facilitating
‘doing’)).
(Matthiessen 2014b:170–1)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 159
Figure 6.3 Matthiessen’s map of the different ‘fields of activity’ (from Matthiessen 2014a:11)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
160 WENDY L. BOWCHER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 161
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
162 WENDY L. BOWCHER
connotative semiotic
stratified context plane expression form
tenor
genre field
mode
discourse phonology/
lexicogrammar
semantics graphology
denotative semiotic
Figure 6.5 Martin’s model of stratified context plane (connotative semiotic) with
language as expression form (denotative semiotic) (from Martin 1999:40)
On the right-hand side of Figure 6.6 are several key terms. ‘Generalized
meaning potential’ refers to the totality of choices possible within a culture.
Within this system are ‘sub-potentials’, such as genres and registers. The
term ‘generalized actual’ refers to the text type, which is also a potential, in
that various individual texts may be classified as a certain type, but unlike
genre and register, which are ‘extra-linguistic’ levels, this term refers to the
categorization of a group of actual texts. The term ‘affording instance’
attempts to capture the idea that the semantic and lexicogrammatical
choices of a text motivate a certain interpretation or reading of a text,
and the notion of ‘subjectified reading’ refers to the way a text is actually
read and understood, deriving from the ‘meaning potential afforded by
individual texts’ (Martin 2008:33). Thus, the hierarchy represents a
narrowing of perspective from potential (most generalized) to instance,
the latter being ‘the reading of a particular text’ (Martin 2008:33).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 163
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
164 WENDY L. BOWCHER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 165
into account ‘pre-text’ context and ‘via-text’ context features. Pre-text con-
textual features may be within the general parameters of field, tenor, and
mode. For instance, prior to a telephone conversation, there are the pre-text
mode features of ‘spoken’ and ‘not-co-present’ already in place. Social
distance features might also be in place prior to a text coming into exist-
ence, and hence be pre-text tenor features, but during the text, such
features may change, such as when a friend who is also one’s boss switches
from friendly chat to outlining some job requirements needed within the
next day or two. Such changes to the social status quo would be considered
via-text contextual features. Berry has developed system networks to
account for these kinds of features (see Berry 2016).
This chapter has provided only an outline of the concepts of context and
register, their development, models, and current research. Context and
register are cornerstones of SFL theory (Lukin et al. 2011) and offer a rich
and varied way into studying language and its relation to society. Readers
are encouraged to consult the reference list to gain a deeper understanding
of the concepts, their theoretical and research value, and their analytical
utility.
References
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
166 WENDY L. BOWCHER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 167
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
168 WENDY L. BOWCHER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
Context and Register 169
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
170 WENDY L. BOWCHER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:37:13, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.008
7
Intonation
Wendy L. Bowcher and Meena Debashish
7.1 Introduction
7.2 Background
1
Although aligning with much of the SFL intonation framework, there are some differences between Tench’s (1996)
description of intonation in English and Halliday’s.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
172 WENDY L. BOWCHER AND MEENA DEBASHISH
The phonological structure of the sentence and the words which comprise it
are to be expressed as a plurality of systems of interrelated phonematic and
prosodic categories. Such systems and categories are not necessarily linear
and certainly cannot bear direct relations to successive fractions or segments
of the time-track instances of speech. By their very nature they are abstrac-
tions from such time-track items. Their order and interrelations are not
chronological.
Meaning was central to Firth’s ideas on language (see Butt, this volume) and
to the study of phonology:
It is not enough to treat the intonation systems as if they merely carr[y] a set of
emotional nuances superimposed on the grammatical and lexical items and
categories. . . . English intonation contrasts are grammatical: they are exploited
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 173
in the grammar of the language. The systems expounded by intonation are just
as much grammatical as are those, such as tense, number and mood, expounded
by other means . . . Therefore, in the description of the grammar of spoken
English, ‘intonational’ and ‘non-intonational’ systems figure side by side.
With regard to meaning, intonation is studied from the point of view that
‘[i]f you change the intonation of a sentence you change its meaning . . .
Intonation is one of the many kinds of resources that are available in the
language for making meaningful distinctions’ (Halliday 1970:21). More-
over, its significance in the act of communicating is indispensable, as Tench
(1996:151) comments:
Over the course of the development of SFL theory, there has been a shift in
the description of intonation in relation to other linguistic categories. For
instance, whereas earlier volumes introducing the SFL model of grammar
described intonation as ‘beside the clause’ (see IFG1 and IFG2),2 later
volumes (IFG3 and IFG4)3 more fully integrate intonation into descriptions
of various features of the grammar, reflecting Halliday’s early assertion
that “intonational” systems operate at many different places in the gram-
mar’, they are not independent but are ‘incorporated throughout the
description wherever appropriate’ (Halliday 1967:10–11).
The different pitch movements – falling, rising, or combinations of
these – contribute to melodic variation in language, and each language
has its own set of tones, a system of tone choices, which contributes to
the meaning-making system of the language. Moreover, as with other
systems and levels of analysis in SFL theory, intonation(al) systems are
interdependent with other systems. This more encompassing view of the
role of intonation in the description of language reflects the SFL multi-
faceted architecture of language. The next section outlines the place of
intonation in the SFL architecture of language.
2
IFG1 and IFG2 refer to Halliday (1985a) and (1994) respectively.
3
IFG3 and IFG4 refer to Halliday and Matthiessen (2004) and (2014) respectively.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
174 WENDY L. BOWCHER AND MEENA DEBASHISH
4
The Cardiff model of language is a bi-stratal model. For a discussion of this in relation to intonation, see Fawcett (2014:
331–4).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 175
Figure 7.1 The relation of realization among the strata of language with tone unit
and information unit shown [note: the downward slanting arrows mean ‘is realised by’]
(cf. Halliday 1992:24)
// 4 ᴧ by the /time the /Great /Central was /built the // 1+ trains could /
manage the /gradients /much more /easily and the //13 Great /Central
/line //. . .
(excerpt from Halliday 1970:127)
Table 7.1 indicates that the first and second tone units encompass more
‘wording’ than the information units they realize due to the tail of the tone
melody extending over a few words belonging to the subsequent grammat-
ical units of clauses or phrases. This wording is shown in italicized text in
Table 7.1. When these tone units are represented as information units, the
‘tails’ are shifted into the information unit, which brings the information
5
The use of the caret ᴧ varies across the SFL literature. Some authors prefer to use a raised caret ^ (e.g. Halliday and
Matthiessen 2014; O’Grady 2017), while others (Halliday 1970; Halliday and Greaves 2008) prefer to use the low
caret ᴧ. In this chapter we use the low caret so as to distinguish this from the raised caret which typically means ‘is
followed by’.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
176 WENDY L. BOWCHER AND MEENA DEBASHISH
Table 7.1 Tone units and information units (tone unit boundaries marked
with double forward slashes)
Tone Unit //by the time the Great Central was built the//
Information Unit by the time the Great Central was built
Tone Unit ___//trains could manage the gradients much more easily and the//
Information Unit the trains could manage the gradients much more easily
Tone Unit ______//Great Central line//
Information Unit and the Great Central line
melody // TONE UNIT // *The tone unit or melodic line is the highest phonological rank.
melodic line *Tone unit boundary shown by double forward slashes.
rhythm // foot / foot / foot / foot // *A tone unit or melodic line of speech may consist of one or more feet.
*A foot boundary is shown by a single forward slash.
Figure 7.2 The constituents of the tone unit showing some analytical conventions
As already noted, the tone unit at the phonological stratum realizes the
information unit in the grammatical stratum. At the grammatical stratum
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 177
The first syllable in the foot is always salient. The salient syllable carries the
beat . . . [and is] followed by one or more non-salient, or weak syllables . . . A foot
may begin with a silent beat, without the rhythm being disrupted or lost.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
178 WENDY L. BOWCHER AND MEENA DEBASHISH
The first syllable in the foot whether silent or spoken is called the Ictus.
Further, this might be the only choice, and thus the foot would be a ‘simple
foot’ which could be silent or spoken. On the other hand, there may be any
number of syllables in a foot, and those syllables following the Ictus,
whether one or more, are known collectively as Remiss. A foot of this latter
type is called a compound foot. The system network for choices in foot
composition and ictus state is shown in Figure 7.3.
At the rank of tone unit, the three main systems of intonation are
tonality, tonicity, and tone (see Figure 7.4). Tonality refers to choices
available for the organization of a discourse into tone units; tonicity refers
to the selection and assignment of prominence within the tone unit; and
tone refers to the choices of pitch movement. Halliday (1963a, 1963b,
1967, 1970) based these systems of English intonation on various samples
of spoken English, the largest consisting of ‘just under 2,000 tone groups’
(Halliday 1967:9).
simple
FOOT
COMPOSITION
compound
+Remiss
Remiss: syllable1-n
Ictus^Remiss
foot
+Ictus
filled
ICTUS Ictus:syllable1-n
STATE
empty
Ictus:silent
Figure 7.3 System network for choices in FOOT COMPOSITION and ICTUS STATE (from Halliday
and Matthiessen 2014:18)
TONALITY
INTONATION TONICITY
SYSTEMS
TONE
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 179
In this sample of reading aloud, we can see a high number of tone groups
across one or two clauses. For example, the second clause ‘when he noticed a
plump hen sitting on the branch of a tree’ is spoken in four tone groups. The
choice of reading the sentence into four tone groups represents selections
from the system of tonality, and the selection of tonic prominences (shown
here by bolded text) reflects choices from the system of tonicity. The choice
of tone groups themselves – tones 3, 1, 1, and 1 – represents choices from the
system of tone.
System networks have also been developed for features at the level of
segmental and word phonology, such as Young’s (1992) work on English
consonant clusters and Tench’s (2014) work on English word phonology. The
concept of system, however, is understood somewhat differently when it is
applied at the levels of word, syllable, and phonemes. In discussing his systems
of consonant clusters, Young explains that while the concept of system net-
work implies sets of choices of meanings, ‘by the time we get . . . [to] segmental
phonology all the meaning choices have been made long ago and everything
now to be selected is predetermined’ (Young 1992:58). Tench reiterates this
view: ‘System at the level of word (and also at the level of groups/phrases) is
rather the specifications of what the speakers of a language recognize as
having been established in, or “chosen” by, the language to represent its
words’ (Tench 2014:274). Tench has developed system networks outlining
the possible syllable structures, peaks, margins, strong and weak vowels,
and syllable initial consonants in English (standard southern English pronun-
ciation) (see Tench 2014). An interesting point made by Young is that even
though system networks (for consonant clusters) display predetermined sets
of choices, ‘they probably have a more positive role to play in decoding, and
they certainly need to be built into a model of English which accounts for the
ability of speakers to add to their vocabulary (for example, by means of foreign
loanwords) words that conform to the phonology of English’ (Young 1992:58).
While system networks represent choices, Fawcett (2014) argues that the
typical SFL system networks for intonation are largely descriptive frame-
works useful for analyzing language instances, or output (2014:325).
Fawcett (2014) thus proposes a generative model of intonation and punctu-
ation. He sets out several concepts and realization rules related to ‘intonation
components’ and ‘punctuation components’ needed for a comprehensive
model of grammar that can be used for generating English text. We do not
take up this argument here, but note that Fawcett (2014) provides a careful
and detailed proposal for ‘construct[ing] a generative systemic functional
grammar of intonation and punctuation for English’ (Fawcett 2014:396).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
180 WENDY L. BOWCHER AND MEENA DEBASHISH
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 181
classical choral music, such as Handel’s Messiah. These studies broaden the
scope of the study of intonation beyond spoken language registers.
Thus, while some inroads into identifying the phonological character of
different registers have been made, further research using the SFL model of
intonation to compare and contrast choices across the various intonation
systems in different situations would help to establish a more comprehen-
sive description of different registers in language by adding a phonological
dimension to register variation and identification.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
182 WENDY L. BOWCHER AND MEENA DEBASHISH
Ictus (^ Remiss)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 183
The terms Ictus and Remiss are functional terms, with the Ictus syllables
playing a role in the rhythm of the discourse; these syllables tend to occur
at fairly equal intervals of time in continuous spoken English. This affects
the syllables which may occur within the feet from one Ictus syllable to
another; that is, in order to maintain tempo, the Remiss or non-salient
syllables (when present) tend to be ‘squashed’ through contraction or
weakening of vowels, particularly when there are multiple syllables.
With regard to the tone unit, paradigmatically there are several choices
of single and compound tones. The tone groups, as carrying the melodic
shape of the language, may be characterized by primary and ‘accompani-
ment’ pitch movements (Halliday and Greaves 2008). In this sense, the
‘functioning elements’ in the tone group are the Tonic and the Pretonic,
‘which supply the framework within which the speaker’s variations in
pitch and loudness are perceived and interpreted by the listener’ (Halliday
and Greaves 2008:42). The major pitch movement of the tone unit is
initiated on an Ictus syllable. This constitutes the tonic syllable, and the foot
in which the tonic syllable is located is called the tonic foot. The tonic foot
may be followed by one or more feet, which continue the pitch movement
initiated in the tonic syllable. For a Pretonic segment to be present, there
must be at least one complete foot prior to the tonic foot and not connected
with the previous tone group. The Pretonic has its own set of pitch contour
patterns, but the pitch movement in the Tonic is defining as far as tone
group choice goes. In other words, the pitch movements in the Pretonic are
determined by those in the Tonic, and each Tonic pitch movement has its
own set of Pretonic pitch movements. Thus, the tone unit structure is
described as having an obligatory Tonic which is optionally preceded by a
Pretonic. In terms of syntagmatic structure, when a Pretonic is present, the
tone unit is realized as Pretonic ^ Tonic. Figure 7.5 displays the tone group
choices and the choices of [with pretonic] or [without pretonic].
Figure 7.5 indicates that a tone unit is realized (&) with a tonic element.
The choices in the system of tonic composition are simple or compound.
In the system of pretonic are the choices [with pretonic] or [without
pretonic]. Moving to the right of the figure we find more delicate choices
available for the tones of English. Here we can see the various secondary
tones that are possible – the indirect secondary tones are choices deriving
from the [with pretonic] system, and the direct secondary tones deriving
from the simple and compound tone systems. The system network also
indicates that there are choices of both indirect and direct secondary tones
for certain tone groups. For example, a narrow Tone 1 [1-] may be spoken on
an even pretonic [.1]. These variations in secondary tones are described in
Section 7.4.2 of this chapter.
We can illustrate the syntagmatic structure of tone units and the paradig-
matic choices at a primary level of delicacy using a clause complex from Halli-
day (1970:120): ‘Not always was the kangaroo as now we do behold him but a
different animal with four short legs.’ This is analyzed in the following way:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
without pretonic
pretonic
with pretonic
+ Pretonic:
Pretonic: foot1-n
Pretonic ^ Tonic
.1 even
-1 bouncing
1+ wide …1 listing
tone 1 fall 1. medium [tone 1]
1- narrow
.2 high
tone 2 (high rise/- 2. straight [tone 2] -2 low
high fall-high rise)
SIMPLE
simple PRIMARY 2- broken
TONE tone 3 low rise .3 mid
[tone 3] -3 low
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
tone group
+Tonic; tone 4 (rise-) fall-rise 4. high
Tonic: foot1-n 4- low
COMPOUND
compound
PRIMARY
+Tonic 2; TONE tone 53 (fall-) rise-fall
Tonic 2: foot1-n plus low rise
Tonic^Tonic 2
Figure 7.5 Choices in the TONE U NIT system showing the more delicate choices available in the pretonic and system of TONE [note: in this
diagram Tonic 2 refers to a Minor Tonic] (from Halliday and Matthiessen 2014:18)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
Intonation 185
Table 7.2 Distribution of tone, tonic, and pretonic elements in the excerpt.
(NB: MT = minor tonic)
The clause complex is spoken on three tone units. The third tone group is
realized with both a pretonic (/four /short) and tonic (/ legs//), therefore the
choice is ‘+pretonic’. The first and the second tone units are both realized
with only tonic segments, i.e. ‘-pretonic’, but with differences between the
two. In the first tone unit, the tonic segment is preceded by an incomplete
foot (// ˄ not /) which does not constitute a pretonic segment. In the first tone
group there are five feet, in the second there are two, and in the third there
are three. Each foot begins with an Ictus syllable, or salient syllable, but in
the first foot in the first tone unit there is a silent Ictus indicated by the caret
symbol. Generally, monosyllabic lexical words and the accented syllables of
polysyllabic words tend to take salience in a tone unit while the monosyllabic
grammatical or function words tend to be non-salient. We can see that the
example reflects this pattern of salience. The rhythm (or beat) is carried by
these salient or Ictus syllables. The sequence of Tone Unit Structures in this
excerpt is [Tonic^Minor Tonic] ^ [Tonic] ^ [Pretonic ^ Tonic], and at the
grammatical level this would mean that there are three quanta of infor-
mation and thus three information foci, with the first quantum of infor-
mation comprising a major and minor information focus. The distribution of
tone, tonic, and pretonic segments for the excerpt is displayed in Table 7.2.
Each tone unit, whether it be of one or more feet, has a distinct pitch
contour and contains a point of prominence known as the tonic syllable.
Tonic prominence is always assigned to a salient syllable in a foot within
the tone group in response to the relevant activating feature(s) of the
context of situation and the co-text. This assignment of tonic prominence
realizes the functional element of New in an information unit, which is a
culmination point or pulse. This relates to the textual function of the
information unit; its role in organizing discourse in terms of the status of
information – information is presented as ‘Given’ or it is presented as
‘New’. Ascertaining where the New element begins is somewhat indeter-
minate because New is realized by sound not written words, and New is a
culmination point of a range of prosodic features. Furthermore, because the
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
186 WENDY L. BOWCHER AND MEENA DEBASHISH
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 187
Table 7.3 Choices in the system of key and their typical meanings (see
Halliday and Matthiessen 2014:169)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
188 WENDY L. BOWCHER AND MEENA DEBASHISH
the clause complex as one quantum of information through the use of one
tone group only, as in:
The assignment of New illustrates how the textual metafunction may work
independently of the hierarchic organisation generated by the experiential
one. In particular, the element New of the information unit is not restricted
to focus (1) on an element of structure selected from within a single clause;
nor (2) on an element of clause structure.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 189
// 3 when he / noticed// 1 a /plump / hen// //1 sitting //1 on the /branch of a / tree//
when he noticed a plump hen sitting on the branch of a tree
Theme Rheme
Given New New New New
Section 7.3.3 shows several instances of marked New. Bowcher (2003, 2004)
describes the relationship between Theme and New choices in excerpts
from radio sports commentaries, and specifically in play-by-play talk of a
Rugby League game. She finds that in this register there is a complementary
relation between information carried by the Theme and that carried by the
New, with informational peaks falling predominantly on players either as
Participants or as Circumstances (destinations of the ball), and that marked
informational prominence is not common in this type of talk. Zhu (2014)
focuses on the relationship between Theme and Information choices in a
BBC news reading. Lukin (2014) also analyzes the relationship between New
and Theme in her study of a news report of the ‘Coalition’s’ war with Iraq in
2003, finding that ‘in the choice of “person” and “place”, the system of IF
[information focus] is reinforcing patterns established via the system of
theme’ (Lukin 2014:65).
Choices from the systems of tonality and tonicity play a role in the
degree to which the boundaries of information units coincide with those of
clauses and what is assigned focus by the speaker. Various researchers
highlight registerial patterns in this regard. For instance, Lukin (2014:63)
notes that there is a ‘higher ratio of tone units to the grammatical unit of
clause’ in broadcast news (see van Leeuwen 1992; Smith 2008). However,
such a claim is not particularly revealing, considering that other registers
also exhibit a high number of tone units per clause, such as reading aloud
children’s stories (e.g. Bowcher and Zhu 2014; Halliday 1970), and indeed,
Lukin effectively acknowledges the too-general nature of her claim when she
asks, ‘Is there a metafunctionally significant pattern in what is selected for
focus by the location of intonational focus?’ This kind of question is of more
value in identifying registerial differences in tonality, tonicity, and
information distribution, and Lukin’s analysis indicates that in her
news broadcast data choices construe largely textual and interpersonal
meanings. Other register-focused findings include Smith’s (2008) work
which includes a description of the way that intonation choices (amongst
others) play a role in shifting the focus of talk taking place in a surgery from
experiential to interpersonal meanings, and Bowcher and Zhu’s (2014:17)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
190 WENDY L. BOWCHER AND MEENA DEBASHISH
study points out that even though a children’s story ‘has a low lexical density
and is grammatically simple, when it is read aloud, it is assigned a high
“informational” density through the resources of the intonation system’.
Of relevance to intonation choices in relation to the textual function is the
higher unit of paraphones, or phonological ‘paragraphs’. Tench (1996:24)
argues that ‘phonological paragraphs’ typically begin with a ‘high pitch on
the onset syllable of the initial intonation unit’ and that the pitch gradually
falls until the final tone unit, wherein ‘the depth of fall in the final unit is
the lowest in the whole paragraph’. He also observes that the tempo tends to
slow down in the final intonation unit and that ‘there is a longer pause than
is normally allowed between intonation units’ between phonological para-
graphs (Tench 1996:24; see also Tench 2014:272–3). Iwamoto (2014) picks up
the idea of phonological paragraphs, calling them ‘paraphones’ after Halli-
day (1961:253 note 30) and proposes that ‘paraphoning’ is a textual process
across all strata of the language system and is semogenic in nature. He
argues that a paraphone is a semantic unit, whose boundaries are realized
by specific phonetic cues, such as pitch levels, and he hypothesizes that
paraphoning differs across registers in that ‘speaker[s] select one way [of
paraphoning] over others according to the context of situation to create
distinctions in meaning’ (Iwamoto 2014:143). Iwamoto’s work leaves open
an enticing area for further research into register variation.
The textual metafunction plays a key role in the construal of the contextual
parameter of mode. We noted in Section 2 Tench’s (1996:151) comment on the
importance of intonation in reading and writing, and several SFL studies have
focused on this very point. These include Davies’ (1989, 1992, 1994a, 1994b,
2014) extensive work on the relation between cohesion, information structure
in written and spoken text, punctuation, and intonation systems for effective
understanding and reading aloud of written text. There is also Cummings’
(2000, 2001, 2014) research into the interpretation of the intonation patterns
relevant to written text, and Bowcher and Zhu’s (2014) study of native and
non-native English speakers reading a children’s story. The relationship
between spoken and written language features is also of critical importance
in Fawcett’s (2014) generative model of English intonation and punctuation.
The next section of this chapter describes in more detail the tones in
English as modelled in Systemic Functional Linguistics. Our aim is to
highlight the basic shapes of the tones of English, as it is the tone group
that is the core around which other choices in the system of intonation
and the system of information operate.
While this section sets out a description of the tones of English, some experi-
ence in listening to the different tones should be gained prior to conducting
one’s own analysis. A good place to gain experience in hearing the tones is
Halliday and Greaves (2008). Additionally, Greaves (2014) is an interactive
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 191
chapter on the SFL system of intonation and includes practice in rhythm and
hearing tones, and in analyzing spoken utterances using Praat.6
According to the SFL model, there are seven primary tones in English, within
which five are simple tones and two are compound tones. We describe each of
the primary tones in turn before turning to the secondary tones.
7.4.1.2 Tone 2
Tone 2 is a sharp rising pitch contour from low or mid low, and it covers a
wide pitch range. The pretonic pitch contour is either high level (as in
Figure 7.9b), or steps down from high or mid high to the point/pitch level
from where the tonic pitch movement begins (as in Figure 7.9a).
Figure 7.9a Tone 2 – the sharp rising tonic with a step down pretonic
Figure 7.9b Tone 2 – the sharp rising tonic with a high level pretonic
6
See www.equinoxpub.com/systemic-phonology-files for the accompanying sound files to Greaves (2014).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
192 WENDY L. BOWCHER AND MEENA DEBASHISH
Will you please finish your // 2 ᴧ will you / please /finish your / work //(step down
work? pretonic)
Have you finished your // 2 ᴧ have you / finished your / work //(high level
work? pretonic)
7.4.1.3 Tone 3
Tone 3 is a level rise from about low or mid low to mid pitch. If there is a
pretonic segment, it has a level contour.
7.4.1.4 Tone 4
Tone 4 is a fall-rise pitch contour with more force on the falling movement.
This tone exhibits a rise-fall hook onset before the main falling-rising pitch
movement; the pitch first rises from mid to about mid high before execut-
ing the fall-rise movement, which is a key feature of the tone. The fall
covers a wide pitch range, and the rise is almost to the same level as the
beginning of the fall.
The pretonic contour steps down from high pitch to around mid pitch
level, i.e. to the level from where the rise-fall hook onset of Tone 4 begins.
He finished his work, but . . . // 4 ᴧ he / finished his / work but // . . . //
If you don’t finish your work in time // 4 ᴧ if you / don’t / finish your / work in / time //. . .
7.4.1.5 Tone 5
Tone 5 exhibits a pitch movement which is in the opposite direction to Tone
4. This is a rise-fall pitch contour with more force on the rising movement.
This tone also exhibits a hook onset which has a fall-rise movement. The
pitch in the tonic syllable first falls from mid-high to mid and then rises to
cover a wide range before executing the fall.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 193
The pretonic steps up from about mid level to mid high pitch level, i.e.
almost to the level from where the fall-rise hook onset for Tone 5 begins.
I’d never seen such a hullabaloo // 5 ᴧ I’d /never /seen such a /hullaba/loo//
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
194 WENDY L. BOWCHER AND MEENA DEBASHISH
//13 Give my re/gards to your /parents /next time you /see them//
(example from Halliday 1970:88)
//53 ᴧ I’d /rather /like one if you /feel you can /spare it//
(example from Halliday 1970:93)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 195
The neutral type [.1] is also referred to as the ‘even’ type, as the pitch
contour is more or less level/even: at mid pitch level for [1.], steps up from mid
low to high for tone [1+], and steps down from mid high to mid low for [1-].
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
196 WENDY L. BOWCHER AND MEENA DEBASHISH
Figure 7.17: Indirect (neutral) secondary tone [.1] ([.1.], [.1+], and [.1-] respectively)
The [-1] ‘uneven’ pretonic has a dipping or bouncing contour from around
mid to a fairly high pitch in each foot. The [. . .1] pretonic has a level rising
contour in each foot.
In the second, the swinging movement in each foot adds to the intensity of
the meaning of the sentence. This pretonic occurs more naturally with a
‘strong’ [1+] tonic, making it more forceful.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 197
These indirect secondary tones [.2] and [-2] can combine with either of the
direct secondary tones, [2.] or [2], but the combination of [-2] with [2]
appears to be rare.
The high level pretonic with a jump down in pitch to a sharp rise in
the tonic [.2.] is the most unmarked way of realizing the yes-no interroga-
tive. When a low level pretonic combines with a neutral tonic [-2.],
the yes-no interrogative acquires an additional meaning of being more
‘involved’.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
198 WENDY L. BOWCHER AND MEENA DEBASHISH
Figure 7.20 Indirect secondary tones of Tone 3 ([.3] and [-3] respectively)
Figure 7.21 Direct and indirect secondary tones of Tone 4 ([4.] and [4] respectively)
//4. ᴧ it’s a / bit /dangerous// (‘I can’t help being worried; . . .’)
(from Halliday 1970:110)
Tone [4.] is the neutral one with a fall-rise pitch contour from mid high, and it
covers a wide pitch range. The pretonic for this tone is a step down contour
from high to about mid pitch level. Tone [4] is the marked variant with the
fall-rise pitched lower. The preceding pretonic contour exhibits a fall-rise
pitch movement in each foot, which seems to be imitating the fall-rise
movement of the tonic segment. This series of low pitched fall-rise move-
ments in the pretonic adds to the intensity of the tone.
//4 not unless he’s /willing to a/pologise // (‘I might see him if he does’)
(Halliday 1970:111)
The Tone [4] variant makes the meaning of fall-rise pitch more intense, and
is accompanied by a distinct voice quality. For instance, if, in a
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 199
Figure 7.22 Direct and indirect secondary tones of Tone 5 ([5.] and [5] respectively)
Tone [5.] is the neutral tone having a rise from about mid to mid high or high
and then a fall to a lower pitch level. The pitch in the pretonic segment exhibits
a step up from about mid low to about mid high. Tone [5] is a low pitched rise-
fall contour at mid low/mid pitch level. This is preceded by a pretonic, with
each foot exhibiting a step down rise-fall movement.
//5. ᴧ I /can’t be/lieve they would /ever have /thought that a/bout her//
//5 ᴧ I can’t be/lieve they could /be so /stupid//
The meaning of the neutral Tone [5.] is related to the rise-fall prosody, i.e.
‘there was some doubt, but all is fine’. On the other hand, Tone [5], which is
lower pitched and usually accompanied by a breathy voice quality, is used
to indicate ‘awe’, and sometimes, depending on the context, ‘sarcasm’ or
‘disappointment’, as in the examples.
7.5 Conclusion
This chapter has provided an overview of intonation within the SFL architec-
ture of language with some mention along the way of specific research that
has been conducted. The second half of the chapter outlined the melodic
shapes and general meanings of the primary and secondary tone groups in
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
200 WENDY L. BOWCHER AND MEENA DEBASHISH
English. Our aim has been to provide a description useful for understanding
the place of intonation in the SFL architecture of language and for conduct-
ing some basic analyses of spoken English, albeit recognizing that practice in
listening to the tone groups would be essential for undertaking such an
analysis. Further, while it is impossible to do justice to all of the areas of
research that have been developed for intonation within the SFL framework
in a chapter of this size, we hope that we have provided sufficient back-
ground to offer readers ideas for possible research directions.
References
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 201
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
202 WENDY L. BOWCHER AND MEENA DEBASHISH
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
Intonation 203
Storynory, n.d. How Love and Peace Came to the Woods. Available online at:
www.storynory.com/2017/04/22/love-peace-came-woods. (Last accessed
27/07/2017.)
Tench, P. 1988. The Stylistic Potential of Intonation. In N. Coupland, ed.,
Styles of Discourse. London: Croom Helm. 50–84.
Tench, P. 1990. The Roles of Intonation in English Discourse. Frankfurt am Main:
Peter Lang.
Tench, P. 1992a. From Prosodic Analysis to Systemic Phonology. In
P. Tench, ed., Studies in Systemic Phonology. London: Pinter. 1–17.
Tench, P., ed. 1992b. Studies in Systemic Phonology. London: Pinter.
Tench, P. 1996. The Intonation Systems of English. London: Cassell.
Tench, P. 2014. Towards a Systemic Presentation of the Word Phonology of
English. In W. L. Bowcher and B. A. Smith, eds., Systemic Phonology: Recent
Studies in English. Sheffield: Equinox. 267–93.
van Leeuwen, T. 1992. Rhythm and Social Context: Accent and Juncture in
the Speech of Professional Radio Announcers. In P. Tench, ed., Studies in
Systemic Phonology. London: Pinter. 231–62.
Young, D. 1992. English Consonant Clusters: A Systemic Approach. In
P. Tench, ed., Studies in Systemic Phonology. London: Pinter. 44–69.
Zhu, S. 2014. Intonation: Signal of Information Peaks. In W. L. Bowcher and
B. A. Smith, eds., Systemic Phonology: Recent Studies in English. Sheffield:
Equinox. 91–115.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 08:26:40, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.009
8
Continuing Issues in SFL
Mick O’Donnell
8.1 Introduction
This chapter will discuss various issues that are not fully resolved within
Systemic Functional Linguistics (SFL). Two main issues will be addressed:
Both of these issues are still under debate within the community, often
leading to divergent approaches, and, if we are not aware of the underlying
differences, result in misunderstanding of the arguments others are
making.
This chapter will refer extensively to the four editions of ‘Introduction to
Functional Grammar’, the first two by Halliday (Halliday 1985, 1994), and
the last two revised by Matthiessen (Halliday and Matthiessen 2004, 2014).
To simplify references, I will refer to these as IFG1, IFG2, IFG3, and IFG4.
My thanks to Tom Bartlett, Margaret Berry, Lise Fontaine, Jim Martin, and Geoff Thompson for comments on this work.
While Butler (2003) is only lightly cited in this work, it has strongly influenced my way of thinking about functional
grammar discussed in Section 8.2.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 205
but see also Hudson (1971); O’Donnell et al. (2008); Fawcett (2009); Tucker
(2014); Gwilliams and Fontaine (2015).
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
206 MICK O’DONNELL
Halliday argues that this intimate relation between grammar and its con-
text of use is not accidental, but rather the result of language (including
grammar) having evolved in its use:
Language has evolved to satisfy human needs; and the way it is organized is
functional with respect to these needs – it is not arbitrary. A functional
grammar is essentially a ‘natural’ grammar, in the sense that everything
in it can be explained, ultimately, by reference to how language is used.
(IFG2:xiii)
The concept of the social function of language is central to the interpretation
of language as a system. The internal organisation of language is not acci-
dental; it embodies the functions that language has evolved to serve in the
life of social man.
(Halliday 1973:44)
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 207
Or in Martin’s words:
This is not to say that grammars can be defined entirely in terms of how
words are used. This would lead to what is called a ‘notional grammar’:
grammar defined entirely in terms of meaning.1 In such an approach, a
‘verb’ might be defined as ‘a word that expresses an event’, and ‘noun’ as ‘a
word that expresses an entity’.
For Halliday, a functional grammar has to relate outwards to the mean-
ings it realizes, and also account generatively for the range of forms that
realize these meanings. He stresses that our grammatical organization
cannot be divorced from the need to account for structural patterning:
All the categories employed must be clearly ‘there’ in the grammar of the
language. They are not set up simply to label differences in meaning. In
other words, we do not argue: ‘these two sets of examples differ in meaning;
therefore they must be systematically distinct in the grammar’. They may
be; but if there is no lexicogrammatical reflex of the distinction, they are not.
(IFG1:xx)
Halliday and Matthiessen raise the point here that, when building a gram-
matical description, evidence from the three viewpoints may conflict, and
the model builder needs to choose what importance they give to each
1
David Rose, in the Sysfling discussion list (09/04/2011), suggested that the term ‘notional grammar’ is appropriate for
grammars which ignore grammatical reactance totally, while ‘functional grammar’ is appropriate for a grammar based
on the identification of recurring grammatical structures which have distinct semantic functions.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
208 MICK O’DONNELL
The fact that the required grammatical reactance can be very indirect
shows that priority is given to similarity of meaning when deciding on
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 209
Being a ‘functional grammar’ means that priority is given to the view ‘from
above’; that is, grammar is seen as a resource for making meaning – it is a
‘semanticky’ kind of grammar. But the focus of attention is still on the
grammar itself.
(IFG3:31)
Rather than talking about the relative importance of the criteria, it may be
better to talk of starting point. In IFG2 (xiv), Halliday contrasts the traditional
approach of starting with modelling the word forms (morphology), building
a syntax on top of that, and only then asking what these forms mean, with
the SFL approach, which starts by interpreting a language as a system of
meanings, and then explores how those meanings can be realized as forms.
A possible critique to this approach is that there are many ways to
organize a language in terms of meaning, and only a subset of these will
allow a simple mapping onto forms. What we ideally want is a meaningful
organization of language which has the strongest correlation with regular-
ities of form. We cannot do this by exploring meaning in isolation from the
forms that realize them.
This is the reverse of the criticism levelled by functionalists and cognitiv-
ists against Bloomfield and his successors, who tried to construct an
autonomous syntax without considering meaning (Tomlin 1990; Newmeyer
1991:62; Halliday in Martin 2013:164). The full answer is that, in the
construction of a language model, we need to consider meanings and forms
at the same time.
Structuralists often apply the principle of Occam’s Razor: the best gram-
matical description is that which uses the least rules to describe the phe-
nomena at hand. The principle can also be applied to the construction of
functional grammars: the best description is the briefest which represents
the meaningful aspects of language use and from which all forms can be
generated (with the mapping of meaning onto form included in the size of
the description).
Unfortunately, SFL grammars (or semantic specifications) are rarely pre-
sented with both system network and realization statements (Hudson 1971
and Matthiessen 1995 being good exceptions), and thus commonality of
meaning may play a bigger role in grammar construction than it should.
Halliday’s verbal processes (see Berry, this volume) offer a good area
through which to demonstrate the problems of grammatical classification.
Exactly what constitutes a verbal process is often debated within the com-
munity, and the four versions of IFG have shifted on the issue over time.
There is a common belief within the community that the test to identify a
verbal process is that there must be projection in clausal form (or at least
potential for clausal projection). This test would result in He said he was going
being classified as verbal, while He talked about the weather would not.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
210 MICK O’DONNELL
However, all versions of IFG include at least two classes of verbal process:
the projecting kind just mentioned, and one which involves a Target (I’m
always praising you to my friends). As there is no structural similarity between
these two types of verbal process which would motivate their grouping, we
must assume they are being grouped totally on notional grounds: they both
express a verbal action.2
We might propose putting targeted verbal processes aside, and say that
the clausal projection criterion applies to the remaining verbal processes.
However, it seems that the SFL community as a whole is divided as to what
to do with processes which involve verbal action, but where clausal projec-
tion is not involved. This involves verbs such as talk, and grumble, and
includes cases where no Matter is specified (We talked for hours), and also
where Matter is specified (He talked about his hometown).
An online survey was conducted in 2004 to test how a range of SFL
practitioners coded various clauses in terms of process types (O’Donnell
et al. 2008). Seventy-five respondents coded thirty-two difficult clauses. The
survey revealed a spread of coding styles, ranging from heavy dependence
on structural criteria, to those who coded largely on the semantics of the
clause. In respect to We talked for hours, 60 per cent of coders placed it as
behavioural, and 35 per cent followed notional criteria, coding it as verbal
(the remaining 5 per cent coded it as material, some indicating they did not
use the behavioural category).
In regards to He talked about his hometown, a surprising result was that
15 per cent of those who had coded the previous sentence as behavioural
swung over to verbal for this case. This suggests that these coders do
require presence of the verbal product to code as verbal, but do not go so
far as to require clausal projection (a similar pattern was shown in the
coding of mental processes). This variation in the coding community dem-
onstrates that the nature and degree of structural reactance needed varies
across the community. A further study reported in Gwilliams and Fontaine
(2015) confirmed these results.
Part of the disparity in coding verbal processes may stem from the
treatment of this area in the four versions of IFG. In IFG1, verbs like talk
were not covered explicitly in the section on verbal processes, although a
later section on Range classified She speaks German and Don’t talk nonsense! as
verbal processes (IFG1:133). Behavioural processes are said to be intermedi-
ate between material and mental processes, which seems to exclude the talk
verbs from this category.
IFG2 however expands behavioural processes to include a ‘near verbal’
category, which includes talk, grumble, and chatter (IFG2:139). The possibility
of Matter with these verbs is explicitly mentioned, so He talked about the
2
Tom Bartlett (personal communication) prefers to phrase the semantic label for verbal processes as ‘transfer of
information from one person (or semiotic object) to another’. This would leave We talked about the weather out of
verbal processes, while leaving He called me a bastard in.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 211
The targeting type of process is again mentioned, with a statement that this
subtype of verbal process does not easily project reported speech. A list of
verbs taking a (nominal) Verbiage includes some which cannot easily pro-
ject: He described the apartment, or He outlined his plan.
In IFG3, representing a revision of IFG2 by Matthiessen, we see a change
back towards more notional coding. To talk to that priest about Kukal is said to
be verbal (IFG3:252). However, there seems to be some inconsistency here,
as grumbled about the food (IFG3:251) is said to be behavioural. I believe this
was a state of transition from Halliday’s original more structural orienta-
tion towards Matthiessen’s more notional orientation.
In IFG4, the talk processes are fully instantiated as a subtype of verbal
process, as shown in Table 8.2. However, behavioural processes still include
verbs such as chatter, grumble, and talk, which appears to be an inconsist-
ency, with grumbled about the food explicitly mentioned as behavioural
(IFG4:302).
Thompson (2015) believes this is not just a problem for verbal processes,
but general across transitivity classification:
Halliday (1994: xix) has consistently argued that ‘all the categories employed
must be clearly “there” in the grammar of the language. They are not set up
simply to label differences in meaning’; and in the case of transitivity certain
key grammatical criteria for categorization (such as preferred tense/aspect,
and the potential to project) have been elaborated (e.g. Halliday 1994:
115–16). However, it has proved difficult to implement the principle of
‘clearly “there” in the grammar’ in all cases: the grammatical criteria by
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
212 MICK O’DONNELL
which one process type can be differentiated from another are not always
precisely definable. As a result, analysts may, implicitly or explicitly, find
themselves forced to fall back on purely semantic criteria.
(Thompson 2015:21–2)
The above discussion has tried to show that in Halliday’s grammar, the
notional criteria dominate over secondary, indirect, grammatical reactance.
Some in the community however stress the importance of the grammatical:
Process types are entirely grammatical categories. The names of the process
types are just aide-memoires that capture only their most common notional
features; they are not useable as criteria for defining on notional grounds.
(Tom Bartlett 2011, sys-func discussion list)
I would strongly encourage holding onto grammatical reactances when
reasoning about process type. These are the grounding strength of our SFL
approach to case relations, compared with work in other models. We should
be enriching our argumentation based on reactances . . . We need to push on
to tackle the challenge of finding distinctive reactances for process types as
we move from language to language – and NOT abandon the criterial argu-
mentation the reactances afford. It is very worrying to think that the power
of the SFL approach (its revelation of the meaningful ways in which lan-
guages construe reality) might become its undoing via a collapse into notion-
alism, or accommodation of notionalist ‘reasoning’ (sic) alongside reasoning
based on grammatical criteria.
(Jim Martin 2011, sys-func discussion list)
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 213
The point here is that, given the different possible expressions of a notional
situation, using notional criteria to classify process types does not seem
promising.
Gwilliams and Fontaine (2015) look at the problem from the other side:
because of ideational grammatical metaphor, similar clausal expressions
can be used for distinct notional situations, for instance, ‘Ivy touched Fred
with a stick’ (representing a notionally material action) and ‘Ivy touched
Fred with her words’ (representing a notionally mental action).
They note that such ambiguous cases give rise to two distinct problems.
Firstly, because these examples allow for two analyses, they introduce the
potential for inconsistent coding amongst analysts, given that some may
favour semantic criteria, and others, syntactic criteria. They point out
however that this can be avoided by explicit direction as to the coding
criteria.
The second problem they think is more important:
The solution they propose is to always allow for two analyses of clauses: a
surface analysis based on syntactic tests, and a deep analysis based on
notional grounds (although in most cases, these would be the same). They
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
214 MICK O’DONNELL
In analyzing Process types and PRs, it doesn’t help to use the realm of
experience as a guide. And the analyst who has been forewarned of this
problem is less likely to fall into the trap of skipping the stage of applying the
tests for the Participant Roles, when trying to establish the Process type of a
clause.
(Fawcett 2009:215)
The point of this discussion has been to show that the issue of relative
importance of notional vs. structural criteria in the grammar is still an
open issue. Surveys of coding practice show practitioners range from more
to less notional in their coding of process type, and the four versions of IFG
are themselves in flux as to the importance of notional vs. structural
criteria.
The issues of what ‘genre’ is, and where it belongs in relation to the other
components of the linguistic model, have long been debated within SFL, and
the debate continues today between different parts of the community. This
section will discuss some of the main issues in this area.
One note on terminology: Halliday uses the term ‘register’ to refer to the
set of linguistic features that realizes a particular configuration of situ-
ational features. Martin, on the other hand, uses the term ‘register’ to refer
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 215
• In terms external to the text, generally the purpose of the text, e.g. to
persuade, to educate, to entertain, etc.
• In terms internal to the text, most typically in terms of common schematic
structures, or linguistic styles, e.g. editorial, narrative, anecdote,
report, etc.
‘Genre’ has been used in various places in both of these senses. Lee (2001:38)
describes the first approach, which makes a distinction between ‘genres’
(defined on external criteria) and ‘text types’ (defined on internal criteria):
In earlier works within SFL, ‘genre’ was not directly covered by the model,
although aspects relatable to external definitions of genre were mentioned,
placed within Context of Situation. This includes Halliday’s inclusion of
‘purpose’ and ‘rhetorical mode’, Ure and Ellis’ ‘role’, and Gregory and
Carroll’s ‘functional tenor’. These aspects will be discussed further below.
More recently, the internal definition of genre has been more prominent
within SFL: genres being defined as groupings of text with common text
structures. Hasan (1978:229), for instance, clearly takes this approach:
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
216 MICK O’DONNELL
Martin (and Martin and Rose) also indicate the use of internal criteria:
precisely because they are common sense everyday labels which aren’t
names of recurrent patterns of meaning (e.g. a poem can be almost any
genre in my terms – anecdote, report, description, narrative, procedure
etc. – where genre is a recurrent pattern of meaning).
(Martin 2015, personal communication)
We use the term genre in this book to refer to different types of texts that
enact various types of social contexts.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 217
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
218 MICK O’DONNELL
THE FIELD: Here we include the subject matter; and also the type of situation
in which language is used, including the purpose – e.g., didactic or explana-
tory, for information, for action, consolation or self-satisfaction.
(Halliday 1965:14)
The field of activity is ‘what’s going on’ in context . . . The activity is either
primarily social or primarily semiotic – i.e. either primarily a process of
interactive behaviour or one of exchanging meaning. To capture this, I have
called this parameter SOCIO-SEMIOTIC PROCESS.
the category used to describe what language is being used for in the situ-
ation. Is the speaker trying to persuade? to exhort? to discipline?
(Gregory and Carroll 1978:53)
Functional tenor was placed under tenor because it relates in effect to how
the interactants relate to each other (as the persuader and persuaded, etc.),
very external criteria. Note however that Gregory and Carroll state that
‘genre’ covers more than just functional tenor, involving also field, personal
tenor, and mode:
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 219
IFG4 notes that some rhetorical modes are oriented towards the Field of the
text (e.g. informative, didactic, explanatory and explicatory contexts), and
others towards the Tenor of the text (e.g. persuasive, exhortatory, horta-
tory, polemic contexts).
As a fourth component: Ellis and Ure (Ellis 1965; Ellis and Ure 1969; Ure and
Ellis 1977) have two distinct components in place of Halliday’s Tenor:
‘formality’ and ‘role’. Role corresponds roughly to genre, being defined as
‘the dimension correlating with the social or other role of the utterance or
text, e.g., conversation, literature, technical writing, etc.’ (Ellis 1965:13).
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
220 MICK O’DONNELL
approach, but students found this difficult, as they had earlier been exposed
to Halliday’s Rhetorical Mode approach (see Martin 2014). Class discussions
explored the cross-component implications of genre, and two class members,
Guenter Plum and Joan Rothery, suggested ‘positioning functional tenor as a
deeper variable, since the purpose of a text influenced all of interpersonal,
ideational and textual meaning’ (Martin 2014:12). The eventual result of this
discussion, under Martin’s leadership, was to rename the deeper variable as
‘Genre’, recognizing it as a stratum separate from Context of Situation.
‘Personal Tenor’ could thus be renamed as simply ‘Tenor’.
Genre in Martin’s approach is described in terms of both a system
network (defining genres and their variants), and a layer of structure, such
that choices in the genre network determine which schematic elements are
realized in the text. In his model, the register of each stage is determined
via interstratal realization:
As part of the realisation process, generic choices would preselect field, mode
and tenor options associated with particular elements of text structure.
(Martin 1992:505)
For Martin it seems, Genre only interfaces with Context of Situation (which
he calls Register), and does not interface directly with lower strata:
If this is so, then there must be some variables in the register which can pass
on the linguistic constraints of the genre to the lower stratum. The linguistic
patterns which were previously activated by functional tenor still need to be
activated by some variables in the Context of Situation (Martin’s Register
layer), and these variables need to be activated by the choice of Genre. The
alternative is to allow selections in the Genre stratum to directly interface
with each of the strata below: limiting the allowable contextual configur-
ations, in addition to activating linguistic possibilities. Both approaches are
viable, although each one has strong consequences for linguistic modelling,
and there should be a clear statement as to which approach is being followed.
Martin’s model of Genre over Register is fairly widely accepted within the
educational side of SFL. His approach is not uncritically accepted however,
particularly in regards to Hasan and those who follow her model (e.g. see
Hasan 1995). She argues that by putting Genre outside the semiotic space,
human interaction is de-humanized:
My own view is that the stratification of genre and register, the collapsing of
the social and the verbal, at both these planes, . . . has a highly deleterious
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 221
effect: It moves the whole issue of text structure and its activation from
active, feeling, reacting participants co-engaged in some interaction to given
forms of talk that represent the ways things are done in our culture, as if the
culture is unchanging and as if the participants are simply pre-programmed.
(Hasan 1995:283)
Martin responds that his Genre is in fact a semiotic system, open to human
choice:
Lukin et al. (2011:189) put forward a different argument, that placing genre
and register together provides a simpler analytical tool:
As a central conceptual tool that does not stratify the relation of genre and
register, Halliday’s notion of register helps us recognize – or at least frame
and test – the idea that recognized social situations might sometimes be the
same register, or identify and evaluate the register differences in what are
normally counted as ‘the same’ social activities: it is a model well suited to
calibrating the shuffling and reshuffling of cultural space-time and its
boundaries.
Martin (1999:505), on the other hand, argues that placing genre within
context of situation is just not feasible:
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
222 MICK O’DONNELL
register is the set of linguistic choices that recurrently occur in the texts
produced in a given situation type. So, by placing generic structure within
the framework of register, generic structures are thus linguistic patterns
that realize particular situation types.
One of the most likely places for generic structure in the linguistic model
is on the Semantic stratum. The closest thing to genre in Halliday’s later
model is ‘rhetorical mode’, which is a component of Mode. And because
Halliday often states that Mode is in most cases realized through the
Textual metafunction, we might infer that, for Halliday, generic structure
is a part of the Textual component of the Semantic stratum, along with
Information Structure and Thematic Structure.
For Hasan also, generic structure is a realization of the Context of Situ-
ation: the structure of a text is determined by the selection of features from
the Context of Situation network, which
can predict the obligatory and the optional elements of a text’s structure as
well as their sequence vis-à-vis each other and the possibility of their iteration.
(Halliday and Hasan 1989:56)
She stresses that one cannot expect elements of the text structure to be
determined by individual situational features, but that they are determined
by the configuration of features selected from Field, Tenor, and Mode (what
she calls a ‘contextual configuration’, or a CC):
We need the notion of CC for talking about the structure of the text because
it is the specific features of a CC . . . that permit statements about the text’s
structure. We cannot work from the general notion of, say, ‘field’ since it is
not possible to claim, for example, that field always leads to the appearance
of this or that element.
(Halliday and Hasan 1989:56)
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 223
The phrase ‘structure of the situation type’ implies that he takes generic
structure to be a structure level of the context of situation (context of
situation thus having system and structure specifications). He confirms this
two pages later:
The realization statements in Figure 6 are inter-axial but intra-stratal: that is, they
relate paradigmatic order to syntagmatic order within the stratum of context.
(Matthiessen 2015:12)
when we describe the linguistic details of texts in close detail, it is rare that an
entire text exhibits precisely the same range of stylistic options. More often we
can locate particular phases or segments of a text showing a relatively homo-
genous range of stylistic options and other segments of the same text that
show different options being taken up. Therefore, a single linguistic text, or
linguistic event, may appeal to several distinct registers while it is unfolding
and yet still be seen as a coherent example of a single ‘type’ of text. . . . Since
texts need not be homogeneous, simple ‘labels’ for registers or genres are
rarely appropriate. . . . Each stage can take on a distinctive register.
For Martin’s model, this is not a problem, as each schematic element can
be related directly to the register selections appropriate for that section.
He says:
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
224 MICK O’DONNELL
Making genre rather than register variables responsible for generating sche-
matic structure makes it easier to handle changes in experiential, interper-
sonal and textual meanings from [one] stage to another in a text . . .
Underlying register, genre can be used to predict these changes, stage by
stage, while at the same time accounting for a text’s overall coherence.
(Martin 1992:506)
For the Halliday and Hasan models, this is more of a problem. For them,
generic structure is created by the selected contextual configuration. But
there is no mechanism to allow for the element of the generic structure to
turn back and change the contextual configuration which would allow
variation in linguistic selection, although Martin (personal communication)
suggests that Hasan’s (2015) ‘ITERATION’ systems, which allow re-entry into
the Field network using a recursive system, might work for this. However, in
that article, the recursive system was applied to modelling situations with
multiple fields, not to modelling the staging of texts/interactions.
When the context is co-operatively negotiated, the text and context evolve
approximately concurrently, each successive message functioning as an
input to the interactants’ definition of what is being achieved.
These words suggest that the Context of Situation is not constant over a
given text or interaction, but can change as the text unfolds. Generic
staging in a text can thus be seen as the result of a sequence of shifts in
the context of situation, a change in what the participants are trying to
achieve at each point of the text (e.g. from motivating a study to detailing
that study). Each shift in Context of Situation is associated with a shift in
the register of the text.
One approach that takes this assumption most seriously is that of Phasal
Analysis (Gregory and Malcom 1981; Malcolm 2010), which allows for
phasal shifts in register throughout a text:
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 225
Phases differ from stages in that phases can overlap; for instance, we might
have a phase of consistent ideational selections (e.g. talking about Lego),
containing shifts in the interpersonal selections (change from monologue to
dialogue).
My own work on dynamic modelling of interaction (O’Donnell 1990,
1999; O’Donnell and Sefton 1995) takes a similar approach to the idea of
context as dynamically mutable, but that work was not trying to explain
the registerial shifts over stages. O’Donnell (2012) looked more deeply at
dynamic shifts in tenor over a text.
In dynamic approaches like these, we might do away with modelling
generic structure as such, seeing the apparent staging as the result of the
dynamic shifts in the Context of Situation, the interactants’ notion of what
is going on. As the context shifts throughout the interaction/text, as a
result, the register shifts as well. We do not need to posit text structure
intermediate between context and text.
We are left however with the problem of modelling the process of how
shifts in the context of situation take place, both in dialogue and in written
text. Cloran (1987) offers an interesting discussion of how contextual shifts
in interaction can be negotiated by the participants. Much work is however
needed to apply this dynamic context perspective to describe registerial
shifts in written texts, such as are usually explained by generic structure.
8.3.7 Summary
The problem of where Genre belongs in the SFL model stems from the
seemingly circular relation between Genre and Context of Situation. On
the one hand, we can say that the Context of Situation determines whether
a given Genre (or generic structure) is appropriate or not, and more deeply,
particular variants of a generic structure potential may be activated or
deactivated by particular contextual features. On the other hand, each
element of a generic structure is associated with distinct language patterns,
so, to this extent, the stages of the genre determine the register used
within them.
In the traditional SFL approach, the Context of Situation, and thus Regis-
ter, is seen as something constant over a text as a whole. And given the need
to account for staging of language patterns over a text, the traditional
approach thus needs to account for this staging outside of context. Hasan
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
226 MICK O’DONNELL
has taken one approach, placing the staging as a realization of the Context
of Situation. But taking that approach makes it difficult to account for the
distinct micro-registers of language use in each stage. Martin has chosen to
place the notion of genre, and thus of staging, above that of Context of
Situation, and thus allows each generic stage to reflect distinct registerial
patterns. An alternative approach avoids the problem by rejecting the
assumption that the Context of Situation is static, allowing for micro-shifts
in register to result from a dynamically shifting context.
This chapter has explored two areas of interest within SFL that are still
unresolved, that of the degree of notionalism in determining grammatical
categories, and also the exact role and nature of genre within the model.
Both of these areas lead to active discussions on the SFL discussion lists, at
conferences, and in publications. Often these discussions contain confu-
sions where participants see the issue through the lens of their own
assumptions, not aware of the underlying issues that lead their very words
to mean different things to different readers. It is the hope of the author
that, by bringing these underlying issues to the surface, future discussion
will be less distracted by mistaken interpretations, and driven more
through mutual understanding of the different sides of the issues.
References
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 227
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
228 MICK O’DONNELL
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
Continuing Issues in SFL 229
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:59, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.010
9
The Cardiff Model
of Functional Syntax
Anke Schulz and Lise Fontaine
9.1 Introduction
Within the theory of Systemic Functional Linguistics (SFL), there is, what
Halliday refers to as the ‘powerhouse’ of the theory, the ‘central processing
unit’ where ‘meanings are created’ (Halliday 1994:15). Despite this central
role in the theory, there have been relatively few developments specifically
related to the lexicogrammar since the 1980s and 1990s. There has been,
however, one concerted effort to promote debate in this area and to suggest
theoretical developments to the grammar by a team of scholars at Cardiff
University (e.g. Fawcett 1980, 2000a, 2008a, 2017; Tucker 1998, 2017; Tench
1990, 1996, 2017). This chapter provides an overview of the Cardiff
approach to syntax within the SFL framework. What we might refer to as
the ‘standard’ model of SFL1 is described in various chapters in this volume,
notably, Berry, Butt, Fontaine and Schönthal, Taverniers, and Webster, and
is represented in various other chapters throughout the rest of the volume.
There is also a very useful comparison between the more ‘standard’ model
and the Cardiff model in Butler (this volume), but see also Butler (2003a,
2003b) and Butler and Gonzálvez-García (2014). While such comparisons
are valid and important, the focus in this chapter is not on comparing the
two models, although where appropriate, important issues are mentioned.
The Cardiff model has its basis in SFL theory and, in particular, in
Halliday’s earlier work. While many of the principles are shared between
the two models, Butler (2003a:153) points out that ‘there are in the Cardiff
account some important differences in the underlying goals, as well as
extensions and simplifications of the grammar itself’. There have been
several key concerns that have driven the model, and while a full
1
For ease of reference, the term ‘standard model’ will be used to refer to the more widely known model of
grammar, e.g. Halliday and Matthiessen (2014) or any of the editions of IFG, Introduction to Functional Grammar
(e.g. Halliday 1994 or Halliday and Matthiessen 2004), also sometimes referred to by some as the Sydney grammar.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 231
All SFL theory can be said to stem primarily from Halliday’s (1961) most
important and influential article, ‘Categories of the Theory of Grammar’.
Butler (2003a:153) acknowledges, as Fawcett himself does, that there
are more similarities than differences between the two models.
While the Cardiff model is very clearly rooted in Halliday’s early work,
there are points of divergence that have shaped the path leading to the
Cardiff model. Fawcett (2010:93) explains the different pathways as
follows: developments in SFL theory, from early work in the 1960s, were
shaped by Halliday’s involvement with the Penman project (see Matthies-
sen and Bateman 1991) and all the work that Halliday has done since then,
whereas the Cardiff model shares the same roots but diverges slightly,
being influenced by Hudson (1971) and work by Fawcett (1973, 1980).
Fawcett considers Halliday’s 1970 paper, ‘Language as Choice in Social
Contexts’, as resembling most the Cardiff Grammar (Fawcett 2010:93).
The Cardiff model was also shaped by a computational implementation,
the COMMUNAL project (e.g. Fawcett et al. 1993). Butler and Gonzálvez-
García (2014:49) suggest that, because of this, the model ‘offers a high
level of explicitness’. However, as Fawcett states (2008a:13), both models
share ‘the same historical roots and they still share essentially the same
basic concepts’.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
232 ANKE SCHULZ AND LISE FONTAINE
The theoretical framework of the Cardiff model has been described in many
different publications but notably in Fawcett (1980, 2000a, 2008a), Tucker
(1998), and Neale (2002), as well as in many articles and book chapters. It
has been and is being developed around the world and in different lan-
guages, for example, in work on Chinese (He 2014), German (Schulz 2008,
2015), and Japanese (Funamoto 2014).
For Fawcett (2000a:34), the basic relationship between meaning and form
in any sign system can be described in terms of realization (i.e. meaning is
realized by form). The relationship between meaning and form is illustrated
in Figure 9.1, where the system networks are a components of the gram-
mar, representing the semantic options available to speakers. The output
of the networks is a set of selection expressions which then becomes the
input to the realization rules; the realization rules and the potential struc-
tures are another component of the grammar, also expressed as potential.2
In this sense, as Butler (2003a:185) explains, ‘The level of form is also
regarded as having a potential, consisting of realisation rules’. The output
from this component is a layer of richly labelled tree structures. As shown
in the diagram in Figure 9.1, there is a loop enabling this process to
continue, where it is possible for a realization rule to state a re-entry rule,
for example, when an element of a unit is ‘filled’ by another unit (this will
be made clear in Section 9.4).
2
The relationship between meaning potential and instance is described by the concept of instantiation, which
operates in a different dimension from that of realization. Although this is an important distinction, it will not be
discussed in this chapter (see e.g. Wegener 2011).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 233
Figure 9.1 The main components of a systemic functional grammar (e.g. Fawcett 2000a:36)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
234 ANKE SCHULZ AND LISE FONTAINE
Figure 9.2 Partial view of the main generative components of the COMMUNAL
computer model
components, the sentence planner produces the best formal and semantic
representation.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 235
This basic logical form is then the input into the next stage of planning,
which is called the microplanner. The microplanner is a component that
handles various algorithms that guide the choices (the selection of options)
in the system networks. The system networks represent the networks of
systems of semantic options, not decision trees. This is a very important
component, yet very little attention has been given to this area in SFL. It is
broadly accepted in natural language generation that such a component
is necessary. However, there are still many unanswered questions as to how
it should work and what parts of the generation process belong in the
microplanner and what parts belong elsewhere. The microplanner is simply
a set of algorithms that determine the selection of options in the system
networks; for example, the selection of Theme or verb tense.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
236 ANKE SCHULZ AND LISE FONTAINE
1. rule number;
2. network feature(s);
3. any conditions on the rules;
4. rule operations.
For example, in the small system network presented in Figure 9.3 below, if
the feature [situation] is selected (as it would need to be in order to
generate a clause), then as Fawcett (2008a:100) explains, the correspond-
ing realization rule is to insert a clause and within the clause to insert a
main verb (see Section 9.4). If the feature [information giver] is selected
then the realization rule specifies that the Subject must be positioned
before the Operator.
Realization rules may be simple or complex. For example, the realization
rule for the feature [thing] in the system network for thing, or referent-as-
thing, is given by Fawcett (1998) as the following:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 237
Figure 9.3 A very small systemic functional grammar for the English clause (Fawcett
2008a:93)
60: thing:
if congruent_thing then ngp,
if minor_relationship_with_thing then pgp.
This example shows the rule number (60), the system feature ([thing]),
and the operation (insert unit, i.e. ‘unit insertion rule’). This rule handles
the difference between examples such as the woman and to the man as
in the woman gave the ticket to the man. In the first case, the woman, rule
60 would insert a nominal group. In the second case, it would insert a
prepositional group for to the man. A more complex rule will have condi-
tions such as in the following example, which covers the realization of the
system feature of [prediction] and [future time reference point] (Fawcett
2008b):
Rule (5) applies when either [prediction] or [future time reference point]
has been selected. It also describes the conditions of realization depend-
ent on whether or not the system feature [negative] has also been
selected. If it has not been selected, the Operator will be expounded by
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
238 ANKE SCHULZ AND LISE FONTAINE
will, and if it has, the Operator will be expounded by wo. This will be
combined with the realization rule for [negative], which is expressed as
follows (and also applies when the feature [confirmation seeker] is
selected):
The final realization then for [prediction] and [negative] will be that the
Operator will be expounded by won’t, as in He won’t phone. It is therefore
the features of the system network and the realization rules which create
the formal representation. As Fawcett (2000a:149) explains,
The first stage is the selection of the features in the system network (i.e., the
creation of a selection expression). The realization rules then integrate the
various partial ‘strands of meaning’ that are represented by these features
into a single functional structure.
As the examples above show, the realization component has the function of
integrating the selected semantic features and operationalizing the realiza-
tion of the form.
It is also possible for a realization rule to specify another semantic
feature that must be selected on a future pass or on the same pass
through the network. In the former, a preference is given for a par-
ticular feature on a future pass. This is one way in which preselection
occurs, where a realization rule preselects other features. According to
Tucker (1998:46), ‘The preselection of one feature from a system is
equivalent to stating that in the relevant context there is no systemic
choice, and consequently no choice in meaning or in structural
realization.’
The first pass through the networks deals with the clause ([situation]),
whereas the second pass typically will serve to fill an element of the
clause such as Subject (see Section 9.4 for the concept of filling). Subject
is most commonly filled by a nominal group ([thing]) and the selection of
[thing] would have been preselected from a realization rule determined
by the selection of certain semantic features on the first pass. Preselec-
tion may also occur if the planner has passed down an instruction to
select a particular feature (e.g. validity assessment). Some pass rules
cover relationships between features which can be thought of as a kind
of co-dependency. For example, certain process types are far more likely
to prefer progressive aspect (e.g. Material or action type processes) and
others are far more likely to disprefer progressive aspect (e.g. Mental
processes). In other words, action type processes in the present tense are
more likely to be progressive than are mental type processes (consider
John is bouncing the ball versus John bounces the ball and John sees the bird
versus John is seeing the bird).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 239
9.3.3 Probabilities
The role of the probabilities in the system networks is basically two-fold.
One is that they reflect tendencies or the frequency distribution of the fea-
tures in a given system. The other is that they can be set (for example, by a
realization rule) to preselect (effectively remove the ‘choice’ from the system)
or exclude a particular systemic feature. In the generation of a referring
expression, for example, the selection of the feature [thing] is preset to
100 per cent. Figure 9.4 illustrates a simplified system network for the ‘infor-
mation’ sub-network of the Mood system network, where the semantic fea-
tures are shown together with their associated probabilities and examples of
the realization for each feature. As Fawcett (2008b) explains, ‘Probabilities . . .
change from one context of situation to another, and furthermore . . . they
are often overridden by the requirements of the performer’s current commu-
nicative purposes.’ Therefore, the probability associated with a given seman-
tic feature is not fixed but will change depending on many other variables.
Figure 9.4 A simplified system network for the ‘information’ sub-network of MOOD
(Fawcett 2008a:157)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
240 ANKE SCHULZ AND LISE FONTAINE
experiential TRANSITIVITY
CIRCUMSTANCES
CONTROL & DISPOSITION
TIME
logical relations COORDINATION
SUBORDINATION
EXTERNAL LOGICAL RELATIONS
interpersonal MOOD
negativity POLARITY
validity BASIC & AUXILIARY VALIDITY
ADJUNCTIVAL VALIDITY
affective AFFECTIVE ADJUNCTS
thematic SUBJECT THEME
NON-S PR AS MARKED THEME
ADJUNCT THEME & INTEGRATION
ENHANCED THEMES
informational RECOVERABILITY
UNMARKED NEWNESS
CONTRASTIVE NEWNESS
INFORMATION STATUS
These system networks are not detailed enough to illustrate the process
of traversing the network, but the full system network is too large to
present here. Instead, as a brief example, we will consider some of the
semantic features that would be selected in order to generate a simple
clause such as I drove the car (invented example), using the system networks
in the computational implementation of the Cardiff Grammar as presented
in Fawcett, Tucker, and Lin 1993. The first selection is the feature [situation]
(other features are possible, such as [thing], but [situation] is the semantic
feature for the clause). This then leads to eight parallel systems,3 corres-
ponding to the eight major strands of meaning in the Cardiff model
(see Table 9.1): transitivity, time specification, situation coordin-
ation, subordination type, mood, polarity, logical circumstance,
and information focus. Each of these systems presents the semantic
options available within that system, and each will lead to further systems.
Selections are made for each system, and this process is continued until a
terminal feature is reached for all relevant systems.
Then in the transitivity system (see Neale 2002, 2017), a process type
must be selected, and in our case it would be an action process type,
specified as driving. In addition to this and as part of the transitivity system,
subject-theme is selected, which must be considered here in order to
3
Names of major systems are given in small caps, features are given in square brackets.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 241
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
242 ANKE SCHULZ AND LISE FONTAINE
This brief example has been used to give readers an idea of the steps
involved and the relationship between the system network and realization.
These are the important concepts which underpin the Cardiff approach to
syntax.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 243
to units (see Fontaine and Schönthal, this volume), the Cardiff model
defines them by their internal structure (see ‘type’ vs. ‘class’ in Halliday’s
early writing; also see Berry 1975; Fontaine and Schönthal, this volume),
i.e. the components that constitute the unit. These components are
referred to as functional elements, which form the basic categories of
the functional syntax in the Cardiff Grammar. They are illustrated in
Figure 9.5. The relationships among the various components are shown
in Figure 9.6. The relationships among them are explained by Fawcett
(2008a:76) as follows:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
244 ANKE SCHULZ AND LISE FONTAINE
Figure 9.5). When an element is not filled by a unit but rather expounded by
an item, the relationship is that of exponence (see Figure 9.6).
Whereas for Halliday and Matthiessen (2004:54) the elements of the
clause are word groups and phrases such as nominal group, verbal group,
prepositional phrase, and adverbial group, in the Cardiff Grammar, the
elements of the clause are functional elements which are either filled by a
unit of structure or expounded by an item. For example, Subject is most
likely to be filled by a nominal group, and the Main Verb is most likely to be
expounded by some lexical verb.
According to Butler (2002:75–6), it is ‘the relationships of filling and
exponence (that) make possible the realization of recursion in the gram-
mar. Recursion is located in the systemic component of the grammar,
leading to realization rules which specify that the system network must
be re-entered.’ It is therefore the re-entry into the system network that
specifies the recursion and specifically filling that captures this process.
There is another important distinction which is inherent in the repre-
sentation of the structure of the clause. This is the role of the formal
structure of the clause. In the Cardiff Grammar, the three main strands
of meaning are integrated into the overall model to such a point that they
have no formal role in the grammar. For Halliday, Subject, for example, is
not an element of structure, although this was its original proposal in
earlier work (notably Halliday 1961). In the standard SFL model, Subject,
Predicator, Complement, and Adjunct are treated as ‘the “secondary” struc-
ture of interpersonal meaning’ (Fawcett 2000b:154).
The full list of the strands of meaning is given in Table 9.1, where each
strand of meaning is associated with the unit which expresses it. The main
elements of the clause are given in Table 9.2. Most of the work done on
these strands of meaning specifically is only available in the form of unpub-
lished reports, but see Huang (2017), who has worked extensively on the-
matic meanings; Neale (2002, 2017) on transitivity; and of course many of
the publications listed in the references by Fawcett and also by Tucker.
Table 9.2 The main elements of the clause in the Cardiff Grammar
Element Example
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 245
(&) (rd) (v) (pd) (qd) (v) (sd or od) (v) (dd) (m)* h (q)*
Table 9.3 Examples taken from Fawcett (2007) of individual elements of the
nominal group [note: element being sampled is in italics; head is
underscored; each example may be composed of more than one element]
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
246 ANKE SCHULZ AND LISE FONTAINE
Figure 9.7 A common configuration of the nominal group in the Cardiff Grammar
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 247
Key
pgp: Prepositional Group
tp: prepositional temperer
p: preposition
cv: completive
Figure 9.11 Full tree diagram showing a nominal group, a quality group, and a
prepositional group
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
248 ANKE SCHULZ AND LISE FONTAINE
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 249
ad am qtf
sixty
about sixty
more than I had expected
Key:
ad = adjustor; am = amount; qtf = quantity finisher
Key:
po = possessor; g = genitive element;
o = own element
Figure 9.14 Diagram showing the generic structure of the deictic determiner
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
250 ANKE SCHULZ AND LISE FONTAINE
Table 9.8 Main processes and participant roles in the Cardiff Grammar
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 251
hair. In this example, the Subject is conflated with Agent and the Comple-
ment is conflated with Affected. Depending on the process type, Subject
will most often be conflated with some Participant Role (although not
always), while Complement is always conflated with a Participant Role.
Adjuncts may be conflated with a Circumstance Role. It is through confla-
tion that the relationship between syntax and semantics is made explicit
in the Cardiff Grammar.
It is also possible for an element to be conflated with more than a single
Participant Role. Both the standard SFL model and Cardiff SFG include
compound Participant Roles where a single element of structure expresses
a compound role in transitivity, although this is not generally accounted for
by conflation as it is in the Cardiff model. An example of this is given in (2),
where one referent expresses a compound role. The key to the abbreviations
is as follows: [S]=Subject, [M]=Main Verb, [C]=Complement, [Ag]=Agent,
[Af]=Affected, [Ca]=Carrier, and [At]=Attribute. Compound roles are hyphen-
ated. Consequently, in (2), Ivan expresses the conflated roles of Affected
and Carrier.
(2) The war [S/Ag] made [M] Ivan [C/Af-Ca] very rich [C/At] (from Fawcett
forthcoming)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
252 ANKE SCHULZ AND LISE FONTAINE
transitivity configuration for each: in (3), two PRs, I and a dog (someone has
something); in (4), one PR, I (someone is having a bath); in (5), three PRs, I,
my keys, and in the desk (someone is putting something somewhere); in (6),
two PRs, I and a cake (someone is baking something); and in (7), we might
debate whether the process test can be used definitively to work out the
‘expected’ PRs, since it is not immediately obvious that bake as a process
requires a third PR (see Fawcett 2009). While the test is not perfect, it does
help provide a framework with which to build the analysis. However, the
process test can only be applied if the main verb has been identified (see
steps 3 to 5 below).4
The Cardiff model has built up a set of guidelines that help to identify the
elements of a clause (see Fawcett 2008a:208–31). Here, we provide a brief,
slightly adapted version of the guidelines to illustrate the main principles in
the syntactic analysis of the clause.
Guidelines for analyzing clause syntax:
4
There is insufficient space here to explain how to identify whether or not the main verb is extended: in e.g. I’m looking
up the answer, is this a process of ‘looking’ or a process of ‘looking up’? See Fawcett (2009; forthcoming).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 253
These steps will be used to analyze the clause given in (8), first by establish-
ing the structural configuration of the clause and then by discussing the
full functional account within the Cardiff model.
(8) my Mom would visit Texas from time to time (Source: Ententen13,
SketchEngine)
In our example (see (8) above), the clause is in the declarative structure (step
1); it has no Linker or Binder (step 2); the main verb is visit (step 3); and there
is a modal auxiliary verb would (step 5). Having identified the main verb, the
process test can be applied (step 4), which provides the basic syntax of the
clause; in a process of visiting we expect someone to be visiting someone/
somewhere, i.e. two PRs. Considering steps 6 and 7, we can conclude that
there are not auxiliary extensions since there is only one auxiliary verb in
this clause, and that the auxiliary is conflated with the Operator since it has
the potential to express mood. There is no Negator element (step 8), and we
can identify the subject (step 9) because the Operator (would) would be to the
left of the subject in an interrogative, e.g. Would my mother visit Texas from
time to time? Having done the process test in step 4, we can identify a
Complement, i.e. the non-subject PR, to complete step 10. So far then, we
have a two-participant process with the basic syntax as follows: S O/X M C,
i.e. Subject, Operator/Auxiliary, Main verb, and Complement. There is one
unit unaccounted for, and so in step 11 we assume that from time to time is
an Adjunct. None of the elements listed in steps 12 to 14 apply, therefore
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
254 ANKE SCHULZ AND LISE FONTAINE
Figure 9.17 Analysis of the syntax and semantics of example (7) in the Cardiff Grammar
(based on Fawcett 2010:148)
Key: Cl = Clause; S = Subject; Ag = Agent; O = Operator; M = Main verb; C =
Complement; Af = affected; A = adjunct.
our final syntax is given in Figure 9.16. While this gives us the basic
structure, we do not yet have a full account of the semantics nor of the
nature of the structural units which fill these elements of the clause. If we
recall the concepts of componence, filling, and exponence as discussed
above (see also Figure 9.6), we can say that this clause is composed of five
elements. The subject, complement, and adjunct elements are filled by
units of structure, in this case by a nominal group for both the subject
and complement and by a prepositional group for the adjunct. The
remaining two elements, operator and main verb, are each expounded, by
a modal auxiliary verb and a lexical verb respectively. However, as shown in
Table 9.1, there are eight strands of meaning yet to be analyzed.
While there is no space here to discuss in detail how the semantic
analysis is handled in the Cardiff model, we will simply briefly highlight
a few important points about it. Figure 9.17 illustrates the analysis of the
syntax and the semantics. In a sense this diagram brings together the key
concepts of the Cardiff model. As Fawcett (2010:148) explains, ‘A clause is
regarded as the realization, in a single, integrated structure, of the various
types of meaning that are modelled in the system networks
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 255
9.6 Summary
In this chapter we have shown how the Cardiff model developed from
Halliday’s early work, diverging from Halliday’s standard model due to
different influences and concerns; however, the core concepts are clearly
shared. The priority given in the Cardiff model to conflating meanings
rather than structures has informed the approach it has developed, and
we have shown in this chapter how that works. In addition to this, the role
of realization is made explicit in the Cardiff model through the relationship
between the system networks, selection expressions, and realization rules
through to the structural representation of these meanings. Here we have
demonstrated how these two key aspects of the model have shaped its
development in terms of how it applies to the English language.
Limits of space have not permitted a more extensive presentation of the
Cardiff model, and that was never the aim of the chapter. We set out to give
some insight into the model and interested readers can continue to read
about it by following up with the various publications referred to in this
chapter.
References
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
256 ANKE SCHULZ AND LISE FONTAINE
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
The Cardiff Model of Functional Syntax 257
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
258 ANKE SCHULZ AND LISE FONTAINE
Lin, Y., R. Fawcett, and B. Davies. 1993. GENEDIS: The Discourse Generator
in COMMUNAL. In A. Sloman, D. Hogg, G. Humphreys, A. Ramsay, and
D. Partridge, eds., Prospects for Artificial Intelligence: Proceedings of AISB 93,
the 9th Biennial Conference of the Society for the Study of Artificial Intelligence
and the Simulation of Behaviour. Amsterdam: IOS Press. 148–57.
Matthiessen, C. M. I. M. and J. Bateman. 1991. Text Generation and
Systemic-functional Linguistics: Experiences from English and Japanese. London:
Pinter.
Neale, A. 2002. More Delicate TRANSITIVITY: Extending the PROCESS TYPE
System Networks for English to Include Full Semantic Classifications. PhD Thesis,
Cardiff University.
Neale, A. 2017. Transitivity in the Cardiff Grammar. In T. Bartlett and
G. O’Grady, eds., The Routledge Handbook of Systemic Functional Linguistics.
London: Routledge. 178–93.
Quirk, R., S. Greenbaum, S. Leech, and J. Svartvik. 1985. A Comprehensive
Grammar of the English Language. London: Longman.
Quiroz, B. 2017. The Verbal Group. In T. Bartlett and G. O’Grady, eds.,
The Routledge Handbook of Systemic Functional Linguistics. London: Routledge.
301–18.
Schulz, A. 2008. Tense, Modality and Polarity: The Finite Verbal Group in
English and German Newsgroup Texts. In N. Norgaard, ed., Systemic
Functional Linguistics in Use. Odense: OWPLC 29. 697–716.
Schulz, A. 2015. Me, Myself and I: A Corpus-based, Contrastive Study of English
and German Computer-mediated Communication from a Systemic Functional
Perspective. PhD Thesis, Technische Universität Darmstadt.
Tench, P. 1990. The Roles of Intonation in English Discourse. Bern: Peter Lang.
Tench, P. 1996. The Intonation Systems of English. London: Cassell.
Tench, P. 2017. The Phoneme and Word Phonology in SFL. In T. Bartlett and
G. O’Grady, eds., The Routledge Handbook of Systemic Functional Linguistics.
London: Routledge. 233–50.
Tucker, G. 1998. The Lexicogrammar of Adjectives: A Systemic Functional
Approach to Lexis. London: Cassell.
Tucker, G. 2017. The Adjectival Group. In T. Bartlett and G. O’Grady, eds.,
The Routledge Handbook of Systemic Functional Linguistics. London: Routledge.
284–300.
Wegener, R. 2011. Parameters of Context: From Theory to Model and Application.
PhD Thesis, Macquarie University.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:03:22, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.011
10
SFL in Context
Christopher S. Butler
This chapter situates SFL in what has been called functional-cognitive space,
a multidimensional space based on a wide range of properties, in which
various functionally and/or cognitively oriented and/or constructionist
approaches to language can be plotted (Butler and Gonzálvez-García 2005,
2014; Gonzálvez-García and Butler 2006). The aim is to highlight significant
similarities and differences between SFL and other functional/cognitive/
constructionist models.1
1
Considerations of space permit only a brief summary of this work here: for a detailed argument, backed up by
copious references to the literature, readers are referred to Butler and Gonzálvez-García (2014).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
260 CHRISTOPHER S. BUTLER
2
For information on these approaches to language see the brief profiles of each in Butler and Gonzálvez-García
(2014: Chapter 2) and the key references given there.
3
The term ‘Sydney approach/model’ is used here as a convenient label for the model of SFL put forward by Halliday
and his colleagues, while ‘Cardiff approach/model’ is used to refer to the model advocated by Fawcett and his
co-workers. My view is that although the two approaches share a large number of aims and assumptions, the Cardiff
model should not be seen simply as a minor offshoot from what some consider to be ‘mainstream’ SFL, but deserves
to be considered on a par with the Sydney model.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 261
correlated significantly with each other and with those for one of the
Cardiff questionnaires, but not with the ratings for either the second
Cardiff questionnaire or those for any other model.4 Neither did the ratings
for the second Cardiff questionnaire correlate significantly with those for
any other model. Using a less conservative technique, but one which is not
backed up by rigorous significance testing, all four sets of SFL ratings
emerged as correlated with each other, and similarities with the LCM
ratings were also revealed, as well as a negative correlation with one of
the MP questionnaires. These results already suggest that SFL is very dis-
tinctive in relation to the other models studied in our research, and we will
see below that this conclusion is corroborated through other techniques.
In Figure 10.1, taken from Butler and Gonzálvez-García (2014:188), is
shown the map produced by the multidimensional scaling programme
PROXSCAL which forms part of the IBM SPSS Statistics package.5,6 This
analyzes the distances between the sets of questionnaire responses and pro-
duces a visual representation of their relationships. Figure 10.1 shows an
analysis along just two dimensions, though higher numbers of dimensions
4
Lack of space precludes discussion of the methodology used in the statistical analyses. For details, the reader is referred
to Butler and Gonzálvez-García (2014: Chapter 4).
5
I am grateful to John Benjamins Publishing Company for permission to reproduce the diagrams labelled here as
Figures 10.1 to 10.4, taken from Butler and Gonzálvez-García (2014).
6
The questionnaires labelled as EG (Emergent Grammar), UBT (Usage-Based Theory), and EXT (Exemplar Theory) are
those which are brought together as EG+ in the discussion and as EG_PLUS in the final analysis illustrated in Figures 10.3
and 10.4.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
262 CHRISTOPHER S. BUTLER
are also possible. It can be seen that the four SFL questionnaires cluster
together in the bottom left quadrant of the plot, with negative values on both
dimensions of the analysis. Dimension 1 is clearly concerned with the extent
to which a model resembles formalist approaches: the questionnaires with
positive values on this dimension (MP2, MP1, FDG3, RRG2, RRG1, SBCG,
FDG1, FDG2, in descending order) are for the most formally oriented models,
while the model with the most strongly negative value (the Emergent Gram-
mar of Hopper: EG) is radically functional, and the others with negative
values, apart from SFL, are cognitively based.7 Dimension 2 of the plot is
harder to interpret linguistically (for further analysis see Butler and Gonzál-
vez-García 2014:189).
7
MP1 and MP2 refer to the two MP questionnaires, FDG1, FDG2, and FDG3 to the three FDG questionnaires, and
so on.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 263
Figure 10.3 Multidimensional scaling plot for the final ratings data
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
264 CHRISTOPHER S. BUTLER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 265
10.4.2 Coverage
We now turn to what aspects of the study of language the model is intended
to cover. An important distinction is that between approaches which aim to
offer as full a model as possible of language as a whole, and those which
aim primarily to offer simply a model of grammar in the wide sense, to
include semantics, pragmatics, morphosyntax, and phonology. The four
SFL questionnaires were unanimous in agreeing that the model goes
beyond the grammar, and this is amply borne out by the voluminous
literature on, for example, context, register, and genre, as well as on the
‘appliability’ of descriptions based on theoretical constructs. The only
models which were given a negative rating for this feature in the final
analysis were FDG, RRG, and CS, all of which were deemed to concentrate
largely on the grammar itself.
A related question is whether the model aims to account for all system-
atic phenomena within its chosen domain, or just a ‘core’ as in formalist
approaches. SFL certainly does not restrict itself to a putative ‘core’, and this
is reflected not only in the unanimous questionnaire responses, but also in
the impressive breadth of coverage evidenced even by standard works such
as Halliday and Matthiessen (2014) and Fawcett (2000, 2008). Halliday
(2009:73) comments that ‘SFL places a high value on comprehensiveness
in description’.
A further important distinction between approaches is whether they
build in models of the processes by which speakers and writers produce
and understand language, as well as of the systematic patterns found in
language. This item in the questionnaire resulted in a split between the
SFL respondents, the Sydney approach experts giving a negative response,
the Cardiff ones a positive score. The SFL literature clearly shows that those
working within the Sydney approach do not consider themselves to be
modelling the minds of speakers and hearers, but rather the options
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
266 CHRISTOPHER S. BUTLER
8
I have argued elsewhere (Butler 2013) that these two orientations are not mutually exclusive, i.e. that both cognitive
and social considerations are important when considering matters of categorization and construal.
9
SFL uses the term ‘nominal group’ in place of the more usual ‘noun phrase’ (also see Fontaine and Schönthal, this
volume).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 267
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
268 CHRISTOPHER S. BUTLER
In the Cardiff model, no fewer than eight major strands of meaning (experi-
ential, interpersonal, thematic, logical relations, polarity, validity assess-
ment, affective, and informational) are recognized at the semantic level
(Fawcett 2008:245).
The final ratings for the questionnaire item concerned with interpersonal
phenomena were equally distributed between positive and negative scores:
for FDG, Givón, EG+, WG, CG, RCG, and LCM, in addition to SFL, this area is
important, while the remaining models pay comparatively little attention
to it. For the importance of information structuring, FDG, RRG, EG+, SBCG,
CCG, LCM, and PA, in addition to SFL, were given positive ratings, the other
models negative ones.
Finally in this section, the questionnaire asked whether detailed atten-
tion was given to the non-discreteness of language. Both Sydney SFL
respondents gave positive answers, as did one of the Cardiff respondents.
The literature shows that Sydney SFL, in particular, recognizes the ‘messi-
ness’ of linguistic categories, but does not deal with it through the concept
of prototype, as in more cognitively oriented models (see Halliday 2003a:2,
28). One important aspect of indeterminacy in SFL is the use of the cline, a
continuum which permits potentially infinite gradation (Halliday
1961:249). This concept is important in, for example, the arrangement of
systems in ordering of delicacy, from quite general ones to the left of the
network to more detailed and specific ones to the right. It is also involved in
the dimension of instantiation, which relates the grammar to other aspects
of language and is modelled as a cline with the whole potential of a
language at one end and specific instances of language use at the other.
A further mechanism for dealing with non-discreteness is to supplement
the typology inherent in system networks with a topological perspective
which operates with the idea of a flexible, multidimensional space in which
categories can be located (Martin and Matthiessen 1991). Added to these
mechanisms is the importance of probability: probabilities are assigned to
choices from systems, and differ with the variety of language under scru-
tiny. This is true for both the Cardiff and the Sydney approaches (see
Fawcett 2008:18–19, 122; Matthiessen 2014). Of the other models investi-
gated in our study, only FDG, RRG, CS, SBCG, and CLS were seen as not
giving detailed attention to non-discreteness. It will be apparent that apart
from SFL, the cognitively oriented models are the ones which pay most
attention to this area. CLS is the exception, because although cognitive in its
orientation, its scope is largely limited to the interplay of grammatical and
lexical phenomena.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 269
tool, it might be expected that the database for description would include
samples of attested naturally occurring language use. We saw earlier that
SFL pays a great deal of attention to the structuring of texts, and it is
therefore not surprising that all four SFL respondents agreed that samples
of authentic text, such as corpora, text collections, or individual texts, do
indeed form an important part of the data used (see Halliday and Matthies-
sen 2014:51–3, 69–74). In the final analysis of the models, taking into
account the literature as well as questionnaire findings, only RRG, WG,
ECG, and PA were rated negatively for this feature.
The next three questions were concerned with the range of languages
from which data are taken and the approach to language typology favoured
by the model. The SFL respondents’ ratings were mixed for the question
concerned with whether, during the development of the approach, applic-
ability to the whole range of language types found in the world was a major
consideration: one Sydney and one Cardiff respondent gave positive ratings,
the other two ratings being negative. However, all respondents agreed that
data from a range of languages were not normally used in arguing for
specific theoretical constructions, but rather that there was a preference
in SFL for describing single languages in some detail and only then making
comparisons. These opinions are amply confirmed by the SFL literature: see
e.g. Caffarel et al. (2004). In the final table of ratings, only SFL, CS, SBCG,
CLS, and LCM were given negative scores for the importance of applicability
to all language types. Ratings for the other two questions were predictably
totally complementary: FDG, RRG, Givón, EG+, CG, and RCG use data from a
wide range of languages in arguing for particular theoretical constructs,
whereas SFL, WG, CS, SBCG, CCG, ECG, FSCG, CLS, LCM, and PA tend to
describe particular languages in detail before making comparisons.
All SFL experts also agreed that data from extended stretches of discourse
are used as well as individual sentences or utterances, as expected from a
text-oriented model. The only other models which were rated positively in
the final analysis were Givón, EG+, CS, and ECG.
Three out of the four SFL respondents expressed the view that data from
sociolinguistic and sociological studies are not widely used in arguing for
particular constructs, the exception being one of the Cardiff linguists. This
is an interesting result in the light of the fact that Hallidayan linguistics
treats language as a social semiotic, and the Cardiff linguists aim to integrate
cognitive and sociological aspects of language and its use. The clue to this
apparent anomaly was given by one Sydney model respondent, who com-
mented that SFL is interested primarily in social practices as construed in
language, so that social matters are approached largely if not entirely
through the analysis of language itself. Of the sixteen models we investi-
gated, only WG and ECG were ultimately rated as making significant use of
the sociolinguistic and/or sociological literature.
The questionnaire also probed the use of data from different synchronic
varieties (dialects, registers) in arguing for particular theoretical constructs.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
270 CHRISTOPHER S. BUTLER
Here, all but one of the SFL respondents agreed that such data were of
importance, in conformity with the long history of engagement with lan-
guage varieties in this model (see Matthiessen 2007:538–40, 2009:29–43).
In the final analysis only one other model, WG, was rated positively for this
feature.
The question on the use of diachronic data was uniformly negatively
rated by the SFL informants, and this view is borne out by a study of the
literature. The models which were finally rated positively for this item were
FDG, Givón, EG+, WG, CG, RCG, and CLS.
Finally, in this section, we asked whether data from psycholinguistic and
psychological studies were used in arguing for specific theoretical points or
for providing a principled account for linguistic generalizations. This again
produced uniformly negative ratings from our SFL respondents. We have
seen that Sydney SFL does not see itself as modelling the mind. Cardiff SFL
is indeed concerned with cognitive modelling, but has made very little use
of the psycholinguistic or psychological literature. For discussion of this
issue, readers are referred to Butler (2008, 2013). The models which were
finally rated positively for this feature were FDG, Givón, EG+, WG, CCG,
ECG, CLS, and PA.
10.4.4 Explanation
We come now to a set of fourteen questions which are concerned with
aspects of explanatory connections between language and the factors which
are held to motivate its structure and functioning. These range over matters
related broadly to cognitive motivation, the relationship between language
and usage, iconicity, the relationship between the form of language and
discourse requirements, sociocultural factors, the question of language
universals, and the relationship between language and its acquisition.
We begin with a question which asks whether knowledge of language
(often called ‘competence’) is seen as intimately related to the use of
language (often termed ‘performance’). All four SFL respondents gave posi-
tive answers to this question. However, Hallidayan SFL does not conceptual-
ize the distinction as one between knowledge and use, but rather in terms
of an opposition between the potential afforded by the linguistic system
and the use which speakers and writers make of this potential in actual
texts (see the interview with Halliday in Parret 1974:84–5, reprinted in
Halliday 1978:38). We saw in Section 10.4.2 that this distinction is captured
in the concept of instantiation. In the final analysis of models in Butler and
Gonzálvez-García (2014), all were given a positive rating for this feature.
A further question related to knowledge of language asks whether this is
seen as different from other types of knowledge or not. Our respondents
were split evenly on this question, one Sydney and one Cardiff expert going
each way. The problem here is partly that because Hallidayan SFL thinks of
language not as knowledge but as a complex set of resources available to
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 271
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
272 CHRISTOPHER S. BUTLER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 273
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
274 CHRISTOPHER S. BUTLER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 275
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
276 CHRISTOPHER S. BUTLER
10.4.6 Applications
The final section of the questionnaire consisted of three items relating to
the application of (descriptions derived from) the models. First, we enquired
whether applicability, for example to educational linguistics, stylistics, or
translation studies, is considered to be a major criterion for the success of
the model. Both Sydney model respondents and one Cardiff expert gave
positive ratings. Halliday makes it very clear that for him the value of a
theoretical approach resides in the use which can be made of it by those
with a practical interest in language: SFL is intended to be an ‘appliable’
model, and tries to answer questions asked by people who are not them-
selves linguists, but are nevertheless professionally engaged with language
(Halliday 1985:7, 2002:2, 2009:61). Halliday also sees SFL as a resource for
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 277
As we saw in Section 10.3, SFL stands out from other models in the way it
patterns overall with respect to the features discussed in detail in Butler
and Gonzálvez-García (2014). In this concluding section, I will highlight
some of the main points of similarity and difference between SFL and other
(groups of ) models.
Let us first turn to the features which SFL shares with most if not all of
the other fifteen models studied in Butler and Gonzálvez-García (2014). Like
all the other models, SFL rejects the autonomy of the language system, aims
in principle to investigate all systematic phenomena rather than just a
‘core’, regards knowledge of language as closely related to use, and offers
a ‘monostratal’ account of linguistic form in the sense that no underlying
formal level, linked to a ‘surface’ level by means of transformation-like
processes, is postulated. Like all other models except SBCG, SFL regards the
communicative function of language as central; like all but PA, it rejects the
autonomy of syntax; in common with all but EG+, it sees syntax as having
real theoretical status rather than being just an epiphenomenon of dis-
course; and similarly to all but PA, it does not postulate universals with
specific linguistic content. Like all but FDG and RRG, it is a model of
language rather than just of the grammar itself, does not make extensive
use of the sociolinguistic or sociological literature, and has little interest in
the process of subjectivization. Similarly to all except Givón, WG, and CLS,
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
278 CHRISTOPHER S. BUTLER
SFL invokes the concept of construal; like all except FDG, RRG, and CS,
it makes use of networks, though in a different way from other models; and
in common with all but Givón, EG+, and CG, it does not claim that the
characteristics of grammars emerge largely from the requirements of dis-
course. Like all models except FDG, RRG, SBCG, and CS, it pays considerable
attention to categorization (though not in cognitive terms); it shares with
all but RRG, WG, ECG, and PA the use of data from attested samples of
language use; like all but FDG, CS, SBCG, and RCG, it recognizes the
importance of innate biological factors in language acquisition; in common
with all but RRG, CCG, ECG, and CLS, it has not tested the learnability, as
such, of the constructs it proposes; and like all but RRG, SBCG, LCM, and PA,
it does not place a high value on elegance and simplicity.
I will now present the features which SFL shares with a maximum of only
four other models, in other words, those for which SFL is either totally
distinctive or in a small group of models which are distinct from the rest.
Only one feature is totally distinctive, namely, the status of being negative
for the dispreferral of empty/invisible categories, due to the importance of
covert categories, or ‘cryptotypes’, in SFL. There are five features which SFL
shares with only one other model: accounting for the properties of whole
texts/discourses (with EG+); the use of data from synchronic varieties (with
WG); the centrality of paradigmatic relations (with CS); rejections of the
claim that linguistic knowledge is no different from other kinds of know-
ledge (with Givón); and the non-recognition of (un)grammaticality (with
CS). There is one feature which SFL shares with just two other models:
rejection of the claim that cognitive mechanisms motivate theoretical
claims (with FDG and SBCG). And there are five features which SFL shares
with three other models: the importance of relationships between language
and sociocultural context (with EG+, WG, and RCG); the postulation of a
single meaning for each formal signal where possible (with WG, CS, and
CG); the importance of pedagogical applications (with EG+, FSCG, and CLS);
the lack of importance of similarities and differences between language
and other cognitive systems (with FDG, CS, and SBCG); and the non-use of
the construction as a packaged pairing of a form with a meaning (with FDG,
Givón, and WG).
It is also of interest here to see which models, within a group of four or
fewer which are distinct from the rest, SFL has most in common with. SFL
shares four features with WG (three positive, one negative) and CS (two
positive, two negative), three with EG+ (all positive) and FDG (all negative),
two with Givón and SBCG (all negative in both cases), and one with RCG, CG,
FSCG, and CLS (all positive). The distribution here is interesting, in that SFL
has links not just with other centrally functional models such as CS and
FDG, but also with what we might describe as functional-cognitive (or
cognitive-functional!) approaches (Givón, EG+, WG), centrally cognitive
models (RCG, CG, FSCG, CLS) and even a non-cognitive, formally oriented
one (SBCG). This pattern is echoed in the more detailed analysis presented
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 279
in Section 10.4, where we saw, for example, that SFL shares with Givón, EG
+, WG, CG, RCG, and LCM the property of paying considerable attention to
interpersonal meanings, and with FDG, RRG, EG+, SBCG, CCG, LCM, and PA
a concern with information structuring. The postulation of a continuum
between lexical and grammatical phenomena is also a feature which con-
nects SFL with a range of cognitive and functional-cognitive as well as
centrally functional models. Finally, we should note that there are cases
where SFL shares with cognitively oriented models an interest in a particu-
lar area, but the ways in which it deals with these areas are distinctive.
This is especially true of the linked areas of categorization and construal:
while cognitive models approach these phenomena from the viewpoint of
cognitive operations, SFL deals with them in terms of choices from system
networks which formalize paradigmatic options within a social semiotic
framework.
Work of the kind summarized in the present article demonstrates that
although SFL is distinctive in its pattern of features, there are nevertheless
some interesting links with a range of other approaches. In Butler (2013) it
is suggested that some of these links could be exploited, with advantages to
both sides, if scholars working in SFL and in other functionally and cogni-
tively oriented models paid more attention to each other’s work.
References
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
280 CHRISTOPHER S. BUTLER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
SFL in Context 281
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
282 CHRISTOPHER S. BUTLER
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 11 Aug 2019 at 10:06:14, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.012
11
Models of Discourse
in Systemic Functional
Linguistics
Tom Bartlett
Many thanks to Jim Martin for comments on an earlier version, including help with references to Halliday and
Hasan’s early work and clarification of his own work and viewpoint. I have incorporated several of Jim’s points in the
chapter but have not been able to cover them all.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
286 TOM BARTLETT
1
The mind is a contentious term in SFL, but, I think, unnecessarily so. Anxious to avoid a dualistic approach separating the
mind and the body, many authors have avoided the word and limited discussion to the brain and neural activity;
however, following Vygotsky (1978), seeing the mind as the socialized and individualized brain, far from being the
oxymoron it first appears, is a crucial conceptual pairing from an SFL perspective.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 287
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
288 TOM BARTLETT
There are several points to note here, most notably the correlation between
cohesion and a text as a coherent unit and the idea that cohesion is ‘the way
in which sentences hang together’. In the following sections I will show
how this basic conception has been extended and elaborated in the work of
Hasan and others but rejected as a ‘grammar and glue approach’ by Martin
(Martin in Andersen et al. 2015:53).3
Developing the sketch set out in Halliday’s Study Paper, in Cohesion in
English Halliday and Hasan (1976) recognize four types of grammatical
cohesion as well as lexical cohesion between sentences (or better, between
clauses). The grammatical relations are reference, substitution, ellipsis, and
conjunction, and lexical relations include repetition, synonymy, and logical
relations such as hyponymy and meronymy, as well as collocational associ-
ation. It is the interweaving of these ties across stretches of language that
creates texture (as opposed to structure, see Sections 11.3.2 and 11.3.4) and
which are at the heart of the juggling act between continuity and develop-
ment that defines textuality in general and which charts the logogenetic
2
Halliday made this paper available for circulation, and my thanks to Martin Davies for sending me an electronic copy
of his typed-up version which also includes some of his own historical annotations.
3
Martin (personal communication) sees Hasan’s approach as a bridge between Halliday’s ‘grammar and glue’
approach and his own Discourse Semantics.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 289
4
The word pragmatic might raise a few SFL hackles here, but I think it can be glossed for the more context-appropriate
‘socially situated’ in this instance. I have deliberately cited Widdowson as a friendly critic of SFL whose observations
need to be addressed, even if not accepted.
5
See Hasan (2016), which brings many of the most significant papers together, and Lukin (2015) for a very useful
overview.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
290 TOM BARTLETT
6
This is of course not quite true, as cohesion is necessarily dependent not only on the reader of a text having certain
linguistic capabilities, but also on a degree of real-world knowledge (e.g. it is necessary to know that ‘house’ and
‘kitchen’ are in a whole-part relationship, and so are superordinate and meronym respectively in linguistic terms, in
order to identify the cohesive link).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 291
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
292 TOM BARTLETT
one form of patterning, the other being the overall structure of the text in
terms of the appearance and sequencing of the different stages, compulsory
or optional, that it moves through in order to achieve its social purpose. For
Hasan (1996:53) the GSP represents ‘the total potential of structures for a
genre, while the schematic structure of any one instance of the G would
represent a particular configuration permitted by the GSP itself’. For example,
Hasan (1996:54) suggests the GSP in Figure 11.1 for the Nursery Tale genre.
This GSP has the added complication that while Sale Request and Sale
Compliance necessarily appear in that order, they interact as a whole unit
with other elements of structure (as signalled by the curly brackets).
Secondly, in all cases the meaning potential for each stage is specified in
terms of ‘its crucial semantic attributes’ (Hasan 1996:58) and from there to
the range of lexicogrammatical patterns which potentially realize these.
Returning to the Nursery Tale, one of the crucial semantic attributes of the
element Placement is ‘person particularisation’, which will be realized
lexicogrammatically by ‘indefinite modification . . . of an animate/quasi
animate noun as Thing’ (Hasan 1996:62) – in other words, the Placement
will include a nominal group such as three little pigs or a beautiful princess.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 293
The full description of a genre thus includes both the structure potential
of the necessary and optional elements and the further specification of
the semantic and lexicogrammatical attributes of these. And with regard
to the context of situation which these genres realize, it can be stated that
it is the least delicate options within the context that generate the overall
GSP, while more delicate features will generate the semantic details of
each element and hence the cohesive texture that runs throughout the
discrete elements like the coloured threads in a tapestry. Thus, the context-
ual features shopkeeper and customer in a transaction would generate the
structure in Figure 11.2, which would hold for all sales transactions
(as generically defined), while the inclusion of further details within the
contextual variable of field might specify that there will be continued
reference to clothes, fabrics, sizes, and colouring across the text, and more
delicate features within the contextual variable of tenor, such as familiarity
or status, would generate more specific ways of making enquiries and
requests. In this way the realization relationship between context and text
is able to capture both important generalizations across instances of a
genre type and the specific features of individual instances of that genre.
For a fuller discussion of texture and structure, see Halliday and Hasan
(1989: Part B).
One potential problem with this approach is Hasan’s (1995:219) con-
cept of relevant context, glossed as those aspects of the non-linguistic
environment that are made relevant through language. In this formula-
tion there is by definition a correlation between text and context, and
this has led to discussions as to whether this pairing is the essence of a
supervenient and non-essentialist conception of context or a circularity
which ignores, or at least marginalizes, how extra-textual features affect
text in less visible ways than direct inscription.7 I will return to these
issues in Section 11.7, but see Bartlett (2017) and Moore (2017) for
exchanges on this point.
7
Recent work on multimodality opens this question up a little, as the borderline between material context and
semiotic system becomes increasingly fuzzy. This is an area which would merit further theoretical discussion.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
294 TOM BARTLETT
8
Note the capital letters to show that these are technical metalinguistic terms, the names of which refer to similar lay
metalinguistic categories.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 295
here
(5) Mother: There aren’t many passionfruit out there at the moment
Report
Stephen: Why?
Mother: Because passionfruit come when it’s warm
Generalization
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
296 TOM BARTLETT
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 297
(8) Quizmaster: In England, which cathedral has the tallest spire dk1
Contestant: Salisbury k2
Quizmaster: Yes k1
(9) Son (doing crossword): Which English cathedral has the k2
tallest spire
Father: Salisbury k1
Son: Oh, right k2f
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
298 TOM BARTLETT
support (ps). Adding these labels to example (9) above provides the complete
metafunctional analysis of the exchange:
Berry (forthcoming) sets out more delicate options within the three
metafunctions, exploring how exchanges can depart from the smooth and
narrow, and discussing the function and variability of side sequences. The
complete framework developed therefore incorporates important insights
from Conversation Analysis and Gricean pragmatics into an SFL frame-
work, though there is not room here to expand on these ideas in depth.
In this way Exchange Structure provides analysts with a more dynamic and
emergent view of discourse as something that happens, and Berry (2016)
provides a systems network that accounts for the different possibilities that
arise at different points in the discourse.
While the formalisms of Exchange Structure all relate to in-text rela-
tions, they provide a useful tool for discourse analysis. As Berry (2016)
explains, labelling for all three metafunctions provides richer analysis of
exchanges, often distinguishing superficially similar structures. Though
the approach was developed within and is particularly suited to educational
settings, it can be applied to a range of contexts to analyze, for example,
how power is played out in various settings in terms of who introduces
the propositional bases that delimit the scope of the conversation and who
act as primary and secondary knowers within these contexts. And while the
approach is based at one level on structure and hierarchy, the localized
range of these hierarchies, which flow from one to the next without
developing into superordinate structures, means the approach can also
be used to analyze less structured genres, including casual conversation.
In Sections 11.5 and 11.7 we will see contrasting approaches to discourse
analysis in these terms: Rhetorical Structure Analysis, which works very
much within the tradition of hierarchicization, combining units of ever-
increasing size to analyze whole texts as single structures, and Phase
Analysis, which emphasizes the flow of texts across the metafunctions
and their emergent properties.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 299
Thompson 1988; Mann and Matthiessen 1991; Mann et al. 1992), a method
for analyzing the logical relations, both syntactic and functional, between
individual messages within what might be classed as a single prolonged
move in Berry’s terms. And while this approach could be extended to cover
dialogic speech, it is therefore most appropriate to monologue. Mann et al.
(1992:43–6) list the basic assumptions underlying RST, which are para-
phrased and condensed here:
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
300 TOM BARTLETT
the evidence relation to be said to hold, the constraints are that the reader
may not believe the proposition in the nucleus; that the reader is likely to
find the satellite credible; and that the reader’s comprehension of the satel-
lite will increase their belief in the nucleus. The effect therefore is that the
reader’s belief of the nucleus is increased. For full examples of analyzed
texts, see Mann and Matthiessen (1991) and Mann et al. (1992).
Given the focus on the recurrent functional relationships between
elements of a text at different scales, Hasan and Fries (1995:xxxiii) suggest
that RST would be better labelled as ‘logical structure theory’. If we take
this conclusion at face value, it could be claimed that RST provides the
missing metafunctional link, adding a description of logical relations across
texts to the experiential (lexical cohesion and cohesive harmony), the text-
ual (RU analysis; Cohesion in English), and the interpersonal (message seman-
tics and exchange structure, though these also consider aspects from the
other metafunctions). In an extension of the original descriptions of
RST that overlays the basic analysis with interpersonal and ideational
features to provide a fuller representation of extended text that is reminis-
cent of Hasan’s conception of register as the accumulation of message
semantics across whole texts, Mann and Thompson (1991) go as far as
to suggest that a correlation between the relations of RST and all the
metafunctions is robust.
In terms of the text/discourse distinction proposed at the beginning
of this chapter, while RST relations are based on predictions of authorial
intention and favoured reading, these are all features that can be read off
from the decontextualized text alone so that, despite the additional analyt-
ical features afforded by RST, it would have to be considered text rather
than discourse analysis according to this definition. Webster et al. (2013) is
an example of RST in analyzing political speeches, while Bateman
(2008) utilizes a somewhat extended version of RST as one of the layers of
description for multimodal static page-based documents.
9
Though Martin sees both perspectives as useful.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 301
(as one dimension of the realization of a genre), with meanings realized within
the text in different ways: periodically and cumulatively across the text, as well
as structurally within the confines of the individual clause or clause complex.
Thus, rather than focusing on realizations of experiential, logical,
interpersonal, and textual meanings at clause rank, Martin (1992) proposes
the following metafunctional categories at the discourse semantic stratum: idea-
tion, conjunction and continuity, negotiation (based largely on Berry’s
exchange structure), and identification, with the further category of texture
‘interleaving’ the features of each together to create texts.
In extended collaborations with Peter White and David Rose, Martin
produced two further landmark books, Evaluation in English (Martin and
White 2005) and Working with Discourse (Martin and Rose 2003, 2007). The
first of these extends the discourse semantic treatment of interpersonal
(including subjective) meaning to comprise three major systems – Attitude,
Engagement, and Graduation – which consider attitudinal language,
interspeaker/intertextual relations, and amplification/moderation respect-
ively. There is no room here to discuss these categories in depth (but see
Martin, this volume). In the second of these books, Martin and Rose update
the analytical framework of Martin’s English Text (1992), re-presenting
the approach from a more text-analytical-up and less theory-down perspec-
tive. The categories presented have now evolved into ideation, conjunction,
appraisal (based on Martin and White’s work on attitude, engagement, and
graduation), negotiation, identification, and periodicity (a revised version of
texture).
An interesting point worth dwelling on here is Martin’s (1992:249–64)
critique of the hierarchical representation foundational to RST, which he
sees as too product-oriented and especially at variance with ‘the dynamics
of text as process, particularly in the spoken mode’ (Martin 1992:258).
In contrast he proposes a linear dynamic which resonates with his
characterization of genre as ‘a staged goal-oriented social process’
(Martin 1992:505). From this perspective, genres unfold as a sequence of
necessary stages, semantically driven and realized through the metafunc-
tionally diversified resources of discourse semantics, lexicogrammar, and
phonology/graphology, to fulfil recognized and recognizable social
activities.
Much of the work of what has come to be termed the Sydney School, after
Martin’s university, has been produced within action research projects
aimed at extending the literacy skills of disadvantaged groups through a
visible pedagogy that focuses on producing highly valued written work
across a range of disciplines at all levels of the curriculum (though Martin
is quick to emphasize this is just one application of the theory, which has a
broader genesis). Within such a framework, a genre can be defined as
situation/language pairings which have become socially recognizable
through repeated association (a point I return to in Section 11.7), and
Martin and Rose (2008) exemplifies this approach through the analysis into
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
302 TOM BARTLETT
named stages of genres across a range of disciplines. While the goal of this
work is to make visible and hence replicable what is hidden to students and
teachers alike, this apprenticeship approach has regularly invited criticisms
of prescriptivism and acculturation within the oft-heated debate on minor-
ity rights in education.
More recently, the work of Martin and his collaborators has moved into
new fields, particularly restorative justice, and this has led not only to
further socially motivated applications of the approach, but also to a
continuing enrichment of the descriptive and theoretical power of their
work. Bednarek and Martin (2010) is a strong testimony to the achieve-
ments of this prolonged collaborative labour between Martin and his
research students, many of whom are now recognized figures in SFL.
A particularly fruitful collaboration has been with the Bernsteinian soci-
ologist Karl Maton, resulting in the 2014 volume Knowledge and Knowers:
Towards a Realist Sociology of Education, which analyzes a range of discip-
lines across the sciences, social sciences, and humanities in terms of the
kinds of knowledge and epistemic stances favoured and their transpar-
ency and hence their potential for transmission through visible pedago-
gies (see Maton 2014).
One continuous thread in Martin’s work is his insistence on a super-
venient model of context and his occasionally scathing rejection of eth-
nography, or at least what often passes as such in applied linguistic
work, which he labels ‘ethNOgraphy’. Martin consistently takes the
line that, within a social realist ontology such as SFL, culture is realized
in text, via genres, semantics, lexicogrammar, and phonology/graphology
and that non-linguistic information, including observation and the views
of insiders, adds nothing to and potentially distorts our understanding
of ‘what’s going on here’.10 In the following section, I take issue
with Martin’s stance, drawing on the work of Michael Gregory and my
own heavily contextualized approach, before suggesting a potential
reconciliation.
10
Martin (personal communication) suggests the following alternative wording: ‘do not provide an answer to
questions addressed in social semiotics analysis informed by SFL (including the multimodal analysis evolving out of
SFL).’ I have decided to keep both my original wording and Martin’s alternative.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 303
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
304 TOM BARTLETT
11
More recent work in the Martinian tradition has acknowledged the emphasis on the system rather than the
instance in textual analysis and is working to develop better descriptions of individual texts as instances.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 305
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
306 TOM BARTLETT
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 307
References
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
308 TOM BARTLETT
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
Models of Discourse in SFL 309
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
310 TOM BARTLETT
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:23:01, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.013
12
Cohesion and
Conjunction
Maite Taboada
Most texts we encounter on an everyday basis are coherent. They make sense
in the situation in which they are presented, and their meaning and inten-
tion can usually be understood. Even in cases where the communication
may be more difficult (a ‘different’ accent; spelling mistakes; complex argu-
mentative structure), we tend to accept texts as being coherent, and make an
effort to grasp their meaning. Coherence is such a fundamental property of
texts and of our communication that it is difficult to conceive of completely
incoherent texts. Consider the two invented examples in (1) and (2). In the
first case, we have a set of sentences, each connected to the previous one
through a lexical item. This is an instance of sets of cohesive links, with
items such as last night – at night, which may have a semantic relation of
repetition in most texts, but which do not here. The passage, however, does
not seem to have a common thread; it is not coherent. Conversely, the two
sentences in (2) are coherent in terms of a thread (dark clouds – rained), but
the sentences are not well related, because the conjunction however does not
usually relate two units in this way. It sets up an unfulfilled expectation, or
one contrary to expectation, but rain following dark clouds is actually not
contrary to expectation. Example (2) is coherent, but fails in the way that
coherence is made explicit through the conjunction however.
(1) I went home very late last night. At night, owls come out and hunt.
Harry Potter uses an owl to have his mail delivered. The mail was
very erratic over the Christmas holidays. The holidays were too
short, and short indeed is this paragraph.
(2) There were dark clouds in the sky today. However, it rained.
I would like to thank Geoff Thompson, for inviting me to contribute this chapter, and for providing very insightful
comments shortly before his untimely death.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
312 MAITE TABOADA
1. Reference
2. Substitution
3. Ellipsis
4. Conjunction
5. Lexical cohesion
1. Conjunction
2. Reference
3. Substitution and ellipsis
4. Lexical organization
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 313
12.2.1 Conjunction
Halliday and Hasan (1976) initially divided cohesion into two types:
grammatical and lexical. Conjunction was listed under the lexical label,
because it makes use of lexical items, i.e. conjunctions such as and,
but, or if. Halliday and Matthiessen (2014) move it out of the lexical
realm, and present it as the first element in the list of cohesive resources.
This is understandable, because, although it deploys lexical items, the
particular items are function words, i.e. not open-class items such as
nouns or verbs. In example (3), we see how the two clauses in the first
sentence are joined by the conjunction but. Conjunction also links sen-
tences, as we can see towards the end of example (3), which uses the
conjunction and to relate the two parallel sentences (I laughed. I cried) to
the last sentence.2
(3) I can rarely say that a movie made me laugh and cry without feeling
like an idiot, but the caliber of this picture is so high that I don’t even
feel embarrassed. I laughed. I cried. And you will too.
1
See www.imdb.com
2
Halliday and Hasan restrict cohesion to links across, not within, sentences (i.e., beyond the clause and clause complex
level). In example (3), then, the connection signalled with and would be cohesive, but not the one indicated by but in
the first clause. Here, I will consider both of these as instances of cohesion, because I believe that cohesion and
conjunction occur across clauses, whether in different sentences or not.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
314 MAITE TABOADA
(4) I remember talking to friends after seeing it and we all had interest-
ing points. I felt the film focuses on Chihiro’s innocence as compared
with the other characters she encounters, but her child like views are
so carefree (and naive at times) and her youthful exuberance really
makes it endearing. Another friend said it was a coming of age and
how Chihiro herself progresses throughout the film. I mean, if you
can find so much insight in a film, you know you have a great film.
12.2.2 Reference
Reference is achieved mostly through relations between a pronoun and an
antecedent, forming a referential chain in the text, which is characterized
as anaphoric reference. Reference links to elements outside the text consti-
tute ‘exophoric reference’, whereas links within the text are ‘endophoric’.
Some reference chains may include both exophoric and anaphoric refer-
ence. For instance, first- and second-person personal pronouns refer to
relationships defined outside the text, but often also create text-internal
relations. In example (5), the personal pronoun I in the first sentence is
exophoric, in that it refers to the writer as somebody outside the text. The
reference chain, however, continues inside the text, with another I in the
second sentence, and the possessive my.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 315
(5) Spirited Away is one of the most perfect movies I have ever seen. The
least I can say about it is that there was not a single moment during
it that my attention wasn’t completely focused.
(6) We discover the world as Chihiro does and it’s truly amazing to
watch. But Miyazaki doesn’t seem to treat this world as something
amazing.
(7) The story is imaginative and the characters and animations endlessly
unique and strange. . . . What I also loved in this film is that the
animation gives it a real sense of cinematography, . . . Another great
point in fact the best part of it, is the fantastic score.
(8) Even if you don’t normally like ‘cartoon movies’, you might give this
one a chance.
Ellipsis makes coherent text possible, and even more so in spoken language.
The simple ‘yes’ and ‘no’ answers that we furnish as answers to questions
would be cumbersome if they were always accompanied by a full answer
which repeats information already present in the question (although
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
316 MAITE TABOADA
(9) I can’t imagine seeing a live action foreign language film dubbed into
another language, but hey, this is a kids cartoon, what does it
matter? Up to a point it didn’t, because I loved the film.
(10) When Chihiro’s parents see the food in the market, they just sit
down and eat but they turn out to turn into pigs.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 317
1. Reiteration
Same word (repetition) Same referent
Synonym Inclusive
Superordinate Exclusive
General word External
2. Collocation
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
318 MAITE TABOADA
Elements in the text related through cohesion establish a ‘cohesive tie’: the
interpretation of one element in the discourse depends on the interpret-
ation of another, whether preceding (anaphoric relation) or following (cat-
aphoric). The fact that the interpretation is successfully established creates
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 319
the cohesive tie between the two elements. As more elements are related to
each other, a ‘cohesive chain’ is created, a series of elements related to one
another. Chains are frequently created through identity of referent (I-I-my
in example (5)), but Hasan (1985:73) clearly states that the relation between
elements in a cohesive tie is semantic, i.e. the elements are tied together
through some meaning relation, even if the relation is expressed through
grammatical resources, such as personal pronouns.
One way of exploring the relationship in a cohesive tie, and ultimately in
cohesive chains, is to measure the distance between components. The
relationship may be immediate (the cohesive element refers to an immedi-
ately preceding one); remote (the referent is more than one clause away); or
mediated (where the ultimate referent is a few clauses earlier in the dis-
course, but has been recaptured in some other element). Although I often
refer to ‘preceding discourse’, this naturally applies only to anaphoric
relations; cataphora works in the opposite direction, by establishing links
that look forward to an element that completes the interpretation. Cata-
phora seems to be extremely rare: in a study of over 11,000 instances of
third-person pronouns, my colleague Radoslava Trnavac and I found only
fifty-seven instances of cataphora (Trnavac and Taboada 2016). Halliday and
Matthiessen (2014:625) also state that cataphora is rare, with the exception
of ‘structural cataphora’, where the reference to a pronoun is resolved
immediately afterwards, in the same clause. Such is the case in (12), where
the referent for those is in the relative clause immediately following (who are
just looking . . .).
(12) Highly recommended to anime fans and those who are just looking
for a film that is unique and interesting.
Cohesive chains run through texts, and provide them with the links to
create texture. Most texts (spoken or written) contain more than one chain.
The short excerpt in (13) contains at least five interrelated chains, as shown
in Table 12.3. The table breaks down the text into units, which are some-
what arbitrary, for ease of presentation, and do not necessarily correspond
to clauses or independent units. We can see that there is a chain relating to
the film under discussion, which includes Spirited Away as the first element
in the chain. The noun group Spirited Away is not technically a cohesive
element yet, as this is the beginning of the text, and the noun group does
not refer to anything preceding, although it naturally establishes links
outside the text proper, on the web page where this review appeared,3
and the web page from which the review is linked.4 The next element in
this chain, the latest, relates to Spirited Away through ellipsis of the Head
noun film. The chain continues with repetition of the noun film. A second
3
See www.imdb.com/title/tt0245429/reviews-25
4
Some of these ties are multimodal, because they relate different modalities. Multimodal cohesion and conjunction
are beyond the scope of this paper, but see Bateman (2008) for an overview of multimodality in discourse.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
320 MAITE TABOADA
chain, related to the first, establishes that this is an animation film, and the
director then naturally an animator. The third chain refers to the director,
through reference in the form of the personal pronoun he in the second
sentence. Finally, two related chains establish Miyazaki as an excellent
director, first in a narrower frame of reference (Japan), and then more
broadly, as the best in the entire world.
Cohesive chains and chain interaction are some of the most interesting
constructs for describing cohesion in text, and how texture is achieved.
Hasan (1984) proposed the idea of ‘cohesive harmony’, a measure of how
well-integrated cohesive chains are (see also Hoey 1991; Khoo 2016; Parsons
1996). Chains do not occur in isolation, but alongside other chains. How-
ever, the mere presence of two or more chains in a text does not guarantee a
cohesive effort. Although chains contribute to cohesion in a text, they need
to be related to each other somehow. This relationship is called ‘chain
interaction’. The relationships are mostly grammatical, as part of the tran-
sitivity structure of the clause, such as the relationship between Processes
and Participants. Hasan establishes a minimum requirement for chain
interaction: at least two members of one chain should stand in the same
relation to two members of another chain. For a better definition of the
interactions, she divides the tokens in a text into three categories:
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 321
• Relevant tokens: all tokens that enter into chains, further divided into:
○ Central tokens: relevant tokens that interact.
○ Non-central tokens: relevant tokens that do not interact.
• Peripheral tokens: Tokens that do not enter into any kind of chain.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
322 MAITE TABOADA
of a cohesive tie, and are at the same time members of the transitivity
structure of the clause. It is the linking of those devices that creates cohe-
sion across clauses. Organic devices, i.e. conjunctive devices, tie whole
clauses rather than clause components. Hasan points out that other devices
of language organization, such as adjacency pairs, are also organic.
Conjunctive devices (conjunctions, some continuatives, and some
Adjuncts) link clauses in logico-semantic relations such as cause, conces-
sion, or condition. They also serve to indicate temporal and additive rela-
tions. In this sense, conjunction is closely related to the clause complexing
system, and moves cohesion outside of the clause proper. Halliday and
Hasan would always have cohesion act outside of the clause, but the view
taken here is more inclusive, whereby cohesive and conjunctive links occur
both within the clause and across clauses. Conjunction serves to indicate
that a relation exists between clauses (clause complexes, sentences, or
entire text passages), and sometimes it provides an indication of the nature
of the relation. This indication can be quite clear, such as the relation
signalled by because, or it can be underspecified, as is the case with and,
which can indicate a variety of relations.
This general idea, that conjunction relates clauses, or propositions, has
been made specific and instantiated under different theories, and different
taxonomies. Halliday and Hasan (1976) proposed a top-level classification
into additive, adversative, causal, and temporal. The classification is exclu-
sively based on the presence of a conjunctive item (e.g. so, consequently, for this
reason, or it follows for cause). In the first edition of Introduction to Functional
Grammar (Halliday 1985), and in subsequent editions, this semantic classifi-
cation (based on how the conjunction specifies the semantic content of the
linkage) is made more abstract, with a higher-level classification based on
how one clause adds to another, and using three types of connection:
elaboration, extension, or enhancement. This classification is based on the
form of the contribution, rather than the semantic meaning that is contrib-
uted, but typically relies on the presence of a conjunctive item.5
Martin (1992) proposed a slightly different classification of what he
named conjunctive relations, which are outside of cohesion proper. Add-
itionally, he expands on the internal/external distinction, relating
to whether relations refer to external relations, in the real world, or to
the internal organization of the events in the text. The latter are more
‘rhetorical’ in nature, in that they have to do with how arguments are
presented. The distinction is quite clear with temporal relations. An exter-
nal temporal relation describes sequences of activities as they occur in the
world. Internal temporal relations, on the other hand, capture time within
the text, i.e. in relation to what is being said and how the text is organized.
The two following examples are from Martin (1992:182), with the words in
5
Halliday does mention non-finite clauses as examples of conjunctive relations which do not have an explicit marker for
the relation.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 323
bold, from the original text, indicating conjunctive items. Example (14)
presents an external temporal relation, where the events are presented as
they unfolded in the world. In (15), an internal relation, the sequence first-
second could have well been presented in the reverse order, depending on
the effect that the writer or speaker wanted to create.
(16) Dominique quit his job because he was tired of the long hours.
(17) Dominique quit his job. He was tired of the long hours.
6
Not strictly ‘classical’ in the sense of Halliday and Hasan, since for them cohesion only takes place across, not within
sentences. For them, this example would be accounted for within clause complexing, not cohesive conjunction.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
324 MAITE TABOADA
(18) While these two themes are very much current in Japan, they are
also universal themes.
(19) Sometimes in real life the most grim moments contain honest elem-
ents of comedy that do not seem out-of-place. But trying to put that
sort of convoluted emotion into a film creates a very thin line that
too many have fallen off of.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 325
7
www.epinions.com/review/Hot_Six_by_Janet_Evanovich_and_narrated_by_Debi_Mazar/2004218900/963557
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
326 MAITE TABOADA
which are related through a Background relation: the fact that the book
under review is the sixth in the series is background information necessary
for understanding the Condition conveyed in Spans 2–3. Here, Spans 2–3,
combined, are the nucleus, whereas Span 1 is the satellite. The excerpt has a
second part, composed of a relation which does not contain a nucleus
and a satellite, like the ones we have seen so far, but which instead is made
up of three nuclei, three units of equal importance (Spans 3–6). Such
relations are referred to as ‘multinuclear’ relations. Together, these three
units constitute a Sequence relation, outlining the steps that the author
thinks the reader should take. Finally, these two macro-units or sequences
of spans (1–3 and 3–6) are joined together into another multinuclear rela-
tion, this time of Contrast. The author establishes a contrast between two
possibilities, not reading the book under review and reading the previous
ones in order.
(20) [This is the sixth book in the Stephanie Plum series.] [If you have
never had the fun of reading a book in this series,] [do not start with
this one.] [Go to the library] [and start with One For The Money] [and
work your way up to Hot Six.]
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 327
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
328 MAITE TABOADA
12.7 Conclusion
8
In addition, for further reading on cohesion, consider the following. The original description of cohesion (in English) is
Halliday and Hasan (1976), and it still remains the most detailed account of the phenomenon, with plenty of examples.
The theory was refined and developed in a book by Halliday and Hasan (1985), and in particular a chapter by Hasan in
that book (Hasan 1985). Several introductions to Systemic Functional Linguistics explain cohesion in very concise
terms, often in a single chapter (Eggins 2004; Thompson 2014), but perhaps the most clear and concise is
Flowerdew’s introduction to SFL (Flowerdew 2013), which emphasizes applications of Systemic Functional Linguistics
to language education. And, of course, Halliday and Matthiessen’s (2014) Halliday’s Introduction to Functional
Grammar remains the ‘official’ version of the theory, with a chapter on cohesion and its place in the functional analysis
of language.
More specialized descriptions focus on cohesion, often with an introduction that then leads to an in-depth study,
typically corpus-based. Tanskanen (2006) does not strictly follow Halliday and Hasan’s classification, but hers is a
thorough corpus-based study. Fox (1987) presents an analysis of both cohesion and coherence in conversational
speech.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 329
References
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
330 MAITE TABOADA
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
Cohesion and Conjunction 331
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
332 MAITE TABOADA
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:51:51, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.014
13
Semantic Networks
Andy Fung and Francis Robert Low
13.1 Introduction
1
While recent years have witnessed a growing trend in analyses of multimodal discourse, this chapter, following Hasan
(2014:3), regards language as the central object of enquiry in SFL.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
334 AN DY F U NG AN D FR ANCI S ROB ERT LOW
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 335
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
336 AN DY F U NG AN D FR ANCI S ROB ERT LOW
The ‘threat’ may be a threat of physical punishment. Here the clause is of the
action type, and, within this, of intentional or voluntary action, not super-
vention (i.e. the verb is of the do type, not the happen type). The process is a
two-participant process, with the verb from a lexical set expressing ‘punish-
ment by physical violence’, roughly that of § 972 (PUNISHMENT) in Roget’s
Thesaurus, or perhaps the intersection of this with § 276 (IMPULSE). The tense
is simple future. The Goal, as already noted, is you; and the clause may be
2
This contrasts with the description of semantics from below, or chooser and inquiry semantics (in Matthiessen’s 1990
terminology), which is typically employed in the model of text generation. Examples include Matthiessen (1988b),
Patten (1988), and Matthiessen and Bateman (1991), to name but a few. See Matthiessen (1990) for details.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 337
Figure 13.1 The semantic network of warning and threat (Halliday 1973:89)
either active, in which case the agency of the punishment is likely to be the
speaker (I as Actor), or passive, which has the purpose of leaving the agency
unspecified.
(Halliday 1973:78, emphases in original)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
338 AN DY F U NG AN D FR ANCI S ROB ERT LOW
3
As illustrated in Hasan’s subsequent work in contextual modelling, mother–child talk is essentially registerially/
contextually inconsistent, entailing frequent reclassifications of con/text as the talk develops (see Cloran 1999; Hasan
1999, 2000 for a detailed discussion on con/textual shift).
4
Though the account of meaning potential is robust, such descriptions, as noted by Hasan, are not yet exhaustive.
Further tests and applications are thus needed.
5
It should be emphasized that the term ‘message’ is also used in another sense in SFL, denoting the textual unit of
meaning (Halliday and Matthiessen 2004:588–9). In this chapter, ‘message’, following Hasan, refers to the semantic
rank scale only.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 339
13.2.2.3 Trinocularity
Following Halliday, each semantic option in Hasan’s message semantics
network attaches fundamental importance to the ‘concept of trinocularity’
(see Halliday 2009:79–80). That is, Hasan’s semantic networks not only
concern the interrelations among semantic options (i.e. whether the options
postulated are internally duplicate or contradictory (Hasan 1996:110)), but
also emphasize the relations with context (i.e. what contextual features are
construed) and lexicogrammar (i.e. what lexicogrammatical patterns are
activated). In other words, the analysis of meaning through utilizing seman-
tic networks not only illustrates the meanings at risk, but also enables
analysts to explain ‘why and how something is said’ and ‘why these patterns
of wordings appear rather than any other’ (Hasan 2009c:170).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
340 AN DY F U NG AN D FR ANCI S ROB ERT LOW
6
Due to space constraints, Figure 13.2 only includes the primary sub-system for a message with the feature
[progressive] under the four metafunctions. For example, the topic indicates the sub-system of CONTINUATION,
selecting between [turn-maintaining] and [turn-changing]. Unfortunately, lack of space precludes a detailed
discussion of each semantic feature. For detail, see Williams (1995); Lukin (2012, 2013).
7
It should be noted that the term AMPLIFICATION was previously used in Martin’s earlier accounts of APPRAISAL (see
Martin 2000). However, it has been re-labelled GRADUATION in Martin and White (2005). Following Hasan, the term
AMPLIFICATION used here refers to the semantic system of logical meanings.
8
Contra Halliday, Hasan separates the experiential metafunction from the logical metafunction, leading to four systems
of meanings.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 341
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
342 AN DY F U NG AN D FR ANCI S ROB ERT LOW
9
‘Meanings’ refers to all modalities of semiosis in Bernstein’s coding orientation. Hasan, by contrast, takes a restricted
view of meaning, with a particular focus on the modality of language. Such a restricted view on coding orientation, in
Hasan’s word, is termed semantic orientation.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 343
Figure 13.3 Options in expressing questions: a simplified fragment (see Hasan 1989:246;
Hasan et al. 2007:713)
Gloss: Each system in the QU ESTION network is labelled. For example, the primary options are
G and H in Figure 13.1, and each of the successive systems is labelled a, b, c . . ., and
finally each of the terms is labelled 1, 2, 3 . . .
10
This short excerpt of interaction is taken from Hasan (2009e).
11
Following Hasan’s recent (2014:17) account, distinctions between the terms (i) option, (ii) choice, and (iii) feature
deserve to be noted. Briefly, the term option is ‘choose-able’ in the sense that it refers to the ‘as-yet-unexplored
property of potential’ in the system; the term choice, by contrast, denotes ‘the option selected for further exploration’;
and the term feature refers to the ‘properties of unit under description’ (see also Hasan 2013 for a distinction
between choice and feature).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
344 AN DY F U NG AN D FR ANCI S ROB ERT LOW
Message
ID Speakers Message
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 345
Lexicogrammatical realization
[confirm: verify: reassure] (i) You love Uncle Matt, don’t you?
(ii) You don’t love Uncle Matt, do you?
[confirm: verify: probe] (i) You love Uncle Matt, do you?
(ii) You don’t love Uncle Matt, don’t you?*
[confirm: enquire: ask: non-assumptive] Do you love Uncle Matt?
[confirm: enquire: ask: assumptive] Don’t you love Uncle Matt?
[confirm: enquire: check: non- You love Uncle Matt?
assumptive]
[confirm: enquire: check: assumptive] You didn’t love Uncle Matt?
* As noted in Hasan (2013), this is normal usage in some varieties of Australian English.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
346 AN DY F U NG AN D FR ANCI S ROB ERT LOW
Lexicogrammatical realization
Semantic
option Systemic realization Structural realization
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 347
12
Since the transcription is not phonologically annotated, one could equally interpret Message 95 as an elliptical
interrogative, with an ellipted Mood element Do you. In this case, Message 95 would be analyzed as [demand;
information: confirm: enquire: ask].
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
348 AN DY F U NG AN D FR ANCI S ROB ERT LOW
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 349
13
Wake (2006) uses the feature [validate] to refer to questions which are realized by clauses preselecting
[declarative: Adjunct right?], as in The price is part of the world price, right?
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
350 AN DY F U NG AN D FR ANCI S ROB ERT LOW
14
Examples of ‘core democratic functions’, as stated by Clayman and Heritage (2002:2) include ‘soliciting statements of
official policy, holding officials accountable for their actions, and managing the parameters of public debate, all of this
under the immediate scrutiny of the citizenry’.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 351
This chapter has demonstrated one claim, namely, that Hasan’s message
semantics network is essentially a discourse analytical tool enabling dis-
course analysts to study language in use in various contexts. In Section 13.2,
we highlighted the theoretical foundations of semantic networks, and more
specifically, the key advancements that Hasan has made, which contribute
to our current understanding of message semantics. Since it is not our
intention to repeat all the ideas which have been discussed in the previous
literature, we have deliberately kept the discussion short and precise.
However, the issues that we have highlighted are sufficient to demonstrate
that Hasan’s message semantics is a powerful tool for discriminating
among the meaningful choices enacted by interlocutors in a dialogue. Given
this significance, it is therefore not surprising that such a paradigmatic
description of semantics has been extensively applied in various discourse
studies. Hasan’s semantic networks are thus not just a tool for semantic
variation or integrated sociolinguistic research (see Section 13.3), but essen-
tially a discourse analytical tool for meaning analysis. We have also
surveyed the use of semantic networks in discourse studies, presenting an
up-to-date review of the different uses of message semantics (see Table 13.4).
Space precludes a detailed discussion of all the domains of application of
semantic networks; the illustrations of pedagogical and journalistic dis-
courses are, we hope, sufficient to exemplify in what ways message seman-
tic networks are employed by discourse analysts to tackle various research
problems.
For Hasan (2005:56), ‘why and how language works’ and ‘the nature of
the relationship between language and society’ are two sides of the same
coin. While one of the primary tasks of discourse analysts is to interpret
and make sense of the meaning of what people say and write and listen to
and read in context, the ability to calibrate context to the semantics and to
lexicogrammar in Hasan’s contextually open semantic networks thus
serves as a powerful analytical tool in analyzing meanings in dialogue,
enabling discourse analysts to understand ‘why and how language
works’.15
References
15
Recent years have witnessed an increase in studies describing context through system networks (e.g. Butt 2004;
Hasan 1999, 2009c, 2014; Bowcher 2007, 2014). Such descriptions enable analysts to integrate network-based
descriptions from context to semantics to lexicogrammar.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
352 AN DY F U NG AN D FR ANCI S ROB ERT LOW
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 353
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
354 AN DY F U NG AN D FR ANCI S ROB ERT LOW
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 355
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
356 AN DY F U NG AN D FR ANCI S ROB ERT LOW
Martin, J. R. and D. Rose. 2007. Working with Discourse: Meaning beyond the
Clause. 2nd ed. London: Continuum.
Martin, J. R. and P. R. R. White. 2005. The Language of Evaluation: Appraisal in
English. London: Palgrave.
Matthiessen, C. M. I. M. 1988a. Representational Issues in Systemic Func-
tional Grammar. In J. D. Benson and W. S. Greaves, eds., Systemic Func-
tional Perspectives on Discourse. Norwood: Ablex. 136–75.
Matthiessen, C. M. I. M. 1988b. Semantics for a Systemic Grammar:
The Chooser and Inquiry Framework. In J. D Benson, M. J. Cummings,
and W. S. Greaves, eds., Linguistics in a Systemic Perspective. Amsterdam:
John Benjamins. 221–42.
Matthiessen, C. M. I. M. 1990. Two Approaches to Semantic Interfaces in
Text Generation. COLING 90: 322–9.
Matthiessen, C. M. I. M. 1993. Register in the Round: Diversity in a Unified
Theory of Register Analysis. In M. Ghadessy, ed., Register Analysis: Theory
and Practice. London: Painter. 221–92.
Matthiessen, C. M. I. M. 1995. Lexicogrammatical Cartography: English Systems.
Tokyo: International Language Sciences Publishers.
Matthiessen, C. M. I. M. 2004. The Semantic System of Relational Expan-
sion: Rhetorical Structure Theory Revised. Unpublished draft, Macquarie
University.
Matthiessen, C. M. I. M. 2007. The Architecture of Language According
to Systemic Functional Theory: Developments since the 1970s. In
R. Hasan, C. M. I. M. Matthiessen, and J. J. Webster, eds., Continuing
Discourse on Language: A Functional Perspective, Volume 2. Sheffield: Equinox.
505–61.
Matthiessen, C. M. I. M. 2009. Ideas & New Directions. In M. A. K. Halliday
and J. J. Webster, eds., Continuum Companion to Systemic Functional Linguis-
tics. London: Continuum. 12–58.
Matthiessen, C. M. I. M. 2012. Systemic Functional Linguistics as Applicable
Linguistics: Social Accountability and Critical Approaches. D.E.L.T.A. 28:
435–71.
Matthiessen, C. M. I. M. 2014. Appliable Discourse Analysis. In Y. Fang and
J. J. Webster, eds., Developing Systemic Functional Linguistics: Theory and
Application. Sheffield: Equinox. 138–208.
Matthiessen, C. M. I. M. 2015a. Halliday on Language. In J. J. Webster, ed.,
The Bloomsbury Companion to M. A. K. Halliday. London: Bloomsbury.
137–202.
Matthiessen, C. M. I. M. 2015b. Modelling Context and Register: The Long-
term Project of Registerial Cartography. Letras, Santa Maria 25(50): 15–90.
Matthiessen, C. M. I. M. and J. Bateman. 1991. Systemic Linguistics and Text
Generation: Experiences from Japanese and English. London: Pinter.
Moore, A. R. 2016. Can Semantic Networks Capture Intra- and Inter-
Registerial Variation? Palliative Care Discourse Interrogates Hasan’s Mes-
sage Semantics. In W. L. Bowcher and J. Y. Liang, eds., Society in Language,
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
Semantic Networks 357
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 27 Aug 2019 at 06:51:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.015
14
Discourse Semantics
J. R. Martin
Discourse semantics is the term used by Martin and his colleagues (after
Martin 1992; Martin and Rose 2003) to refer to the stratum of meaning
interfacing lexicogrammar with context (register and genre) in SFL. It
comprises six major systems, organized into three metafunctions: ideation
and conjunction (ideational); negotiation and appraisal (interpersonal); and
identification and periodicity (textual). As reviewed in Martin (2014), this
work reinterprets Halliday and Hasan’s model of cohesion (e.g. Halliday and
Hasan 1976) from the perspective of Gleason’s work on text semantics (e.g.
Gleason 1968) – as a set of text-forming resources realized through gram-
mar, lexis, and intonation. From this perspective, Halliday and Hasan’s
cohesive ties are reinterpreted as discourse structures; and the resources
which Halliday (e.g. 2009:85) positions as non-structural components of the
textual metafunction in grammar are reinterpreted at a deeper level of
abstraction as discourse semantic systems. This reinterpretation fore-
grounds meaning beyond the clause as fundamental to semantic analysis
in SFL and can be usefully compared with the clause semantics research foci
inspired by Halliday and Matthiessen (1999), Hasan (2009), or Fawcett
(2008). Halliday and Hasan (1985:82 in particular) can be read as developing
work on cohesion in a similar direction.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 359
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
360 J. R. MARTIN
together under the heading collocation in Halliday and Hasan (1976). The
discourse structures afforded by these ideation relations are termed lexical
strings.
Conjunction integrates earlier work on cohesion (Halliday and Hasan
1976) and expanding clause complexes (Halliday and Matthiessen 2014)
through consideration of semantic relations between figures, whether real-
ized within or between clause complexes – i.e. cohesively between clause
complexes, paratactically or hypotactically between clauses in a clause
complex, or via logical metaphor (Halliday 1998) as a participant, process,
or circumstance within a clause. Four main types of conjunctive relation
are recognized: additive, comparative, temporal, and consequential. Halli-
day and Hasan’s important distinction between external and internal con-
junctive relations is sustained. And the model recognizes the possibility of
implicit conjunctive relations abduced in the interpretation of adjacent
figures (with the constraint that any relation so abduced could be made
explicit). The discourse structures afforded by these conjunctive relations
are modelled as reticula, elaborating on Gleason’s (1968) notion of an
‘event-line’.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 361
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
362 J. R. MARTIN
In this section I will try and illustrate the discourse semantic resources
introduced above and address the practical concerns of this handbook by
engaging in an exercise I will refer to as text synthesis. Text synthesis
proceeds by building up a coherent text and thus contrasts with text
analysis which starts with a finished product and breaks it down. Synthesis
is rhetorical in orientation; it aims to demonstrate how linguistic resources
can be deployed. As such it reflects a long-standing but until recently under-
utilized tradition in language teaching – one which features in my grand-
father’s language textbooks in late nineteenth-century rural Canada, but
had to be re-introduced in the genre-based literacy programmes designed
by the ‘Sydney School’ (Rose and Martin 2012). Synthesis disappeared, one
has to presume, because the knowledge about language available to stu-
dents and teachers across sectors in education became so impoverished that
it could not sustain rhetorically oriented text construction. Now, thanks
particularly to SFL, the knowledge about language we need is readily avail-
able. So let us put it to use.
As our starting point, consider the following string of alphabetically
listed clauses. How might we begin to compose a text from these?
Let us begin with ideation. Nuclear relations have been provided for us, in
order to make this exercise fit into a short chapter of this kind (for a
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 363
[2] The timber roof on the House of Meleager was designed for light-
weight roofing. The roof could not support the weight of the tiles.
The roof collapsed. Inexpensive mortar has also been used. This was
to protect ancient stonework. Over time this mortar has cracked.
This allowed water and vegetation to penetrate. çè Over thirty
different varieties of weed have been identified, including ivy,
fennel, and fig. The roots grow. The roots open up further cracks.
This allows even more weeds in. çè No walkways for viewing
platforms have been constructed. Tourists walk along ancient paths.
Tourists enter buildings that are not roped off. In some places
ancient lead water pipes have been exposed. çè Damaged paths
and walls have not been repaired. Frescoes have not been preserved.
Mangy dogs roam the site. Available finance has not been managed.
No conservation and interpretation programme has been put in place.
The principal lexical strings underpinning this construal of the field are
outlined in Table 14.1 – with headings highlighting the kinds of relation
involved as mostly based on composition (co/meronymy)1 or mostly based
on classification (co/hyponymy). In general terms then our gaze shifts from
construction to vegetation to vantage points to aspects of the site overall.
[3] The timber roof on the House of Meleager was designed for light-
weight roofing. The roof could not support the weight of the tiles.
The roof collapsed. Inexpensive mortar has also been used. This was
1
In Table 14.1 the ‘co/meronymy’ headings indicate that lexical relations involving both meronymy (part/whole
relations) and co-meronymy (part/part relations) are found; the ‘co/hyponymy’ heading indicates that both hyponymy
(class/subclass relations) and co-hyponymy (subclass/subclass relations) are found.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
364 J. R. MARTIN
At this point we can bring identification into the picture and consider
anaphoric identity chains introducing (technically speaking ‘presenting’)
and tracking (technically speaking ‘presuming’) the entities introduced as
the timber roof on the House of Meleager, poor quality mortar, over thirty different
varieties of weed, and tourists. Each of these entities was fully lexicalized in
Text 3; this redundancy is adjusted as in Text 4 below.
[4] The timber roof on the House of Meleager was designed for light-weight
roofing. It could not support the weight of the tiles. It collapsed.
Inexpensive mortar has also been used. This was to protect ancient
stonework. Over time this mortar has cracked. This allowed water and
vegetation to penetrate.
Over thirty different varieties of weed have been identified, including
ivy, fennel, and fig. The roots grow. They open up further cracks. This
allows even more weeds in.
No walkways for viewing platforms have been constructed. Tourists
walk along ancient paths. They enter buildings that are not roped off.
In some places ancient lead water pipes have been exposed.
Damaged paths and walls have not been repaired. Frescoes have
not been preserved. Mangy dogs roam the site. Available finance has
not been managed. No conservation and interpretation programme
has been put in place.
The identity chains at issue here are outlined in Table 14.2. The mortar,
weeds, and tourists chains are initiated non-phorically (via non-specific
the timber roof . . . inexpensive mortar over thirty different varieties . . . tourists
it this mortar the roots they
it they
even more weeds
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 365
nominal deixis); the roof chain on the other hand involves both esphora (the
timber roof pointing forward in its nominal group to its Qualifier on the House
of Meleager) and homophora, since the House of Meleager is treated as assumed
knowledge in this field. Thereafter pronouns and specific deixis are used to
track entities (it, it; this mortar; the roots; they); and a ‘sub-entity’, the roots of
the weeds, is introduced via bridging (taking advantage of the part/whole
ideation relation between weeds and roots). The weeds chain also includes a
comparative reference, introducing an additional set of weeds beyond the
weeds initiating the chain.
As we can see, even in a few phases of discourse of this kind, as far as
identification is concerned there is a lot going on. This can be especially
challenging for speakers coming from a language that manages identifica-
tion differently from English. The lack of an obligatory presenting/presum-
ing reference distinction in many languages is especially troubling (as
teachers and supervisors of academic writing well know); and this problem
may be exacerbated by the fact that in such languages the distinction
between specific and generic reference is not explicitly grammaticalized.
We should also note at this point three instances of text reference
(technically ‘extended reference’) in Text 4. With text reference indefinitely
long phases of meaning can be presumed (one or two sentences worth in
the examples below). As the term implies, what is identified is phases of
unfolding discourse rather than specific discourse semantic entities that
are introduced and tracked as part of the construal of a field.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
366 J. R. MARTIN
[5a] The timber roof on the House of Meleager was not designed to
support the weight of the tiles and collapsed. Inexpensive mortar
has also been used to protect ancient stonework. Over time this
mortar has cracked, allowing water and vegetation to penetrate.
Over thirty different varieties of weed have been identified, includ-
ing ivy, fennel, and fig. As the roots grow they open up further
cracks, allowing even more weeds in.
No walkways for viewing platforms have been constructed, so
tourists walk along ancient paths and enter buildings that are not
roped off. In some places ancient lead water pipes have been
exposed.
Damaged paths and walls have not been repaired, frescoes have
not been preserved, and mangy dogs roam the site. Available finance
has not been managed, and no conservation and interpretation pro-
gramme has been put in place.
These relations are specified below; the specification (-) indicates that a
comparative, temporal or consequential conjunctive relation cannot be
made explicit between these two clauses (by convention, implicit additive
relations are left unspecified, by way of lightening the workload for
text analysts). Note that the use of non-finite clauses (e.g. allowing
water and vegetation to penetrate) and branched paratactic clause
complexes (e.g. so tourists walk along ancient paths and enter buildings that
have not been roped off) means that certain entities are ideationally
implicit, and are thus tracked here via ellipsis as far as identity chains are
concerned. This further reduces the ideational redundancy apparent in Text
2 above.
[5b] The timber roof on the House of Meleager was not designed
(explicit purpose)
to support the weight of the tiles
(explicit additive, implicit causal)
and collapsed.
(-)
Inexpensive mortar has also been used
(explicit purpose)
to protect ancient stonework.
(implicit succession)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 367
At this point we can turn to periodicity and consider how to scaffold the
waves of information in the text. Text 5’s topical Themes are highlighted in
bold below for each of its finite ranking clauses;4 (“) there indicates that the
orientation to the field is being sustained through ellipsis (a more common
pattern in many languages than in English, both within and between clause
complexes).
[5c] The timber roof on the House of Meleager was not designed
to support the weight of the tiles
2
Over time is interpreted here as a circumstance of Location, ideationally construing the time frame over which the
stonework and mortar problem arose, rather than as a temporal connector realizing conjunction.
3
The dependent non-finite clause, including ivy, fennel, and fig, grammatically elaborates the varieties of weed that have
invaded; accordingly it is not interpreted here as a distinct figure conjunctively related to the identification of the weeds.
4
Theme has not been analyzed in non-finite ranking clauses, with their non-finiteness interpreted as downgrading the
figure as far as textual and interpersonal meaning are concerned.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
368 J. R. MARTIN
[5d] The timber roof on the House of Meleager was not designed
to support the weight of the tiles
and collapsed.
Inexpensive mortar has also been used
to protect ancient stonework.
Over time this mortar has cracked,
allowing water and vegetation to penetrate.
Over thirty different varieties have been identified,
including ivy, fennel, and fig.
As the roots grow
they open up further cracks,
allowing even more weeds in.
5
Following Martin and Rose (2003) Topical Theme in declarative clauses is analyzed up to and including the Subject.
6
Since intonation does not specify how much of the clause beyond the constituent containing the tonic syllable is
involved, this reading has to be undertaken in relation to the ‘point’ of each phase – by including information relevant to
expanding the field.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 369
[6] Much of the restoration work on Pompeii has been done by local
firms with no knowledge of restoration techniques.
For example, the timber roof on the House of Meleager was not
designed to support the weight of the tiles and collapsed. Inexpensive
mortar has also been used to protect ancient stonework. Over time this
mortar has cracked, allowing water and vegetation to penetrate.
Damaged paths and walls have not been repaired, frescoes have not
been preserved, and mangy dogs roam the site. Available finance has
not been managed, and no conservation and interpretation pro-
gramme has been put in place.
7
Writers who proceed from a carefully constructed plan lean towards front-loading of this kind, since they know where
they are going as they write; writers who figure out what they want to say as they work through successive drafts may
find periodic summarizing, via hyper-News, more appropriate.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
370 J. R. MARTIN
[7] To begin, much of the restoration work on Pompeii has been done
by local firms with no knowledge of restoration techniques. For
example, the timber roof on the House of Meleager was not designed
to support the weight of the tiles and collapsed. Inexpensive mortar
has also been used to protect ancient stonework. Over time this
mortar has cracked, allowing water and vegetation to penetrate.
Second, the incursion of uncontrolled weeds has hastened the
decay of the ruins. Over thirty different varieties have been identi-
fied, including ivy, fennel, and fig. As the roots grow they open up
further cracks, allowing even more weeds in.
In addition, Pompeii’s position as an international tourist attrac-
tion brings half a million visitors each year. No walkways for viewing
platforms have been constructed, so tourists walk along ancient
paths and enter buildings that are not roped off. In some places
ancient lead water pipes have been exposed.
Finally, there seems to be no overall management plan for the site.
Damaged paths and walls have not been repaired, frescoes have not
been preserved, and mangy dogs roam the site. Available finance has
not been managed, and no proper conservation and interpretation
programme has been put in place.
[8] To begin, much of the restoration work on Pompeii has been done
by local firms with no knowledge of restoration techniques. For
example, the timber roof on the House of Meleager was not designed
to support the weight of the tiles and collapsed. Inexpensive mortar
has also been used to protect ancient stonework. Over time this
mortar has cracked, allowing water and vegetation to penetrate.
A second factor is the incursion of uncontrolled weeds which
have hastened the decay of the ruins. Over thirty different varieties
have been identified, including ivy, fennel, and fig. As the roots grow
they open up further cracks, allowing even more weeds in.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 371
Let us now bring periodicity back into the picture, and cook up a rhetorical
sandwich for the phases as a whole – introducing a macro-Theme to predict
the hyper-Themes of each phase and a macro-New to consolidate their News.
macro-Theme
Since its discovery, Pompeii has been damaged as an archaeological site,
affected by the quality of restoration work, vegetation, tourism, and site
management issues.
[phases 1–4]
Damage to the site remains a key issue for archaeologists and administrators.
There are ongoing concerns arising in relation to restoration work, uncon-
trolled vegetation, tourism, and management which need to be addressed if
the site is to be preserved for future generations of research and public access.
macro-New
8
We can compare this selection with some other possibilities: thus (conjunction only); as a result (circumstance of
cause now lexicalized as a cohesive conjunction); as a result of this (conjunction plus text reference); as a result of
these factors (conjunction plus text reference plus metadiscourse).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
372 J. R. MARTIN
=
As a result of these factors, damage to the site remains an issue for
archaeologists and administrators. There are decisions which
need to be made in relation to restoration work, uncontrolled
vegetation, tourism, and management which need to be addressed
if the site is to be preserved for future generations of research and
public access.
This bring us to appraisal, and the question of how and how far to make
explicit the evaluative stance appropriate to a history text of this kind. One
fundamental value in historical axiology is the preservation of archaeo-
logical sites, and so propagating a negative prosody is relevant here –
including negative judgements of individuals, agents, and agencies failing
to preserve the site, and negative appreciations of phenomena involved.
Propagation of this prosody is highlighted in Text 10.
[10] To begin, much of the restoration work on Pompeii has been done by
local firms with no specialized knowledge of restoration techniques.
For example, the timber roof on the House of Meleager was so poorly
designed it could not support the weight of the tiles and collapsed.
Poor quality mortar has also been used to protect ancient
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 373
stonework. Over time this mortar has cracked, allowing water and
vegetation to penetrate.
A second factor is the incursion of dangerous weeds which have
hastened the decay of the ruins. Over thirty different varieties have
been identified, including ivy, fennel, and fig. As the roots grow they
open up further cracks, allowing even more threatening weeds in.
In addition to this, Pompeii’s position as an international tourist
attraction brings half a million visitors each year. No special walk-
ways for viewing platforms have been constructed, so tourists walk
along ancient paths and enter buildings that are not roped off. In
some places ancient lead water pipes have been carelessly exposed.
Finally, there seems to be no overall management plan for the site.
Damaged paths and walls have not been repaired, frescoes have not
been preserved, and mangy dogs roam the site. Available finance has
been poorly managed, and no proper conservation and interpret-
ation programme has been put in place.
Here, three negative judgements have been inscribed, the first of which is
intensified (so poorly designed, have been carelessly exposed, has been poorly
managed); and there are six negative appreciations, the fourth quantified (no
specialized9 knowledge of restoration techniques, poor quality mortar, dangerous
weeds, even more threatening weeds in, no special walkways, no proper conser-
vation and interpretation programme). This puts us in a position to predict the
negative prosody of conservation problems in the text’s macro-Theme
(adversely affected by a number of conservation problems) . . .
As a result of these problems, damage to the site remains a key issue for
archaeologists and administrators. There are important decisions which
need to be made in relation to poor restoration work, invasive vegetation,
insensitive tourism, and neglectful management which need to be
addressed if the site is to be preserved for future generations of research
and public access.
9
Strictly speaking there is an interaction of engagement (no realizing contract: deny) and attitude (specialized realizing
appreciation: valuation) enacting the negative evaluation here (as for no special walkways, no proper conservation and
interpretation programme as well).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
374 J. R. MARTIN
[11a] Outcome
Since its discovery, Pompeii has been damaged as an archaeological
site, adversely affected by a number of conservation problems,
including the quality of restoration work, vegetation, tourism, and
site management issues.
Factor 1
To begin, much of the restoration work on Pompeii has been done by
local firms with no specialized knowledge of restoration techniques.
For example, the timber roof on the House of Meleager was so poorly
designed it could not support the weight of the tiles and collapsed.
Poor quality mortar has also been used to protect ancient stonework.
Over time this mortar has cracked, allowing water and vegetation to
penetrate.
Factor 2
A second factor is the incursion of dangerous weeds which have
hastened the decay of the ruins. Over thirty different varieties have
been identified, including ivy, fennel, and fig. As the roots grow they
open up further cracks, allowing even more threatening weeds in.
Factor 3
In addition to this, Pompeii’s position as an international tourist
attraction brings half a million visitors each year. No special walk-
ways for viewing platforms have been constructed, so tourists walk
along ancient paths and enter buildings that are not roped off. In
some places ancient lead water pipes have been carelessly exposed.
Factor 4
Finally, there seems to be no overall management plan for the site.
Damaged paths and walls have not been repaired, frescoes have not
been preserved, and mangy dogs roam the site. Available finance has
been poorly managed, and no proper conservation and interpret-
ation programme has been put in place.
Reinforcement of factors
As a result of these problems, damage to the site remains a key issue
for archaeologists and administrators. There are important decisions
which need to be made in relation to poor restoration work, invasive
vegetation, insensitive tourism, and neglectful management which
need to be addressed if the site is to be preserved for future gener-
ations of research and public access.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 375
Pompeii has been described as a victim of state neglect and indifference and
an archaeological catastrophe of the first order. Its ongoing destruction
since its discovery in the 1590s has arguably resulted in a greater disaster
than its initial destruction by the eruption of Mt Vesuvius one and a half
millennia earlier.
From another perspective, the ideation at risk here could have been used to
construe a classifying report listing the ways in which archaeological sites
are potentially at risk. A suitable macro-Theme might be the following:
There are several types of damage that can affect archaeological sites. These
include the quality of restoration work, vegetation, tourism, and site man-
agement issues.
In a genre of this kind, appraisal has a much quieter role to play since the
focus is on classifying and describing types of phenomena (in this case types
of activities that are nominalized as abstract things). The macro-Theme
classifies the types of damage, and the remainder of the text describes each
type in detail. For relevant discussion of genre typology and topology, see
Martin and Rose (2008).
Returning to our factorial explanation, it is salutary to keep in mind that
most of the scaffolding we have introduced whereby Text 11a in effect
announces its genre is invisible to an untrained eye. In everyday terms,
only the paragraphing is visible (Text 11b below) as a reflection of all that is
going on. Bringing the discourse semantics of genres to consciousness, and
using this knowledge to design and inform teaching practice, has been an
ongoing concern in SFL since the inception of the genre-based literacy
pedagogy of the ‘Sydney School’ (Rose and Martin 2012).
10
Note how expanding engagement resources have been used here to position the inscribed evaluation as contestable:
has been described as, arguably.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
376 J. R. MARTIN
collapsed. Poor quality mortar has also been used to protect ancient
stonework. Over time this mortar has cracked, allowing water and
vegetation to penetrate.
A second factor is the incursion of dangerous weeds which have
hastened the decay of the ruins. Over thirty different varieties have
been identified, including ivy, fennel, and fig. As the roots grow they
open up further cracks, allowing even more threatening weeds in.
In addition to this, Pompeii’s position as an international tourist
attraction brings half a million visitors each year. No special walk-
ways for viewing platforms have been constructed, so tourists walk
along ancient paths and enter buildings that are not roped off. In
some places ancient lead water pipes have been carelessly exposed.
Finally, there seems to be no overall management plan for the site.
Damaged paths and walls have not been repaired, frescoes have not
been preserved, and mangy dogs roam the site. Available finance has
been poorly managed, and no proper conservation and interpret-
ation programme has been put in place.
As a result of these problems, damage to the site remains a key
issue for archaeologists and administrators. There are important
decisions which need to be made in relation to poor restoration
work, invasive vegetation, insensitive tourism, and neglectful man-
agement which need to be addressed if the site is to be preserved for
future generations of research and public access.
The sentence would then be read aloud, and its elements identified and
elaborated as follows:
11
For access to this pedagogy, see Rose and Martin (2012) and Rose’s R2L website: www.readingtolearn.com.au.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 377
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
378 J. R. MARTIN
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 379
References
12
Lemke (1985:287–8) refers to these semantic relations as covariate relations, contrasting them with the multivariate
and univariate relations constituting lexicogrammatical structure.
13
Equally confounding is the practice in much SFL research of treating field, tenor, and mode as realized directly by
lexicogrammatical systems, bypassing discourse semantic systems altogether in text analysis. The final chapter of
Martin et al. (2010) explores an alternative rite of passage.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
380 J. R. MARTIN
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
Discourse Semantics 381
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:24:47, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.016
15
Appraisal
Susan Hood
15.1 Introduction
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 383
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
384 SUSAN HOOD
1
For further insights into the complementary system of N EGOTIATION , see Martin and Rose (2007: Chapter 7).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 385
Figure 15.1 An outline of the system of appraisal (from Martin and White 2005:38)
affect
category judgement
appreciation
positive
ATTITUDE
‘vibe’
negative
inscribe
mode of
realization
invoke
Figure 15.2 Simultaneous choices in identifying instances of ATTITUDE (adapted from Liu
2017:74)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
386 SUSAN HOOD
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 387
for other than a few core lexical expressions. For a more detailed linguistic
justification for the categorizations in Martin and White (2005), see Martin
(2017:33–6). In the interest of clarification, Martin (2017) does propose a
renaming of the sub-category of triggered insecurity, from ‘surprise’ to
‘perturbance’.
Martin (2017) explains the development of this typological perspective on
affect as an evolutionary process:
I was parenting a small child at the time and suggested categories based on
my reading of his emotional repertoire in relation to his parents coping (or
not) with his moments of distress – basically asking whether he was
unhappy because he wanted his mother or father (contented sociability), or
because he wanted the comfort of his security blanket (which he called
‘baggy’), or because he wanted the satisfaction of his bottle (‘bopple’). This
gave us the [unhappiness/happiness], [insecurity/security] and [dissatisfaction
/satisfaction] oppositions.
(Martin 2017:31)
He continues:
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
388 SUSAN HOOD
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 389
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
390 SUSAN HOOD
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 391
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
392 SUSAN HOOD
Figure 15.5 A network of modes of realization for ATTITUDE (from Hood and Martin
2007:746)
found that rather than mirroring the amplified levels of emotive intensity
on the part of customers, the agents made frequent use of concessive
contractors (and sometimes silence) to rein in or defuse the attitudinal
intensity. For example, an expression such as just a few days on the part of
an agent could defuse an amplification of extended time by the caller.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 393
assume a positive charge, and for yet another it might be read as purely
categorical, as describing, for example, a room with desks in rows and a
blackboard at the front of the room.
(3) Milford (2000) showed that the rate of success fell with age.
Values also propagate across longer phases of text. In example (4), the first
clause constitutes a hyper-Theme (underlined) in which the ideational
focus, ‘methodology’, couples with positive appreciation in ‘refinements’.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
394 SUSAN HOOD
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 395
well as options where sources are implied and processes are metaphorized,
as in examples (5) to (10):
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
396 SUSAN HOOD
Figure 15.6 The system of ENGAGEMENT (from Martin and White 2005:134)
Given the broad and rapidly expanding scope of studies drawing on the
system of appraisal, it is impossible to provide anywhere near a comprehen-
sive account of existing literature. Distinct categorizations are also a chal-
lenge. Contributions from applied studies are organized loosely around
dimensions of fields, features, and questions explored.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 397
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
398 SUSAN HOOD
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 399
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
400 SUSAN HOOD
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 401
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
402 SUSAN HOOD
As identified in this chapter, for well over a decade SFL discourse scholars
have applied the tools of appraisal with more or less attention to particular
dimensions of the system, to who is doing the appraising, to the audiences
with whom they are negotiating values, and to the influence of genre, field,
and mode on the potentials for evaluation. Studies have addressed them-
selves to diverse questions in diverse fields, with diverse kinds of data,
including different semiotic modes, and in a growing number of languages.
Analytical perspectives have varied from static to dynamic, and multiple
methods of analysis have been deployed. I conclude the chapter with some
insights into one front of knowledge that is of current interest to those
working to extend SFL theorizations of interpersonal meaning in discourse
semantics. This refers to the modelling of the dynamic complexity of social
interactions and relations as the semiotics of identity and affiliation.
Here, a number of additional dimensions of SFL theory need some intro-
duction. First is the concept of ‘instantiation’ that models how the meaning
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 403
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
404 SUSAN HOOD
References
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 405
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
406 SUSAN HOOD
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 407
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
408 SUSAN HOOD
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
Appraisal 409
White, P. R. R. 1998. Telling Media Tales: The News Story as Rhetoric. PhD
Thesis, University of Sydney.
White, P. R. R. 2003. Beyond Modality and Hedging: A Dialogic View of the
Language of Intersubjective Stance. Text 23(2): 259–84.
White, P. R. R. 2012. Exploring the Axiological Workings of ‘Reporter Voice’
News Stories: Attribution and Attitudinal Positioning. Discourse, Context
and Media 1: 57–67.
Yu, L., X. Peng, Z. He, Y. Liu, R. Zhang, X. Tan, and Y. Wang. 2017. Generic
Distributions of English Appraisal Categories Based on Appraisal Corpus.
In J. J. Webster and X. Peng, eds., Applying Systemic Functional Linguistics:
The State of the Art in China Today. London: Bloomsbury. 169–83.
Zappavigna, M. 2012. Discourse of Twitter and Social Media: How to Use Language
to Create Affiliation on the Web. London: Continuum.
Zappavigna, M., D. Cléirigh, P. Dwyer, and J. R. Martin 2010. Visualising
Appraisal Prosody. In A. Mahboob and N. Knight, eds., Appliable Linguistics:
Texts, Contexts, and Meanings. London: Continuum. 150–67.
Zappavigna, M., P. Dwyer, and J. R. Martin. 2008. Syndromes of Meaning:
Exploring Patterned Coupling in a NSW Youth Justice Conference. In
A. Mahboob and N. Knight, eds., Questioning Linguistics. Newcastle upon
Tyne: Cambridge Scholars Publishing. 164–85.
Zhang, M. and T. Guo. 2014. Event-related Brain Potentials Differentiate
Three Types of Emotional Words Categorized from Linguistic Perspective.
Journal of Neurolinguistics 31: 17–27.
Zhang, M., Y. Ge, C. Kang, T. Guo, and D. Peng. 2018. ERP Evidence for
the Contribution of Meaning Complexity Underlying Emotional Word
Processing. Journal of Neurolinguistics 45: 110–18.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 06 Aug 2019 at 14:29:28, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.017
16
SFL and Diachronic
Studies
David Banks
What Lyons says about dead languages is true of former stages of still living
languages too. However, it might be felt that such studies would involve at
least implicit contrast with the present-day language for a contemporary
reader. For example, the modern Anglophone reader of a purely synchronic
account of, say, Middle English, would inevitably contrast it with his own
language which has developed from Middle English. Of course, such con-
trasts can be made explicit, which would draw the study in the direction of
I would like to thank Geoff Thompson for his insightful and helpful remarks and suggestions on earlier drafts of this
chapter, which he made only shortly before his untimely death in November 2015. It goes without saying that I am
solely responsible for any shortcomings that may remain.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 411
diachrony. Thus, there seem to be at least four points on the scale from
historical synchronic to diachronic: synchronic study of a now dead lan-
guage; synchronic study of the former state of a still living language, with
implicit contrast with its current state; synchronic study of the former state
of a still living language, with explicit contrast with its current state; and
study of the development of a language over time, which is diachronic
study properly so-called; and of course, it is easy to imagine finer gradations
between these four points.
Of course, no one would deny that natural languages are in a constant
state of evolution; they only stop evolving when they are no longer used as a
means of communication, and are thus dead. So, in a sense, since languages
are in a constant state of evolution, a language in its synchronic state is
something which does not exist in the real world. Treating a language
synchronically is a tactic, like taking a photograph or freeze-framing some-
thing in motion, to make it easier to study. This is quite reasonable, but it
must be accepted that a language in its synchronic state is a human
construct, not something which exists in that state. It is the tactic that
has been adopted by Systemic Functional Linguists in general, and, indeed,
by those working within many other frameworks as well, to the extent that
historical, or diachronic, linguistics is often treated, and felt, as a quite
different branch of linguistic studies. As a result, diachronic studies in a
Systemic Functional Linguistics (SFL) framework are relatively rare, and SFL
excursions into the diachronic field are fairly sporadic.
What I shall aim to do in this chapter is to outline these SFL incursions
into the diachronic field. The most important of these, which can be seen as
diachronic studies properly speaking, occur in the work of Michael Halliday
(notably Halliday 1988, 1993), where he studies a number of scientific texts
over a wide period of time, showing such features as the development of
grammatical metaphor, and how this is used in thematic structure to build
up an argument. Michael Cummings (1995, 2010) has used the SFL frame-
work as a basis for his description of Old English (OE). Although dealing
with a single period, in so far as he contrasts OE with present-day English
this might be considered marginally diachronic, or historical, with explicit
contrast to contemporary English. My own work comprises studies which
are both fully diachronic and historical with implicit contrast with present-
day language. In the former category (mainly Banks 2008, 2017), I studied a
corpus of articles from the Philosophical Transactions covering the period
1700–1980. In the latter category, I have more recently considered the
seminal period of 1665–1700, contrasting the first two academic periodic-
als, the Journal des Sçavans in French, and the Philosophical Transactions in
English. To this body of research can be added the work of Martínez-Insua
(2013) on there-constructions, Starc (2010) on advertisements, O’Halloran
(2005) on mathematics, and Urbach (2013) on register variation.
Although diachronic studies using the SFL framework are not numerous,
there is quite a lot of work which, while not strictly speaking SFL, is
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
412 DAVID BAN KS
The work of Michael Cummings on Old English (OE) (1995, 2010) is basically
a description of the language of the OE period. It falls into the class which is
a synchronic study, with explicit contrast to the modern state of the
language. To that extent, this work does not perhaps strictly count as
diachronic linguistics; however, it can be included in this survey, at least
marginally, for two reasons. First, the OE period itself extends over almost
five centuries, from the end of the seventh to the early twelfth century, so it
can be assumed that a description of OE over this period must necessarily
include a certain degree of change over the period, though admittedly this
is not something that Cummings himself explicitly goes into. Secondly, and
more importantly, in addition to describing OE, Cummings sets out to show
the differences between that variety and present-day English: ‘The systemic
approach also highlights areas in which significant changes have taken
place in the transition from Old to modern English’ (Cummings 2010:2).
One of the main areas where present-day English differs from OE is in the
interpersonal metafunction. In OE, the items which function as modal or
perfective operators still retain their original lexical content. This means
that in combination with another lexical verb, they are in the process of
transition between a lexical verb combined with an infinitive or participle
and an operator proper with a lexical predicator. In the case of the peri-
phrastic perfect with the verb habban (to possess), the lexical meaning is
already diminished in OE, as in the example from Beowulf in Table 16.1.
All of the verbs which would evolve into the modal auxiliaries of present-
day English could still be used as full lexical verbs, but also with highly
after him the creator condemned had amongst the kin of Cain
Subject Predicator Finite
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 413
Gehyrest þu Eadwacer?
Forðan ælc þæra þe ongean þæt to ælc þæra bið Antecrist genamod
swyðe deð oððerne ongean
þæt læreð þe his
cristendome to gebyreð
For each of those who sins too each of those is Antichrist called
greatly against that or who
teaches another contrary to
what belongs to his
Christianity
Theme Rheme
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
414 DAVID BAN KS
Cummings also notes a number of differences within the noun group. For
example, when a noun group containing a determiner functions as the
Complement of a prepositional phrase, in OE, but not in present-day Eng-
lish, the determiner is frequently placed before the preposition; this is the
case of Him in the example in Table 16.5 from Beowulf.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 415
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
416 DAVID BAN KS
points out, this was to become ‘an essential resource for constructing
scientific discourse. We see it emerging in the language of this period,
when the foundations of an effective register for codifying, transmitting
and extending the “new learning” are rapidly being laid down’. With this
device, instead of clauses dominated by a process, we can have clauses
which express relations between processes, either of an external type, ‘a
causes x to happen’, or of an internal type, ‘b causes me to think y’. Halliday
and Martin (1993) give the following examples from Newton:
(1) The explosion of gunpowder arises therefore from the violent action
whereby . . .
(Halliday and Martin 1993:67)
(2) Now those Colours argue a diverging and separation of the heteroge-
neous Rays . . .
(Halliday and Martin 1993:65)
He points out that in example (1) there are two nominalized processes,
explosion and action, and these are linked by the verb arises, expressing the
fact that one is caused by the other. In example (2), evidence (the colours)
leads the observer to deduce (they ‘argue’) a possible explanation.
Halliday finds that the use of this type of grammatical metaphor (nom-
inalized processes) has been taken even further in Priestley’s The History and
Present State of Electricity, with Original Experiments, published in the 1760s.
The ‘nominal elements in the clause are gradually taking over the whole of
the semantic content, leaving the verb to express the relationship between
these nominalized processes’ (Halliday 1988:171). This process has gone on
developing up to the present day, so that Halliday can set up a schematic
representation of the progress of this phenomenon. This has both external
and internal forms. Externally the schema is as follows:
Internally it is as follows:
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 417
We used an ISOL, and thus could experiment even where it was difficult
By using an ISOL we solved the difficult parts of the experiment
Our using an ISOL resolved the difficulties of the experiment
The experimental difficulties were resolved by the use of an ISOL
The resolution of the experimental difficulties came in the form of
an ISOL
This does not mean that this process is complete and that the final form is
the one we will inevitably encounter in contemporary scientific texts.
Probably all the schema could be found somewhere in texts today. What
Halliday is saying is that over time there is a general trend towards the later
more metaphorical forms, so that there will be a tendency, as time goes on,
to find more of the later forms.
He takes these points up again in Halliday and Martin (1993), where he
uses them to give a detailed analysis of the final section of Darwin’s On the
Origin of Species by Means of Natural Selection, published in 1859. He shows
that grammatical metaphor is not only important at the level of clause, but
also plays a significant role in thematic progression, and hence in the flow
of information and the cohesion of the text as a whole. Items which appear
in non-metaphorical form in rhematic sections, where they are treated as
New, are taken up again in nominalized (grammatical) form as Theme and
Given in a following clause. The argument can thus be built up in this way
from Rheme to Theme. Not only does this give the text a tight cohesion, but
the nominalized process is now backgrounded as Given in thematic pos-
ition. As a noun, it is presented as something fixed and solid, whose
existence cannot be questioned, thus ‘the nominalization picks up
the preceding argument and presents it in this “objectified” form as some-
thing to be taken for granted’ (Halliday and Martin 1993:98). Hence, the
thematic progression moves from the dynamic presentation of the process
as Rheme and New to its presentation in nominal form, and thus as an
objectified item, as Theme and Given. This can be seen in example (3) from
an article in the Philosophical Transactions for the year 1900 (taken from
Banks 2008:153):
(4) But having beheld it with a Telescope, I soon said, that it was joyned
with two small Stars, whereof one was pretty bright, which I had
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
418 DAVID BAN KS
already seen, on February 28, and 29. And this conjunction gave the
Comet that brightness, as it happens to most of the Stars of the fifth
and sixth magnitude . . .
Here, the verbal form was joyned is taken up by the nominalized form
conjunction as theme of the following clause.
Halliday’s study of the rise of grammatical metaphor, notably in the form
of nominalized processes, in scientific writing and the link between gram-
matical metaphor and thematic progression may well be the most signifi-
cant contribution that SFL has made to date in the area of diachronic
linguistics.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 419
(5) The ivory was supported horizontally by a stand made of the pre-
pared wood. When the glass was made a little warmer than the
external air, my finger rubbed that side thereof which was furthest
from, and opposite to the ivory.
(Philosophical Transactions 1780, from Banks 2008:102)
(6) This change has been observed in some of the bird tribe, but princi-
pally in the common pheasant.
(Philosophical Transactions 1780, from Banks 2008:103)
(7) In order to do that, I tied a stick of sealing wax to a silk string about a
yard long, and after having excited it very powerfully with flannel,
I plunged it in a tin vessel full of water, and immediately drawing it
out, brought a very accurate electrometer near it, and observed, that
at first it shewed no sign of electricity; but in about half a minute’s
time it manifested a small but very sensible degree of negative
electricity.
(Philosophical Transactions 1780, from Banks 2008:115)
(8) This was a Sight I little expected to meet with; and being aware how
much Imagination has frequently had to do with microscopical
Observations, I distrusted my own eyes.
(Philosophical Transactions 1740, from Banks 2008:116)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
420 DAVID BAN KS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 421
This appears to reflect the major change in focus, mentioned above, from
experimentation to mathematical modelling which takes place in the phys-
ical sciences, but not in the biological sciences, at the beginning of the
twentieth century.
In Banks (2017) I have concentrated on the much shorter time scale of
1665–1700. This was a particularly important period for the establishment
of academic writing since the first two academic periodicals both appeared
in 1665. The first was the Journal des Sçavans in Paris, and it was followed
two months later by the Philosophical Transactions in London. The Journal des
Sçavans, founded by Denis de Sallo, had state support, covered the full range
of new knowledge, and was mainly made up of book reviews (Morgan 1928).
The Philosophical Transactions was founded by Henry Oldenburg as a private
venture; it was mainly restricted to science and technology, and was based
on Oldenburg’s voluminous correspondence (Hall 2002). The cut-off point is
1700 because until that point the Académie Royale des Sciences only published
its papers in luxurious limited editions, which were considered the per-
sonal property of the French monarch, Louis XIV. He used them as gifts
for illustrious visitors, but they were not easy to come by otherwise
(Liccope 1994, 1996). At the end of the seventeenth century, the decision
was made to alter this state of affairs and to publish more widely; the
first of these volumes, that for 1699, actually appeared in print in 1702.
Thus, the period from 1665 to 1700 was peculiar, with the Journal des
Sçavans and the Philosophical Transactions as the only real outlets for aca-
demic writing in French and English. From 1700 onwards this situation
changed radically with the appearance of the Mémoires de l’Académie Royale de
sciences.
As might be expected, Subjects function as Theme in the majority of
cases; however, this is more frequently the case in the Journal des Sçavans
(73 per cent) than in the Philosophical Transactions (62 per cent). Corres-
pondingly there are more adjunct Themes in the Philosophical Transactions
(31 per cent) than in the Journal des Sçavans (22 per cent). Thematic
structure functions in virtually the same way in French and English,
and there is no evidence to suggest that differences between the two
languages might account for the differences in these figures. While the
figures for the Journal des Sçavans are fairly stable over the period, in the
Philosophical Transactions the percentage of subject Themes increases and
that of adjunct Themes falls over the years 1665 to 1695. In both journals,
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
422 DAVID BAN KS
roughly half of the adjunct Themes are clausal in nature. The use of
linear progression rises over time in the Journal des Sçavans (38 per cent
to 45 per cent), but it is stable in the Philosophical Transactions at around 43
per cent. In the French journal the most common semantic category of
Theme is that of humans other than the author (37 per cent), who are
frequently the author of a book under review; the object of study
accounts for a further 22 per cent, and other texts, usually a book under
review, 16 per cent. Use of humans other than the author as Theme is
stable after 1675; that of the object of study shows an increase in 1695;
and that of other texts falls over the period. In the English journal, the
object of study is by far the commonest type of Theme, accounting for 45
per cent overall and rising from 32 per cent to 57 per cent over the period.
Reference to the author as theme accounts for a further 14 per cent, and
humans other than the author, 13 per cent. Hence, it can be seen that the
thematic interest of the Journal des Sçavans is directed primarily towards
humans, and secondarily to objects of study, whereas in the Philosophical
Transactions, this is reversed with objects of study being by far the major
interest, and humans a minor secondary interest. The following examples
show humans other than the author functioning as theme in the Journal
des Sçavans, and the object of study functioning as theme in the Philosoph-
ical Transactions.
The finite verbs were analyzed in terms of process type, using a system of
five processes: Material, Mental, Relational, Verbal, and Existential
(Banks 2005, 2016). In both journals, Relational process accounts for
about 30 per cent of the finite verbs. However, while this is the common-
est category in the Journal des Sçavans, this is not the case in the Philosoph-
ical Transactions, where Material process accounts for 35 per cent, as in
example (16):
(16) I bought a female Rabbet and let it take Buck 3 times in my presence,
(which was quickly done) and then killed it, but did not open the
womb, till a quarter of an hour after; about an Inch from the
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 423
In the Journal des Sçavans, Material process accounts for 27 per cent and
Verbal process a further 24 per cent. In the Philosophical Transactions, Verbal
process accounts for only 13 per cent and falls over the period considered.
The sentence in (17) is an example of Verbal process in the Journal des
Sçavans:
(17) Car il pretend qu’il ne l’auoit pas escrit pour estre publié. (Journal des
Sçavans 1665; from Banks 2017:104) [For he claims that he did not
write it in order to be published.]
(18) Pour voir si les astres peuvent causer dans les hommes quelques
inclinations, il recherche la cause des differentes humeurs. (Journal
des Sçavans 1675; from Banks 2017:134) [To see whether the heav-
enly bodies can produce certain inclinations in men, he looks for the
cause of different humours.]
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
424 DAVID BAN KS
Mental and Verbal processes account for 24 per cent and 26 per cent of
nominalized processes respectively in the Journal des Sçavans, while the
corresponding figures for the Philosophical Transactions are 18 per cent and
14 per cent. There is no distinct pattern of development over the time
period considered. Once again, the greater leaning of the English journal
towards action and event, and the interest of the Journal des Sçavans in
questions of communication are brought out by these figures.
The linguistic features which are highlighted by this study can be shown
to be the direct result of the editorial decisions made by de Sallo and
Oldenburg, in de Sallo’s case to cover the whole range of disciplines and
to print mainly book reviews, and in Oldenburg’s case to concentrate on
science and technology based on his correspondence. These decisions them-
selves can be understood in the context of the differing historical situations
of France and England in the late seventeenth century. France was at that
time the economic and cultural centre of Europe, with Louis XIV on the
throne, and Colbert as his first minister. Colbert wanted to control every-
thing including new knowledge, which he found potentially dangerous.
Since new knowledge was to be found in books, it was natural that the
French periodical should concentrate on book reviews, and since control
was the object of the exercise, no disciplines could be excluded from its
scope. England on the other hand was basking in the new-found hope of the
Restoration, after decades of chaos. The impoverished crown was unable to
subsidize the Royal Society, and Oldenburg, as one of its secretaries, had to
find supplementary sources of income. Hence, he used his voluminous
correspondence to create a bulletin for his potential readership, the
members and friends of the Royal Society, who were interested in scientific
matters. Thus, it can be seen that the linguistic features found in these two
periodicals derive from the editorial decisions made by de Sallo and Old-
enburg, and these decisions themselves can be seen as being determined by
the historical context in which they were made. This does not mean, of
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 425
course, that things could not have been otherwise, but it does mean that the
linguistic features can be seen as consistent with the historical context. This
seems to be an excellent illustration of the SFL contention that there is an
intimate link between language and the context in which it is produced.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
426 DAVID BAN KS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 427
with it. However, it should be noted that contributions from outside of SFL
tend to concentrate on changes in form, whereas, as the discussion so far
will have made clear, SFL, being a social semiotic approach, sees the
changing form at the discourse level within its social and historical context.
Among these non-SFL contributions one might include the vast literature
devoted to ‘grammaticalization’. Traugott (1988:406) defines grammatica-
lization as ‘the dynamic unidirectional historical process whereby lexical
items in the course of time acquire a new status as grammatical, morpho-
syntactic forms’. Although developed against a generative grammar back-
ground, the evolution of grammaticalization has moved away from that
basis (Hopper and Traugott 2003), and the later findings could easily be re-
expressed in SFL terms. This emerges particularly clearly in the association
that is often noted between grammaticalization and subjectification – in
SFL terms, a move from experiential to interpersonal meaning. Present-day
modal auxiliaries are a prime example since they derive from lexical verbs
in Old English (Hopper and Traugott 2003).
The work of Salager-Meyer on hedging and related matters is basically
within a Swalesian tradition (Swales 1990, 2004), but shares with SFL an
interest in discourse. In Salager-Meyer (1999), she studies references in a
corpus of English medical articles covering the period 1810 to 1995. She
distinguishes between references which are critical and those which are
not. Critical references account for 41.6 per cent of the sample for the
period 1810–1929, but only 17.4 per cent for the period 1930–1995. This,
however, is not due to a fall in the incidence of critical references, but to a
sharp continuous increase in non-critical references from 1930 on. More-
over, although critical references are fairly stable over the period con-
sidered, before 1930 they tend to be pointed and personal, as in the
following:
Mr. Brodie objects to my experiments that they were not exact repetition of
his, and therefore not entitled to much consideration in estimating the
causes of animal heat . . . I cannot conceive, I am afraid, on what Mr. Brodie’s
opinion is founded. (1823)
(from Salager-Meyer 1999:20)
We have carried out both the test of Akerfeldt and Gibbs and have been
unable to confirm the findings of either investigator. (1960)
(from Salager-Meyer 1999:22)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
428 DAVID BAN KS
references (here termed direct academic conflicts) in the French corpus, and
more non-critical references (indirect academic conflicts) in the English
corpus. In the period before 1930, however, direct criticism in the English
articles is frequently accompanied by gentlemanly polite remarks; these are
absent in the French texts. In the period after 1930, where the English lose
their aggressive tone, the French texts, even if they become a little less
aggressive, remain fairly hard-hitting, until the end of the twentieth cen-
tury, when they start to resemble the English texts on this point. For those
working within an SFL framework, these processes can be seen in terms of
changes in the interpersonal metafunction; and Salager-Meyer’s emphasis
on discourse categories as subject to diachronic change is very much in line
with SFL approaches.
Valle’s (1999) monograph is also in this line of enquiry. It considers the
development of scientific discourse from a sociolinguistic point of view,
focusing on citation practices, and using a corpus of texts dealing with the
life sciences from the Philosophical Transactions for four periods: 1665–1669,
1765–1768, 1865–1869, and 1965–1966. Among her many results, she
notes that the use of short quotations increases over the period, and that
of long quotations falls – to almost zero by the twentieth century. The
majority of quotations in the eighteenth century are for creating debate,
juxtaposition, comparison, and polemic; by the nineteenth century this
category has fallen, and is equalled by those used for knowledge-
embedding, historical background, or filling gaps. By the twentieth century
these have both fallen, and have been overtaken by those used for social
reasons, gaining support or giving credit, and assigning priority. As in SFL
studies, these changes in discourse are seen as reflecting and construing
social and cultural changes across the period.
Biber and Gray in a series of articles (2010, 2011, 2013), again not SFL,
confirm and build on some of Halliday’s findings. They agree with Halli-
day’s contention that in academic writing there has been a move from a
more elaborate style depending on complex clause structures to one which
is more condensed with a relatively simple clause structure but incorpor-
ating complex nominal groups. Their evidence indicates that this is basic-
ally a twentieth-century change, with its origins in the late nineteenth
century. They note that in a corpus of news reporting texts taken
from Time and The New York Times, while direct and indirect quotes and
nominal modifiers have increased over the twentieth century in both
publications, the increase is greater in The New York Times. Similarly,
use of passives and of-genitives has decreased, but the decrease, particularly
of passives, is greater again in The New York Times. In academic writing, the
use of nouns is steady in history texts from the nineteenth century
onwards, increases in the social sciences, and increases dramatically in
the hard sciences. In making these observations, Biber and Gray warn
against the danger of underestimating register (and indeed sub-register)
differences.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 429
I should like to round off this survey by returning to the work of Michael
Halliday. In a recent work, in collaboration with Jonathan Webster (Halliday
and Webster 2014), the authors hypothesize that over the last thousand years
English has been moving from a transitive to an ergative system. This idea is
mentioned in earlier writing by Halliday (e.g. Halliday 1985:146), but has,
I think, never been expressed with such clarity or force as here. A transitive
system sees things in terms of action and extension, whereas an ergative system
sees them in terms of cause and effect. Thus, in a clause such as Harry opened the
door, the door functions as Goal in a transitive analysis, but as Medium in an
ergative analysis: Goal is the second participant to which the process extends in
a transitive structure, and Medium is the participant which is centrally
involved in, or actualizes, the process. As Medium, door is the participant which
functions as Subject, when there is only one participant (The door opened), thus
showing that it is the participant that is essential for the process to take place.
Traditionally, English verbs are seen as forming a cline from those that are
exclusively transitive, like throw or dig, to those that are exclusively intransitive
like swim or crawl. Halliday and Webster claim that there is a gradual drift, in
modern English, towards the centre of the cline where verbs can function both
transitively and intransitively, like open or ring. What functions as Subject was,
they claim, in earlier periods, selected on the basis of transitivity: that is, the
unmarked choice for Subject was the participant which functioned as Actor in
the transitivity system. But today selection is on the basis of thematic structure:
that is, whatever component the speakers choose as their starting point can be
selected as Subject. The move to an ergative conception of clause structure
makes this easier. They thus reinterpret Modern English as having a cline of
possibilities from verbs which are always non-middle (like throw), to those
which are always middle (like swim). The fact that some verbs function as both
active and passive, such as derive (x derives from y or x is derived from y) is seen as a
natural consequence of the move towards an ergative system. It would seem
that from this point of view English is more ‘mixed’ than many other lan-
guages, and they speculate that a mixture of this kind is unstable. At all events,
this situation, they claim, ‘is the result of extensive change over the last
millennium, the change being away from the transitive type towards the
ergative’ (Halliday and Webster 2014:40), and ‘[t]he ergative interpretation
seems to reflect the direction in which the language is changing’ (Halliday
and Webster 2014:42). It is even possible that we are returning to a state which
obtained in the Indo-European in which English ultimately has its roots.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
430 DAVID BAN KS
the SFL framework. The main contribution that SFL has made to diachronic
studies is the work in the area of the development of scientific writing. This
is basically the work of Michael Halliday, to which I would like to think my
own work provides a suitable supplement. Systemic Functional Linguistics
contributions to the diachronic study of other text types, while perhaps
significant in their own right, remain fairly piecemeal, and cannot be said
to add up to a sizeable body of work. Nevertheless, these contributions,
which could form the basis of future studies, include the work of Starc on
advertisements, O’Halloran on multimodality in mathematics, and Urbach
on context in journalism. In the area of general English, Cummings has
provided an extensive study of Old English, and to this might be added the
work of Martínez-Insua on there-constructions.
Even though the sum of these contributions is not large, they are
sufficient to show that SFL is capable of providing interesting and signifi-
cant insights into the field of diachronic studies. As has been shown,
there are a number of studies which, although outside of SFL proper, are
nevertheless compatible with it, and which, indeed, could have been carried
out in an SFL framework. Systemic Functional Linguistics, however, does
bring a specific point of view to bear on the questions it treats, and it can
only be hoped that in the years to come, linguists working within the SFL
tradition will turn their attention to questions of language development, so
that the particular insights which SFL is capable of offering will enhance
the field of diachronic studies, as they have so many other fields of
linguistics.
References
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
SFL and Diachronic Studies 431
Biber, D. and B. Gray. 2011. Grammatical Change in the Noun Phrase: The
Influence of Written Language Use. English Language and Linguistics 15(2):
223–50.
Biber, D. and B. Gray. 2013. Being Specific about Historical Change: The
Influence of Sub-register. Journal of English Linguistics 41(2): 104–34.
Bynon, T. 1977. Historical Linguistics. Cambridge: Cambridge University
Press.
Crystal, D. 1980. A First Dictionary of Linguistics and Phonetics. London: André
Deutsch.
Crystal, D. 1999. The Penguin Dictionary of Language. 2nd ed. London:
Penguin.
Cummings, M. 1995. A Systemic Functional Approach to the Thematic
Structure of the Old English Clause. In R. Hasan and P. H. Fries, eds., On
Subject and Theme: A Discourse Functional Perspective. Amsterdam: John
Benjamins. 275–316.
Cummings, M. 2010. An Introduction to the Grammar of Old English: A Systemic
Functional Approach. Sheffield: Equinox.
Hall, M. B. 2002. Henry Oldenburg: Shaping the Royal Society. Oxford: Oxford
University Press.
Halliday, M. A. K. 1985. Introduction to Functional Grammar. London: Arnold.
Halliday, M. A. K. 1988. On the Language of Physical Science. In M. Gha-
dessy, ed., Registers of Written English: Situational Factors and Linguistic
Features. London: Pinter. 162–78.
Halliday, M. A. K. and J. R. Martin. 1993. Writing Science: Literacy and Discur-
sive Power. London: The Falmer Press.
Halliday, M. A. K. and C. M. I M. Matthiessen. 2014. Halliday’s Introduction to
Functional Grammar. 4th ed. London: Routledge.
Halliday, M. A. K. and J. J. Webster. 2014. Text Linguistics: The How and Why of
Meaning. Sheffield: Equinox.
Hopper, P. J. and E. C. Traugott. 2003. Grammaticalization. 2nd ed. Cam-
bridge: Cambridge University Press.
Licoppe, C. 1994. The Crystallization of a New Narrative Form in Experi-
mental Reports (1660–1690): The Experimental Evidence as a Transaction
between Philosophical Knowledge and Aristocratic Power. Science in Con-
text 7(2): 205–44.
Licoppe, C. 1996. La formation de la pratique scientifique: Le discours de l’expéri-
ence en France et en Angleterre (1630–1820). Paris: La Découverte.
Lyons, J. 1968. Introduction to Theoretical Linguistics. Cambridge: Cambridge
University Press.
Martínez-Insua, A. E. 2013. There-constructions as a Choice for Coherence in
the Recent History of English. In L. Fontaine, T. Bartlett, and G. O’Grady,
eds., Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge
University Press. 207–25.
Morgan, B. T. 1928. Histoire du Journal des Sçavans depuis 1665 jusqu’en 1701.
Paris: Presses Universitaires de France.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
432 DAVID BAN KS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:10:54, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.018
17
SFL and Multimodal
Discourse Analysis
Kay L. O’Halloran, Sabine Tan, and Peter Wignell
17.1 Introduction
1
See web.archive.org/web/20150620204949/http://apps.who.int/ebola. (Last accessed: 07/08/2017.)
2
See multimodal-analysis.com/products/multimodal-analysis-image. (Last accessed 07/08/2017.)
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
434 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
(e.g. Iedema 2001, 2003), and the major theoretical and analytical chal-
lenges facing multimodal analysts, together with the associated vision for
future research in the field, are explored in detail.
This section presents the theoretical foundations that inform the SF-MDA
approach introduced in this chapter. This approach builds upon Michael
Halliday’s Systemic Functional Theory (SFT) (e.g. Halliday 1978, 1985; Halli-
day and Matthiessen 2014), where semiotic resources are conceptualized in
terms of the functions they serve in society. From an SF-perspective, culture
is defined as a network of semiotic systems, that is, ‘a set of systems of
meaning, all of which interrelate’, and which ‘taken all together, constitute
human culture’ (Halliday and Hasan 1985:4). The practice of modelling the
meaning potential in culture as system networks reflects Halliday’s (1994:
xiv) understanding of systemic theory as ‘a theory of meaning as choice, by
which language, or any other semiotic system, is interpreted as networks of
interlocking options’, whereby the particular choices that are made are not
to be viewed as the result of conscious decisions, but rather as ‘a set of
possible alternatives’ (Halliday 1994:xiv–xxvi).
Although initially applied to language, SFT has since been adapted and
extended to the study of multimodal texts and artefacts to account for
the ways in which linguistic and non-linguistic resources (e.g. spoken
and written language, image, gesture, sound, music, film, page layout,
website design) combine and interact in the communication of meaning
(e.g. Bateman 2014a, 2014b; Bateman and Schmidt 2012; Iedema 2001,
2003; Jewitt 2014; Kress and van Leeuwen 2001, 2006; Machin 2007;
O’Halloran 2004, 2008a, 2008b; O’Toole 2011; Royce 2007, 2015; van
Leeuwen 1999, 2005, 2012).
Because of its adaptability and amenability (e.g. Martin 2002), SFT is
regarded as particularly well suited to provide the theoretical foundation
for SF-MDA (see Jewitt et al. 2015: Chapter 3), as demonstrated, for example,
by Kress and van Leeuwen’s (2006) and O’Toole’s (2011) frameworks for
analyzing images and displayed art (paintings, sculptures, architecture)
respectively. It must be emphasized, however, that while the higher-level
principles of SFT can be applied to the analysis of multimodal texts, the
systems for visual images and other semiotic resources are not the same as
the systems for language. Systemic Functional MDA extends beyond the
simple adaptation of established ‘SF-approaches which were largely
developed for modeling discourse and grammatical systems in language’
(O’Halloran 2008a:446), and requires the development – and integration –
of different, yet complementary, models and approaches for the study of
multimodal semiosis. For this reason, the term SFT is used here rather
than SFL.
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 435
In what follows, key concepts from SFT that form the guiding principles
for modelling language and other semiotic resources and their interactions
are discussed in detail, followed by practical examples of how aspects of the
SF-MDA model can be applied to the WHO Ebola webpage. The issues which
arise when language and other semiotic resources are considered as inter-
related systems of meaning (i.e. ‘semiotic resources as system’) and as
multimodal texts (i.e. ‘semiotic resources as text’) (see Halliday 2008) are
central to this discussion.
17.2.1 Metafunctions
One of the key tenets in SFT is Halliday’s metafunctional principle, which
posits that language and other semiotic systems are structured to make
three kinds of meanings simultaneously: (a) ideational meaning for constru-
ing our experience and knowledge of the world (i.e. experiential meaning)
and making logical connections in that world (i.e. logical meaning); (b)
interpersonal meaning for enacting social relations and expressing attitudes;
and (c) textual meaning for organizing meanings into coherent messages
(e.g. Halliday 1978; Halliday and Matthiessen 2014).
In language, the three metafunctions are mapped on the structure of
the clause by specifying the grammatical systems through which these
metafunctions are realized. For example, experiential meaning is realized
through the grammatical system of Transitivity which accounts for the differ-
ent types of process that are found in a language, and the structures through
which they are expressed (Halliday and Matthiessen 2014:211–335).
A process potentially has three components: the process itself; participants
in the process; and circumstances associated with the process. The following
process types are found in English: Material Processes of doing and acting;
Mental Processes of thinking, feeling, and perceiving; Relational Processes
of classifying and describing (i.e. having attribute) and identifying; Verbal
Processes of saying and meaning; Behavioural Processes of human physio-
logical behaviour (e.g. smiling); and Existential Processes of existing and
happening; as well as the functional participant roles and circumstances
that are associated with these processes. Logical meaning is concerned
with the relations between happenings at the clause and discourse level,
realized through conjunctions (e.g. if, so, moreover). Logical meaning is
mapped through the systems of logico-semantic relations (i.e. nature of the
semantic relations, which is either to expand meaning or to project what is
said or thought) and taxis (i.e. clause dependency relations). Interpersonal
meaning is realized through the grammatical system of Mood, which – at
the level of the clause – realizes ‘meaning as an exchange’ (Halliday and
Matthiessen 2014:135–9), with choices for giving and demanding informa-
tion (statements, questions), and for giving and demanding goods and ser-
vices (offers, commands); and the system of Modality (expressions of
probability, usuality, obligation, and inclination). Textual meaning is
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
436 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 437
Figure 17.1 Register and genre (reproduced from Martin and White 2005:32)
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
438 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Figure 17.2 Language strata (reproduced from Martin and White 2005:9)
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 439
Figure 17.3 Adapted from O’Halloran’s (2008a) SF-MDA framework work for language
and visual imagery
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
440 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
In the preceding section, key concepts from SFT that inform SF-MDA have
been outlined. These concepts are now applied for the analysis of texts,
images, and their relations in the WHO Ebola webpage. While one aim is to
discover how the webpage functions to create meaning, a further and wider
aim is to explore the implications of the SF-MDA approach for the analysis
of such complex multi-semiotic texts, and to demonstrate the possibilities,
strengths, and limitations associated with this approach.
The language in the example texts is analyzed using SFL (e.g. Halliday and
Matthiessen 2014; Martin and Rose 2007), while the visual texts are
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 441
3
See multimodal-analysis.com/products/multimodal-analysis-image. (Last accessed 07/08/2017.)
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
442 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Semiotic
Resource
Metafunction
Rank System Description
Text
EXPERIENTIAL
Clause Processes; Participant Roles; Happenings, actions, and relations
Circumstance
INTERPERSONAL
Clause Speech Function Exchange of information (e.g.
statements and questions) and
goods and services (e.g. commands
and offers)
TEXTUAL
Clause Information Focus Organization of information, with
Discourse points of departure for what follows
Semantics
Image
EXPERIENTIAL
Work Narrative Theme; Nature of the scene
Representation; Setting
Episode Processes; Participant Roles; Visual happenings, actions, and
Circumstance relations
Figure Posture; Dress Characteristics of the participants
INTERPERSONAL
Work Angle; Shot Distance; Lighting Visual effects
Episode Proportion in Relation to the Happenings, actions, and relations
Whole Image: Focus; with respect to the whole image
Perspective
Figure Gaze-Visual Address Direction of participant’s gaze as
internal to image or external to
viewer
TEXTUAL
Work Compositional Vectors; Framing The organization of the parts as a
whole, with the visual marking (e.g.
framing) of certain parts
Episode Relative Placement of the Position of the happenings, actions,
Episode; Framing and relations in relation to the whole
image, and the visual marking of
certain aspects
Figure Relative Placement of the Position of the figure in relation to
Figure within the Episode; happening, action, or relation,
Arrangement; Framing and the visual marking of certain
aspects of the figure
(Figure 17.4 right) are ordered neither hierarchically nor sequentially, but
comprise the generic mix by which the webpage presents information and
engages the reader. The different discourse types found on the Ebola web-
page are Reporting (4), Information (5), Promotion (6), and News (7), respectively
(see Sharoff 2010 for a discussion of genres on the web).
In terms of the registerial mix deployed, the most global organizational
level (Figure 17.4 left) realizes mostly textual or compositional meaning,
whilst the second layer of organization, which contains different discourse
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 443
Figure 17.4 Constituent levels of the Ebola webpage organized in terms of hypertext
components (left): Masthead/Banner (1), Content (2), Bottom Sitemap (3); and discourse
types (right): Reporting (4), Information (5), Promotion (6), News (7)
types (Figure 17.4 right), realizes ideational (i.e. experiential and logical),
interpersonal and textual meaning. In terms of the tenor relations estab-
lished on the webpage, the different sections and subsections are concerned
largely with presenting and reporting information about Ebola, except for
the section entitled ‘Get Involved’, which draws on promotional discourse
(see Section 17.3.5 below).
Each of the webpage’s sections and subsections can be broken down
further into constituent parts and elements (see Figures 17.6 and 17.11
below, for example). In what follows, examples of image-text combinations
found in different sections of the webpage (as displayed in Figure 17.4) are
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
444 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Figure 17.5 Photograph-text complex from the webpage’s main visual display
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 445
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
446 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 447
from the text why they are smiling. The semiotic selections in the photo-
graph and the surrounding contextual factors influence how we perceive
and understand those processes. In this case, the text and image co-
contextualize each other: the main visual participants are a happy smiling
family, and after setting up an expectation of success via the headline
(Despite Ebola, vigilance and hope prevail in Forecariah) the linguistic text pro-
vides the reasons for that happiness.
Although the text and the photograph are drawing on textual, interper-
sonal, and ideational choices, the choices are made from different sets of
systems with different sets of options. Moreover, the selections function
differently in terms of structure: whereas language structures thought and
reality as ordered sequences of events, thought and reality in the photo-
graph are ordered in terms of a part-whole relationship, which makes it
possible to see the family in relation to each other and the immediate
context which is depicted.
Intersemiotic relations are also established within and across ranks (see
O’Toole 2011:11–31). For example, what happens at clause rank in the
transitivity also happens at the rank of Figure in the photograph: i.e.
Participants engage in Processes. Also, what happens at the rank of Episode
in the photograph appears to align with what happens in clause complexes
and in larger pieces of text: multiple participants engage in multiple pro-
cesses. There is, however, a key point of difference: in text, someone (or
something) can be a Participant in only one Process at a time, while in a
photograph the same person or thing can be a Participant in a number of
Processes simultaneously, such as the family members in the photograph,
who are portrayed as Behavers in the Behavioural Processes of smiling,
looking, and sitting.
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
448 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Figure 17.6 ‘Current Situation’ section of webpage with constituent parts outlined
which functions as a macroTheme for the section. The two action buttons
in the bottom right corner ‘Ebola data and statistics’ and ‘Ebola Situation
Report – 24 June 2015’ function as a kind of macroNew. The smaller
heading ‘Cases in the most affected countries’ functions as hyperTheme
for the graphs and written text combined. Each of the three graphs has its
own heading structure which identifies the country (i.e. Guinea, Liberia, and
Sierra Leone) and the sources of information for the graphs. Within each
graph, mouse-over pop-ups bring to the foreground information on the
number of cases at a particular time and make that information prominent
by having it stand out from the rest of the graph. The information in these
pop-ups is aggregated and used in the text underneath the graphs, where it
is typically made thematic as part of the Theme (see Figure 17.7). That is,
topical Themes in the text are mainly numbers of cases and countries, and
changes over time and more specific locations within Guinea and Sierra
Leone are found in the New. There is a strong textual link between infor-
mation which is made prominent in the graphs and the Themes in the
linguistic text underneath the graphs.
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 449
implicit addition connecting the three graphs into a semantic unit, as each
graph displays the same categories of information but for different coun-
tries, thus providing a visual comparison. This comparison can be made by
looking at the graphs in any order. Each graph shows the name of a
country, number of cases, and dates for that country. The relationship
between the graph complex and the subsequent text could be seen as
implicit similarity, since the graphs and the text present comparatively
similar, but not identical, information.
The x axis of each graph represents time (e.g. x1) and the y axis represents
number of cases (e.g. y1), forming a Relational process with a Token and
Value respectively (see Figure 17.7). The relationship between time and the
number of cases (e.g. (x1, y1) and (x2, y2)) is viewed as a pop-up which
becomes visible through mouse-over (see Figure 17.7). The series of rela-
tions is represented by points on the graph. Each relationship and its point
of intersection can be regarded as a visual Relational Identifying Process,
with the relationship between number and time as the Value/Identifier (x, y)
and the point of intersection as the Token/Identified (the point). However,
this visual relational process is implicit, and is based on knowledge of
mathematics and graph theory, in particular, line graphs. In this case, the
mathematical relation (x, y) is resemiotized as the visual Participant (•),
representing a situation where there is an intersemiotic downranking of a
mathematical identifying process (x corresponds to y) to a visual entity (the
point). Each graph, then, encodes a large number such metaphorical rela-
tions simultaneously (in this case, the number of cases and the number of
dates on which data was entered). While the whole graph encodes these
relationships simultaneously, the mouse-over function allows them to be
viewed one at a time (see Figure 17.7). From the series of identifying
relations which are resemiotized as points, a new visual entity in the form
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
450 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 451
each other by bold horizontal lines and by a straight blank space vertically
down the centre, the five parts of the infographic are unified by having
much the same general layout of a bold, large heading in the upper left,
smaller supporting text, and prominent, stylized images which use the
same colours and basic shapes. The position of the headings suggests a
reading/viewing path based on written English, where the headings serve
as Themes for their respective parts, and the accompanying text functions
as New to its respective Theme. The accompanying text ranges from a list of
symptoms in the top left part, two elliptical clauses in the top right part,
and text consisting of full clauses in the other three parts. Each image
resemiotizes some of the information in the written text, reinforcing and
highlighting its function as New information. The combination of linguistic
elements and images in the composition of the individual parts and in all of
the parts together creates a visual cohesion which assists the reader/viewer
in following the flow of information.
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
452 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Figure 17.10 Intersemiotic connections between the infographic and other parts of the
webpage
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 453
(see Figure 17.10, right). The video itself contains sequences which illustrate
how to wash hands and how to hand-rub, which are also mentioned in the
text of the infographic. In this instance, the same action is resemiotized in
different media.
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
454 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 455
objectively depicts the current funding scenario, reduces the need for
further appeals to the viewer, apart from the demands for goods and
services in the heading (‘Get Involved’) and the action button (‘View More’).
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
456 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 457
marked 3). The featured video ‘Hand hygiene in Ebola care facilities’ con-
nects to information in the infographic, while information in the info-
graphic connects to the photograph-text complex in the main visual
display (Figure 17.14, marked 4 and 5) as discussed in Section 17.3.4 above.
Lastly, the tweet displayed in the tweets-window shows part of the map re-
tweeted from the ‘Ebola data and statistics’ page of the Ebola website, and
functions to re-contextualize the place names in the graph and text in the
‘About Ebola’ section (Figure 17.14, marked 6). There are further connec-
tions to other sections of the Ebola webpage, largely through place names, as
a number of the image-text complexes in the main visual display are located
in places shown on the map and places mentioned in the text of this section.
There are, of course, other connections outside the ‘News’ section, but
this section seems to provide a focal point for intersemiotic connections on
the webpage. Its position as the final section of the webpage suggests that it
functions as macroNew for the webpage as a whole, although this function
is different from that of a macroNew in a written text. Martin and Rose
(2007) describe the function of macroNew as ‘distilling’ information. This
section does link back to other sections, but it also has the additional
function of providing a springboard to other and new (as in both most
recent and previously unseen) sources of information, in which case it
functions as macroTheme for the material to which it links.
The discussion above shows the significance of SFT for the analysis and
interpretation of complex multi-semiotic texts. As the above analysis has
shown, certain aspects of systemic theory can be generalized across semi-
otic resources, while others are more problematic, as summarized below.
Ideationally, configurations of Participants, Processes, and Circum-
stances realized through language can be realized (albeit, in a different
form and substance) in images, and multi-semiotic configurations. Like-
wise, identifying and attributive relationships realized in language can be
realized in images (again, in a different form and substance) and, even more
efficiently, in mathematical tools such as graphs.
Logically, texts constructed from different semiotic resources can be
connected both to each other and as parts of a larger text. For example,
conjunctive relations like those found in language appear to also apply to
multi-semiotic texts, especially when a whole text is made up of or includes
parts which are constructed from different semiotic resources. The ‘Cur-
rent Situation’ section is a good example of this. Conjunctive relations can
also apply across texts, as is the case in the photograph-text complex
discussed in Section 17.3.2 above, although the application of systems of
conjunction to photographs is more problematic. In language, systems of
conjunction link events logically as they unfold in text. While this staging
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
458 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
of events can also happen visually (e.g. scientific diagrams, visual instruc-
tions), an image (typically) functions by situating happenings, actions, and
relations as parts of a whole, although visual choices make particular
relations more salient or prominent than others.
Ideationally, the combination of different semiotic resources and the
connections between and among them contribute to structuring thought
and reality in new ways, as illustrated in the preceding discussion. For
example, the potential to visually show participants engaged in multiple
processes simultaneously can be advantageous. A graph (or graphs), for
instance, can be used to show patterns of multiple (theoretically infinite)
relationships, which can be compared according to different parameters (in
this case, location, time, and number of cases). Language can then be used
to single out and discuss examples of those relationships considered to be
most important. Accompanying text can isolate which of those configur-
ations is most relevant to the context, to construct further information
beyond what is portrayed visually.
Compositionally, the overall text organization is achieved visually
through spatial layout and framing devices, accompanied by linguistic
headings for each section. Waves or layers of information, realized through
the concepts of Theme and New, are encoded in language. In an image,
while the viewer perceives the whole image at once, there are compos-
itional, ideational, and interpersonal elements which combine to make
some figures or episodes prominent, and which are designed to guide the
viewer through the image to create particular meanings.
In each of the instances discussed above, texts constructed from different
semiotic resources work together to both expand and constrain the range of
possible meanings. As Lemke (2005) explains, multimodal semiosis multi-
plies the semantic possibilities, from which certain selections are made in
order to constrain the possible meanings made within any one instance.
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 459
References
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
460 K AY L . O ’ H A L LO R A N , S A B I N E T A N , A N D P ET E R W I G N E L L
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
SFL and Multimodal Discourse Analysis 461
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 22 Jul 2019 at 01:04:22, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.019
18
SFL and Critical Discourse
Analysis
Gerard O’Grady
18.1 Introduction
I’d like to express my gratitude to the editors, and especially Geoff Thompson, for their thoughts, queries, and corrections.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 463
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
464 GERARD O’GRADY
1
Bernstein’s claim that code correlates with social class has been considered by some to be controversial. See Jones
(2013) for a discussion of the pros and cons of Bernstein’s approach.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 465
In example (1), the lead sentence of a news article, the text producer has
selected passive voice and so for whatever reason obscured the identity of
the killers. The use of the verb killed carries less negative prosody than
synonymous verbs such as murdered or slaughtered would have. After all,
we speak of the killing of, though not the slaughtering or murdering of,
cells, bacteria, and trees. Finally, the information is presented as hearsay
and not as an evidential fact. The victims are presented as the Goal of the
killing without any Interpersonal evaluation. An alternative representation
is, Named participant murders eight innocent people in a shooting etc.
(2) Syriza’s victory in the Greek elections at the end of January gave
hope to some that the eurozone would change its economic policy.
But a month and many hours of painful diplomatic arm-wrestling
later, those hopes have clearly been dashed.
Even if Friday’s agreement between Greece and its lenders is
approved by eurozone ministers on Tuesday, it decides very little
apart from ensuring that the next four months will be a battle of
attrition. The eurozone will look to keep the new government in
Athens in check, and the Syriza-led coalition will try to eke out fiscal
space for some of the policies it has promised.
Politically, the four-month extension will be a challenge rather
than a blessing for the government. Prime minister Alexis Tsipras
will have to ditch plans to increase low-income pensions and other
similar measures that would affect the country’s fiscal balance.3
In example (2), which is the opening of an opinion piece, the text producer’s
linguistic selections represent the newly elected Greek government as
essentially powerless. Their election is represented as giving [false] hope.
The identity and number of recipients of the hope is left unspecified. The
details of Syriza’s election victory are backgrounded. An alternate repre-
sentation could have been, The greatest number of Greek adults voted for Syriza.
In the second clause, the text producer states that these hopes have clearly been
dashed. The text producer’s selection of the passive voice obscures the
identity of the unwritten Actor of the clause, and more importantly why
and how the Actor dashed those hopes. The use of the modal adjunct clearly,
which represents the author’s summation of the likelihood of the propos-
ition being true, adds to the representation of the Greek government as
powerless. In the second paragraph lexical parallelism of the processes will
look to keep . . . in check and will try to eke out further emphasizes the
2
www.theguardian.com/world/2015/feb/24/czech-republic-restaurant-shooting-multiple-deaths-reported. (Last
accessed 24/02/2015.)
3
www.theguardian.com/commentisfree/2015/feb/24/greece-syriza-victory-euphoria-gone-reforms. (Last accessed
28/02/2015.)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
466 GERARD O’GRADY
(3) Miss Royal showed ‘that she has absolutely no concrete solutions to
respond to the problems of the French people’.5
4
An objection raised against the efficacy of CDA is that it is incapable of describing what was not said or written. But in
this case the text producer’s focus on the negative impact on the country’s fiscal balance rather than on the economic
and physical health of its citizens is in and of itself highly revealing of the author’s ideology.
5
www.telegraph.co.uk/news/worldnews/1543398/Royals-spirits-raised-by-record-for-TV-debate.html. (Last accessed
24/02/2015.)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 467
it able to account for the effect of language upon society. Consequently, the
critical linguistics framework was merged with social and critical theory
and re-emerged as CDA. Slembrouck (2001:35)6 notes the influence on the
formation of CDA of the work of Stuart Hall and his colleagues at the
Birmingham Centre for Contemporary Cultural Studies, who engaged in
influential critical analyses of the social, economic, and political changes
caused by the emergence of the Thatcherite consensus. Hall and his col-
leagues gave a central place to the close study of symbolic practices as a
means for understanding how social relationships were transformed in
everyday social practice. This, as Slembrouck notes, not only accorded well
with the close linguistic analyses practised by the critical linguists, but also
allowed CDA theorists to develop a theoretical approach based on social
theory aimed at explaining the relationship between linguistic practice and
changes in the world.
The publication of Fairclough’s (1989) seminal book Language and Power is
considered to mark the birth of the CDA programme (Blommaert 2005:23).
The following paragraphs will sketch the growth of CDA as an academic
discipline, though I will naturally slant the discussion towards CDA work
which has been more overtly influenced by SFL and illustrate how SFL has
been used to advance the CDA programme. I will postpone discussion of the
various criticisms which have been levelled against SFL-flavoured CDA until
the next section.
Fairclough (1989, 2015) identified three dimensions of discourse. The
first is discourse as text or product. This dimension refers to the formal
linguistic features of the text such as wordings, transitivity choices, modal-
ity, cohesion, and text structure. Readers will have noted that it was this
first dimension of discourse which the critical linguists focused on. The
second dimension is discourse as discursive practice. This dimension refers to
the interactive nature of discourse and describes how discourse is pro-
duced, circulated, and consumed. Analysts examining discourse in this
dimension examine the aspects of the text that link it to its wider social
context, such as speech acts, intertextuality, and coherence. The final
dimension is discourse as social practice. This dimension refers to the ideo-
logical effects and hegemonic practices pre-existing in the context in which
the discourse is produced. For instance, the prevailing ideology may nor-
malize a text as common sense or label it as outside the norm.
Fairclough (1989:26, 2015:58–9) proposed a three-stage methodology to
enable CDA to account for the three dimensions of discourse. Firstly,
analysts must describe discourse in terms of its formal properties. While
engaged in the description of discourse, analysts must not only adopt the
participants’ perspective but also attempt to make their description expli-
cit. Secondly, they must interpret the discourse as interaction in order to
6
It should be noted that Slembrouck’s account of the birth of CDA is slanted heavily towards Fairclough’s approach.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
468 GERARD O’GRADY
7
The potential for feedback is clearly related to the power relations existing between the participants and thus their roles
in the discourse.
8
It is not entirely clear to me whether in an increasingly digital world, where large amounts of the day are spent on social
media, the distinction between written and spoken forms of the language is an entirely accurate way of capturing the
distinction between language intended to have a less permanent inscription from one which was intended to have a
more permanent inscription.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 469
(4) Syriza’s victory in the Greek elections at the end of January gave
hope to some that the eurozone would change its economic policy.
But a month and many hours of painful diplomatic arm-wrestling
later, those hopes have clearly been dashed.
Even if Friday’s agreement between Greece and its lenders is
approved by eurozone ministers on Tuesday, it decides very little
apart from ensuring that the next four months will be a battle of
attrition. The eurozone will look to keep the new government in
Athens in check, and the Syriza-led coalition will try to eke out fiscal
space for some of the policies it has promised.
Politically, the four-month extension will be a challenge rather
than a blessing for the government. Prime minister Alexis Tsipras
will have to ditch plans to increase low-income pensions and other
similar measures that would affect the country’s fiscal balance.
Some on the left wing of the party clearly find this difficult to
swallow. Veteran leftist MEP Manolis Glezos has already likened
the party leadership’s achievements as ‘renaming meat fish’ and
apologised for contributing to the ‘illusion’ that Syriza would change
anything. His comments were politely dismissed by key figures in
the party, and it is clear that there is a damage-limitation exercise in
operation in the wake of the agreement.
The first question probes the Field, the following two the Tenor, and the
final one the Mode.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
470 GERARD O’GRADY
conforms to the website’s style sheet. It complies with the expected word
count. The writer, Nick Malkoutzis, is presented as the possessor of specific
cultural knowledge, in this case an account of whether the newly elected
Greek government has honoured its electoral pledges and whether its
actions will benefit or harm the Greek people. The semantic domain is
specialized, and the author presents information from the fields of politics
and economics.
Summary
Mr Malkoutzis, while notionally presenting his views as opinion, is in fact
construing a world where his view is the authoritative truth. He is the
expert, and the views he represents are unarguable.
Texts and discourses are not created in isolation. They exist in a dialogic
chain with both previously produced texts and potentially produced future
texts (Voloshinov 1973). In order to interpret the meaning of the text, one
must decide to which discursive chain a text belongs, and therefore what is
presupposed as common ground between the writer and the reader (Fair-
clough 1989:152, 2015:164). Assertions in the example text rely on presup-
positions such as increasing public spending is bad and it is sensible to balance
national budgets. While these project an assumption that readers already
know this, it is impossible for a writer in the mass media to know what his/
her individual reader’s intertextual experiences are. Thus, what is pre-
sented are the presumed intertextual experiences of an ideal reader. This
Fairclough (1989:152, 2015:164) reminds us is a powerful weapon in the
9
The reader is not informed that Kathimerini is a highly partisan Conservative newspaper which keenly supports the
former right-wing governing party New Democracy.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 471
armoury of the mass media. They are free to present contestable or dis-
puted propositions as uncontested or leave them unstated as presupposed
backgrounded common sense which underlie their assertions.
Presuppositions do not of course exist in the text but rather are created by
the readers’ interactions with the intertextual context. As such, prime facie,
SFL might not appear the best tool to unpack pragmatic meaning. However,
Martin and White’s (2005: Chapter 3) Appraisal framework provides a valu-
able social-dialogic framework which can be used to explicate whether or not
alternative positions are acknowledged in texts. As noted in the discussion of
Tenor, Mr Malkoutzis’ discourse is predominantly monoglossic. Compare the
opening clause of the third paragraph Politically, the four-month extension will be
a challenge rather than a blessing for the government with possible dialogic
alternatives Commentators expect/predict/believe/speculate that the four-month exten-
sion will be a challenge rather than a blessing for the government.
However, Mr Malkoutzis intervenes in the text on three occasions, dem-
onstrating his stance through the use of the modal clearly on two occasions
and the modalized metaphor it is clear on one occasion. Halliday and
Matthiessen (2014:190) label clearly as a comment adjunct which occurs in
declarative mood. In their network it is an asseverative subclass of obvious
and functions to assert that it is so. Martin and White’s (2005) framework
labels Mr Malkoutzis’ modality choices as proclaim: pronounce which is a
contracting strategy designed to project a claim as self-evident.
This can be seen particularly distinctly in the following heteroglossic prop-
osition: Some on the left wing of the party clearly find this difficult to swallow. This
example, while notionally up for discussion, is immediately followed by an
illustrative example of a left-wing member who is quoted as apologizing for his
part in the creation of the false hope. Veteran leftist MEP Manolis Glezos10 has
already likened the party leadership’s achievements as ‘renaming meat fish’ and apolo-
gised for contributing to the ‘illusion’ that Syriza would change anything. As such the
text construes a reader who may not necessarily share Mr Malkoutzis’ evalu-
ation of Syriza – one who is perhaps undecided and in need of further evidence.
The effect of the further evidence is to support the contraction of the dialogue
and present Mr Malkoutzis’ view as incontestable common sense.
One further stage remains in Fairclough’s methodology, namely, explan-
ation. The objective of this stage is to portray discourse, itself a social
practice, as part of a social process, illustrating how it is determined by
social structures and explicating how the reproduction of discourses
reinforces or weakens these structures (Fairclough 1989:163, 2015:172).
This stage naturally draws much less from linguistic theory and much
more from social theory and other relevant disciplines – see Fairclough’s
(2009:163) point that an effective CDA must be transdisciplinary. Dis-
courses are examined as part of a social struggle and/or contextualized in
10
Manolis Glezos is a noted World War II resistance fighter famed for his part in removing the swastika from the Acropolis
during the German occupation of Greece.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
472 GERARD O’GRADY
11
The observant reader will have noted that all of these scholars’ thinking has to a greater or lesser extent been
influenced by Marxist thought and a view that knowledge and power are formed discursively. I will argue in the
following sections that it is precisely this Marxist orientation that makes SFL congruent with these various approaches.
Fairclough and Graham (2010:340) classify Marx as a discourse theorist based on their reading of his body of work as
one which contains a discursive view of language as an element of social life.
12
A very similar argument could be couched in terms of symbolic capital and its inscribed effect on the habitus
(Bourdieu 1991). But for present purposes the details of the argument are not of relevance. What is of relevance is to
illustrate how critical theory can be employed alongside detailed linguistic description in analyzing a text and in
showing how it fits into a chain of discourse in order to trace its effect on social practice.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 473
To conclude this section, we have seen that explanation rests upon solid
linguistic description and analysis, and that for CDA to be truly effective it
must relate the semiotic to social practices and social practices to societal
structures. In order to do this, an analyst must engage in transdisciplinary
work, fusing linguistics with other relevant approaches. In short, CDA must
not be exclusively a semiotic approach.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
474 GERARD O’GRADY
Analysis does not, in other words, make predictions about how individuals
will respond to a text. The explanation stage goes further and shows how
different positions emerge historically.
O’Halloran (2003), using insights from Relevance Theory (Sperber and
Wilson 1995), proposes an IR (idealized reader) analysis of news text. This is
potentially a fruitful first step in examining how individual readers interact
with a text by examining differences in terms of reading effort, cognitive
bias, and previous experience with the topic. But, the key point for SFL-
flavoured CDA is that such an interpretation should be grounded in rigor-
ous linguistic description.
The SFL view of context has been robustly criticized by van Dijk (2008).
He argues that it is too linguistic, not cognitive enough, and based on a
limited social theory of language. In the paragraphs that follow, I will
address these issues in turn. As SFL grew out of a sentence grammar
approach, namely, scale and category grammar, it is unsurprising that it
heavily focused on the clause as the unit of analysis. And, indeed, if one
wishes to incorporate a rigorous descriptive linguistic analysis into CDA, a
clause-based grammar, especially for pre-planned formal texts, is a power-
ful device. This of course does not entail that CDA grammatical description
must necessarily restrict itself to the level of the clause.
Systemic Functional Linguistics as a social semiotic theory does not focus
on the individual speaker. Halliday, criticizing the individualist philosophy
underpinning much of linguistics, commented: ‘Creating language and
creating through language, are essentially interactive processes; they can
never take place inside one individual’s skin’ (Halliday 2007:56). This does
not presume that CDA or indeed other forms of language study could not
usefully be informed by insights from the psychological literature. Systemic
Functional Linguistics itself has usefully incorporated Vygotsky into studies
of child language development, e.g. Halliday (1975) and Hasan (2005). In
short, van Dijk seems not to recognize that the issue is not whether SFL
needs to develop a cognitive dimension, but rather whether it, as a semiotic
theory, is compatible with cognitive approaches.
Van Dijk (2008:38) states that, despite what he claims to be the anti-
mentalist view of context inherited by SFL from Firth and Malinowski,
Halliday’s description implies cognitive notions (Halliday 1978). It is indeed
true that Halliday’s view implies cognition, but van Dijk is mistaken in
labelling SFL as anti-mentalist. Thibault (2011) is an SFL description of
semiosis as a dynamic biocultural process distributed across brains, bodies,
and aspects of the social and cultural world. He shows that while semiotic
processes are grounded in the ‘signifying body’, semiosis cannot be reduced
to bodily processes (Thibault 2011:52). A social semiotic approach is entirely
compatible with the theory of distributed cognition (Hutchins 1995).13
13
Fawcett (1980) and the Cardiff Grammar in general is a far more individual speaker/hearer centred version of SFG.
But as I know of no CDA work grounded in the Cardiff Grammar I shall not mention it further.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 475
The social structure and the State are continually evolving out of the life-
process of definite individuals, but of individuals, not as they may appear in
their own or other people’s imagination, but as they really are; i.e. as they
operate, produce materially, and hence as they work under definite material
limits, presuppositions and conditions independent of their will.
(Marx 1932)
14
Bartlett’s approach is prima facie not dissimilar from the Discourse Historic Approach advocated by Ruth Wodak. This is
a problem-oriented approach that examines changes in discursive practice over time. Wodak draws on a range of
resources, such as text linguistics, argumentation theory, and the Frankfurt School of Critical Theory in analyzing
discourse (Reisigl and Wodak 2009).
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
476 GERARD O’GRADY
15
See, for instance, the dynamic CDA work being produced under the cognitive linguistic influenced CADAAD network
http://cadaad.net/.
16
www.gov.uk/government/speeches/prime-minister-a-speech-on-the-economy. (Last accessed 06/03/2015.)
17
For reasons that are not clear, Hart has altered the order of the speech. In Cameron’s version, the words despite
headwinds . . . were produced after the other extracts.
18
In a world where anthropogenic climate change is changing weather, Cameron could perhaps have opted for an
alternate construal!
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 477
19
It remains to be seen how applicable conceptual metaphor theory is outside first world and perhaps Indo-European
contexts. See below for Blommaert’s critique of CDA as closed to non-Western societies.
20
This is a mechanism that shows how emotion guides rational thought. In other words, rationality cannot be
disentangled from emotion.
21
It goes without saying that no form of CDA can analyze absent discourses, by which I mean ones that are not available
for recording and hence for analysis.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
478 GERARD O’GRADY
Critical Discourse Analysis aims not only to explain the world but also to
effect a positive change, no matter how limited, in it. An SFL study which
does precisely that is Martin et al.’s (2013) innovative multimodal investi-
gation of identity construction within the genre of youth justice counsel-
ling, a form of restorative justice. Martin et al. note that the genre idealizes
young offenders who construe themselves as both remorseful and rational
about their futures. Such a construal creates the space that permits the
genre to transition from redemption to reintegration. Through detailed
analysis of the meanings created by the coupling of verbal interaction and
gesture, Martin et al. show how bonds are created and maintained over the
interaction. While Martin et al.’s study is not couched in terms of CDA, by
mapping the multimodal generic potential of the youth justice counselling
sessions, they are in a position to advise, i.e. directly influence, the social
practice of participants involved in managing the youth counselling system.
Part of the innovation in the work of Martin and his colleagues is that they
focus not only on what is wrong in discourse but what is right. Their point
is that changes in social practice may come from the spreading of what is
positive in a discourse rather than exclusively from trying to suppress what
is wrong (see Martin and Rose 2003).
In today’s world, ideology is increasingly disseminated by a combination
of hybrid semiotic modes in fields such as advertising and new media
platforms. While there has been a technological revolution since the publi-
cation of Fairclough (1989) which has profoundly changed the way humans
interact, learn, and consume information, CDA has from its inception
recognized that discourse is not formed exclusively from verbal language
(Fairclough 1989:27).
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 479
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
480 GERARD O’GRADY
References
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 481
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
482 GERARD O’GRADY
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
SFL and Critical Discourse Analysis 483
Lakoff, G. and M. Johnson. 1999. Philosophy in the Flesh. New York: Basic Books.
Leckie-Tarry, H. 1995. Language and Context: A Functional Theory of Register.
Edited by D. Birch. London: Pinter.
Machin, D. 2013. What Is Multimodal Critical Discourse Studies? Critical
Discourse Studies 10(4): 347–55.
Mahboob, A. and N. Knight. 2010. Appliable Linguistics. London: Continuum.
Martin, J. R. and D. Rose. 2003. Working with Discourse: Meaning beyond the
Clause. London: Continuum.
Martin, J. R. and P. R. R. White. 2005. The Language of Evaluation: Appraisal of
English. Basingstoke: Palgrave.
Martin, J. R., M. Zappavigna, P. Dwyer, and C. Cléirigh. 2013. Users in Uses
of Language: Embodied Identity in Youth Justice Conferencing. Text and
Talk 33(4–5): 467–94.
Marx, K. 1932. A Critique of the German Ideology. Available online at: www
.marxists.org/archive/marx/works/download/Marx_The_German_Ideology.
pdf. (Last accessed 05/03/2015.)
Maton, K. 2015. Knowledge and Knowers: Towards a Realist Sociology of Educa-
tion. Abingdon: Routledge.
Miller, D. R. and J. H. Johnson. 2013. Register-idiosyncratic Evaluative
Choices in Congressional Debate: A Corpus-assisted Comparative Study.
In L. Fontaine, T. Bartlett, and G. O’Grady, eds., Systemic Functional Linguis-
tics: Exploring Choice. Cambridge: Cambridge University Press. 432–52.
O’Grady, G. 2011. The Unfolded Imagining of Segolene Royal. Journal of
Pragmatics 43: 2489–500.
O’Grady, G., T. Bartlett, and L. Fontaine, eds. Choice in Language: Applications
in Text Analysis. Sheffield: Equinox.
O’Halloran, K. 2003. Critical Discourse Analysis and Language Cognition. Edin-
burgh: Edinburgh University Press.
Reisigl, M. and R. Wodak. 2009. The Discursive Historical Approach. In
R. Wodak and M. Meyer, eds., Methods of Critical Discourse Analysis. 2nd
ed. London: SAGE. 87–121.
Rosch, E. 1975. Cognitive Reference Points. Cognitive Psychology 7: 532–47.
Slembrouk, S. 2001. Explanation, Interpretation and Critique in the Analy-
sis of Discourse. Critique of Anthropolology 21: 33–57.
Sperber, D. and D. Wilson. 1995. Relevance: Communication and Cognition.
2nd ed. Oxford: Blackwell.
Taylor, C. 2017. Reading Images (Including Moving Ones). In T. Bartlett and
G. O’Grady, eds., The Routledge Handbook of Systemic Functional Linguistics.
Abingdon: Routledge. 575–90.
Thibault, P. 2011. First-Order Languaging Dynamics and Second-Order
Language: The Distributed Language View. Ecological Psychology 23: 1–36.
Thompson, G. and S. Hunston. 2006. System and Corpus: Exploring Connections.
Sheffield: Equinox.
van Dijk, T. 2008. Discourse and Social Context. Cambridge: Cambridge
University Press.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
484 GERARD O’GRADY
Varoufakis, Y. 2013. The Global Minotaur: America, Europe and the Future of the
Global Economy. 2nd ed. London: Zed.
Voloshinov, V. N. 1973. Language Speech and Utterance. In V. N. Voloshi-
nov, ed., Marxism and the Philosophy of Language. Cambridge, MA: Harvard
University Press. 65–82.
Webster, J. J. 2007. M. A. K. Halliday: The Early Years 1925–1970. In
R. Hasan, C. M. I. M. Matthiessen, and J. J. Webster, eds., Continuing
Discourse on Language: A Functional Perspective, Volume 1. Sheffield: Equinox.
3–14.
Widdowson, H. 1995. Discourse Analysis: A Critical View. Language and
Literature 4: 157–72.
Widdowson, H. 1998. The Theory and Practice. Applied Linguistics 19:
136–51.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:01:33, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.020
19
Language Development
Geoff Williams
19.1 Introduction
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
488 GEOFF WILLIAMS
how a child is learning all the time through language: how the
microsemiotic exchanges of family and peer group life contain within them-
selves indices of the most pervasive semiotic patterns of the culture.
(Halliday 2004:111, emphasis in original)
While the claim has been very productively explored through case studies,
findings are structurally limited by the fact that they are about individual
children interacting with caregivers in just one socio-semiotic position.
A different perspective on Halliday’s general claim can be achieved by
drawing on samples of interaction between young children, now able to
speak their mother-tongue, and their caregivers in contrasted socio-
semiotic locations. This is a strategy used by Hasan (2009), and subse-
quently by two of her graduate students, Cloran (1994) and Williams
(1995). Neither one perspective on language development is intrinsically
better than the other: they are different ways of investigating the same
phenomenon, provided that phenomenon is construed at a high enough
level of abstraction, i.e. child language development during 0–5 years. Since
there are good, recent introductory overviews of Halliday’s early child
language development theory and research, together with the further
research it has produced (e.g. Painter et al. 2007; Torr 2015), it is possible
to introduce both research approaches here.
There are many advantages to discussing the two perspectives in one
paper, though this does not appear to have been done previously.1 Most
importantly, it is possible to see how some of Halliday’s innovative meth-
odology in Learning How to Mean has been either extended, replicated, or
replaced in subsequent work. Halliday has characterized his research as a
‘diary-based case study’. While this is of course accurate, his methodology
does actually involve much more than is usually understood by that term
because of the new techniques he introduced to ‘map’ meaning-making
development. There is a literal as well as a metaphorical meaning to ‘map’
in play here, which has proved highly significant for SFL explorations of
discourse.
A further important methodological and theoretical feature of Halliday’s
work in his child language development research that has been taken up in
subsequent work is the dialogue he established between linguistics and
other fields, particularly sociology. Even at this early stage of the develop-
ment of SFL theory and research, Halliday adopted a transdisciplinary
approach, not merely by supplementing linguistic research with other
compatible perspectives, but rather by enabling deeper theoretical linguis-
tic questions to be raised through interaction with other disciplines, and
also by developing a sympathetic critique of theory in other disciplines in
which language use plays some role. His dialogue with Bernstein’s theory of
1
One drawback to establishing this scope is that closely related work on infant language development by Painter (1984)
and Torr (1997), though it is significant both as independent support for Halliday’s findings and as sources of new
insights, can only be referred to briefly.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 489
19.2 Protolanguage
2
There is a parallel limitation to the above. Only brief reference can be made to Cloran’s and Williams’ research, though it
supports and complements Hasan’s work in significant ways. Cloran’s work on gender adds the analysis of effects
of this variable operating within the contrasted family social positions.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
490 GEOFF WILLIAMS
Halliday’s data, the second diary entry illustrates the point succinctly,
perhaps also rather surprisingly.
When he was two months old, Nigel used to greet me when I came home
from work. He would give a long gurgling account of the day’s events, always
in a cheerful tone. Then I told him how I had got on at the college, and
whether the train had been crowded at Charing Cross.
One day he told me a very sad tale. His face was frowning, his tone was
mournful, and he was barely holding back his tears.
What happened today, I asked his mother.
‘He had his first injections’, she replied.
(Halliday 1984:1)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 491
Somehow, the child moves from the one to the other, from his own system to
that of the adult; and our hypothesis must be such as at least to show that it
would have been possible for him to make the transition.
(Halliday 2004:70)
Instrumental ‘I want’
Regulatory ‘do as I tell you’
Interactional ‘me and you’
Personal ‘here I come’
Heuristic ‘tell me why’
Imaginative ‘let’s pretend’
Informative ‘I’ve got something to tell you’
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
492 GEOFF WILLIAMS
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 493
there is only content with respect; that is, with respect to the functions that
language serves in the life of the developing child . . . the content of an
utterance is the meaning that it has with respect to a given function, to
one or other of the things that the child is making language do for him. It is a
semiotic act which is interpretable by reference to the total range of semiotic
options, the total meaning potential that the child has accessible to him at
any moment.
(Halliday 2004:67–8, emphasis in original)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
494 GEOFF WILLIAMS
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 495
Nigel expressed the systematic distinction between the two major modes by
means of the phonological opposition between falling and rising tone: prag-
matic (‘response required’) as rising tone, mathetic (‘response not required’)
as falling tone. The falling tone was a direct continuation of the dominant
intonation pattern of Phase 1, where all tones were falling except in one
interactional system, that of individualized greetings.
(Halliday 1975:87)
Halliday notes that it was the mathetic function that provided entry to
classes of objects, properties of objects, then circumstantial elements (e.g.
toothpaste òn . . . toothbrush; bumblebee on tràin),3 and the pragmatic was the
source of the key interpersonal resource of the mood system. However,
there was no simple, neat distribution of developmental responsibility such
that one macrofunction maps directly onto one metafunction – the process
was more intricate and complex. Halliday provides, for example, interest-
ing observations and suggestions about how the elements noted above tend
to appear first in the mathetic function, but verbs (processes) tend to appear
first in the pragmatic function, which is also possibly the source of agentive
constructions in ergative processes (Halliday 1975:106–8).
In overview, Halliday’s case study demonstrates the value of a sustained
focus and systematic mapping of an individual’s development. Through
such a detailed approach we can see, almost moment by moment, how
Nigel begins to develop a rudimentary understanding of the culture in
which he is located. For example, features of the material environment –
from bananas to pantographs, family roles and food categories – all become
part of everyday experience. In a paper published contemporaneously with
Learning How to Mean, Halliday remarked:
The learning of language and the learning of culture are obviously two
different things. At the same time, they are closely interdependent. This is
true not only in the sense that a child constructs a reality for himself largely
through language, but also in the more fundamental sense that language is
itself a part of this reality. The linguistic system is part of the social system.
Neither can be learnt without the other.
(Halliday 2004:281)
3
The downward sloping symbol on the words òn and tràin indicates a falling tone.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
496 GEOFF WILLIAMS
social system is realized through (inter alia) the linguistic system’ (Halliday
2004:281–2, emphases in original). Following Hasan, we can interpret this
to mean that co-occurring features of the social system typically activate
features of the linguistic system, and co-occurring features of the linguistic
system construe features of the social system (Hasan 2010).
However, there is a crucial caveat. The social system of a given culture is,
clearly, not a seamless unity. It is fractioned by class, gender, age, and so on.
So, closely following Halliday’s line of argument, is it possible that children
developing language in contrasted social locations within these intersecting
fractions within the same culture might learn to mean in somewhat different
ways? Or do they learn broadly the same ways of meaning across a culture,
with merely non-systematic variation as a result of idiosyncratic features of
their social and personal environments?
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 497
to which I refer now. She therefore literally lived intricate cultural and
intercultural differences in ways of saying and meaning. Second, while she
was in Britain, she worked in Bernstein’s Sociological Research Unit at the
University of London around the time he was redeveloping his theory of
social transmission and reproduction, including especially his ideas about
how language might function in these processes. The redevelopment
involved maintaining his earlier emphasis on the significance of language
use, but developing a semantic rather than a structural account of its
function (Williams 2005). In fact, as Bernstein warmly acknowledged,
Hasan herself played a key role in that redevelopment soon after her arrival
(see, for example, Bernstein 1973:5, 7). The title paper in her Ways of Saying,
Ways of Meaning, which was first published in 1984, demonstrates her
meticulous use of these two features of her experience (Hasan 2015b).
As the second feature implies, Hasan also adopted a transdisciplinary
approach to her language development and semantic variation research. In
addition to sociology, Hasan looked for a theory of cognitive development
that could be brought into productive dialogue with both linguistic and
sociological theory. This she found in the work of Vygotsky and his student
and collaborator, Luria (e.g. Vygotsky 1978, 1986), though her engagement
with this work was decidedly not uncritical, in large part because she
considered Vygotsky’s proposals in need of extension through linguistic
and sociological theory (Hasan 2005, especially Chapter 3). Nevertheless,
her account was deeply appreciative. In reflecting on her research approach
in an interview with David Butt and Jennifer Yameng Liang towards the end
of her life, she commented:
[I]t seems to me, that to study Basil [Bernstein GW] without Vygotsky, or
Vygotsky without Basil, is to read only half the story: the link between the
social and the psychological in ways of saying and ways of meaning is
provided by these two scholars.
(Butt and Liang 2016:394)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
498 GEOFF WILLIAMS
The simpler the social division of labour, and the more specific and local the
relation between an agent and its material base, the more direct the relation
between meanings and a specific material base, and the greater the probabil-
ity of a restricted coding orientation. The more complex the social division of
labour, the less specific and local the relation between an agent and its
material base, the more indirect the relation between meanings and a
specific material base, and the greater the probability of an elaborated
coding orientation.
(Bernstein 1990:20)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 499
Hasan took this complex theoretical proposal and created from it a par-
ticipant selection principle that was specific enough to enable empirical
linguistic research. The alternative, and admittedly much easier, strategy
would have been to opt for ‘socioeconomic status’ as the selection
principle, but this commonly used measure, however defined, would have
prevented her from engaging in a transdisciplinary theoretical as well as
empirical dialogue with Bernstein’s sociology (Hasan 2005: Chapter 2).
The linguistic significance of her findings would also have been markedly
diminished.
The participant selection principle was the degree to which a person’s
position in the social division of labour would typically allow that person to
make decisions that would be brought into effect by others. A higher
autonomy professional (HAP) would thus have considerably more discre-
tion in making workplace decisions than a lower autonomy professional
(LAP), though neither would be either fully autonomous or completely
lacking in autonomy. To exemplify from Williams’ (1995) research, some
occupations of the main breadwinners in the HAP families were engineer,
financial consultant, and barrister, while some in the LAP families were
paint batcher, soldier, and loader driver. It is not the occupations them-
selves that are important – they do not necessarily directly indicate a
family’s social positioning – but they are theoretically well-grounded view-
points from which to explore language in use. Additional to ascertaining
occupational data, it proved necessary to check whether a person had
voluntarily taken a lower autonomy position, such as a person wanting to
have a particular work schedule for domestic reasons or to allow further
study. For further discussion of the selection principle operating in prac-
tice, see Hasan (2009:90–1) and Williams (1995:94–9). Achieving well-
contrasted samples of participants using this principle, while maintaining
respectful recruiting processes, has proved quite feasible, though complex
to manage.
Hasan’s research focus was on features of talk between children and
their mothers. Mothers were selected because they were almost exclu-
sively at this age the primary carers of young children. Audio record-
ings of naturally occurring, everyday interaction between the mothers
and their children, aged approximately 3–5 years were made by the
mothers themselves using small, powerful audio recorders. Mothers
were asked to turn on the recorder while they went about everyday
activities with the focal child that involved talking with him or her, and
only to turn it off when the child had switched to another activity. They
were invited to erase any recording that they did not wish to be heard
outside the family. This approach yielded natural unselfconscious data,
which was crucial to the purpose of the project. It would have been
difficult for the mothers to be self-conscious about their worded mean-
ings while talking with their children and simultaneously completing
domestic tasks. Additionally, Hasan comments that ‘their little children
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
500 GEOFF WILLIAMS
insisted on the same maternal “face” with which they were familiar’
during the recorded interaction (Hasan 2009:92). Recordings were sub-
sequently transcribed and then explored through detailed linguistic
analysis.
The analysis employed a message semantic approach, developed by
Hasan (2009; see also Low and Fung, this volume). The approach
involved segmenting the interactive discourse into individual messages
and then describing the selection of features for each message using
semantic networks. Messages are defined as the smallest unit at the
semantic stratum, typically realized by a clause at the
lexicogrammatical stratum, with one major exception, one which
proved to be key to the outcomes of the research. To exemplify, Text
19.1a presents a sample of mother–child interaction while reading a
book about a family visit to the beach, and Text 19.1b presents the talk
in this stretch analyzed into messages. Message boundaries are indicated
by bracketed numerals.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 501
The basis for this approach lies in the SFL concept of logical metafunction.
I don’t know projects the second clause and is thus considered a
metarepresentation rather than a direct representation. From a semantic
perspective, then, projecting clauses are considered to be part of a message
that also includes the projected clause. It is possible that more than one
projecting clause, and therefore more than one metarepresentation, might
contribute to a message, as in I remembered she had said she was thinking that
she’d go. However, treating these instances as one message does not mean
that the semantic contribution of projecting clauses is ignored: in fact, it
emerges that this contribution is one of a set of features that play a key role
in semantic variation, as will emerge shortly.
There is one further basic descriptive move for messages, which is import-
ant because it makes a substantial difference to interpretation. Hasan dis-
tinguished between ‘progressive’ and ‘punctuative’ messages because it is
only the first type that open the possibility of a full range of metafunctional
meaning. In contrast, punctuative messages are somewhat akin to gram-
matical continuatives. To exemplify, Text 19.2a introduces the concluding
stretch from one of Michael’s joint reading times with his mother, and Text
19.2b presents an analysis of this short stretch into messages.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
502 GEOFF WILLIAMS
Messages (1) and (3) to (11) are [progressive], but (2) is [punctuative].
Punctuative messages are typically formulaic greetings, hesitations,
addresses, and reactive expressions. Viewed at the lexicogrammatical
stratum, they do not select for Predicator. It is perhaps useful to note that
message (11), though realized by a single word and so perhaps appearing to
be [punctuative], is treated as an elliptical form of They are called anemones
and is therefore analyzed as [progressive].
How then did Hasan analyze her data from a semantic perspective? There
was a rich array of approaches to semantic analysis from which to choose,
but for Hasan’s purposes none provided a sufficiently detailed analysis of
interpersonal, logical, textual, and experiential meanings. What she did
instead was to take up an analytic initiative first introduced by Halliday and
to reconfigure and expand it so it could be used to explore the development
of children’s ways of meaning through language.
Halliday’s initiative was the introduction of semantic networks to map
meaning-making in a specific context type (Halliday 1973). At this point we
can see that the two apparently discrete methodologies in SFL child lan-
guage development research are actually developments of the same basic
techniques. Semantic networks are an extension of Halliday’s approach to
mapping the development of protolanguage, discussed earlier in this paper,
albeit with some major qualitative differences. As in the protolanguage
research, the orientation is to describe what a speaker ‘can mean’ based
on observations of what they ‘do mean’. But a semantic network is also
categorically different from the small, functionally derived protolanguage
networks because it describes language itself, which implicates multifunc-
tionality, i.e. multiple, simultaneous features of meaning-making, together
with multi-stratal interrelationships, because semantic features are related
through realization to lexicogrammar and are themselves realizations of
features of context. The complexity of semantic networks is therefore much
greater.
Halliday’s context-specific semantic network mapped meaning resources
habitually used in maternal control of young children. The relationship
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 503
between ‘can do’ and ‘do do’, which was to prove so crucial to subsequent
work in semantic variation, was succinctly summarized by Halliday in the
paper in which semantic networks were first introduced:
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
504 GEOFF WILLIAMS
reassure 1
b
verify
1 probe 2
a
ask 1
G c
enquire check 2
2
demand d
d assumptive 1 e
d
nonassumptive } simple 1
info
2
alternative 2
explain prompted 1
b c
1
precise
1 specify unprompted 2
a
apprize 2
circumstance 1
H f
tentative event 2
2 g
actant
3
Figure 19.2 Hasan’s semantic network fragment: making choices in demands for
information (reproduced in Williams 1995:158)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 505
[confirm] major:indicative
[apprize] major:indicative:interrogative:nonpolar
For example, in Text 19.1b message (3), Doesn’t it look great? selects [confirm]
at the primary level, and in Text 19.2b [apprize] is selected by messages (1)
What are those? and (10) Well why do the enemies close up?
Moving toward the right of the networks, each more specific (or delicate)
feature is realized through a set of more delicate lexicogrammatical fea-
tures. For example, in Text 19.1b, message (3) Doesn’t it look great? selects
[confirm], and further selects the features [confirm:enquire:ask:assumptive:
simple]. In Text 19.2b message (1) What are those? selects [apprize:precise:
specify:unprompted:actant:nonspecific].
To emphasize, Figure 19.2 presents just one network fragment. To gain
some sense of the scope and detail of the analysis, it is necessary to envisage
similar network fragments across all four metafunctions, complemented by
interrelations between them. Each message is therefore analyzed on
approximately seventy variables.4
More typically in child language development research, analysis of
discourse data is carried out through a content analysis technique, often
complemented by digital resources such as NVIVO. However, message
semantic analysis has the obvious advantages that, comparatively, it pro-
vides a much more detailed exploration of the interactive language based
on explicit criteria for the recognition of semantic features, which are the
realization statements exemplified above. This is particularly important for
exploring semantic variation since any variation is most likely to be found
in configurations of the more specific semantic features: i.e. not just use of
‘yes/no’ or ‘wh/’ questions, for example, but variation in specific types of
question, the logico-semantic relations between questions and other mes-
sages, patterns of response and responses to responses to different types of
questions, and so on. The equally obvious disadvantage is the time involved
in conducting the analyses, but that is a common feature of any scientific
approach to language.
This brief methodological discussion makes it possible to return to the
initial questions about semantic variation in relation to Halliday’s proposal
concerning ‘learning language, learning culture’, raised at the end of
Section 19.2: is it possible that children developing language in contrasted
social locations within intersecting fractions within the same culture might
learn to mean in somewhat different ways? Or do they learn broadly the
4
Realization statements for all features presented in Figure 19.2 are available in Williams (1995:159), quoting from
earlier work by Hasan that is not readily accessible.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
506 GEOFF WILLIAMS
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 507
methodology. In this research both the feature [prefaced] and more delicate
features of it were analyzed, e.g. ‘modal’ projections such as I think or Do you
think, and ‘knowledge’ projections such as He didn’t know how they were
hooked on or Do you know what that is? As in Hasan’s research, [prefaced]
was selected differentially to a statistically significant extent within the
contrasted social positions. There were also marked differences in the
frequencies of selection of ‘modal’ and ‘knowledge’ projections (for further
discussion, see Williams 1995:271–5, 2012).
From the perspective of child language development, what is particularly
interesting is that in the children’s talk similar semantic features as in the
mothers’ talk were loaded onto PC1, including [prefaced]. There was again a
strong correlation between children’s scores on PC1 and their social pos-
ition (HAP children>LAP children: p<.009). Again, consistent with Hasan’s
findings, a statistically significant difference was found by Williams
(1995:272). It would appear that by the fourth year of life the children were
taking up different semantic variants of the mother tongue consistent with
features selected differentially by their mothers, and dependent on the
social positioning of their families.
To reiterate, the feature [prefaced] is only one of several which loaded
strongly onto PC1. It was chosen primarily to illustrate the nature of the
findings produced by this SFL approach to child language development
research. However, its selection was not entirely arbitrary. It is now pos-
sible to briefly note a broader research issue, perhaps potentially one of the
most interesting prospects in this theoretical research tradition.
In some of her later work, Hasan advanced some more general pro-
posals based on observations about the semantic functioning of [pref-
aced] and a feature frequently selected in LAP, but not HAP, discourse
which Hasan argues is ‘diametrically opposed’ to [prefaced]. This is the
feature [assumptive], exemplified by such locutions as Didn’t you see me go
out? and Why didn’t you tell me about it? (The location of [assumptive] in
the [demand; information] semantic network can be seen in Figure 19.2.
It is realized lexicogrammatically by an interrogative with negative
polarity.) Selection of [assumptive] implies that the speaker already
knows what an answer should be, and it is ‘diametrically opposed’ to
[prefaced] in the sense that questions with this feature do not carry such
an assumption, i.e. that they are neutral with respect to the addressee’s
knowledge. Thus, if [assumptive] is selected in discourse frequently,
[prefaced] is much less likely to be selected since each implies a different
discourse logic.
Hasan suggested that [assumptive] and [prefaced] function like nuclei in
sets of closely interrelated semantic features, indicated by high-scoring
features on the PCs. These sets of features, she further argued, can be
understood as ‘formative motifs’ in the discourse of caregivers and chil-
dren. She defines a ‘formative motif’ as ‘a cluster of semantic features
which are related to each other by a logic that underlies their configurative
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
508 GEOFF WILLIAMS
rapport’ (Hasan 2004:173), the latter term being derived from Whorf’s use
of it, here reconfigured to semantics. She comments:
Such clusters are made up of meanings that are held together by a logical
necessity, each cluster being built around one strong node, which bears some
contextual presupposition giving rise to a set of implications. For a cluster to
function as a formative motif, it acts not just as some localized meanings –
e.g. the meaning of items such as did you know, do you think, do you remember –
that are relevant but rather what is implied by their use as [preface]: it is the
implied meanings that appear to be most relevant to a semantic cluster’s
capacity to function as formative motif.
(Hasan 2009:449)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 509
It is hard for any of us to keep language in the focus of attention for very
long: we tend to fly off from it in all directions, to study thought processes,
behaviour patterns, aesthetic values and so on. But I shall try to resist this
tendency and shall ask you to think linguistically, that is, to use your
conscious and unconscious understanding of language as a means of think-
ing about the world, and in particular – since this is my unifying concern in
these lectures – to use language as a tool for exploring how people learn.
(Halliday 2016:1)
References
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
510 GEOFF WILLIAMS
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
Language Development 511
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:42:23, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.021
20
Applying SFL for
Understanding and
Fostering Instructed
Second Language
Development
Heidi Byrnes
20.1 Introduction
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 513
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
514 HEIDI BYRNES
1
Comprehensive discussion that also provides some historical depth is offered in Lafford (2007) and Hulstijn et al.
(2015).
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 515
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
516 HEIDI BYRNES
nature of that goal have been well exposed by researchers like Cook (2008)
under the construct of ‘multicompetence’. Not only does multicompetence
nullify notions of permanently deficient L2 learning efforts; it reverses
them by asserting that multilingual users must be seen for what they are
in their own right, not under the restrictive lens of monolingualism.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 517
be posed might be: What are the goals of language development, what counts
as language development in the micro-environment of classroom interaction,
and what would it look like along the extended trajectory of an entire
instructional programme? How might development be ascertained in a prin-
cipled and replicable way that is both attentive to the formal manifestations
of language and unequivocally meaning-oriented? How might it be fostered
and sustained in educational settings, with their relatively proceduralized
assumptions about L2 learning, in light of considerable variation in learning
success among individual learners and, indeed, entire cohorts of learners?
And how might an instructional context prepare learners not only to function
but, ultimately, to thrive in a multilingual environment with its ever-shifting
and complex demands for oral and written literacies?
This section explicates constructs and processes in SFL that are particularly
pertinent to addressing such questions.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
518 HEIDI BYRNES
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 519
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
520 HEIDI BYRNES
SFL answers this key question on two levels. On the first and more abstract
level, it proposes the dynamically double-faced construct of the ‘cline of
instantiation’, imagined as a continuum between the instance and the
system. Coming from the system side, a particular instance realizes certain
subpotentials for meaning-making that the full system holds. Coming from
the opposite direction, a particular text is educationally motivated (as con-
trasted with merely being ‘interesting’) to the extent that it well models a
broader instance type, in short, a particular genre. One might visualize this
cline with the construct of genre occupying the middle ground, as it were,
between the instructional instance and the semiotic system (see Byrnes
2015). By ensuring that its instructed language users will, over the duration
of their studies, encounter an increasingly broader range of major genres, a
programme can lay the groundwork for long-term L2 development. Indeed,
how a programme’s educators imagine that trajectory well describes the
central task of curriculum development (Byrnes 2014a, 2017).
On a second, more directly language-based level, SFL invokes the stratal
quality of language through the ‘hierarchy of stratification’, according to
which language shows ‘a split of content into meaning (semantics) and
wording (lexicogrammar) and a split of expression into phonology and
phonetics’ (Matthiessen 2007:519; also see Tarverniers, this volume). That
stratal quality is further characterized in terms of two axes that operate at
each stratum: the syntagmatic axis is concerned with phenomena of
‘chaining’; by contrast, the paradigmatic axis addresses possible and per-
missible contrasts in a network of options within a stratum – in other
words, it addresses ‘choice’ toward meaning-making.
Not surprisingly, SFL favours the paradigmatic axis of choices, in contrast
to the primarily syntagmatically (or rule-)oriented approaches prevalent in
linguistic analysis and SLA research. Relating that preference to L2 develop-
ment, we might then describe the learner’s real task as that of negotiating –
and repeatedly renegotiating in expanding acts of meaning-making – the
realizational space between the paradigmatic and the syntagmatic axes,
between choice and chain.2 Given that different learners are likely to want
to mean different things even in the relatively staged environment of the
language classroom, a kind of bounded variation characterizes the entire
enterprise of instructed L2 learning and development.
2
For a visualization of these constructs as applied to curriculum building, see Byrnes (2015).
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 521
means in two phases, having a distinct phase of ‘wording’ serving as the base
for the construction of meaning. In other words, its ‘content plane’ contains
a ‘grammar’ as well as a ‘semantics’ . . . It is the presence of a grammar that
gives such a system its unique potential for creating (as distinct from merely
reflecting) meaning.
(Halliday 1996:5)
Speech and writing will appear, then, as different ways of meaning: speech as
spun out, flowing, choreographic, oriented towards events (doing, happening,
sensing, saying, being), processlike, intricate, with meanings related serially;
writing as dense, structured, crystalline, oriented towards things (entities,
objectified processes), productlike, tight, with meanings related as components.
(Halliday 2002:350)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
522 HEIDI BYRNES
3
From the extensive literature on GM use, I highlight Byrnes (2009), Byrnes et al. (2010), and Schleppegrell
(2004a) for L2 learning; Christie and Derewianka (2008) for L1 literacy development; and Bernstein’s sociology of
knowledge as linked to SFL in Martin (2007).
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 523
On that basis, we are now ready to explore how core SFL constructs have
guided research and L2 educational practice. Rather than following more
customary divisions by syntax or vocabulary or different modalities of
language use, I have chosen foci derived from SFL constructs in order to
showcase its unique capacity to address long-term development. Also, to
the extent possible, I have chosen studies that involve different languages,
different learner ages, different educational settings, different performance
levels, and different facets of L2 use.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
524 HEIDI BYRNES
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 525
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
526 HEIDI BYRNES
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 527
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
528 HEIDI BYRNES
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 529
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
530 HEIDI BYRNES
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 531
References
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
532 HEIDI BYRNES
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 533
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
534 HEIDI BYRNES
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
Applying SFL for Understanding and Fostering Instructed L2 Development 535
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
536 HEIDI BYRNES
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:47:41, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.022
21
Language and Education
Learning to Mean
Peter Mickan
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
538 PETER MICKAN
places ‘meaning’ at its core. Halliday’s concept of register (or text type)
gives a means of understanding why certain grammatical choices are made
in a text, and relates text type to context, thus providing an explanation for
Register thus provides a language variation across societies (see Bowcher, this volume). Register thus
link between social provides a link between social context and text. Further, the theory recog-
context and text.
nizes and accounts for complementarities in language and different
instances of language use. For instance, proposing a trinocular perspective,
Halliday (2008:141) explains:
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
540 PETER MICKAN
Figure 21.1 Genre teaching and learning cycle (Rothery 1994; also see Rose and
Martin 2012)
for learners’ composing their own texts and managing discourses for suc-
cess in written work in schools.
In broadening a focus from writing instruction to reading comprehension,
Martin and Rose (Martin and Rose 2008; Rose and Martin 2012; Rose 2011a,
2015) have developed a reading to learn pedagogy with a focus on strategies
for reading texts as preparation for students’ comprehension and writing of
texts. In an interaction cycle, teachers scaffold preparation for reading, for
task engagement, and for elaboration. In the preparation phase, teachers
talk with students about the language features of texts for identification and
comprehension of content so that students gain control of patterns of
discourse and of the subject field which they can apply in their own writing.
The approach assists students’ understanding of the grammatical features of
written texts for making sense of the content or subject matter of texts. The
associated teacher professional development programme (Rose 2015) details
strategies to support learners’ independence and success in reading and
writing for meaning across curricula and year levels.
Text-based curriculum
(Mickan 2013, 2017) situates Pedagogy based on authentic texts has been applied to text-based syllabus
texts in communities of and methodology (de Silva Joyce and Feez 2012). Text-based curriculum
practice and assumes people’s
participation in social practices
(Mickan 2013, 2017) situates texts in communities of practice and assumes
with semiotic resources, as people’s participation in social practices with semiotic resources, as illus-
illustrated in Figure 21.2. trated in Figure 21.2. Learning to use
Learning to use the resources of a language system is a process of social- the resources of a
ization which depends on students’ intense and rich engagement with language system
is a process of
texts. By the time children begin schooling, they are familiar with many socialization
different text types and have a register knowledge for working with and which depends
on students’
intense and rich
engagement with
textsof use,
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 541
responding to texts. The focus of text-based pedagogy (Mickan 2013, 2017) The focus of text-based
pedagogy (Mickan 2013,
is meaning-making, with language a primary semiotic resource for stu- 2017) is meaning-making,
dents’ development of discourses for engagement in curriculum practices. with language a primary
PREPARING
semiotic resource for students’
STUDENTS Students build on their prior knowledge of language in use with analysis of
development of discourses for
FOR THE context, text types, and lexicogrammar with teachers’ scaffolding support engagement in curriculum
EXPRESSION
(Gibbons 2006). Observing the features of selected text types and respond- practices.
OF MEANING
IN THEIR ing to the content of texts prepares students for the expression of meanings
OWN TEXTS in their own texts.
multisemiotic Text analysis also encompasses multimodality. Ventola and Moya Guijarro
approach (2009:1) write that ‘language rarely stands alone in written and spoken
discourses, that is, mono-modally, and that we urgently need to sharpen our
tools in analysing discourses multisemiotically. We cannot continue analys-
ing language alone, but need an integrated multisemiotic approach.’ SFL
studies reveal the multisemiotic nature of texts and illustrate image–text
interactions in children’s reading and comprehension of texts (Chu 2016;
Painter 2007). The perspective extends to multidimensional analyses of sub-
jects in the curriculum. Examples include mathematics and accounting
(O’Halloran 2004, 2007; Alyousef and Mickan 2016).
A challenge in education is to map and report language learning across
years of schooling to achieve continuity, to verify cumulative learning for
reporting, and to validate measures for assessment (Macken-Horarik et al.
2011). Systemic Functional (SF) linguists have tracked language develop-
ment across levels of schooling. Christie and Derewianka’s (2008) descrip-
tions of writing development across school subjects from familiar informal
and spoken-like language to technical language identify and map grammat-
ical features in different subjects. Christie (2012) gives a broad account
based on a functional linguistic framework of language development across
years of schooling. In early childhood education, children build on spoken
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
542 PETER MICKAN
and informal discourses, which they adapt and develop for participation in
curriculum genres encompassing class practices increasingly focused on
literacy and the genres of school subjects. Schooling in early adolescence –
upper primary to junior secondary – reveals the importance of literacy
skills with expectations of student management of new genres and regis-
ters. Students develop knowledge, procedures, and practices of subject
specialisms. They encounter the interplay of verbal and non-verbal
resources, for example, in mathematics and science subjects.
In mid-adolescence students are dealing with abstract knowledge: ‘The
language students must read and write becomes dense, its grammatical
organization more noncongruent, increasingly unlike the more familiar
congruent expressions in which much early common-sense experience is
expressed’ (Christie 2012:105–6). Teaching involves genre analysis and
understanding of language choice in scaffolding students’ insights into
the ways language functions in different text types in subject areas such
as history, English, mathematics, and science. In late adolescence to adult-
hood students engage with theoretical knowledge and abstract phenomena
and ideas for the application of language in the ‘distinctive method of
inquiry and knowledge creation’ (Christie 2012:149). In higher education
the technicality of discourses is embedded in disciplinary practices (Mickan
2013), which is a further extension of students’ construction of the
meaning-making resources for operating in different educational domains.
SF linguists have a special interest in literacy research and teaching in
all levels of education (McCabe et al. 2009; Ravelli and Ellis 2005; Uns-
worth 2000; Whittaker et al. 2006). Hasan and Williams (1996), for
instance, documents the literacy demands on school students and the
relevance of understanding and explaining subject-specific linguistic fea-
Hasan (2011a:173) refers to tures of texts for teaching literacy. Hasan (2011a:173) refers to literacy as
literacy as ‘language-based ‘language-based semiosis’, defining literacy as acts of meaning associated
semiosis’, defining literacy as with a semiotic system. Students experience language variation in their
acts of meaning associated
with a semiotic system.
speech communities and come to school with differing discursive abilities
and ways of saying and doing. For some students the different home and
community socialization experiences prepare them for managing the
literacy expectations of schooling tasks. For other students this is not
the case.
‘Recognition literacy Hasan (2011a:179–206) distinguishes three conceptions of literacy: recog-
pedagogy nition literacy, action literacy, and reflection literacy. ‘Recognition literacy’
is teaching words and is teaching words and ‘language as inventory of forms’ (Hasan 2011a:178).
‘language as inventory of
She claims that this is ‘a conception of language far removed from language
forms’ (Hasan 2011a:178).
as social semiotic practice’ (Hasan 2011a:178), even though it remains a
popular idea of what literacy is all about. According to Hasan the teaching
of spelling, punctuation, and grammar based on set rules without explan-
ation of the relationship of lexicogrammar and semantics inculcates con-
formity amongst students: ‘We know that the norms of language children
are taught in school are precisely those which conform to the practices of a
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 543
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
544 PETER MICKAN
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 545
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
546 PETER MICKAN
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 547
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
548 PETER MICKAN
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 549
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
550 PETER MICKAN
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 551
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
552 PETER MICKAN
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 553
2012; Christie and Derewianka 2008) offers a foundation for valid testing.
Given the immense complexity of measuring the multiple linguistic fea-
tures of texts (Mickan 2003) and the need for multidimensional frameworks
to capture learners’ meaning-making in multimodal texts (Alyousef and
Mickan 2016; Macken-Horarik et al. 2017; O’Halloran 2007; Unsworth
2001), the use of concordancing and computing programmes for analyzing
learners’ achievements is open to further study, along with ICT networks
with options for students to engage with language and other modes of
communication in acts of meaning (e.g. Jewitt 2006).
Teachers and linguists working within the SFL model have both raised
awareness and provided a robust theory of language learning as a life-long
experience for entry and participation in new sociocultural domains, with
language as a meaning-making resource with the capacity to explore new
spheres of human experience which come with social changes in societies
(Hasan 2011b; Mickan and Lopez 2017). Halliday (1996:363–4) comments
that ‘[l]iteracy today includes many contexts . . . To be literate is to operate
in such complex, multiple contexts, to write with many voices, still ending
up with a text, and to read such texts with kaleidoscopic eyes’. The strength
of applying SFL in educational contexts can help us to make sense of human
relationships, of cultural change, of the material world, and of the quality
of peoples’ lives. SFL analyses and applications can increase students’ and
teachers’ awareness of the power of language in daily life and of the
potential for acting to change cultural conditions.
References
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
554 PETER MICKAN
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 555
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
556 PETER MICKAN
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 557
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
558 PETER MICKAN
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
Language and Education 559
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
560 PETER MICKAN
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:44:17, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.023
22
Systemic Functional
Linguistics and
Computation
New Directions, New Challenges
22.1 Introduction
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
562 J O H N B AT E M A N E T A L .
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 563
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
564 J O H N B AT E M A N E T A L .
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 565
22.2 Parsing
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
566 J O H N B AT E M A N E T A L .
and other levels of description are rich and multi-layered in ways that differ
from many other theoretical accounts.
More specifically, the descriptive power of a Systemic Functional Gram-
mar (SFG) lies to a considerable extent in its separation of descriptive work
across ‘structure’ (i.e. syntagmatic organizations) and ‘choice’ (i.e. paradig-
matic organizations). This comes at the cost of high computational com-
plexity, which still presents today the biggest challenge in parsing broad
coverage texts with full SFGs. O’Donnell and Bateman (2005) discuss how
each successive attempt to construct parsing components using SFL then
necessarily led to the acceptance of limitations either in grammar size or in
language coverage in order to proceed.
A parsing process for full SFGs needs then to derive both syntagmatic (e.g.
constituency structure) and paradigmatic (i.e. selections from the system
networks) descriptions. Providing a syntagmatic description is crucial for
parsing, as it is this organizational frame that serves as an anchor for
structured paradigmatic details – that is, it is not sufficient to know that
some feature has been selected; we also need to know precisely which
grammatical unit that feature constitutes. Moreover, we need to be able
to derive constraints on structure that are given by compatible feature
selections and ruled out by incompatible feature selections. This latter task
is a major source of computational complexity and, as Bateman (2008b)
explains, brings with it significant theoretical implications for the construc-
tion of SFL theory as well as of computational systems. Today, it is common
for parsers to rely on simpler syntactic trees (or other non-SFL grammars) as
starting points for the parsing process. First, a syntagmatic organization, or
‘structural backbone’, would be defined, followed by an enrichment by
paradigmatic selections. This technique was subsequently adopted as a
beneficial heuristic for reducing complexity by most attempts to parse with
SFGs (Kasper 1988; O’Donnell 2005; Costetchi 2013).
The first attempt to achieve larger-scale parsing capabilities for SFG was
that of Robert Kasper (Kasper 1988). The structural backbone employed
was provided by a context-free Phrase Structure Grammar (PSG), similar to
Chomsky’s use of a PSG to generate kernel sentences that would subse-
quently be subject to transformations (Chomsky 1957). In Kasper’s case,
each phrase-structure rule was given additional information for mapping
the phrase structure onto a parallel systemic tree. After all possible sys-
temic trees had been created, they were further enriched using informa-
tion from the Nigel Grammar (the SFG generation grammar developed
within the Penman system – see Matthiessen 1985). This process was
extremely slow and worked only on a limited-size grammar because it
involved ‘multiplying out’ all of the combinatorial possibilities inherent
in the grammar. O’Donnell (1993) and Weerasinghe (1994) subsequently
wrote parsers that attempted to build systemic functional constituency
trees directly, avoiding the need to construct full phrase structure gram-
mars. Again, however, the production of phrase-structure rules for
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 567
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
568 J O H N B AT E M A N E T A L .
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 569
Bayesian (Kersting and De Raedt 2007) and Markov Logic (Richardson and
Domingos 2006; Domingos et al. 2010). They might therefore also offer
good candidates for expressing system networks together with those net-
works’ constraints on syntagmatic structure, especially given that highly
efficient (polynomial and even log linear) learning and inference algorithms
already exist for them (Guo and Hsu 2002).
Another direction worth exploring is building the syntactic backbone
as native SFG constituency structures. For this task, an SFG corpus might
be built either from scratch, such as the one in Fawcett’s (1993) COMMU-
NAL project, or by transformation of existing corpora (Honnibal 2004;
Honnibal and Curran 2007). The latter option is more feasible and in line
with the idea of cross-theoretic transformations across distinct grammat-
ical accounts defended above. Such resources could then be employed
in well-established practice from computational linguistics as training
data sets for machine learning algorithms, which we will return to
briefly below.
Automated parsing with functional categories promises a number of
important applications, both within and outside of research settings. Parsed
data can help in determining the register dimensions of a text, assisting in
document classification or analysis of diachronic change. These approaches
can also help develop linguistic theory – in the case of SFL, automated
frequency counting is perhaps the only feasible way of accomplishing the
‘grammarian’s dream’ (Hasan 1987) whereby grammatical distinctions and
lexical alternatives become one unified resource. Such counting would
enable accounts to systematically extend characterizations from system to
instance while respecting statistically derived probabilities for given
contexts.
Functional linguists approach corpora both from above (i.e. looking at
collections of texts as assemblages of registers) and from below (i.e. by
building profiles of lexicogrammatical frequencies). Teich et al. (2016), for
instance, use register theory and selected elements of SFG to analyze a
large, metadata-rich corpus of scientific writing. Automated tagging and
manual annotation are used in tandem to extract frequency counts for
various lexicogrammatical features. Statistical modelling is then used to
model phylogenetic change, as well as disciplinary specialization. Findings
show that disciplines differentiate themselves not only through experien-
tial choices but through differing probabilities within tenor and mode as
well. Teich et al. point out that tenor and mode are often neglected in tool
and method development within NLP, and SFL approaches could then help
achieve more inclusive accounts.
Using similar methods, McDonald and Woodward-Kron (2016) investigate
language change over the course of membership in an on-line support
group for bipolar disorder. Over time, members’ talk increasingly comes
to align with a biomedical ideology: members’ Mood choices shift in order
to allow advice and foreground declaratives to provide hedged advice that
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
570 J O H N B AT E M A N E T A L .
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 571
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
572 J O H N B AT E M A N E T A L .
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 573
on SFL principles. The application domain for this system is that of the
above mentioned ‘intelligent wheelchair’. Such wheelchairs primarily play
a role in interactive exchanges of goods and services. In particular, they
offer a set of services, including going or taking someone to specified
destinations and recharging themselves at an appropriate docking station.
When interacting with such a wheelchair, a human normally gives the
wheelchair a sequence of tasks for it to perform. These tasks are executions
of the wheelchair’s services. The context of a wheelchair offering services
consequently motivates particular interpretations and descriptions of the
utterances occurring. When humans and wheelchairs exchange services
linguistically, the human takes the role of a service client and the wheel-
chair takes the role of a service provider. The human tells the wheelchair
what to do and so is the source of the request; the wheelchair is the
request’s ‘destination’.
This can then be captured in terms of functional configurations in the
linguistic analyses that a dialogue system needs to perform when participat-
ing in natural dialogue in this scenario. Functional linguistic analyses appro-
priate for the context and constructed automatically on the basis of
situation-specific resources are illustrated in Tables 22.1 and 22.2; in this
scenario, the wheelchair goes by the name ‘Rolland’. In Table 22.1, what is
transferred from the service client (Agent) to the service provider (Medium)
is a request, whereas in Table 22.2 what is described is a service of the service
provider. Selecting the appropriate analysis for these grammatically very
similar cases demands that a language analyst knows about the respective
capabilities of the represented participants: only the wheelchair is capable of
moving the client, whereas only the client is capable of calling for services.
Following this line of discussion further: not every process represented in
human–wheelchair interaction is a service, and so the system, and the
underlying theoretical description, must be able to tease these distinct
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
574 J O H N B AT E M A N E T A L .
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 575
linkages between language and situation is in fact not only of interest for
applications; it is also a highly effective strategy for pushing theory devel-
opment. When a computational system is brought to the level of explicit-
ness and completeness that it can actually produce behaviour (at any levels
of linguistic abstraction), weaknesses or gaps in the theories implemented
can become glaringly evident in a way that is simply not accessible when
considering the theories ‘on paper’. When the produced behaviour does not
meet expectations, this is a good indicator of places where theoretical
frameworks may need refinement.
22.4 Multimodality
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
576 J O H N B AT E M A N E T A L .
Much has been written elsewhere about the tools for supporting multi-
modal analysis (see e.g. O’Halloran et al. 2014b); in the following, therefore,
we focus on the application of computational techniques to analyzing
multimodal data.
Despite the computational emphasis, the underlying theoretical frame-
work of O’Halloran’s projects draws heavily on what O’Halloran (2008)
conceptualizes as Systemic Functional-Multimodal Discourse Analysis
(SF-MDA). Following the social semiotic and systemic-functional approaches
to multimodality (e.g. Kress and van Leeuwen 2006; O’Toole 2011), SF-MDA
considers language and image as resources for meaning-making, building
on the rich theoretical framework provided by systemic-functional theory.
This involves, for instance, applying the concepts of metafunctions and
rank to visual images, so as to provide an integrative framework for
describing multimodal data. With this kind of framework at hand, another
question quickly emerges from the computational perspective: namely,
how to move beyond hand-picked examples and bring the powerful theor-
etical apparatus of SFL to bear on large volumes of multimodal data. Indeed,
this question is well known and long-discussed in non-SFL computational
corpus analysis.
Similar questions are posed in the emerging field of digital humanities,
which studies how computational methods and techniques can help to
answer research questions raised in the humanities (see e.g. Berry 2012;
Schreibman et al. 2016). Multimodality, however, has not been theorized
extensively in the digital humanities, except in discussions of developing
new ways of representing and disseminating research results (Svensson
2010). O’Halloran et al. (2014a:565) consequently characterize their work
as a further extension into what they define as ‘multimodal digital human-
ities’, which involves collecting large volumes of linguistic and visual data,
which are then aggregated for analysis and converted into interactive
visualizations for exploration.
O’Halloran et al. (2014a) provide a good example of this approach at
work, exploring the dynamics of urban life in Singapore. To study large
volumes of data collected from various social media services, O’Halloran
et al. (2014a) examined interpersonal meanings across two different semi-
otic resources by using computational techniques. For written language,
O’Halloran et al. (2014a:572) evaluated the sentiment of 98,733 Twitter
messages sent from specific locations by calculating a lexicon-based emo-
tion vector for each message, which captured the basic emotions of happi-
ness, sadness, fear, anger, disgust, and surprise. For photographs, they
applied automatic face detection to 301,865 images retrieved from
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 577
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
578 J O H N B AT E M A N E T A L .
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 579
Better engagement with CL and NLP has a number of benefits for SFL. First,
computational methods can facilitate automatable and reproducible work.
The large amounts of time taken for manual annotation of data mean that
many SFL projects face time and cost constraints. Heavily automated work-
flows, on the other hand, can be deployed on new data at little to no cost.
This seems to be a practical path to Matthiessen’s notion of language as ‘an
assemblage of registers’ (Matthiessen 2015b:44): the same set of routines,
automatically applied to corpora of different domains, could provide an
insightful account of how Field, Tenor, and Mode influence the probabil-
ities for content-stratum phenomena.
At the same time, computational approaches make it possible to empiric-
ally test key components of an SFG. Automated text processing, for
example, may be able to shed light on the oft-noted difficulty of process-
type identification: if a model trained on large, well-annotated collections of
process types cannot accurately predict process-type labels in unseen data
from a similar text type, we may have an indication that our current
understanding of experiential semantics is incomplete. Moreover, as Halli-
day discovered, the number of words needed to collect a quantitatively
useful sample grows exponentially with the delicacy of the phenomenon
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
580 J O H N B AT E M A N E T A L .
of interest. While only 2000 main clauses are needed to create a profile of
Mood or Process Type, hundreds of thousands of words (or a smaller, highly
specific sample) may be needed in order to develop frequency profiles for
lexical alternatives (Matthiessen 2015a). Realizing the statistical component
of SFG at the grammatical pole of lexicogrammar is therefore dependent on
computational methods.
Another important benefit of combining SFL and NLP is that the high-
quality data produced by human annotators with detailed training in SFG
could be effectively repurposed as training data for machine learning algo-
rithms. As noted above, a consistent, well-annotated data set is the main
prerequisite for the development of a high-quality statistical parser.
Finally, for NLP practitioners, SFL provides several well-articulated
hypotheses concerning the relationship between critical linguistic ques-
tions, such as the relationship between text and context, lexis and gram-
mar, and words and meaning. In clearly differentiating between form,
function, words, and meaning, SFG may be able to avoid the pitfalls that
limit the utility of more popular computational grammars for functional-
semantic tasks. It is certainly possible that current limitations in NLP are
not the result of limitations in statistical methods, but in the grammars
accepted by algorithms as input and output. Put another way, even per-
fectly accurate automatic annotation may have limited usefulness if what is
being annotated does not correspond to meaningful distinctions within the
grammar of a language. Goals that are still distant in NLP, such as semantic
parsing and discourse-level annotation, could foreseeably stand to benefit
from the relatively holistic account provided by SFL.
22.6 Conclusion
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 581
References
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
582 J O H N B AT E M A N E T A L .
Domingos, P., S. Kok, D. Lowd, and H. Poon. 2010. Markov Logic. Journal
of Computational Biology: A Journal of Computational Molecular Cell Biology
17(11): 1491–508.
Fawcett, R. P. 1988. Language Generation as Choice in Social Interaction. In
M. Zoch and G. Sabah, eds., Advances in Natural Language Generation.
London: Pinter. 27–49.
Fawcett, R. P. 1989. Towards a Systemic Flowchart Model for Discourse
Analysis. In R. P. Fawcett and D. Young, eds., New Developments in Systemic
Linguistics: Theory and Application. London: Pinter. 116–43.
Fawcett, R. P. 1993. The Architecture of the COMMUNAL Project in NLG
(and NLU). In The Fourth European Workshop on Natural Language Generation.
Pisa.
Fawcett, R. P. 2008. Invitation to Systemic Functional Linguistics through the
Cardiff Grammar. Sheffield: Equinox.
Fischer, K. 2016. Designing Speech for a Recipient: The Roles of Partner Modeling,
Alignment and Feedback in So-called ‘Simplified Registers’. Amsterdam: John
Benjamins.
Grinstead, C. M. and J. L. Snell. 2012. Introduction to Probability. Providence:
American Mathematical Society.
Guo, H. and W. Hsu. 2002. A Survey of Algorithms for Real-time Bayesian
Network Inference. In The Joint AAAI/KDD/UAI02 Workshop on Real-time
Decision Support and Diagnosis Systems. Edmonton, Canada.
Guo, L. 2004. Multimodality in a Biology Textbook. In K. L. O’Halloran, ed.,
Multimodal Discourse Analysis: Systemic Functional Perspectives. London:
Continuum. 196–219.
Haegeman, L. 1991. Introduction to Government and Binding Theory, Volume 2.
London: Blackwell.
Halliday, M. A. K. 1996. On Grammar and Grammatics. In R. Hasan,
C. Cloran, and D. Butt, eds., Functional Descriptions: Theory in Practice.
Amsterdam: John Benjamins. 1–38.
Halliday, M. A. K. and C. M. I. M. Matthiessen. 2014. Halliday’s Introduction
to Functional Grammar. 4th ed. Oxford: Routledge.
Hasan, R. 1987. The Grammarian’s Dream: Lexis as Most Delicate
Grammar. In M. A. K. Halliday and R. P. Fawcett, eds., New Developments
in Systemic Linguistics, Volume 1. London: Frances Pinter. 184–211.
Hiippala, T. 2016. Semi-automated Annotation of Page-based Documents
within the Genre and Multimodality Framework. In Proceedings of the
10th SIGHUM Workshop on Language Technology for Cultural Heritage, Social
Sciences, and Humanities. Berlin: Association for Computational Linguistics.
84–9.
Honnibal, M. 2004. Converting the Penn Treebank to Systemic Functional
Grammar. In Proceedings of the Australasian Language Technology Workshop
(ALTW04). Available online at: www.alta.asn.au/events/altw2004/publica
tion/04-27.pdf. (Last accessed 12/09/2017.)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 583
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
584 J O H N B AT E M A N E T A L .
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
SFL and Computation 585
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
586 J O H N B AT E M A N E T A L .
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:32:32, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.024
23
Clinical Linguistics
Elissa Asp and Jessica de Villiers
23.1 Introduction
In this chapter, we first outline the goals and scope of clinical linguistics
and then review the work that has been undertaken using Systemic Func-
tional Linguistics (SFL) models. Our approach to SFL contributions is organ-
ized by diagnostic groups and, within groups, by metafunctions, linguistic
levels (‘strata’ in SFL terms), and domains as relevant. One of our goals will
be to clarify the function(s), domain(s), strata, and disorders which have
(or have not) been addressed by SFL work in clinical linguistics. This forms
the background for assessing existing and potential contributions of SFL to
clinical linguistics and ‘state of the art’.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
588 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 589
Crystal (2002, 2013) and Perkins (2011) review the publication history of
CLP and point to accomplishments but also ‘gaps’ in clinical linguistics
research. For example, while the ‘successes’ of clinical linguistics are
reflected by more publications and practitioners over the years, Crystal
(2002, 2013) observes that most work is focused on English: only
12 per cent of articles published between 1987 and 2001 addressed other
languages. The English focus may be because of the well-developed descrip-
tive resources characterizing for English. However, at ICPLA 2016,
28 per cent of presentations addressed languages other than English. So,
the English focus may be shifting. Crystal and Perkins’ reviews also note a
predominance of work (over 60 per cent of all papers) on phonetics and
phonology to the detriment of virtually all other areas of inquiry. Perkins
(2011) attributes this to the case loads for (UK) speech therapists, which are
similarly skewed towards developmental disorders of speech. Crystal also
notes the rarity of work on grammatical and lexical impairments and
especially lamented the even rarer appearance of semantic, discourse, and
pragmatic studies. Of particular interest for our current purpose is Crystal’s
(2013:239–41) call for more ‘routine analysis of text, spoken or written,
from a pragmatic point of view’, with pragmatics defined as ‘the study of
the choices we make when we use language, of the intentions behind those
choices and of the effects that those choices convey’.
Despite the predominance of aspects of phonetics and phonology, the
other areas addressed by clinical linguistics are very broad, and extend
to disorders where speech and language are not primary presenting signs,
but communication may be more or less subtly affected. Thus, the areas
studied in clinical linguistics include fluency disorders (e.g. stuttering,
dysarthria), developmental disorders such as autism spectrum disorder,
(specific) language impairment, speech sound disorder, apraxia of
speech, Down syndrome and Williams syndrome, and acquired disorders
resulting in aphasic and/or motor symptoms from brain injuries, stroke,
the progressive aphasias, tumors, infections, and other causes. However,
there are also studies of discourse or language (dys-)functions arising
from neurodegenerative disorders such as mild cognitive impairment,
Alzheimer’s disease, and Parkinson’s disease, as well as studies addressing
psychiatric disorders such as schizophrenia, anxiety, depression, Atten-
tion Deficit Hyperactivity Disorder (ADHD), and Post-Traumatic Stress
Disorder (PTSD). In some of these, the function of clinical linguistic ana-
lyses is not typically targeted towards speech-language therapy, but vari-
ously towards aiding diagnosis, monitoring treatments and evaluating
therapeutic interventions, improving understanding of how disorders
affect other functional abilities, how such effects may be compensated
for, and also as providing ‘windows’ on cognitive and emotional states of
speakers. Additionally, there is now a large independent field that exam-
ines clinical ‘doctor–patient’ and other interactions. This work is import-
ant but beyond the scope of this review.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
590 ELISSA ASP AND JESSICA DE VILLIERS
The calls from Crystal – for clinical linguists to engage in more systematic
descriptions of discourse, to take on semantics, grammar, and lexical ana-
lyses more extensively, and especially to consider the choices speakers
make in discourse as indices for understanding ‘why’ speakers exhibit
specific patterns – suggest that SFL models of language and descriptive
techniques should provide excellent frameworks and a toolkit for clinical
linguistics. SFL’s functional grounding, its focus on text/discourse as the
object of linguistics, and its organization of grammars as system networks
that characterize semiotic potential as choices that may be probabilistically
weighted for demographic, contextual, and textual factors recommend it
for clinical work.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 591
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
592 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 593
Cohesion
Baltaxe and D’Angiola (1992) examined cohesion in one-hour discourse
samples of ten children with autism, eight children with specific language
impairment (SLI) and eight typical developing (TD) children, recorded
during semi-structured conversational play. They analyzed both correct
and erroneous uses of cohesive categories and found that all three groups
used cohesive tie types (reference, ellipsis, and conjunction) in the same
descending order of frequency. However, the group with autism made
significantly more errors than the SLI and TD groups. The SLI and TD
groups did not differ from each other significantly.
Fine et al. (1994) studied cohesion in conversational interviews of
speakers with Asperger Syndrome (AS), speakers with high-functioning
autism (HFA), and a control group. Blind coders rated subtypes of exopho-
ric references to the physical environment and endophoric references,
where antecedents are found within the discourse. The study found
speakers with AS produced more unclear references than other groups,
and that speakers with HFA had fewer ties to previous stretches of mutual
conversation than other groups, using more exophoric reference, a pat-
tern that Fine and colleagues suggest indicates a reduced connection to
the co-constructed verbal dialogue. The approach differentiated HFA
from AS, which at the time were separate diagnostic categories in the
autism spectrum.
Intonation
Fine et al. (1991) also looked at stress and intonation (also see Bowcher and
Debashish, this volume) use in speakers with HFA, AS, and a control group.
They compared how the different groups used intonation, and how they
used stress to indicate Given and New information. Specifically, they
marked intonation boundaries, and coded whether they occurred at appro-
priate or inappropriate locations. They coded and quantified the distribu-
tion of stress to signal Given and New information within tone groups
according to six types: unmarked stress – with stress occurring on the last
content word in a tone unit – and five categories of marked stress, including
(a) stress on a content word that is not final; (b) equal stress on two words
within a tone unit; (c) stress on a function word that is final within a tone
unit; (d) marked stress on a function word that is non-final in a tone unit;
and (e) marked stress on a function word that is non-final in a tone unit that
has two stresses (Fine et al. 1991).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
594 ELISSA ASP AND JESSICA DE VILLIERS
Case Studies
Several case studies have highlighted communicative patterns associated
with ASD. Contributions from Fine (in press), de Villiers and Szatmari
(2004) and de Villiers (2005, 2006a, 2006b, 2011) examine semantic patterns
in different functional domains, including those that are characteristic of
pedantic and echolalic speech. De Villiers (2006a, 2009, 2011) considers
aspects of organization, fluency, and information density that interact in
talk that is considered ‘pedantic’. De Villiers (2011) applied ‘phasal analysis’
(e.g. Gregory 2002) to illustrate coherence and interactional strengths that
were difficult to discern in the conversational and narrative discourse of a
boy with ASD. The study included a tri-functional analysis of linguistic
features and considered dimensions such as context, social distance, and
register, as well as the function of specific speech acts in global interaction.
It identified patterns of topic development and interaction that, while
atypical, appeared functionally appropriate from the perspective of the
speaker with ASD, and consistent with his purposes and interests. This
study is one of a small number of micro-analytic studies of conversational
discourse in ASD that highlight communicative strengths and consider
normative expectations and assessments of successful communication in
clinical contexts (see also Asp and de Villiers 2010).
Fine (in press) also examines speech functions and exchange structure
patterns in two clinical data sets to characterize the transitions between
interlocutors’ speech. In one of these, the interactions of eight children
with autism who had echolalia in their speech were studied. Echolalia was
more or less frequent depending on the information demands of the pre-
ceding utterance. Prior turns in which the demands of the interaction were
less clear were more likely to be followed by echolalic responses.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 595
Rating Scale
In a larger study of forty-six children with ASD, de Villiers et al. (2007) used
SFL description and theory combined with neuropsychological evaluation
to develop a scale for assessing conversation skills in ASD. The study
identified atypical intonation, semantic drift, terseness, pedantic speech,
and perseveration as broad areas of conversational behaviour that were
characteristic of speakers across the ASD spectrum. The scale is specific to
ASD, has good inter-rater reliability, and reflects categories that are rela-
tively independent of IQ or language competence based on standard meas-
ures. The scale may be clinically useful in evaluating conversation skills
over time, and treatment response. It also highlights areas of language
behaviour where treatment programmes may focus attention.
Textual Function
Two controlled studies of children with SLI used SFL techniques and found
limited use of textual resources in the discourse of children with SLI.
Analyzing conversation samples of fifty-seven children with SLI and sixty-
seven TD children, Adams and Bishop (1989) included cohesion analysis as
part of a larger index of conversational skills. The cohesion index was lower
for the group with SLI, though a subgroup of fourteen children with
semantic-pragmatic disorder were not limited in their use of cohesion,
but violated turn-taking rules more than usual.
Thomson’s (2005) study of twenty-five children with SLI and twenty-five
TD children analyzed their use of marked and unmarked Theme in per-
sonal narratives and story retells. In SFL, thematic analysis refers to the
organization of information at the onset of clauses in terms of markedness.
For example, declarative clauses are unmarked when they begin with the
subject element in English but marked if they begin with objects. Similarly,
clause onsets may have interpersonal or textual elements such as unfortu-
nately and moreover whose default position is clause-initial, and so thematic-
ally signal unmarked aspects of interpersonal and textual meaning
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
596 ELISSA ASP AND JESSICA DE VILLIERS
(see Berry, this volume). The two groups had similar proportions of
marked theme in their discourse. However, the TD group had a signifi-
cantly higher proportion of multiple themes and used theme to develop
their discourse coherently as a linear progression, where new information
in one clause is represented as old thematic information in subsequent
clauses, particularly in the story retell condition. Most of the multiple
themes were a combination of textual and topical elements, and there
were no cases of multiple themes with three elements in the narrative
retells produced by the children in the SLI group. The results suggest that
the children with SLI were not as able as the TD group to exploit the more
varied and linear model of theme in the story retell. Thomson suggests
that restricted access to and/or restrictions in selection of lexicogramma-
tical resources may have been a factor in the reduced production of
multiple themes by children with SLI in the retell condition. She further
suggests that this might be a target for therapy, where scaffolding and
examples could enhance performance.
Grammatical Intricacy
Mathers (2005) investigates whether eleven children with ADHD differ
from eleven control children in grammatical intricacy (GI) across text
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 597
Treatment Effects
Tannock et al. (1995) use detailed cohesion and speech function analyses to
describe and measure dose-related effects of methylphenidate (MPH) in
ADHD. MPH is a stimulant medication that Tannock et al. note can be
therapeutic, but can also produce a cognitive effect of over-focusing. Con-
versational interactions between two boys and an adult interlocutor were
analyzed to investigate the effect of three different doses of MPH in a
randomized placebo-controlled double-blind single case study. Frequencies
of the main categories of speech function, namely, initiations, responses,
and continuations, were calculated relative to the total number of speech
functions produced by the child, and the proportion of adult questions that
were followed by a response speech function was calculated. Cohesion was
also analyzed.
Overall, MPH decreased talkativeness as measured by total number of
utterances and number of turns. This was reflected in a reduction of the
proportion of speech functions used to initiate interactions, and an increase
in continuations, which were often used to extend a topic. The children also
responded to more of the adult’s questions in the medicated conditions.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
598 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 599
Cohesion
Unsurprisingly, early SFL-inspired work on discourse in AD investigated
cohesion with differential diagnosis as a primary goal. Ripich et al. (2000),
following up on earlier work (Ripich et al. 1983; Ripich and Terrell 1988),
conducted a study of cohesion in the conversation of sixty people with
mild/moderate AD contrasted with forty-seven demographically matched
control participants, and followed a subset of twenty-three participants
with AD for eighteen months. They were interested in differential diagnos-
tic potential compared with healthy elderly, and in the use of cohesion as a
potential monitoring aid. Participants with AD produced more referent
errors than control participants but otherwise used cohesive devices simi-
larly. In the longitudinal follow-up, use of conjunctions and ellipsis became
significantly less frequent, and use of all cohesive devices declined as
participants’ conversational contributions decreased in length and fre-
quency. They conclude that although referencing errors did distinguish
mild/moderate AD participants’ conversation from that of control partici-
pants, the monitoring potential of cohesion analysis in conversational
samples alone is limited by decreasing speech output associated with dis-
ease progression.
The finding that people with AD use cohesive devices similarly to control
participants, but with more referential errors, has been replicated, for
instance, in Dijkstra et al. (2004). However, Chapman et al. (1995) find no
difference in referential errors between people with mild AD and old (>80)
healthy elderly people, though both groups differ from younger (48–78)
controls. Both studies examine cohesion as one of several factors contrib-
uting to discourse coherence, which is consistently found to be impaired
relative to control participants. They evaluate discourse coherence ratings
of local and global continuities and disruptions of topic, though they
operationalize topic analysis differently.
Lock and Armstrong (1997) investigated the potential for differential
diagnosis of people with AD and with anomic aphasia using cohesion
analysis, since word-finding difficulties are common to both. The anomic
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
600 ELISSA ASP AND JESSICA DE VILLIERS
aphasic speakers produced twice as many ‘disrupted cohesion ties’ (e.g. use
of pronouns without referents) as people with AD, but they had fewer
misrepresentations of content. Lock and Armstrong suggest that, in the
presence of many other linguistic similarities between the discourses of the
two groups, cohesion analysis could be diagnostically useful.
Transitivity
The breakdown in experiential semantics in AD has been explored in only
one SFL-based case study (although see Asp and de Villiers 2010 for ‘SFL-
informed’ analyses). Mortensen (1992) describes the use of transitivity
options by a woman, L.M., with moderate AD in terms of process types,
participants, and circumstances and the amount of information in phrases.
The approach replicates findings from conventional studies: L.M.’s use of
processes is varied and accompanied by appropriate participant roles, but
her phrase structures are simple and show limited lexical variation which,
together with repetition, incomplete clauses, and pronouns without ante-
cedents limits the informativeness of her discourse to ‘simple core’
information.
Interpersonal Interactions
In a series of case studies, Müller and colleagues have described interper-
sonal and conversational skills of speakers with AD resident in long-term
care homes and have shown that despite cognitive and linguistic limitations,
residents mostly use interpersonal resources effectively in conversations.
For example, analyses of speech functions and moves show a ninety-seven-
year-old man with AD leading a conversation with a younger man, complete
with advice and opinions as befit his role and face needs (Müller and Wilson
2008). Similarly, simple counts of turns, types and numbers of moves and
speech functions, and address terms and naming patterns show extended
conversations (up to forty-five minutes in length) amongst residents in
which they successfully enact social roles, flirt, show affection, and so on
(Müller and Mok 2012; Mok and Müller 2013). While the finding of pre-
served interactional abilities in people with AD is not novel, Müller and
colleagues’ point is that encouraging (and even ‘staging’) interactions
amongst residents in long-term care may improve their quality of life by
reducing their social isolation.
Modalization Battery
Though not presented as systemic functional per se, Asp and de Villiers
(2010) offer an interpersonally grounded ‘modalization battery’ for analyz-
ing and coding epistemic stance in AD discourse. The battery is intended as
AD-specific in foregrounding the resources available for speakers to circum-
navigate the epistemic challenges created by their episodic and semantic
memory impairments. It differs from SFL treatments of appraisal (Martin
and White 2005) in so far as the latter see modalization as almost
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 601
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
602 ELISSA ASP AND JESSICA DE VILLIERS
Cohesion
Cohesion in TBI has been studied by investigating the number or proportion
of cohesive ties, cohesive adequacy, and cohesive harmony, often in com-
bination with measures of discourse productivity and story grammar
(Coelho et al. 1995). Yet there is not a clear-cut consensus about cohesion
and coherence problems in TBI (Davis and Coelho 2004). Some studies
have found people with closed head injuries (CHI) produce lower percent-
ages of cohesive ties in discourse than control groups (Mentis and Prutting
1987; Coelho et al. 1991a; Hartley and Jensen 1991; Jorgensen and Togher
2009), but others have not found such differences (in narrative tasks)
(Liles et al. 1989; Coelho 2002).
Cohesive adequacy has also been reported to be lower in texts produced
by individuals with CHIs than in those of control participants, with CHI
participants using incomplete ties that control participants did not use
(e.g. Mentis and Prutting 1987), or using inappropriate exophoric references
significantly more frequently than control participants in tasks requiring
frequent use of exophoric references (McDonald 1993). In a longitudinal
case study, Coelho et al. (1991a) looked at story grammar and cohesion in a
picture-based story generation task in two young adults with a CHI. The
approach identified cohesion problems in both participants, whose scores
were below those of control participants, and demonstrated different recov-
ery patterns, with one participant improving over time. However, in a large
study, Coelho (2002) compared fifty-five adults with CHI with forty-seven
non-brain damaged (NBD) adults on story generation and story retelling and
found the groups were not significantly different in cohesive adequacy.
Additionally, Coelho et al. (1995) and Coelho (2002) studied executive func-
tion (EF) in relation to cohesive adequacy in the CHI group just described
and in thirty-two adults with TBI. Both studies used the Wisconsin Card
Sorting Task to measure EF and found no significant correlations between
the EF measure and cohesive adequacy.
McDonald (1993) also looked at ‘cohesive harmony’ using Armstrong’s
(1987) cohesive chain index, a measure of cohesion based on lexical rela-
tions of co-referentiality, co-classification, or co-extension (Halliday and
Hasan 1976; Hasan 1985), but found head injured participants were similar
to controls in this analysis. Interestingly, this study also used rating scales
based on Grice’s (1978) maxims and found that raters perceived the head
injured speakers as sounding disorganized and confused, but these judge-
ments did not correlate with cohesive harmony scores.
The lack of a clear consensus about cohesion findings in TBI may be
attributed to the small sample sizes (some with fewer than four clinical
participants), and even in larger studies, the heterogeneity of TBI, together
with methodological differences, makes studies hard to compare (Coelho
2002). However, one fairly consistent finding has been that the use of ties
varies according to task type (Mentis and Prutting 1987; Liles et al. 1989;
Hartley and Jensen 1991; Coelho et al. 1991b; Davis and Coelho 2004;
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 603
Interpersonal Analyses
In a substantial body of research, Togher and colleagues have used GSP and ES
analyses to study the interactions of people with TBI in a variety of contexts
and situations. Studies analyze the telephone interactions of individuals with
TBI talking with different interlocutors (Togher et al. 1997a; Togher and Hand
1998) and in different role relationships (Togher et al. 1997b; Togher and
Hand 1998; Guo and Togher 2008), and one compared monologic and jointly
produced discourse in speakers with TBI (Jorgenson and Togher 2009). All
these studies noted some difficulties with aspects of interaction and global
structure in TBI and emphasized the role of communicative partners in
shaping opportunities for communication. For instance, in telephone inter-
actions, service providers adjusted their discourse behaviour when conversing
with people with TBI, by asking for information they already had, by fre-
quently checking information (Togher et al. 1997b), or by giving and asking
for less information than they did with control participants (Guo and Togher
2008:84), all of which limited successful communication.
Subsequently, Togher and colleagues used these results in training and
intervention studies focused on improving the communication of commu-
nicative partners of individuals with TBI, rather than those of TBI partici-
pants, during service encounters (Togher et al. 2004) and everyday
conversations (Togher et al. 2009, 2013). In each of these controlled studies,
the interactional and conversational performance of TBI participants
improved. In the intervention study (Togher et al. 2013), improvements
were maintained six months after training.
Experiential Analysis
Rigaudeau-McKenna (2005) describes language dysfunction in the produc-
tion of clauses based on language sampling of adolescents with TBI and
uninjured children. The study categorized clause simplex and complex
failures in terms of where hesitations, as well as repaired or incomplete
clauses, occur, and what kinds of repairs (if any) are attempted in order to
establish a template for the analysis of ideational dysfunction. This
approach treats dysfluencies as signs of ideational difficulty, rather than
as evidence of executive (e.g. planning) dysfunction.
Aphasia
Aphasia is a generic term for linguistic signs and symptoms of neurological
dysfunction arising commonly from stroke but also from many other
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
604 ELISSA ASP AND JESSICA DE VILLIERS
Ideational Function
Ideationally, aphasia is characterized by lexicogrammatical disruption such
as word-finding difficulties, reduced syntactic complexity, or other
(morpho-)syntactic problems and restrictions in lexical diversity, although
the extent of lexical and grammatical dysfunction varies with aphasia type
and severity (Armstrong 2000, 2005a). Two SFL studies have explored how
diminished ideational resources in aphasia, reflected in vague or generic
lexis and frequent fillers, may limit participation in complex genres such as
arguments (Armstrong et al. 2013) and written correspondence (Mortensen
2005). In the latter study, aphasic writers were found to use fewer optional
generic elements such as ‘orientation’ than either TBI or control partici-
pants. Mortensen speculates that this reflects lexicogrammatical ideational
limitations rather than limited genre knowledge.
Armstrong (2001) also investigates variation in the use of process types
in elicited recounts of four speakers with fluent aphasia and four control
speakers. Verbs were categorized by superordinate process type as mater-
ial, relational, mental, verbal, or behavioural (Halliday 1994). Two of the
people with aphasia (PWA) had fewer mental and relational processes
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 605
Cohesion
Investigations of cohesion in aphasia have been fairly consistent in
reporting poor cohesion (Piehler and Holland 1984; Glosser and Deser
1990; Coelho et al. 1994; Lock and Armstrong 1997; Armstrong 2000;
Olness and Ulatowska 2011; Andreetta et al. 2012). Longitudinal investi-
gations of cohesion and cohesive chains in aphasia discourse have noted
recovery over time (Coelho et al. 1994; Piehler and Holland 1984;
Armstrong 1997).
Additionally, two studies in the 1980s explored cohesive harmony. Bot-
tenberg et al. (1985) compared oral stories of ten PWA to those of control
participants. The study found more variability in the texts of the PWA, as
well as significantly fewer chain interactions. Armstrong (1987) used cohe-
sive harmony to analyze texts produced by three fluent PWA. Lexical chains
were identified and extracted from each text, and the number of lexical
items in the chains was compared to the number of lexical items overall to
create a score or ‘index’. In addition, six listeners rated the texts for coher-
ence. Each of the participants had low scores on the cohesive harmony
index, with almost all the texts falling below 50 per cent, the postulated
threshold for textual coherence. This was significantly positively correlated
with coherence ratings (see also Andreetta et al. 2012). However, Linnik
et al. (2016) observe that five out of six studies that address cohesion and
coherence in aphasia do not find such correlations.
Interpersonal Function
Interpersonal aspects of communication are often fairly intact in aphasia.
Ferguson (1992) studied speech functions in conversational data of five
PWA who were selected to represent a fairly homogenous group and found
they retained access to all options for expressing speech functions and
could modulate their choices by selecting questions instead of commands
in the interest of interpersonal politeness. Mortensen (2005) also found
PWA retained interpersonal resources in her study of written letters.
However, against this positive picture, Armstrong et al. (2011) and
Armstrong and Ulatowska (2007) note that some resources for evaluative
meaning are preserved, yet simplified, in aphasia. Armstrong and
Ulatowska (2007) look at stories about an emotional topic produced by
three PWA and find they use fewer evaluative and emotion words and show
less varied expressions for emotion than control participants. Armstrong
(2005b) reports that some PWA have difficulty with mental and relational
verbs. Five mildly severe to moderately severe PWA use fewer evaluative
relational verbs and fewer mental processes than control participants in an
elicited discourse task. These process types are involved in explicit evalu-
ation and expressions of opinion, and Armstrong suggests the findings,
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
606 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 607
Lee et al. (2015, 2016a, 2016b) adopt a design that built on Spencer and
colleagues’ work, assessing similar features but adding appraisal and a
longitudinal follow-up. The baseline study (Lee et al. 2015) included
twenty AWS/twenty ANS, and assessed volubility, grammatical complex-
ity (including GI but not Theme), modality, and appraisal. It replicated
Spencer et al.’s (2009) results, adding that AWS also use fewer appraisal
resources for expressing opinions. Post-treatment (Lee et al. 2016a), par-
ticipants stuttered less, and their volubility, grammatical complexity, and
modality use increased. There was also a trend towards increased expres-
sion of affect. At twelve months, Lee et al. (2016b) assessed ten AWS and
six ANS (those available for follow-up). Seven of the ten participants
maintained the therapy effect for reduced stuttering at twelve months,
though the group analysis did not show this. Participants also maintained
increases in other measures, most robustly in modality and appraisal, on
which their scores were not different from control participants. The
interpretation of the results, as in Spencer’s work, is that reductions in
stuttering frequency allow participants to engage more actively in conver-
sations and that these beneficial effects are maintained for most partici-
pants over twelve months.
23.4 Discussion
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
608 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 609
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
610 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 611
References
Aboraya, A., C. France, J. Young, K. Curci, and J. LePage. 2005. The Validity
of Psychiatric Diagnosis Revisited: The Clinician’s Guide to Improve the
Validity of Psychiatric Diagnosis. Psychiatry (Edgmont) 2(9): 48–55.
Adams, C. and D. V. Bishop. 1989. Conversational Characteristics of
Children with Semantic-Pragmatic Disorder. I: Exchange Structure, Turn-
taking, Repairs and Cohesion. British Journal of Disorders of Communication
24(3): 211–39.
American Psychiatric Association. 1968. Diagnostic and Statistical Manual of
Mental Disorders. 2nd ed. Washington, DC: American Psychiatric
Association.
American Psychiatric Association. 1980. Diagnostic and Statistical Manual of
Mental Disorders. 3rd ed. Washington, DC: American Psychiatric
Association.
American Psychiatric Association. 1994. Diagnostic and Statistical Manual of
Mental Disorders. 4th ed. Washington, DC: American Psychiatric
Association.
American Psychiatric Association. 2013. Diagnostic and Statistical Manual of
Mental Disorders. 5th ed. Washington, DC: American Psychiatric
Association.
Andreetta, S., A. Cantagallo, and A. Marini. 2012. Narrative Discourse in
Anomic Aphasia. Neuropsychologia 50: 1787–93.
Armstrong, E. 1987. Cohesive Harmony and Its Significance in Listener
Perception of Coherence. In R. H. Brookshire, ed., Clinical Aphasiology.
Minneapolis: BRK Publishers. 210–15.
Armstrong, E. 1991. The Potential of Cohesion Analysis in the Analysis and
Treatment of Aphasic Discourse. Clinical Linguistics & Phonetics 5(1): 39–51.
Armstrong, E. 1997. A Grammatical Analysis of Aphasic Discourse: Changes in
Meaning-making over Time. PhD Thesis, Macquarie University.
Armstrong, E. 2000. Aphasic Discourse Analysis: The Story so Far. Aphasiol-
ogy 14(9): 875–92.
Armstrong, E. 2001. Connecting Lexical Patterns of Verb Usage with
Discourse Meanings in Aphasia. Aphasiology 15(10–11): 1029–45.
Armstrong, E. 2005a. Language Disorder: A Functional Linguistic Perspec-
tive. Clinical Linguistics & Phonetics 19(3): 137–53.
Armstrong, E. 2005b. Expressing Opinions and Feelings in Aphasia: Linguis-
tic Options. Aphasiology 19(3–5): 285–95.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
612 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 613
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
614 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 615
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
616 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 617
Mok, Z. and N. Müller. 2013. Staging Casual Conversations for People with
Dementia. Dementia 13(6): 834–53.
Mortensen, L. 1992. A Transitivity Analysis of Discourse in Dementia of the
Alzheimer’s Type. Journal of Neurolinguistics 7(4): 309–21.
Mortensen, L. 2005. Written Discourse and Acquired Brain Impairment:
Evaluation of Structural and Semantic Features of Personal Letters from a
Systemic Functional Linguistic Perspective. Clinical Linguistics & Phonetics
19(3): 227–47.
Müller, N., M. J. Ball, and B. Rutter. 2008. An Idiosyncratic Case of /r/
Disorder: Application of Principles from Systemic Phonology and Sys-
temic Functional Linguistics. Asia Pacific Journal of Speech, Language and
Hearing 11(4): 269–81.
Müller, N. and Z. Mok. 2012. Applying Systemic Functional Linguistics
to Conversations with Dementia: The Linguistic Construction of Rela-
tionships between Participants. Seminars in Speech and Language 33(1):
5–15.
Müller, N. and B. T. Wilson. 2008. Collaborative Role Construction in a
Conversation with Dementia: An Application of Systemic Functional
Linguistics. Clinical Linguistics and Phonetics 22: 767–74.
Olness, G. S. and H. K. Ulatowska. 2011. Personal Narratives in Aphasia:
Coherence in the Context of Use. Aphasiology 25: 1393–413.
Perkins, M. 2011. Clinical Linguistics: Its Past, Present and Future. Clinical
Linguistics & Phonetics 25(11–12): 922–7.
Piehler, M. F. and A. L. Holland. 1984. Cohesion in Aphasic Language. In
R. H. Brookshire, ed., Clinical Aphasiology: Conference Proceedings.
Volume 14. Seabrook Island: BRK Publishers. 208–14.
Quirk, R., S. Greenbaum, G. Leech, and J. Svartvik. 1985. A Comprehensive
Grammar of the English Language. London: Longman.
Reilly, S., B. Tomblin, J. Law, C. McKean, F. K. Mensah, A. Morgan, S. Gold-
feld, J. M. Nicolson, and M. Wake. 2014. Specific Language Impairment:
A Convenient Label for Whom? International Journal of Language & Commu-
nication Disorders 49(4): 416–51.
Rigaudeau-McKenna, B. 2005. Towards an Analysis of Dysfunctional Gram-
mar. Clinical Linguistics & Phonetics 19(3): 155–74.
Ripich, D. and B. Terrell. 1988. Patterns of Discourse Cohesion and Coher-
ence in Alzheimer’s Disease. Journal of Speech and Hearing Disorders 53:
8–15.
Ripich, D. N., B. Y. Terrell, and F. Spinelli. 1983. Discourse Cohesion in
Senile Dementia of the Alzheimer Type. In R. H. Brookshire, ed., Clinical
Aphasiology Conference Proceedings. Volume 13. Minneapolis: BRK Publish-
ers. 316–21.
Ripich, D. N., B. D. Carpenter, and E. W. Ziol. 2000. Conversational Cohe-
sion Patterns in Men and Women with Alzheimer’s Disease: A Longitu-
dinal Study. International Journal of Language & Communication Disorders
35(1): 49–64.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
618 ELISSA ASP AND JESSICA DE VILLIERS
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
Clinical Linguistics 619
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:16:58, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.025
24
Language and Science,
Language in Science,
and Linguistics as Science
M. A. K. Halliday† and David G. Butt
24.1 Introduction
†
2018
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 621
relations had a bearing on the classical Greek versions of what we now refer
to as science – e.g. cosmology; medicine; physical theory. We note here the
enormous project of Geoffrey Lloyd (1970, 2002) in investigating the roots
of sciences in classical Greece. We also note the insights of Edward
Grant (1996, 2007) and others in relation to the actual breadth of ‘natural
philosophy’ as European science in the Latin Middle Ages.
But our focus is on the linguistic patterns that are produced by the shift
towards enquiry articulated and argued in the vernacular languages of West-
ern Europe, and more precisely by the shift ‘within’ English in the seventeenth
century. This is when science undergoes a transition in the contexts in which it
is conducted, and when there is a shift in the relation between novelty,
knowledge, and authority. For instance, scientists moved away from the
assumption that science was just the rediscovery of what had been previously
known, but then hidden, by unimpeachable authorities in eras of revelation.
In addressing our second question, we take up the specifics of what can
be regarded as a grammatical ‘drift’ in English, as described in Sapir’s work
(1921:154–5). It constitutes a convergence of features that, together, can be
thought of as an instance of ‘configurative rapport’ – a notion proposed by
Whorf (1956) for an alignment of grammatical categories which appear to
facilitate a specific variety of socio-semantic activity, a register in this case.
The main purpose of Section 24.2 is to illuminate the features of that drift
and their characteristic direction with respect to the purposes of science
from the seventeenth to twenty-first century.
Our third goal is to consider the place of linguistics and its terminologies
and arguments – our language for talking about language – within the drift
discussed in Section 24.3, ‘Language in Science’. As summarized by Wino-
grad (1983), linguistics has hunted down and borrowed its terminology from
diverse cultural activities: for example, from law; chemistry; atomic physics;
mathematics; logic; philosophy; music; and literary theory. Linguistics is not
unique in these analogical borrowings; but it may be unusual in the degree
to which linguistic thought has been controlled by the metaphors drawn
from outside the specific works of the subject (consider the role of logic as an
influence on semantics). In addition, there is in Western science a double
fold in terminology: as Halliday (2004c) points out, we use and mix both
Greek and Latin terminologies, with Greek typically the term that is higher
up in a scale of abstraction (see Section 24.2 below).
At different points in this discussion, we elaborate our views of an
important analogy between ‘verbal science’ and ‘verbal art’. In both verbal
art and verbal science, there is a ‘sideways’ move in the way the practitioner
utilizes the linguistic system: there is a step into an established yet
malleable metaphorical architecture, a tradition that presents opportunity,
constraint, and challenge. The opportunity is in the potential for
innovation – adding to the range of representations and their relevance to
linguistic activities. The constraint is in the formal alignment with previous
practice – how newly introduced metaphors are connected with those
already established and legitimized by a tradition (either formalities
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
622 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 623
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
624 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 625
From this drift in the contexts of scientific debate, we can see the conse-
quences in new pressures on the discourses of science. With such new
pressures to explain the world, it is not difficult to understand that scien-
tists have often expressed dissatisfaction with the forms of language they
have inherited and, in some cases, dissatisfaction with the idioms that
they have (unconsciously) created for themselves. Debate has often con-
cerned precision in language and drawn in major figures like Leibniz and
Bentham. Bentham extolled the solidity of nouns and decried the slipperi-
ness of verbs (Kline 1953:312). Bertrand Russell in the twentieth century, as
well as the logical positivist movement and the Vienna Circle, still worked
for a similar purging of misleading grammatical constructions, namely,
those which looked like propositions but which implied the existence of
the non-existent: e.g. The king of France is bald. There has been, since the
Greeks, a preoccupation with ‘what is’, with ontology and the allocation of
‘be-ing’ (see Kahn 2009).
In both useful and eccentric ways, scientists have had much to say about
the status of language in science: for instance, Einstein, Heisenberg (1958),
and Bohm (1980) in physics; Waddington (1977) and Medawar (1984) in
biology; Whitehead in philosophy/logic; both Whorf and Chomsky (in dif-
ferent ways) in relation to grammar; and Firth in phonology and in linguis-
tics more generally (see Firth 1957, 1968; and also Butt, this volume).
The views are polarized: poetry was dismissed as ‘ingenious fiction’
(by Newton’s teacher Barrow); but in the twentieth century, the poet
Shelley’s sense of imagination (poiesis) was carefully aligned with the innov-
ations of science by Medawar – a biologist with little patience for anything
less than a strong Popperian approach to falsification (Medawar 1984:50–2).
Waddington (1977) and Bohm (in biology and physics respectively), and
Whitehead (logic and philosophy) have all expressed Whorfian-like con-
cerns about the narrowness of scientific habits of representation, for
instance, in relation to the inadequate representation of ‘process’. By con-
trast, Heisenberg was disgusted by the rival version of quantum mechanics
proposed by Schrödinger because it moved away from mathematical for-
malism to an unworthy tendency to visualization (Cassidy 1992:215).
By contrast, in Crosby’s study of measurement – in relation to the
development of sciences and European imperial expansions – it is to the
visualization of knowledge that the colonialist dominance of Europe is
attributed – from ballistics to bookkeeping in commerce (Crosby 1996).
Halliday has discussed two technical problems that are relevant to the
role of natural language as an instrument of science, and as an object of
scientific enquiry – hence, to the two main issues of this chapter. Halliday
(2002) notes that, firstly, the meanings of grammatical categories are
‘ineffable’, or unresolvable to any origin in a non-semantic plane of experi-
ence. That is, they cannot be referred to anything but another version of
themselves. And, second, that the linguistic categories which have to do
service in a metalinguistic statement are dominated by traditions which are
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
626 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 627
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
628 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 629
The rate of crack growth depends not only on the chemical environment but
also on the magnitude of the applied stress. The development of a complete
model of the kinetics of fracture requires an understanding of how stress
accelerates the bond-rupture reaction.
In the absence of stress, silica reacts very slowly with water.
(Michalske and Bunker 1987)
Here are instances of some of the features that express the syndrome
referred to above:
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
630 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
understand; and with the projected clause how stress accelerates . . . func-
tioning, by rank shift, in the Qualifier;
(11) the clause stress accelerates the bond-rupture reaction, with finite verbal
group accelerates as the relationship between two things which are
themselves processes: one brings about a change in an attribute of the
other, agnate to makes . . . happen more quickly;
(12) the simple structure of each clause (three elements only: nominal
group + verbal group + nominal group / prepositional phrase) and the
simple structure of each sentence (one clause only);
(13) the relation of all these features to what has gone before in the
discourse.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 631
(1) expanding the noun as a taxonomic resource: this was the goal of
language planning in the 1600s, especially in England and France (see
Salmon 1979);
(2) transcategorizing processes and qualities into nouns, relators into
verbs, etc., with resulting semantic junction;
(3) compacting pieces of the argument to function (e.g. as Theme of the
clause) in an ‘information flow’ of logical reasoning;
(4) distilling the outcomes of (2) and (3) to create technical taxonomies of
abstract, virtual entities;
(5) theorizing by reconstruing of experience as in (1) to (4), with a
‘favoured clause type’ in which virtual entities participate in virtual
processes based on logical-semantic relations that bestow a linear and
causal plausibility on the phenomena under investigation (see Halliday
and Matthiessen 2014).
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
632 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
(1) A developmental loss of intrinsic reparative capacity and (2) the inhibitory
environment in injury and disease contribute to (3) regenerative failure in
the central nervous system (CNS).
(Cameron and Goldberg 2016:30, emphases added)
The pathway as Head noun is qualified by a nominal group (Head word: cells),
and then qualified separately by a clause with new nominal (Head noun:
factor), which is itself extended by a group which elaborates factor with
another complex group with the Head noun molecule. The management of
the web of life demands an isomorphic web of intricate verbal taxonomy,
each node of which is legitimated by logical relations (which are often
muted or neutralized with forms like involves, mediates, contributes to, emerges,
depends on, comes mainly from, etc.).1
The grammar, as a stratified system of contexts, semantics, and lexico-
grammar, sets up categories and cross-stratal relationships which have the
effect of transforming experience into meaning. In creating a formal dis-
tinction such as that between verb and noun, the grammar is theorizing
about processes: that a distinction can be made, of a very general kind,
between two facets – the process itself, and entities that are involved in it.
But, as remarked above, since the grammar has the power of construing,
by the same token (that is, by virtue of being stratified) it can also decon-
strue, and reconstrue along different lines. Since stratification involves
mapping meanings onto forms, ‘process’ onto verbal, and ‘participant’ onto
nominal, it also allows remapping – say, of ‘process’ onto a nominal form: a
process can be re-packaged as a nominal group so that a bus driver does not
drive but owns the process: viz. the driver’s overrapid downhill driving of the
bus. The experience has now been retransformed – in other words, it
has undergone a process of metaphor. It can now be predicated with a
1
A crucial point in the next decades of science teaching and learning may be that such taxonomies in Chinese, for
example, do not take their morphology from an arcane, classical, synthetic tradition. A fifteen-year-old student learning
science in China deals with combinations of common-sense meanings in trying to understand the complexifications of
biology or physics. We may need to take stock of the language of science teaching along with its current creative
visualizations and online presentations.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 633
(1) verb: active (Actor) noun: ‘one who / that which . . .-s’
ποιέω make ! ποιητής maker
πράσσω do ! πράκτωρ doer
(2) verb: passive (Goal) noun ‘that which is . . .-n’
be made ! ποίημα thing made
be done ! πρᾶγμα thing done, deed
(3) verb: middle (Medium) noun ‘. . .-ing’ (abstract)
make ! ποίησις making
do ! πρᾶξις doing, action
(4) adjective: noun of quality / degree ‘being . . .; . . .?’
μέγας big ! μέγαθος size; greatness
βαθύς deep ! βάθος depth, deepness; altitude
The Greek forms provided the model for scientific terminology in Europe.
In scientific discourses, the semiotic power of referring is being further
exploited so as to create technical taxonomies: constructs of virtual objects
that represent the intensification of experience (typically experience that
has itself been enriched by human engineering, in the form of experiment
and/or measurement). The semiotic power of expanding begins by relating
one process to another by a logical-semantic option. This potential can be
exploited so as to create chains of reasoning: drawing conclusions from
observation (often observation of experimental data) and construing a line
of argument leading on from one step to the next.
Grammatically, these two discursive processes, which lead out of the
daily language into a mode of systematic theory, both depend first and
foremost on the same basic resource: the metaphoric compression of a
clausal into a nominal (thingifying) mode of construal (see Figure 24.1
and Table 24.1). The novel ‘things’ of science populate a world of semiotic
artifice, no less than is true for the metaphoric strategies of verbal art. The
depiction of reality in both cases is through an oblique, derivative semiotic
technique. They start from two variants of experience, however. Verbal art
might find its conventions in clusters of metaphors (consider, for example,
the European Middle Ages and the ‘expected’ values expressed in the
motifs: rose, thorn, blood, and knight errant).
But science, typically as natural philosophy, has had to thingify by creat-
ing a more plausible architecture of smaller components, with or without a
God as ‘prime mover’. Reduction has been the dominant pathway of sci-
ence, whether the existence and relevance of the particular elements pro-
posed could be supported by evidence or not. This semiotic vector has
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
634 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 635
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
636 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
culminating perhaps in the Greek kinesis. Note that the most distilled terms
in English at its most theoretical level tend to be those from Greek (e.g.
ornitho- for ‘bird’).
Let us consider then:
(1) What is the payoff? That is, what effect has such reconstrual on the
construction of the discourse?
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 637
(2) What different kinds of metaphorical shift take place, and is there any
general principle lying behind them?
(3) What are the systemic consequences? To put this in other terms, in
what way is ‘regrammaticizing’ experience also ‘resemanticizing’ it?
In saying that these are of the ‘favoured clause type’, this is not asserting
that they are the most frequent (there would be no sensible way of estimat-
ing this; at the least, they are certainly very common). But they are the most
critical in the semantic load that they carry in developing scientific argu-
ment. What is interesting about them is that their clause structure is
extremely simple: typically one nominal group plus one verbal group plus
a second nominal group or else a prepositional phrase. But packed into this
structure there may be a very high density of lexical matter; again, compare
the tendency of spoken discourse with the ten lexical words to one verbal
group within the third clause (referred to by the number [3] as cited above).
So, a scientific theory is a specialized, semi-designed subsystem of a
natural language; constructing such a theory is an exercise in lexicogram-
mar. Science and technology are (like other human endeavours) at one and
the same time both material and semiotic practices; knowledge advances
through the combination of new techniques with new meanings. Thus
‘reconstruing experience’ is not merely rewording (regrammaticizing); it
is also resemanticizing. The languages of science are not saying the same
things in different ways (although they may be appropriated for this pur-
pose by others wishing to exploit their prestige and power). What is
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
638 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
With our third question put forward at the outset of the article, we empha-
size how scientific thinking must include thinking about language. Much
as an observation requires understanding of light and the effect of lenses,
language is not only an object of enquiry for natural sciences, it is the
medium and channel by which the knowing is secured for the thinking of
scientists. While mathematics has come to be a lingua franca or, perhaps
more narrowly, an esperanto for science (with its own registerial or generic
variants), it is still a requirement of verbal science that quantification
be demonstrably relevant to verbal explanation: i.e. that numbers be
explained. While linguistic sense may need to introduce the role of numer-
ical sense as protection against certain forms of nonsense, all claims come
back to the court of collective consciousness – the language of a
community.
Where there appear to be exceptions – for instance, in Heisenberg’s
formalism in quantum theory, and in quantum entanglement – human
thinking remains restless for a consistency with our common speech and
its common-sense reasonings (see Brooks 2016b; Ananthaswamy 2013). In
quantum weirdness, for example, we might say that the semogenic
resources have been extended to create a new complex meaning; and this
meaning can be thought of, as mentioned above, as an extension of the
texture of language. A comparison with the higher-order process that we
2
See Lucretius’ (1992) 7,000 line poem, De Rerum Natura (On the Nature of Things); and also Mann (2000) on
Aristotle, The Discovery of Things.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 639
call ‘verbal art’ arises at this point in our discussion. Through the tradition
of specialist discussions and mutual exchanges – sometimes face to face –
both scientists and poets create semiotic novelty by reworking the meaning
potential of the collective. In some cases, new symbolic resources are
introduced – for instance, the uncommon sense of calculus with Leibniz’s
infinitesimals and Newton’s fluxions; more recently with the counter-
intuitive behaviours of the very big and the very small in physics. The
physicist Bohm asks us to become conscious of the symbolic process of
making statements in science much in the way that the Russian Formalist
critics of the early twentieth century spoke of defamiliarizing our habitual
modes of describing experience (Bohm 1980:32).
Certain developments may be characterized as forms of semantic canal-
ization (analogous to the biologist’s chreods in an epigenetic landscape
(Waddington 1977:106–7)). These might be efficacious ways of saying novel
ideas as the result of the latent potential of a habitual speech pattern, or a
generic structure. An example may be the power of the definite article in the
grammar of classical Greek, mentioned above. Along with the semantic
‘figure’ of dialogic enquiry (viz. Socrates/Plato), the definite article facilitates
broad, abstract questioning: the essentialist, generalizing, and defining
orientation of sciences (as well as the potentially misleading reification, we
should add). As pointed out by Snell (1953), a Greek philosopher could easily
take discourse from the concrete behaviours of those around to the question
of what defined ‘the good’, and hence on to essentialist discussions of good,
beauty, freedom, power. Kappagoda (2004) gives another important
instance of what might be called a semiotic canalization – a semanticization
of a mode of understanding previously outside the habits of the community.
He shows how Thucydides applies a forensic perspective to the sequence of
events in the plague in Athens. He argues that such a potential grew out
from the metaphoric form As to the leaves of Autumn are human beings.
While there has been a shared origin in the measurement of language in
verse and the measures of music, with Pythagoras first showing the relation
between measurement and human perception (the harmonic scale based on
systematic divisions of string lengths), we will focus on the emergence of
evolutionary ideas in Europe to review the paradoxical status of language
studies, their hidden role at the forefront of what are tendentiously referred
to as the ‘hard’ sciences. As suggested above, the innovations of such sciences
were prepared for by progressive cultural increments in other ideational
domains, domains today not necessarily thought of as objective sciences.
The genealogical ideas of Indo-European and other language families
were adumbrated by William Jones, and authorities preceding Jones, more
than a century before the publication of Darwin’s ‘descent with modifica-
tion’ – the genealogical principle in nature. We would emphasize the
important parallel that Darwin urged between a hypothetical knowledge
of the ancestry of all languages, with all the intervening dialectal variants,
and the ‘economy of nature’, with its slow variants in which ‘natura non
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
640 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
facit saltum’ (‘nature does not make leaps’), and the ‘community of des-
cent’ in relation to species and morphology (Darwin 1998:562–4). Without
any knowledge of genes, the language analogy became Darwin’s most
concrete instance of a mechanism of the genealogical principle. This is
particularly because, in the opening chapters of On the Origin of Species,
Darwin takes such care in distinguishing between selection under nature
and selection by humans, for instance, by animal breeders. Dawkins too, in
his explanation for children of crucial rational ideas – The Magic of Reality –
emphasizes the analogy between natural species and languages (Dawkins
and McKean 2011). But both Darwin and Dawkins pull up short of claiming
a homology, not just an analogy.
Research by Dunn et al. (2011) used mathematical methods for what the
Prague School referred to as ‘characterology’ – a comparison across typo-
logical features between languages. The authors offered the same Darwin-
ian conclusion: the ancestry of a language is the best predictor of the
affinities and contrasts between languages. It is therefore curious to us,
given the role of proto-social sciences in the crowning works of physics and
biology, why so much about language has been excluded from the scientific
conversation. Bohr himself urged the analogy between an anthropologist
disturbing the community under observation and the physicist intervening
in a subatomic system (Bohr 1961); and the literary potential of point of
view, relativity, and complementarity have been utilized by writers and
critics (see, for example, the broad cultural study by Albright 1997). There
was nascent ‘population thinking’ in Defoe’s A Journal of the Plague Year (see
Defoe 2001) as there is a non-fiction journal in Mailer’s Armies of the Night
(1968), and much about point of view and the philosophies of physics and of
rhetoric are illustrated in Sterne’s The Life and Opinions of Tristram Shandy
(1759–1767) (hence the fascination the work produced among the Russian
Formalists, who sought to make a verbal science of the study of verbal art).
The forensic analysis of social complexity in the opening chapter of Mid-
dlemarch (Eliot 1871), and the forensic criticism of Napoleon’s military
strategies in the epilogue of Tolstoy’s War and Peace (1869) tend to support
the idea argued by Snell (1953): namely, that the very forms of literary
genres supplied modes of transitional thinking in the development of
classical Greek science. Since the mobilization of language to represent
human purposes necessarily draws one into “choosing” one thing . . . rather
than the other’ (Halliday 2005), verbal art is a laboratory for thought
experiments on relativity and physics/physis (Butt 2007). Consider how
prose accounts of experience and reported speech commit us to choices of
point of view and deixis.
One problem for the development of any science of language was and is
the ineffability of grammatical categories cited above (Halliday 2002).
While the earliest Greek and Latin grammarians purported to be offering
a theory of natural kinds in the classifications of verbs and nouns, and
other parts of discourse (viz. the very word ‘genres’, or types), it becomes
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 641
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
642 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 643
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
644 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 645
‘Was this the last article that Michael penned? Would this be the last that
he published?’ When the editors asked these two reasonable questions
following Michael Halliday’s death on 15 April 2018, I was concerned that
Michael’s career of assiduous attention to his own textual crafting of
articles might be diminished by my role as a co-author. Certainly, Michael
launched the chapter with typed pages of its plan and coverage of subject
matter; and, yes, for a number of years, Michael and I had discussed
linguistics as a science; the language of science; evolutionary theory; the
semiotic character of debate in physics; the relation between recent neuro-
science and the claims of cognitive science; and the connection between
Ruqaiya Hasan’s work in verbal art and his own work in what he called
‘verbal science’. Ultimately, however, as Michael neared the end of his life,
he was not sufficiently robust to complete the work as he would normally
do – as the keen-eyed writer and editor we knew him to be. The process
made me acutely aware of my own shortcomings as a co-author.
In fact, after many attempts to reduce and represent Michael’s articles on
‘grammatical metaphor’, I had a fresh sense of the profound nature of his
mode of thought and expression: each article added new angles and set out
abstract meanings with the concreteness of the great teacher that he was.
In the end, I opted to include passages from his work rather than mangle
his expository strategies.
By March 2016, Michael had completed the preface to the translation of
some of his papers into Spanish (published 2017). He also penned a series of
gnomic headings concerning the integration of his theory of metaphor with
Ruqaiya’s work, in particular, in relation to the development of higher-
order thinking in a culture. I will pass on to his readers what I can develop
of some challenging words and instructions that he left through his con-
versations, when I can do them justice. For now, as I reflect on Michael’s
lexicogrammatical choices, I see the fulfilment of a goal attributed by Luria
to Marx: the importance of ‘ascending to the concrete’. To end on an
instance, recall:
There can be no semiotic act that leaves the world exactly as it was before.
(Halliday 2002(1994): 2.254)
David G. Butt
References
Albright, D. 1997. Quantum Poetics: Yeats, Pound, Eliot, and the Science of
Modernism. Cambridge: Cambridge University Press.
Ananthaswamy, A. 2013. Quantum Shadows. New Scientist 217(2898):
36–9.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
646 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 647
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
648 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
Language and Science 649
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
650 M . A . K . H A L L I DAY A N D DAV I D G . B U T T
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 08:39:56, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.026
25
Language and Medicine
Alison Rotha Moore
25.1 Introduction
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
652 A L I SO N R OT H A M O O R E
There are well over 100 publications and theses on language and medicine
that use SFL principles, covering many healthcare contexts and analytical
foci. A good way of grouping such publications is around their stated or
implied ‘intervention points’ – the physical or conceptual places within a
biological or institutional system where pressure can be applied to disrupt
existing function and promote change (Reinsborough and Canning 2010).
The present review starts outside the health system, with the everyday
construal of pain, then moves to the ‘core’ of the healthcare system, namely,
spoken interaction between clinicians and patients. We then move to con-
texts that support and/or shape this core, such as interpreting, and inter-
actions within clinical teams. Finally, we consider broader institutional and
cultural contexts in which healthcare is situated, suggesting there is
untapped potential for SFL here.
This notion of intervention is important for evaluating the impact of SFL
medical linguistics and considering where future efforts are best directed
because healthcare is a relatively weak determinant of health (see Figure 25.1),
perhaps as low as 15 per cent (McGinnis et al. 2002). Note that some projects
discussed involve multiple sites of engagement/intervention.
Figure 25.1: The main determinants of health (after Dahlgren and Whitehead 2007)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 653
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
654 A L I SO N R OT H A M O O R E
treatment (Moore et al. 2001; Moore 2004). Some health outcomes may
stem directly from the therapeutic value of the communication itself in
physical conditions (Street et al. 2009) as well as in the context of
psychotherapy.
Probably the first use of SFL for extended analyses of clinician–patient
interaction was Mishler’s Discourse of Medicine (1984). Critiquing the then
mainstream quantitative methods of studying clinical interaction, Mishler
argues that they ignore the problems of transforming speech to written
transcripts as ‘data’, which tends to strip away meaning in a quest for
objectivity. Mishler, a social psychologist, uses all three metafunctions of
SFL but focuses in particular on cohesion, adapting Halliday and Hasan’s
(1976) approach to suit dialogue. He identifies clusters of structural, seman-
tic, and grammatical patterns that tend to be treated as routine and
unproblematic in medical interaction, which he identifies as ‘the voice of
medicine’. These contrast with and are typically used to interrupt ways of
speaking known as ‘the voice of the lifeworld’, which is dominated by
features such as temporal rather than causal organization. Mishler’s work,
which largely aimed to get clinicians to encourage not silence ‘the voice of
the lifeworld’, has had a strong influence on how clinical interaction is
studied and taught, in no small part due to Mishler’s position and influence
at Harvard Medical School.
Close on Mishler’s heels was Cassell’s Talking with Patients (1985). Cassell
stresses that clinicians need a solid grounding in the systems of language
underlying clinical communication before ‘communication skills’ are
taught, just as one would ‘never dream of teaching physical diagnosis to
students lacking a background in anatomy and pathology’ (Cassell 1985:5).
Having recorded hundreds of hours of consultations, Cassell exhorts clin-
icians to study their own dialogue with patients and learn to spot subtle
features, such as how ‘people shift to impersonal pronouns when they
describe their illnesses or unpleasant events’ (Cassell 1985:8) and how
patients attach meaning to symptoms and illnesses.
To my knowledge, the first description of generic structure for medical
consultations was given by Halliday (2005), comprising ‘opening’, ‘investi-
gation’, ‘examination’, ‘diagnosis’, and ‘suggested treatment’, which
includes ‘negotiation’ and ‘reassurance’ (a structure for his single example
text, not a Generic Structure Potential). Interestingly, the ‘treatment phase’
is seen as a typical manifestation of the complex power relationship
between professionals and clients, with its grammatical shifts in mood
and modality. Such an account can be linked with Halliday’s discussions
elsewhere of the relation between domains of activity and registerial
boundaries, or points on the cline of instantiation, since it emphasizes
similarities in the registerial settings of medical and other professional
discourse (see Moore 2017).
Following in this vein, much of the SFL literature on medicine and
language describes generic structure and/or explores patients’ construals
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 655
of their experience and values, and how these are taken up or not by
clinicians in consultation settings. This body of work includes descriptive
studies of specific sites (e.g. general practice, emergency medicine) and
conditions (e.g. HIV disease) and has generated interventions such as prac-
tical handbooks and other professional development material for clinicians.
1
For additional analysis of this data, see Halliday and Greaves (2008:80–94).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
656 A L I SO N R OT H A M O O R E
GPs and patients bond over the patient’s values rather than the doctor’s,
although other studies suggest that this might depend on the presenting
condition and length of clinical relationship (Moore 2004).
Patient-initiated humour is a further resource for building empathy
(Eggins 2014). Whereas clinical discourse can stigmatize patients or treat
them as ‘non-persons’ (Goffman 1963), patient-initiated humour encour-
ages clinicians to depart from the ‘professional’ script and use more inclu-
sive, egalitarian modes of everyday interaction. Eggins’ subjects were
hospital inpatients, but her findings probably extend to primary care set-
tings. Patient-initiated humour seems to differ in function from humour
between clinicians, which can promote solidarity or enforce existing hier-
archies (see Eggins and Slade 2015).
25.2.2.3 Oncology
Several studies have used SFL to explore clinical interaction in oncology
settings, including breast cancer (Lobb et al. 2006; Moore and Butt 2004;
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 657
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
658 A L I SO N R OT H A M O O R E
25.2.2.6 Nursing
Several studies discussed here use SFL tools to analyze nurse–patient inter-
action (e.g. Chandler et al. 2015; Eggins et al. 2016; Kealley et al. 2004;
Kealley 2007; Slade et al. 2015a; Wyer et al. 2017). Additionally, Candlin
(2000, 2002) uses Hasan’s cline of dynamism and coins the term ‘compre-
hensive coherence’ to describe how superficially ‘casual’ conversations
with patients constitute professional nursing expertise. Lassen and
Strunck (2011) show how nurses invoke a ‘positive’ discourse (Martin and
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 659
2
Women attending FP clinics are called ‘clients’, which is arguably consistent with values of patient/client autonomy and
feminism that inform sexual and reproductive healthcare.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
660 A L I SO N R OT H A M O O R E
Cohesion and cohesive harmony (Hasan 1984) have been key to this work,
with Butt et al. (2010) exploring the semantic fragmentation and fusion
that characterizes dissociative episodes among patients with BPD,
and tracking their possible resolution in therapeutic interaction. Butt’s
linguistic concepts of motivated selection, semantic drift, and instantial
weight have been deployed, and analogies between psychotherapy, verbal
art, and science are drawn (Butt et al. 2013). The linguistic notion of
cohesion has become a central metaphor in the Conversational Model’s
theory of BPD and how it arises (Meares et al. 2013).
Khoo (2013) takes a closer look at cohesion in her study of psychothera-
peutic discourse, including the relative merits of quantitative and qualita-
tive analyses, and the iconicity of cohesive harmony. She gives examples of
texts with poor numerical measures of cohesion that are judged more
therapeutically valuable than others with high scores (Khoo 2016).
Using a multi-stratal approach and fine-grained analyses of agency and
appraisal, Henderson-Brooks (2006a, 2006b) examines claims about three
conversation types observed in consultations with patients with BPD.
These linguistically distinguishable text types represent shifts between an
alienated or truncated self, construed through negative capacity, little
agency, and ineffectual verbal action (Chronicles and Scripts), and an
expanded self, construed through features such as real and hypothetical
action on others, positive mental action and verbal action (Narratives).
Other clinical concepts such as the contrastive ‘linear/non-linear speech’
are associated with logico-semantic complexity.
The appraisal system has also been used to explore the extent and nature
of depression in hospitalized patients via their discourse semantics. Using
interview data, Tebble (2012) concluded that familiarizing clinical
staff with key appraisal systems could help identify undiagnosed depres-
sion among inpatients, with the aim of improving their treatment experi-
ence, prognosis, and quality of life. Related research by Caldwell et al.
(2006) reports on appraisals of well-being among a non-depressed compari-
son group.
Korner (2015) draws on SFL and anthropomorphic measurement (heart
rate, skin conductivity, etc.) to examine ‘self’ and ‘person’ as the embodied
flux of feeling in a symbolic, acculturated personal context, or what he calls
a system of self and other in psychotherapeutic discourse.
‘Formulation’ (synopses of a patient’s presenting condition) is another
aspect of psychiatry studied using SFL tools. Formulations produced within
intrapsychic models have been found to be more highly nominalized
than those produced within intersubjective models, reducing the sense of
patient agency, and representing patients as being influenced by ‘unseen’
forces (Korner et al. 2010). Walsh et al. (2016a) address the need for teach-
ing the genre of formulation to mental health professionals. They examine
lexical relations, nominalization, and conjunctions, showing how clin-
icians’ talk shapes their developing understanding into a logical
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 661
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
662 A L I SO N R OT H A M O O R E
Greetings^Introductions^(Contract)^Stating/Eliciting Problem^Ascertaining
Facts^(Diagnosing Facts)^Stating Resolution/Exposition^(Decision by Client)
^Clarifying Residual Matters^Conclusion^Farewell
Such a model allows interpreters to map their ‘location’ and pace their
energy around the critical parts of the consultation.
Interpersonal meanings are a focus in interpreting, and Tebble (1999)
theorizes the teaching of interpreters to ‘read’ the tenor of physicians’
consulting styles, using ‘Exposition’ moves in two specialisms. Appraisal
has been deployed (Willis 2001; Hirsh 2001) including studies of depressed
patients (e.g. Tebble 2012). A German study (Bührig 2004) highlights the
textual function, showing how a doctor and an untrained interpreter used
different ‘linguistic action patterns’ for obtaining informed consent (see
also Torsello 1997).
3
Tebble extends her model to ‘dialogue interpreting’ including legal and bureacratic contexts.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 663
so they can control tenor as field varies within texts, using an authoritarian
tone when necessary (Do not take this medicine if you are pregnant), but not
across all sections. Compared with CMIs from other English-speaking
nations, Australian CMIs showed disrupted cohesive harmony (Moore and
Wegener 2010; Moore 2010). This is partly because writers assemble CMIs
from pre-written paragraphs, but the implicit goals of such texts constitute
another factor; these goals include protecting drug companies (who develop
such documents) from legal harm. While their findings informed docu-
ment redesign (Aslani et al. 2010), one aspect considered too controversial
to include in the national recommendations was the observation that
patients/readers interpreted statements of drug purpose (e.g. Lipitor is used
in people with high cholesterol who have high blood pressure and coronary heart
disease or are at risk of a stroke . . .) as statements of likely benefit, thus
overestimating the chance of personal health benefit (Moore 2010).
Written information can also be about healthcare processes. Kealley et al.
(2004) examine a pamphlet aimed at empowering patients and relatives in a
critical care unit to be active in the healthcare process. Contrary to its
purpose, the pamphlet depicts staff as retaining great authority in a way
that restricted relatives’ actions and interactions, thus reinforcing passive
and compliant behaviour among relatives and patients. This study is one of
the few that does not assume the neutrality of ‘information’ for patients
and relatives and uses linguistic concepts to explore its value.
Eckkrammer (2004) examines medical self-counselling texts and hyper-
texts, showing that layers of intersemiosis were already present in late
fifteenth-century texts, and discussing the specific affordances of hypertext
for this register. One finding of interest is that most diagrams in self-
counselling texts had an illustrative function only.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
664 A L I SO N R OT H A M O O R E
linguistic research has been done on this, but see Bloor (2016) on potentially
misleading construals of terminal illness in online prognostic information.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 665
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
666 A L I SO N R OT H A M O O R E
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 667
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
668 A L I SO N R OT H A M O O R E
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 669
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
670 A L I SO N R OT H A M O O R E
Who are you? ^ What is a gene? ^ [What forecast? (^Go back and talk)]
4
‘Framing effects’ for risk statistics (Tversky and Kahneman 1981) are, however, well known within health
communication research.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 671
• The rationale for this move is posed by the counsellor rather than the
patient5 (which we could perhaps gloss as you need to understand ‘inherited
tendency’ in genetic terms);
• The move comes as a long monologic bloc (punctuated by patient
backchannelling);
• The conceptual sequence of the move focuses on biological processes
rather than, say, social or statistical groupings;
• The move draws heavily on naturalizing metaphors (instructions and
brakes, whose functions are commonly known) but does not de-automatize
or limit these metaphors; importantly brake failure in a car invariably
generates some observable problem, preferably a gentle roll but often a
fatal crash, whereas carrying BrCa1 or BrCa2 mutations only leads to
breast cancer in approximately 40 to 85 per cent of women, and to early
death only in a proportion of those, so the analogy is problematic;
• Semantic tendencies in the spoken move are reinforced by the images
used (see below).
At the level of abstraction below ‘Move’ (see Figure 25.3 – roughly the
stratum/rank of Butt’s (2000) ‘Argument’, Halliday and Matthiessen’s
(1999) ‘Sequence’, and Cloran’s (1994) ‘Rhetorical Unit’), the most crucial
features are as follows.
5
The woman in this consult is not strictly speaking a patient, but this term is used for clarity rather than ‘woman’ or
‘woman from a high breast cancer risk family’, etc.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Network I: Key options in the Genes Talkfest Move (Phase 2 in Breast Cancer Genetic Counselling GSP)
David Butt and Alison Moore, CLSL, Macquarie University 2001
Rationale for what is coming:
Why you need to know/hear this Posed by Patient
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Social elements (family, community, "you" "us")
Conceptual Taxonomic
Sequence Biological elements (cells, genes, etc.)
Biological process
Naturalise analogy only (e.g., a gene is a little bit like a brake; lightning doesn't strike twice)
Pincer De-automatize - make strange (e.g., actually lightning can sometimes strike twice)
Figure 25.2 Network of key contextual options in Phase 2 of genetic counselling (after Moore and Butt 2004)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
Figure 25.3 Network fragment: key semantic options in ‘rational strategy’ within Phase 2 of genetic counselling (after Moore and Butt 2004)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
674 A L I SO N R OT H A M O O R E
Figure 25.4a Original image used in decision aid for genetic counselling (after Lobb
et al. 2002)
Figure 25.4b Revised image used in decision aid for genetic counselling (after Lobb
et al. 2002)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 675
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
676 A L I SO N R OT H A M O O R E
References
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 677
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
678 A L I SO N R OT H A M O O R E
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 679
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
680 A L I SO N R OT H A M O O R E
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 681
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
682 A L I SO N R OT H A M O O R E
Iedema, R., ed. 2007. The Discourses of Hospital Communication: Tracing Complex-
ities in Contemporary Health Organisations. Basingstoke: Palgrave Macmillan.
Iedema, R., N. Mallock, R. Sorensen, E. Manias, A. Tuckett, A. Williams,
et al. 2008. Final Report: Evaluation of the National Open Disclosure Pilot
Program (Report for the Australian Commission on Safety and Quality in Health
Care). Sydney: University of Technology, Sydney.
Iedema, R., C. Jorm, J. Wakefield, C. Ryan, and S. Dunn. 2009. Practising
Open Disclosure: Clinical Communication and Systems Improvement.
Sociology of Health and Illness 31(2): 262–77.
Iedema, R., S. Allen, K. Britton, D. Piper, A. Baker, C. Grbich, A. Allan,
L. Jones, A. Tuckett, A. Williams, E. Manias, and T. Gallagher. 2011.
Patients’ and Family Members’ Views on How Clinicians Enact and How
They Should Enact Incident Disclosure: The ‘100 Patient Stories’ Qualita-
tive Study. BMJ 343: d4423.
Iedema, R., C. Ball, B. Daly, J. Young, T. Green, P. Middleton, C. Foster-Curry,
M. Jones, S. Hoy, and D. Comerford. 2012. Design and Evaluation of a New
Ambulance-to-Emergency Department Handover Protocol: ‘IMIST-
AMBO’. BMJ Quality and Safety 21(8): 627–33.
Jordens, C. 2002. Reading Spoken Stories for Values: A Discursive Study of Cancer
Survivors and Their Professional Carers. PhD Thesis, University of Sydney.
Jordens, C., J. Little, K. Paul, and E. Sayers. 2001. Life Disruption and Generic
Complexity: A Social Linguistic Analysis of Narratives of Cancer Illness.
Social Science and Medicine 53(9): 1227–36.
Jordens, C., B. Morrell, P. Harnett, et al. 2010. Cancergazing? CA125 and
Post-treatment Surveillance in Advanced Ovarian Cancer. Social Science
and Medicine 71(9): 1548–56.
Kaplan, S. H., S. Greenfield, and J. E. Ware. 1989. Assessing the Effects
of Physician–Patient Interactions on the Outcome of Chronic Disease.
Medical Care 27: 110–27.
Kappagoda, A. 2004. Semiosis as the Sixth Sense: Theorising the Unperceived in
Ancient Greek. PhD Thesis, Macquarie University.
Karimi, N., A. R. Moore, A. Lukin, A. Walczak, and P. Butow. 2018. Life
and Death: Consultations between Oncologists and Patients with
Advanced, Incurable Cancer. In A. Sellami-Baklouti and L. Fontaine,
eds., Perspectives from Systemic Functional Linguistics. London: Routledge.
315–37.
Kealley, D. J. 2007. ‘I Can’t Find a Pulse But That’s OK’: Nursing in Context:
A Systemic Functional Linguistic Examination of Nursing Practice. PhD Thesis,
University of South Australia.
Kealley, D. J., C. Smith, and B. Winser. 2004. Information Empowers
But Who Is Empowered? Communication and Medicine 1(2): 119–29.
Khoo, K. 2013. Cohesive Harmony: Exploring a Measure of Cohesion and Coherence
in Psychotherapy. BA Honours Thesis, Macquarie University.
Khoo, K. 2016. ‘Threads of Continuity’ and Interaction: Coherence, Texture
and Cohesive Harmony. In W. Bowcher and J. Liang, eds., Society in
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 683
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
684 A L I SO N R OT H A M O O R E
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 685
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
686 A L I SO N R OT H A M O O R E
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 687
eds., Talk, Work and the Institutional Order: Discourse in Medical, Mediation and
Management Settings. Berlin: Mouton de Gruyter. 1–57.
Sheehan, A. and W. Bowcher. 2017. Messages to New Mothers: An Analysis
of Breast Pump Advertisements. Maternal and Child Nutrition 13(2): 1–13.
Sinclair, J. and M. Coulthard. 1975. Towards an Analysis of Discourse. Oxford:
Oxford University Press.
Slade, D. and S. Eggins. 2015. Communication in Clinical Handover:
Improving the Safety and Quality of the Patient Experience. Journal of
Public Health Research 4: 666.
Slade, D., H. Scheeres, M. Manidis, C. M. I. M. Matthiessen, R. Iedema,
M. Herke, J. McGregor, R. Dunston, and J. Stein-Parbury. 2008. Emer-
gency Communication: The Discursive Challenges Facing Emergency
Clinicians and Patients in Hospital Emergency Departments. Discourse
and Communication 2(3): 289–316.
Slade, D., H. Scheeres, H. de Silva-Joyce, J. McGregor, N. Stanton, M. Herke,
E. Weisberg, and D. Bateson. 2009. Developing Effective Communication
between Doctors and Clients: Sexual and Reproductive Health Consultations.
Research Report to Family Planning NSW.
Slade, D., E. Chandler, J. Pun, M. Lam, C. M. I. M. Matthiessen, G. Williams,
E. Espindola, F. Veloso, K. Tsui, S. Tang, and K. Tang. 2015a.
Effective Clinician–Patient Communication in Hong Kong Accident and
Emergency Departments. Hong Kong Journal of Emergency Medicine 22(2):
69–83.
Slade, D., M. Manidis, J. McGregor, H. Scheeres, E. Chandler, J. Stein-
Parbury, R. Dunston, M. Herke, and C. M. I. M. Matthiessen. 2015b.
Communication in Hospital Emergency Departments. Berlin: Springer.
Smith, S. K., L. Trevena, D. Nutbeam, A. Barratt, and K. J. McCaffery. 2008.
Information Needs and Preferences of Low and High Literacy Consumers
for Decisions about Colorectal Cancer Screening: Utilising a Linguistic
Model. Health Expectations 11(2): 123–36.
Sontag, S. 1978. Illness as a Metaphor. New York: Farrar, Straus and Giroux.
Sorensen, R. and R. Iedema. 2006. Integrating Patients’ Nonmedical Status
in End-of-life Decision Making: Structuring Communication through
‘Conferencing’. Communication and Medicine 3: 185–96.
Stenglin, M. and M. Foureur. 2013. Designing Out the Fear Cascade to
Increase the Likelihood of Normal Birth. Midwifery 29(8): 819–25.
Street, R. L., G. Makoul, N. K. Arora, and R. M. Epstein. 2009. How
Does Communication Heal? Pathways Linking Clinician–Patient
Communication to Health Outcomes. Patient Education and Counseling 74:
295–301.
Tebble, H. 1993. A Discourse Model for Dialogue Interpreting. In Proceedings
of the First Practitioners’ Seminar of AUSIT, November 1992. Canberra: NAATI.
1–26.
Tebble, H. 1996a. Research into Tenor in Medical Interpreting. Interpreting
Research 6: 33–45.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
688 A L I SO N R OT H A M O O R E
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
Language and Medicine 689
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:08:57, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.027
26
Language and Literature
Donna R. Miller
26.1 Introduction
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 691
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
692 DONNA R. MILLER
focus is seen as being on the message for its own sake (Jakobson 1960).
The decade 1929 to 1939 marked the golden years of the Prague School.
Jakobson was collaborating with the Prague Structuralists, notably
Mukařovský, and the two were aiming at identifying the formal and func-
tional linguistic mechanisms responsible for articulating the aesthetic
motivation they both saw as the impetus of literature. They worked
together towards the theorization of what would become seminal and
durable stylistic concepts such as deviation – which Mukařovský saw as
fundamental to the creation of the defamiliarizing effect also central to
literary language – and grammatical parallelism (henceforth GP) – the
structural patterning in texts that Jakobson theorized as the empirical
linguistic evidence of his ‘poetic function’ and which he proposed also
generated defamiliarization. Mukařovský (1977, 1978) moved on to
hypothesizing the role of foregrounding, and Jakobson (1966) to that of
pervasive parallelism, but theoretical intersection endured, despite the
undoubtedly greater ‘success’ of foregrounding – not only in SSS, but in
stylistics tout court. In addition to Halliday and Hasan, mainstream stylisti-
cians in the last decades of the twentieth century, such as Leech and Short
(2007), were convincingly demonstrating the essential role of foreground-
ing (or in the Mukařovský-esque terms preferred by Halliday (2002a:131),
the de-automatization of grammar) in literary interpretation. And fore-
grounding has even been related directly to the concepts of figure and
ground in recent work in cognitive stylistics. Its mileage is impressive,
but perhaps Jakobson’s reach was just as significant, if not as wide-ranging,
as I hope to show.
The Second World War meant a close to Jakobson’s collaboration with
the Prague Structuralists. For a while he was an itinerant scholar, settling
in the USA in 1941, a move that proved decisive for the subsequent spread
of his ideas in America but also in Europe, where work on style in literature
was of course also being done – chiefly by the Austrian philologist, Spitzer,
on the literature of the Romance languages, privileging ‘objective’ over
‘impressionistic’ methods (see Wales 2001:296–7), as well as by scholars
like Auerbach, Bally, and Guiraud, whose work would then impact on the
development of the French analyse de texte.
26.2.2 Moving On
Jakobson’s move to the USA was also significant to the development, in
America and Britain respectively, of the New Criticism and Practical Criti-
cism movements, both of which employed techniques of the ‘close reading’
of literature, a practice which was widespread for decades (1930s–1960s).
The first of these focused on the description of the aesthetic qualities of a
literary text, while the latter, developed by I. A. Richards, engaged with the
psychological aspects of how readers comprehend texts – what the first
school called the ‘affective fallacy’ (Wimsatt and Beardsley 1949). Richard’s
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 693
1
These are intricate Marxist-based critical theory issues that are brilliantly elucidated from a diachronic perspective by a
semi-anonymous author, a certain ‘Paul’, available online at: http://herrnaphta.wordpress.com/2011/01/09/
reification-and-american-literature. (Last accessed 14/09/2017.)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
694 DONNA R. MILLER
including Leech and Short (2007) and Toolan (2001). But Hasan’s framework
is always, conspicuously, ‘missing’. Undoubtedly this has something to do
with Halliday being simply harder to ignore, but also, and primarily, with
the ease with which certain aspects of the SFL model can be integrated into
a non-SFL stylistics text book, without, that is, needing to deal with a more
unwieldy, holistic model of the kind Hasan proposes. But Fowler warrants
further attention.
In what lies the opposition? Surely not in their respective views re the
importance of culture and community. Indeed, Hasan says that ‘the litera-
ture text . . . embodies precisely the kind of “truths” that most communities
are deeply concerned with’ (Hasan 1989:100). She also recognizes that the
social impacts on verbal art: indeed ‘perhaps the most critical part it plays is
in the shaping of the ideological orientations of those who write and those
who read literature’ (Hasan 2007:25). And a text’s endurance as art will
always hinge on the value which is awarded it by successive generations of
readers: ‘The challenge for the creator of verbal art is that the symbolically
articulated Theme has to be capable of striking a chord in the reader over
substantial distances in time and space’ (Hasan 2007:25). But in spite of
these correspondences, the sum and substance is a clash in standpoints.
Hasan does not invest the power for ultimately deciding what is or is not
literature in the community, whether it be that of the time and place of the
text’s creation or at a further semiotic social distance (Hasan 2007:34). And,
ultimately, as we will see below, her definition is eminently ‘essentialist’.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 695
It is not that there is art [somewhere ‘out there’, so to speak], and the job of
language is simply to express it; rather it is that, if there is art, it is because of
how language functions in the text . . . in verbal art the role of language is
central. Here language is not as clothing to the body; it is the body.
(Hasan 1989:91; also see Hasan 1964, 1971)
In 1982, Halliday drolly makes the point for the somehow ‘different’ nature
of literature as text with the quip: ‘the paradox of “poetic” language [is] that
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
696 DONNA R. MILLER
there is no such thing . . . but we can all recognize it when we see it’ (Halliday
2002a:134, emphasis added).
In 2002, already engaging with Hasan’s work, Halliday (2002b) draws a
parallel between Mukařovský’s foregrounding and the motivated ‘promin-
ence’ of grammatical features. He suggests two levels of meaning, both
grammatically realized, but one ‘underlying’ and ‘deeper’ than the first,
‘immediate’ level, also glossed as ‘subject matter’ (2002b:118–20 et passim).
The deeper semantic meanings are said, significantly, ‘to serve a vision of
things . . . The vision provides the motivation for their prominence’. It will
be primarily Hasan’s task in the course of the 1970s to better delineate this
‘vision’.
In Halliday (2002a:131, emphasis added) the term foregrounding, as
noted above, is explicitly replaced by ‘de-automatization’, conceivably with
a view to highlighting the distinct if complementary roles of these two
semantic levels from an SFL perspective:
What is in question is not simply prominence but rather the partial freeing of
the lower-level systems from the control of the semantics so that they become domains
of choice in their own right. In terms of systemic theory the de-automatization
of the grammar means that grammatical choices are not simply determined
from above: there is selection and pre-selection. Hence the wording becomes
a quasi-independent semiotic mode through which the meanings of the
work can be projected.
I have presented this vital thought of Halliday’s here since the two distinct
semiotic modes of literature meaning-making he hypothesizes can with
impunity be said to be analogous to the scaffolding of the doubly articu-
lated framework for the analysis of verbal art which Hasan began to model
in 1971, and which will be properly described and illustrated below.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 697
(1) for SSS, literature is a ‘special’ kind of text, requiring a special type of
descriptive and analytical framework for its study, which we will take
a close look at below;
(2) the two-tiered approach should not, indeed cannot, be applied to the
investigation of other text types: rather than extending its brief to any
and all registers, as mainstream stylisticians (see Simpson 2014) tend
more and more to do, it sees literature as its sole object of study;
(3) the approach ought not to be promiscuous; i.e. it should remain
unadulterated; it demands exclusivity, philandering with incompatible
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
698 DONNA R. MILLER
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 699
His monoglossically stated contentions here are fully consistent with his
position on the practice of stylistics needing to extend to any and all text
types. But do they hold up? One might suggest that to maintain the
distinctiveness of literature, as SSS does, is not to deal with clichés, nor to
apply automatic strictures to artistic expression, nor to necessarily argue
for a ‘literary’ register, not least because, as we will see below, the starting
point for the analysis of verbal art is identical to that for any other text
(Hasan 1989:92). Moreover, Hasan, in suggesting a way to distinguish
between what is verbal art and what, most likely, is not, actually pits
‘literary’ against ‘literature’:
If the patterning [foregrounding] of patterns is consistently utilised for a
second-order semiosis . . . then the text in question is a literature text. If,
however, such a role is not played by the patternings, then we have a literary
text. The recognition of this distinction is important, not least because the
techniques for the study and evaluation of the two are not identical.
(Hasan 1989:101)
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
700 DONNA R. MILLER
The words I have italicized in the segment, however, quite clearly tell us
that there is more to the analysis of literature than there is to that of other
uses of language. And both the first level of analysis, along with that ‘more’,
can be visually grasped straightaway in Figure 26.1 below.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 701
A full account of the figure can be had elsewhere (primarily in Hasan 1989,
with synopses in Miller 2010, 2012, 2013, 2016a), but we can easily recognize
the preliminary level, what is labelled the semiotic system of language, as the
multiple coding system which, from the SFL perspective, is the valid entry
point for the analysis of any text. The system itself is made up of strata, each
of which is realized, i.e. becomes accessible to us, through the one below it. In
short, semantics, or meanings, are realized in lexicogrammar, or wordings,
which become accessible to us in phonology, or soundings, or, in the case of a
written text, in graphology, in written symbols.
But SSS is not content with merely identifying isolated patterns of lan-
guage in verbal art and the meanings they reveal; rather, it insists that such
patternings be demonstrated, at a higher order, to be ‘significant’. This
involves applying Mukařovský’s (1964) notion of ‘foregrounding’ as ‘motiv-
ated’ contrast (also see Halliday’s (2002a:131) concept of the ‘de-automatiza-
tion of grammar’); it involves a second order of meaning, where first-order
meanings are repatterned; it involves this model of what Hasan has called
‘double-articulation’ (Hasan 2007).
The higher, additional level is the semiotic system of verbal art. Its first
stratum, Verbalization, includes all of the lower system and builds upon it –
hence the broken line. This it does by means of the ‘symbolic articulation’
of the ‘Theme’. Looking at it from ‘below’, symbolic articulation is the place
where the basic first-order meanings are expanded upon, or ‘enriched, or
‘deepened’, and are made ‘art’. How it accomplishes this for Hasan is
through foregrounding (or the patterning of patterns). This is the process
by which the meanings emerging through analysis of the semiotic system
of language are symbolically turned into signs, for the purpose, the aes-
thetic motive, or ‘artistic intention’ (Mukařovský 1977) of expressing a
Theme: ‘what the text is about when dissociated from the particularities
of that text. . . . very close to a generalisation, which can be viewed as a
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
702 DONNA R. MILLER
hypothesis about some aspect of the life of social [wo]man’ (Hasan 1989:97).
Perhaps, as Hasan (1989:97) also notes, this is akin to what Aristotle had ‘in
mind when he declared that art is truer than history’. Be that as it may,
without this ‘Theme’, this reflection on humanity, and its symbolic articu-
lation, there is, for Hasan, no verbal art. I now aim at illustrating the
application of the framework just sketched.
I will now offer a brief sample of SSS analysis of the functions of fore-
grounded patterning in a short poem by D.H. Lawrence (see Miller 2007).
The text is the following:
I begin with the semiotic system of language, and, in the stratum of lexico-
grammar – the wordings instantiated in the poem that make its semantics
accessible to us – with transitivity. Table 26.1 shows how the –er roles, or
Doers, in the poem are divided between animate somebody and poetic voice I,
on one side, and inanimate natural phenomena, trees, moonlight, and the living
cosmos, on the other. People/human conversation are present, but substantially
construed as embedded Qualifiers, and so downranked grammatically and
demoted as participants (see Berry, this volume; Fontaine and Schönthal, this
volume) – e.g. the petrol fumes (of human conversation).
But what do the Doers do? In Stanza 1, a no better identified somebody
complain[s] of being lonely or, in American, lonesome. Apparently it is a
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 703
Table 26.1 Taxis, –er roles, Verbal Groups and process types2
1 I hear ment:perc.
embedded somebody complain verbal
2 I wonder and wonder mental:cog.
embedded they mean verbal
3 they mean verbal
3a they are a great deal alone relational:att.
4 what is lovelier than ”
embedded - to be alone ”
” - escaping material
” - be alone relational:att.
” - be alone ”
” - feel mental:perc.
” trees growing behavioural
” - be alone relational:att.
” - see mental:perc.
” moonlight [that is] outside, white and busy and relational:1 circ.
silent + 3 att.
” - be quite alone relational:att.
” - feel mental:perc.
” living [that is] softly rocking, soothing, behavioural
cosmos restoring, healing
5 elliptical [I] [am] Soothed, restored, healed relational:att.
5a I am alone ”
5b there is no grating of people… existential
embedded presences gnawing material
reiterative verbal process that I hear[s] now and again and which causes
him/her, somewhat surprisingly, to wonder and wonder what they mean, with I
as perplexed Senser of the queried embedded Phenomenon with they as
Sayers.
Line 4 stands alone graphically; it consists of a rather ingenuous rhet-
orical question, once again concerning they’s meaning, one that functions to
highlight the response the poetic voice proceeds to provide in the second
rhetorical question (Line 5) and then to elaborate on in the rest of the poem.
Somebody/they then exits from the poem, and the I too fades, to reappear
however, symmetrically, in the poem’s final stanza.
Relational processes enter the text with the second interrogative and
reign uninterruptedly through Stanza 3 Line 9, excepting the non-finite
material Process, escaping, in Line 6. What is being escaped affords the first
appearance of the demoted participants signalled above: the petrol fumes [of
human conversation] and the exhaust smell [of people] – more fittingly com-
mented in terms of evaluation, below.
2
I read VG mean as agnate to want to say, a VGC with a desiderative α verb and verbal β verb.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
704 DONNA R. MILLER
Even more crucially, along with the interrogatives, the Attribute alone
materializes as well. Its seven instances dominate the poem from Line
4 through Line 15, four times as Complement to the non-finite base
form of be. Real time withdraws as the relationals continue and embedd-
edness multiplies: there are no finite verbs from Line 6 to Line 15 and
the present participle assumes command. However, even finites in the
poem (first three stanzas and last) are not typical simple presents or
present-in-presents but rather whenever presents – spanning past/present/
future.
The three mental processes of perception in Stanza 4, also in base form,
likewise work to construe a sense of timelessness: feel the trees silently
growing; see the moonlight outside, white and busy and silent; feel the living cosmos
softly rocking, soothing, restoring, healing. Trees are earthly, natural phenom-
ena, though not human, while moonlight, and even more so the living cosmos,
are preternatural, uncanny Doers. I will take up the special salience of this
fourth stanza again below.
The final stanza – an incomplete clause complex, unless one provides the
‘Subject + Finite’ in Line 14 – reproposes the doings of the cosmos, now
Attributes of the symmetrically reintroduced I. The enhancing temporal
clause also links back symmetrically to that in the first stanza. The existen-
tial clause concluding the poem eliminates the negative presence of the
already established antagonist, people, again grammatically downgraded to
Qualifier, the negated –er role assumed by their presences, injuriously
gnawing.
But it is time to take a closer look at the evaluation enacted in the poem,
from the perspective of appraisal systems (Martin and White 2005; also
see Hood, this volume). Briefly, and appropriating Thompson’s (2014:50)
neat summary:
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 705
insincerity that those who know Lawrence’s work would never assume.
Regarding being alone, we can immediately posit +ve judgement with impun-
ity. Yet familiarity with the writer is hardly requisite, as the question itself,
taken seriously, implies such evaluation. In addition, the fast-accumulating
supporting evidence compounds and fully endorses this reading.
Non-Finites are of course ambiguous in their logico-semantic relations,
and the indistinctness here before ‘escaping’ is reinforced by the non-finite
base form be in Line 8. I hypothesize a Cause:Purpose enhancing relation
such as so that [you’re escaping/you can escape]; thus escaping is also being
evaluated in terms of +ve judgement and may also be seen as an indirect
expression of the +ve judgement of being alone, and an illustration of
what Thompson (2014) has called the ‘Russian doll dilemma’ in appraisal
analysis.3 As one typically expects, what is escaped is inherently negative in
assessment: the noxious car odours. However, the human Qualifiers con-
strue a paradox that enacts further -ve appreciation; this evaluation too
feeds back into the overriding +ve judgement of being alone.
The Cause:Purpose interpretation of escaping with reference to being
alone above is not unlike the meanings of Stanza 4’s looser paratactical
and linking the base forms of be with those of the Processes of percep-
tion. Being alone affords awareness of those uncanny personified par-
ticipants: +ve appreciation is enacted of moonlight outside, white and busy
and . . . silent; whereas I read +ve judgement of those more patent doings
of this inhuman but living cosmos softly rocking, soothing, restoring, healing.
And these assessments once again token the overall +ve judgement of
being alone.
In the first line of the fifth and final stanza these present participles
become past, and function as Attributes of (once again explicit) I, now
personally alone with the silent great cosmos, which as Thing is explicitly
evaluated with +ve appreciation – and which is a further token for the
global +ve judgement of being alone. The final Lines assert the now in-
Existent grating of people with their presences gnawing at the stillness of the air,
with negative soundings enhancing their meanings. Enacted here is
another -ve judgement on the behaviour of people and their presences, one
which again indirectly expresses that overarching +ve judgement of
being alone. The dolls might also be visually represented as in (my admit-
tedly over-simplified) Figure 26.2.
At this point a first formulation of the Theme of the poem can be
advanced. Solitude as bliss, diametrically opposed, or contratextual (Martin
3
As Thompson puts it (2014:49), ‘This relates to the way in which an expression of one category of attitude may
function as a token (an indirect expression) of a different category; and that token may itself function as an indirect
expression of yet another category, and so on. This raises the question of how many of these layers, one inside the
other, should be included in an analysis – and how to code the different layers.’ This ‘refinement’ of the systems is
remarked here, as the poem presents a noteworthy case in point but, given the not-necessarily expert-in-SFL nature of
our audience, is not analyzed in-depth.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
706 DONNA R. MILLER
BEING ALONE I -soothed, restored, healed -alone with the silent great
cosmos
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 707
In point of style, fault is often found with the continual, slightly modified
repetition. The only answer is that it is natural to the author; and that every
natural crisis in emotion or passion or understanding comes from this
pulsing, frictional to-and-fro, which works up to culmination.
The final stanza slows things down again, and winds things up.
In order to test the validity of the Theme and complete its formulation,
the three ingredients of the context of creation of verbal art need to be
probed. We have just now glossed the first of these: the language of its
writer. His position re the artistic conventions of his time can be said to be
‘contra’ as well: beginning as a Georgian poet, he fast rebelled against
allowing any formal strictures to his art. With reference to the worldview
of Lawrence, we have seen his contratextual stance re solitude, but the
poem gives us more.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
708 DONNA R. MILLER
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 709
findings. Here too, incidentally, the name of Hasan only appears linked
to that of Halliday, and in terms of cohesion.
The essential query is whether foregrounding – at least at a preliminary
investigative stage – is in fact quantifiable. We recall Halliday’s (2002b:102–3)
own explicit caveat that prominence as motivated foregrounding is not
tantamount to mere statistical frequency. Also to be recollected is that auto-
matic analyses cannot give us the findings we need at the ‘higher’ levels of
semantics and context (Halliday and Matthiessen 2004:48–9) – no matter
what the text type. Yet also to be recalled is that Halliday did see ‘counting’
the linguistic choices of a writer as a step towards establishing the potential
prominence of patterns: for determining what features deserve further inves-
tigation re motivation. Indeed, as Miller and Luporini (2015, 2018) suggest,
the extent to which foregrounding is quantifiable would seem to be inscrut-
able without the assistance of corpus linguistic methods. But their findings
show that it is not enough; it plays but a supporting role, guaranteeing data
accuracy and statistical significance that cannot be manually achieved in
longer texts, but by no means supplanting the labour-intensive manual
analysis of co-textual logogenesis that SSS requires.4
26.6 In Closing
4
Re the accuracy of CL findings, one needs to bear in mind that in relying excessively on frequency, one risks missing
significant non-frequent foregrounded items, and so vital parts of the full picture. There are those who imagine a brave
new digital world in which the future powers of algorithms would make Fillmore’s (1992) armchair linguist redundant.
I am not among them.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
710 DONNA R. MILLER
References
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 711
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
712 DONNA R. MILLER
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
Language and Literature 713
Miller, D. R. 2010. The Hasanian Framework for the Study of ‘Verbal Art’
Revisited . . . and Reproposed. In J. Douthwaite and K. Wales, eds., Stylistics
and Co. (Unlimited): The Range, Methods and Applications of Stylistics. Textus 23
(1): 71–94.
Miller, D. R. 2012. Slotting Jakobson into the Social Semiotic Approach to
‘Verbal Art’: A Modest Proposal. In F. Dalziel, S. Gesuato, and M. T.
Musacchio, eds., A Lifetime of English Studies. Essays in Honour of Carol Taylor
Torsello. Padua: Il Poligrafo. 215–26.
Miller, D. R. 2013. Another Look at Social Semiotic Stylistics: Coupling
Hasan’s ‘Verbal Art’ Framework with ‘the Mukařovský-Jakobson Theory’.
In C. A. M. Gouveia and M. F. Alexandre, eds., Languages, Metalanguages,
Modalities, Cultures: Functional and Socio-discoursive Perspectives. Lisbon:
BonD. 121–40.
Miller, D. R. 2016a. Jakobson’s Place in Hasan’s Social Semiotic Stylistics:
‘Pervasive Parallelism’ as Symbolic Articulation of Theme. In W. L. Bow-
cher and J. Y. Liang, eds., Society in Language, Language in Society: Essays in
Honour of Ruqaiya Hasan. London: Palgrave Macmillan.
Miller, D. R. 2016b. On Negotiating the Hurdles of Corpus-assisted Appraisal
Analysis in Verbal Art. In S. Gardner and S. Alsop, eds., Systemic Functional
Linguistics in the Digital Age. Sheffield: Equinox. 211–28.
Miller, D. R. and A. Luporini. 2015. Social Semiotic Stylistics and the Corpus:
How Do-able is an Automated Analysis of Verbal Art? In A. Duguid,
A. Marchi, A. Partington, and C. Taylor, eds., Gentle Obsessions: Literature,
Linguistics and Learning: In Honour of John Morley. Roma: Artemide Edizioni.
235–50.
Miller, D. R. and A. Luporini. 2018. Software-assisted Systemic Socio-
Semantic Stylistics: Appraising tru* in J.M. Coetzee’s Foe. In R. Wegener,
A. Oesterle, and R. Neumann, eds., On Verbal Art: Essays in Honour of
Ruqaiya Hasan. Sheffield: Equinox.
Miller, D. R. and M. Turci, eds. 2007. Language and Verbal Art Revisited:
Linguistic Approaches to the Study of Literature. Sheffield: Equinox.
Mukařovský, J. 1964. Standard Language and Poetic Language. In P. L. Garvin,
ed. and trans., A Prague School Reader on Aesthetics, Literary Structure and Style.
Washington, DC: Georgetown University Press. 17–30.
Mukařovský, J. 1977. The Word and Verbal Art: Selected Essays by Jan Mukař-
ovský. Edited and translated by J. Burbank and P. Steiner. New Haven:
Yale University Press.
Mukařovský, J. 1978. Structure, Sign and Function: Selected Essays by Jan Mukař-
ovský. Edited and translated by J. Burbank and P. Steiner. New Haven:
Yale University Press.
Selden, R. 1989. The Cambridge History of Literary Criticism, Vol. 8: From Formal-
ism to Poststructuralism. Cambridge: Cambridge University Press.
Simpson, P. 2014. Stylistics. 2nd ed. London: Routledge.
Stankiewicz, E. 1983. Roman Jakobson, Teacher and Scholar. In A Tribute to
Roman Jakobson 1896–1982. Berlin: Walter de Gruyter. 17–26.
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
714 DONNA R. MILLER
Downloaded from https://www.cambridge.org/core. Columbia University - Law Library, on 11 Aug 2019 at 09:27:07, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.028
27
Language and Social
Media
Enacting Identity through Ambient Affiliation
Michele Zappavigna
27.1 Introduction
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
716 M I C H E L E Z A P PAVI G N A
The main type of texts shared with ambient audiences using social net-
working services are ‘posts’: multimedia content typically arranged into
chronological ‘streams’ or ‘feeds’. In the case of microblogging services
such as Twitter, these posts are short character-constrained messages
that originally functioned predominately as status updates relating
to users’ activities. These have now taken on a broad range of communi-
cative functions, from broadcasting ideational ‘content’, to sharing feel-
ings. An example of a post, in this case a tweet, with a prominent
interpersonal function is the following comment on the 2014 Australian
federal budget in (1):
Figure 27.1 An example of an Instagram feed (left) and an individual post (right)
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 717
data source, for instance, information about the author of a social media
post or the location where it was created. While metadata has a long history
in the domain of information management, this is the first historical period
where we see it so closely tied to enacting social relations, having extended
its semiotic reach as an information-organizing tool to a social resource for
building relationships and communities.
A popular form of ‘social metadata’, or ‘social tagging’, is the hashtag,
indicated with a # symbol followed by a keyword or concatenated phrase
or clause (for an analysis of the linguistic functions of hashtags, see
Zappavigna 2015, 2018). The tweet presented in the previous section
contains two instances of social tagging in the form of hashtags (#Three-
WordBudget and #Budget2014) used to indicate both the semantic domain
of the post, and to designate the post as part of a larger Twitter meme.
Hashtags of this kind are also able to realize a range of complex interper-
sonal and textual functions beyond such topic-marking. Hashtags as a
form of conversational tagging enable individuals to search social media
discourse to find out what people are saying about particular domains, or
to share feelings and opinions with like-minded users (or argue with those
who do not share your worldview). In this way, social metadata supports
forms of ambient communion that arise out of the ability to search for
and engage with other people’s posts in ‘real-time’ within the social
stream.
What makes social metadata particularly interesting to linguists is its
capacity to infiltrate the linguistic structure of the texts that it seeks to
annotate. While traditional metadata is typically hidden from the view of
users of an information system, or separated from the main body of a text
in some systematic way, social metadata is incorporated into social media
communication, and can perform a wide range of functional roles in the
discourse itself (Zappavigna 2015). While social media services collect
many forms of traditional metadata about the networks, users, and con-
tent that they manage, most of this information is not presented within
the main content of a post and certainly does not form part of the
linguistic function and structure of a post. In stark contrast, social meta-
data is user-generated and typically acts as a kind of ‘in-text’ tagging
with a range of novel communicative functions. For example, Zappavigna
(2014b) has explored how hashtags support attitudinal alignment around
iconized dimensions of experience, such as a positive attitude regarding
coffee:
The tweet above presents a coupling of positive affect with #coffee, as well as
using related coffee tags to identify the social media user as a ‘coffeelover’.
It enacts alignment around coffee as a ‘bonding icon’ (Stenglin 2008) and
foregrounds coffee as part of this user’s identity performance, concepts that
we will explore in the second half of this chapter.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
718 M I C H E L E Z A P PAVI G N A
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 719
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
720 M I C H E L E Z A P PAVI G N A
the social media domain (e.g. Campoy and Espinosa 2012; Coupland 2007;
Eckert 2000). Other approaches adopt frameworks from related linguistic
areas such as Page’s (2012b) work applying narrative theory to social media
story-telling, considering dimensions such as the use of hashtags.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 721
Figure 27.2 The individuation cline (adapted from Martin et al. (2013:490))
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
722 M I C H E L E Z A P PAVI G N A
hierarchy from the perspective of both allocation, that is, how semiotic
resources are distributed amongst users, and affiliation, that is, how semi-
otic resources are deployed to commune (Martin et al. 2013:490). The intent
is to model the relationship between personae, sub-cultures, master iden-
tities, and, at the most generalized end of the cline, culture as a system of
meaning potential. However, work on the exact nature of this hierarchy,
and what constitutes meaningful units of analysis, is still in its infancy (e.g.
see Zappavigna and Martin 2014).
Martin grounds his perspective on individuation in the Bernsteinian
notion of culture as a ‘reservoir’ of meanings from which an individual
can mobilize a particular ‘repertoire’. This repertoire arises out of the
registers and genres to which they have been exposed. He quotes the
following passage from Bernstein in order to illustrate this point:
I shall use the term repertoire to refer to the set of strategies and their
analogic potential possessed by any one individual and the term reservoir to
refer to the total of sets and its potential of the community as a whole. Thus
the repertoire of each member of the community will have both a common
nucleus but there will be differences between the repertoires. There will be
differences between the repertoires because of the differences between
members arising out of differences in members’ context and activities and
their associated issues
(Bernstein 2000:158)
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 723
of the self has been undertaken by Zappavigna and Zhao (2017) and Zhao
and Zappavigna (2017), and by Zappavigna (2016) in relation to the repre-
sentation of subjectivity in Instagram images and Tumblr posts. This work
has considered the different options for visually representing the self in
social media images, from the choice to explicitly represent the self in the
form of a self-portrait, or ‘selfie’, or the choice to imply or infer the
presence of the self through compositional choices (e.g. a cup of coffee near
the front of an image indicating the photographer outside the frame), or
through inclusion of part of the photographer’s body in the image (e.g. a
hand holding a coffee cup). Zappavigna (2016) models these semiotic
choices as a system of ‘subjectification’. The major choice in this system
is between ‘as photographer’ perspective, where the photographer’s sub-
jectivity is represented, inferred, or implied through various choices in
visual structure, and ‘with photographer’ subjectivity, where the photog-
rapher’s perspective is unmarked (Figure 27.3).
Drawing on this initial modelling, Zhao and Zappavigna’s work aims to
show that the function of the selfie as a multimodal genre is not solely to
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
724 M I C H E L E Z A P PAVI G N A
represent ‘the self’ but rather to enact intersubjectivity, that is, to generate
various possibilities of relations between perspectives on a particular topic,
issue, or experience and hence to open up potential for negotiating differ-
ent points of view.
Selfies, as performances of identity, have been widely criticized in news
and entertainment media as an inherently narcissistic practice. One of the
assumptions at the core of this criticism is the position that the naturalized
reading of a selfie is ‘Look at me’. It is a proposition difficult to refute, as the
self/subject is the primary representational object in selfies as a visual
media (Pham 2015). However, if we adopt a social semiotic approach that
factors in the importance of interpersonal meaning, we can propose a
complementary reading: that the visual structure seen in selfies fore-
grounds the perspective of the photographer on a particular object, phe-
nomenon, or issue. For instance, in the domain controversially known as
‘mommyblogging’, selfies taken by mothers can be read as ‘Look, it is my
perspective on motherhood’ or ‘Let’s look at motherhood through my per-
spective’ (Zappavigna and Zhao 2017). This subtle shift of focus from ‘me’ to
‘my perspective’ is a very important consideration, as it affords a shift in
analytical focus from the ideational to the interpersonal.
Zappavigna (2014b) has explored microblogging identities in relation to the
types of bonds that are enacted within different communities of users. For
example, this work focused on a cluster of bonds that occur in microblogging
discourse about motherhood, but which also appear to be generalizable across
different types of personae and communities. The first bond is a ‘self-depre-
cation bond’, where the social media user can be read as laughing off (Knight
2010), or lightly and irreverently mocking, a stereotype such as the concept of
the ‘perfect mother’ (Lopez 2009), an example of which is given in (3):
(3) Just spilt wine on son’s homework diary. #badmother
A co-occurring bond is the ‘addiction bond’, as in example (4), in which the
user rallies around everyday bonding icons such as coffee, wine, or technology:
(4) Need.Wine.
And finally a frazzle bond (example (5)), whereby the user relates fatigue or
exasperation resulting from engaging in the core activity of a particular
community of fellowship, for instance:
(5) What a spectacularly horrendous night. Max awake for 3 hours,
wanting ‘iPad! IPAD!!’ I think NOT baby. Then twins snuck into
bed @ 4am. Ugh
This interplay or ‘complex’ of bonds is also seen in personae enacted in
other domains, such as the world of computer coding: for instance,
example (6) is a parallel to the above frazzle bond:
(6) Getting ready to go to bed after a long night of coding up some
custom classes to handle xml parsing and database interaction.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 725
Section 27.4 details how we may identify these types of bonds by the way in
which they are realized as particular couplings of attitude and experience
in discourse.
Feelings are meanings we commune with, since we do not say what we feel
unless we expect the person we are talking with to sympathize or empathize
with us. We express feelings in order to share them . . . to build relationships;
where we misjudge the situation and get rebuffed, then a sense of alienation
sets in.
(Martin 2002:196)
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
726 M I C H E L E Z A P PAVI G N A
Martin et al. (2013) represent this type of coupling with the following
notation:1 [ideation: she / evaluation: negative judgement]. In any given
encounter we are always proposing and reacting to bonds as we negotiate
couplings in discourse. It is in this sense that we can view ‘persons and
personalities as active participators in the creation and maintenance of
cultural values’ (Firth 1957:186). While the term negotiation characterizes
this as a very deliberate activity, the practices at work always involve a tacit
dimension: just as we do not usually consciously attend to our linguistic
choices (Zappavigna 2013), we are rarely directly aware of the patterns of
bonds that we propose and react to in an interaction.
Current work on affiliation is focused on how bonds pattern into higher-
order complexes, that is, how they ‘cluster as belongings of different orders
(including relatively “local” familial, collegial, professional and leisure/rec-
reational affiliations and more “general” fellowships reflecting “master
identities” including social class, gender, generation, ethnicity, and dis/
ability)’ (Martin et al. 2013:490).2 The aim is to use semiotic rather than
common-sense criteria to account for the kinds of memberships that
emerge when we study coupling patterns. According to this perspective,
identities are patterns of meaning inflected by membership in networks of
fellowship. In other words, they can be thought of as the disposition to
enact particular configurations of couplings that realize particular config-
urations of bonds. This disposition is informed by a persona’s particular
semiotic ‘repertoire’ that arises out of the potential semiotic ‘reservoir’
available via their membership in a given community (Bernstein 2000).
1
Other notation systems and ways of visually representing how couplings function in interactions are currently being
developed.
2
For a related approach which focuses instead on the tenor dimension, see Don (2012).
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 727
This kind of data pushes our social semiotic modelling of affiliation into
considering how bonds can function outside exchange structures.
Appending a hashtag to a post presupposes that there is an ambient audi-
ence who may share or contest the values construed.
3
For gesture analysis see Martin et al. (2013); for image analysis see Caple (2010).
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
728 M I C H E L E Z A P PAVI G N A
4
For an example of a Twitter study using the concept of ambient affiliation to explore attitudinal alignments during the
2011 Japan nuclear crisis, see Inako (2013).
5
Indeed Boyd and Crawford (2012:663) point out that ‘some of the data encompassed by big data (e.g. all Twitter
messages about a particular topic) are not nearly as large as earlier data sets that were not considered Big Data (e.g.
census data)’.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 729
particular research question at hand, and technical issues about the prac-
tical mechanics of gathering the data, which in the case of social media
services, usually involves some proficiency in using APIs (Application
Programming Interfaces) for ‘scraping’ data from the service’s databases
(see Zappavigna 2012 for an overview of the key issues relating to API use).
An important methodological concern is sampling, that is, determining the
scope and criteria for collecting texts. Many of the kinds of difficulties that
may be encountered when sampling and collecting social media texts are
not unique to electronic discourse (think, for example, of the multitude of
considerations that must be accounted for when dealing with spoken
discourse that is to be transcribed); however, the relative novelty of social
media platforms means that standards for sampling are still being
developed, and we are yet to even clearly establish what the important
issues and pitfalls are. One popular form of sampling that has received
some criticism is hashtag studies where a tag is used to collect some form of
experience (Bruns 2013). However, as Crawford (2013) has noted in relation
to the general field of big data analysis, this type of perspective only
captures a slice of the potential voices involved in a dimension of social life
at any point in time. It may factor out people who are not using social
media technologies, possibly the most disenfranchised within a commu-
nity. As such it might create what she terms ‘algorithmic illusions’ that
pollute our data analysis.
A prerequisite for usefully accumulating social media data for linguistic
analysis is being able to specify both the unit of analysis (what we are
drawing conclusions about) and the unit of observation (the kinds of textual
data that need to be sampled in order to draw these conclusions). This is the
case for both qualitative and quantitative studies. For example, a study
might investigate the social practices involved in Facebook posts (unit of
analysis), realized in specific patterns of linguistic features in these posts
(unit of observation), identified by applying any of the multifaceted types of
analysis made available by SFL theory. Social media texts, and electronic
discourse more broadly, however, problematize some of the practices lin-
guists may have used in the past for denoting what constitutes a text to be
analyzed. While all texts enter into heteroglossic relations with other texts
(Bakhtin 1981), and can be approached from both dynamic (text as process)
and synoptic (text as artefact) perspectives, determining the ‘bounds’ of a
networked, electronic text is particularly challenging. For example, High-
field and Leaver (2014) have noted the instability of Instagram posts due to
the dynamic nature of this media, which allows users to go back to posts
and add or delete comments on the image. In this case the text is, from the
perspective of mode, never complete as a communicative unit since it can
always be modified and added to.
In addition, in terms of multimodal dimensions such as presentation (e.g.
font, colour choice, layout), the text is also not static since it may be
syndicated to different kinds of devices (e.g. a smart phone) where it may
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
730 M I C H E L E Z A P PAVI G N A
6
Scripts are small programmes using a scripting language to automate text-processing tasks that would otherwise have
to be done manually by a linguist.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 731
In addition, time is important for working out the difficult issue of how we
know where a social media interaction begins and ends. Some of the
parallel parameters we find in other modes, such as ‘turn-taking’ (Sacks
et al. 1974), that might be used to characterize an exchange are not neces-
sarily present in social media interactions. An interaction will look differ-
ent depending on the ‘node’ in the social network from which it is
observed.8 Some ability to track which user is replying to, or rebroadcast-
ing, a particular user is typically available in social media metadata.
However, representing the discourse structure of non-linear, multicast
(many-to-many broadcast) interactions is a problem that has yet to be
solved. Indeed, linguists are still working on how to represent dynamic
exchanges in face-to-face interaction.
7
A timestamp is a record, usually encoded in a standard format, marking when an event occurred.
8
Study of the properties of social media networks is undertaken in an area known as Social Network Analysis (SNA).
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
732 M I C H E L E Z A P PAVI G N A
References
Albury, K. 2015. Selfies, Sexts and Sneaky Hats: Young People’s Understand-
ings of Gendered Practices of Self-Representation. International Journal of
Communication 9: 1734–45.
Argamon, S., M. Koppel, J. Pennebaker, and J. Schler. 2007. Mining the
Blogosphere: Age, Gender, and the Varieties of Self-expression. First
Monday 12(9). Available online at: http://firstmonday.org/ojs/index.php/
fm/article/view/2003/1878. (Last accessed 15/08/2018.)
Bakar, K. A. 2014. ‘A (Sensitive New Age Guy) with Difference’: Gendered
Performances in Online Personal Advertisements. Linguistics and the
Human Sciences 9(1): 5–34.
Bakar, K. A. 2015. Identity in Online Personal Ads: A Multimodal Investi-
gation. Asian Social Science 11(15): 313.
Baker, P. 2006. Using Corpora in Discourse Analysis. London: Continuum.
Bakhtin, M. 1981. The Dialogic Imagination: Four Essays. Edited by M. Holquist.
Austin: University of Texas Press.
Bamman, D., J. Eisenstein, and T. Schnoebelen. 2014. Gender Identity
and Lexical Variation in Social Media. Journal of Sociolinguistics 18(2):
135–60.
Bednarek, M. 2009. Corpora and Discourse: A Three-Pronged Approach
to Analyzing Linguistic Data. In M. Haugh, K. Burridge, J. Mulder,
and P. Peters, eds., Selected Proceedings of the 2008 HCSNet Workshop on
Designing the Australian National Corpus. Somerville: Cascadilla Proceedings
Project.
Bednarek, M. 2010. The Language of Fictional Television: Drama and Identity.
London: Bloomsbury.
Bednarek, M. 2013. ‘There’s No Harm, Is There, in Letting Your Emotions
Out’: A Multimodal Perspective on Language, Emotion and Identity in
MasterChef Australia. In N. Lorenzo-Dus and P. Blitvi, eds., Real Talk:
Reality Television and Discourse Analysis in Action. London: Palgrave Macmil-
lan. 88–114.
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 733
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
734 M I C H E L E Z A P PAVI G N A
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 735
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
736 M I C H E L E Z A P PAVI G N A
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
Language and Social Media 737
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
738 M I C H E L E Z A P PAVI G N A
Downloaded from https://www.cambridge.org/core. The Librarian-Seeley Historical Library, on 20 Aug 2019 at 08:20:57, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.029
28
Theorizing and Modelling
Translation
Erich Steiner
1
See Steiner (2005, 2015a) for more elaborate accounts of the history of SFL Translation.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
740 ERICH STEINER
term by ethnographic analysis, that is, by placing it within its context of culture,
by putting it within the sets of kindred and cognate expressions, by contrast-
ing it with its opposites, by grammatical analysis and above all by a number
of well-chosen examples – such translation is feasible and is the only correct
way of defining the linguistic and cultural character of a word.
(Malinowski 1935:17, emphases in original)
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 741
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
742 ERICH STEINER
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 743
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
744 ERICH STEINER
28.2.3 Equivalence
‘Equivalence’ as a key notion of translation has an important place in SFL
modelling, though more diversified than in most non-SFL approaches, and
depending on the relationships between source context and target context
(see Yallop 2001; for a summary discussion across schools of translation
studies, see Halverson 1997). ‘Equivalence’ between some source and target
text can be sought and privileged at different linguistic levels (phonological,
lexicogrammatical, semantic, and contextual) and with different emphases
on ideational, interpersonal, or textual meaning. Any of these can be
privileged depending on the purpose of the translation at hand, even if a
sort of default case will usually be equivalence on the semantic level, and
here particularly with regard to ideational meaning. An optimized solution
in any translation task will be a compromise approximation, rather than
simply ‘correct’ or ‘incorrect’. The following illustrates how at the lexico-
grammatical level and between ideational, interpersonal, and textual func-
tions, originals and translations may diverge substantially in an SFL
analysis (Steiner 2004:150–61, example from Doherty 1991):
(1) The suspicion that volcanic eruptions are the primary source of
aerosols in the upper atmosphere has been around for many years.
(original, New Scientist 21 January 1982: 150)
(2) Seit vielen Jahren vermutet man schon, dass die Aerosole in den höheren
Schichten der Atmosphäre vor allem aus Vulkanausbrüchen stammen.
(German translation, Doherty 1991)
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 745
(3) For many years one has been assuming already that the aerosols in the upper
atmosphere derive above all from volcanic eruptions.
Simplified Analysis
English Source Text:
• Experiential: Carrier (embedded: Token – Process:relational – Value) –
Process:relational:locational – Location – Circumstance of Time
• Logical: Carrier (Head ! [[projected fact]])
• Interpersonal: Indicative/Non-modalized
• Textual: Theme: unmarked, on Subject ‘The suspicion that . . .
atmosphere’
• Information: NEW strongly on ‘primary source . . . Atmosphere’ and
weakly on ‘for many years’
German Translation:
• Experiential: Circumstance – Process: mental: cognition – Processor Cir-
cumstance Idea (Carrier – Process:relational – Locational)
• Logical: α: Head-clause – projection – β: projected clause
• Interpersonal: Indicative + indicative/Non-modalized
• Textual: Theme unmarked, on ‚Seit vielen Jahren . . .‘
• Information: NEW on ‚vor allem aus Vulkanausbrüchen . . .‘
As we can see, there is no isomorphism and in this sense no equivalence
between experiential and logical structures at the lexicogrammatical level
between source and target. The configurations of clause functions vary, and
the mapping of semantic functions onto lexicogrammar varies too – with
the exception of the NEW-element.
In terms of a clause semantics (Halliday and Matthiessen 1999), however,
both source text and target text encode the following ideationally equiva-
lent ‘sequence of figures’ (states of affairs):
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
746 ERICH STEINER
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 747
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
748 ERICH STEINER
Figure 28.1 Direction of metaphorization [note: numbers refer to examples in Figure 28.3]
(Halliday and Matthiessen 1999:264)
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 749
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
750 ERICH STEINER
2
Section 28.3.1 is a revised version of Steiner (2004:25–43).
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 751
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
752 ERICH STEINER
3
Halliday and Hasan (1976:229) include these ‘text-type’ options under ‘mode of discourse’, realized through the
textual component of the lexicogrammar. We interpret them here as more directly deriving from the semiotic goals
under ‘field’, diverging from a strong version of the ‘context-metafunction-hook-up hypothesis’, also problematized in
Thompson (1999).
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 753
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
754 ERICH STEINER
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 755
face interaction, for example, limits the use of gesture and body-language in
communication, but may support other (variants of ) semiotic systems
instead, such as parallel display of texts, icons, etc. A recognition of the
importance of channel constraints helps the translator give due weight to
conventions of punctuation and paragraphing when using any of the print
channels. Again, languages show considerable variation here. A historical
example of the pervasive influence of the channel (and medium) of dis-
course is the invention and spread of printing and its influence on standard
languages in particular. The spread of electronic channels may currently be
causing another wave of textual and cultural change. Relevant textual
realizations, possibly indicating such changes, are the increasing frequency
of multiple and de-personalized authorship with all its linguistic effects, the
use of hypertext techniques in addition to, or instead of, continuous linear
text, and consequently, new patterns of cohesion and coherence.
In ‘medium of discourse’, the major variables probably still are spoken
vs. written, although electronic channels might conceivably lead to a new
medium of discourse. Patterns of realization involved are the use of pro-
nouns vs. full words, exophoric vs. endophoric reference, types of cohesion
in general, certain types of clause complexity, grammatical metaphor, etc.
English and German, for example, differ in interesting ways as to how they
employ lexicogrammatical resources to structure information in written
and spoken language, for example, through types of clefting, extraposition,
inversion, and related phenomena – in general thematizing and focusing
devices. A model of translation needs to offer scope for deciding whether
the medium of discourse remains constant in the translation of a given text,
and what lexicogrammatical configurations are appropriate in the target
language for realizing the medium.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
756 ERICH STEINER
(5) The suspicion that volcanic eruptions are the primary source of
aerosols in the upper atmosphere has been around for many years.
(original, New Scientist 21 January 1982: 150)
(6) Seit vielen Jahren vermutet man schon, dass die Aerosole in den höheren
Schichten der Atmosphäre vor allem aus Vulkanausbrüchen stammen.
(German translation, Doherty 1991)
The translation procedures in Figure 28.4 find their more global counter-
parts in what Newmark (1988:45–53) calls ‘translation methods’, and what
SFL integrates into its overall model as ‘translation strategy’ (Teich
2001:212–13), globally applied to entire texts. They move between ‘free/
target-culture oriented’ translation privileging contextual equivalence, and
‘literal/source-culture oriented’ privileging grammatical equivalence.
Translation strategies at the level of cohesion are exemplified for English-
German in Steiner (2015b).
4
A fuller analysis of the grammatical relationships in (1) to (2) can be found in Steiner (2004:157–8).
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 757
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
758 ERICH STEINER
Prerequisites for work with electronic corpora include tools for semi-
automatic annotation, such as the UAM corpus tool (O’Donnell 2008). On
the other hand, Bateman (2008), Hiippala (2012), and Taylor and Baldry
(2001) suggest approaches and initial implementations for the analysis and
translation of multimodal documents, strongly influenced by SFL
theorizing.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 759
source and target texts, the translator has to decide more consciously what
his/her own agentive role is. Should the translator become visible and create
something that is recognizable as a translation, or should s/he remain
invisible and strive for maximum adaptation to target-cultural norms?5
Likewise, ‘social roles’ have become more diversified, and so have the
means of realization. The possible degrees of levels of expertise have
become more differentiated. There are nowadays between the extremes of
the absolute layperson and the absolute expert all shades of semiotic
possibilities as encoded in texts. Social hierarchies are often expressed
textually. Along with a broader range of social roles, we find a broader
range of ‘social distance’. The possibility of styles from absolutely frozen to
intimate has to be mastered by the present-day translator, alongside cultur-
ally specific changes in these scales. Finally, the various ways of realizing
affect, both as positive or negative and as self-oriented or other-oriented (see
Martin 1992; Munday 2012) are of increasing importance, especially in fields
such as advertising and in direct interactions (such as in interpreting), but
also in the various forms of synchronization and subtitling of films.
It may be in the ‘mode of discourse’ of translation that some of the most
far-reaching changes are taking place. Considering ‘language role’, we find
all possible shades between the poles of ancillary and constitutive. At the
ancillary end, there are registers where language plays a rudimentary role
only, being very much subservient to visual communication (see Taylor and
Baldry 2001; Bateman 2008). Other instances of this include software local-
ization, with ‘canned language’ in menu-guided user interfaces used for
many types of current software. However, there are still numerous registers
in which language is absolutely constitutive, as in the translation of fiction,
(popular) science, and interactions within large international organizations
such as the UN and the EU. As for the ‘channel of discourse’, we have to be
aware that whereas until about 1980 the dominant channel for most types
of translation may have been paper, along with conventional sound waves
for interpreters, this traditional channel has become accompanied, and
occasionally overtaken, by the electronic channel. The translator nowadays
has to be aware of the possibilities and limitations of the major channels of
communication, and of the tools for working in these. Systems for termin-
ology management, machine-aided translation, and above all for straight
text handling and word processing are gaining ground in most areas of
translation and multilingual text production. With these processes, the
possibilities of using intertextual relationships between versions of texts
have expanded, and translation memories for the production, evaluation,
and modification of existing texts are being used in professional practice.
These developments change the notion of ‘a text’, and ‘a translation’, from
a static to a much more dynamic and changeable entity. Finally, the
5
See House’s (2015:63–70) distinction between ‘overt’ and ‘covert’ translation and Venuti’s (1995) discussion of the
visibility of the translator.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
760 ERICH STEINER
‘medium of discourse’: the variation between spoken and written may still
be the primary one, and it remains to be seen whether the electronic
channels ultimately lead to the creation of a major new medium – however,
electronic channels are certainly creating new sub-varieties, transcending
the binary spoken–written divide. Hypertext, inside or outside of the Inter-
net, allows new ways of creating non-linear discourses, as well as new ways
of exploring language-specific systems of conjunctive relations and other
forms of cohesion. Translators have to work increasingly web-based, and
the potentials for multilingual text creation and management have to be
thoroughly familiar to them.
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 761
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
762 ERICH STEINER
References
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 763
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
764 ERICH STEINER
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
Theorizing and Modelling Translation 765
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
766 ERICH STEINER
Downloaded from https://www.cambridge.org/core. Newcastle University, on 27 Aug 2019 at 06:59:53, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.030
29
Language Typology
Abhishek Kumar Kashyap
29.1 Introduction1
1
Please see the end of the chapter for a list of abbreviations used in the examples in this chapter.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
768 ABHISHEK KUMAR KASHYAP
The term ‘typology’ derives from the Greek noun týpos and refers to the
study of types, or ‘a system for dividing things into different types’ (Hornby
2005:1656; Collins English Dictionary 2017). Retaining its meaning of classify-
ing things into categories according to their types, the term is used in
several disciplines such as anthropology (classification of cultures and races
and sociocultural norms), psychology (the classification of different human
personalities and personality traits), archaeology (the classification of
things according to physical characteristics), and theology (the relationship
of Old Testament to New Testament with respect to religious beliefs, events,
persons, or statements).
In linguistics, typology refers to the classification of the world’s lan-
guages according to similarities and differences in their linguistic struc-
tures and genetic relationships. Language typology, therefore, is essentially
comparative and cross-linguistic. That is, a typological analysis obligatorily
involves data from multiple languages, either of different language families
or of the same family, for comparison, and proposes generalizations on the
basis of the analysis.
Language typology has developed from language description, which has a
long history of work going back at least to the time of Pānini (fifth century
_
BC) and other Indian grammarians (such as Yāska, Kātyāyana, and Patañjali)
of ancient India and of which we have written record (see Shukla 2006;
Kiparsky 2009). Initially, the focus was on the description of just one
language and the primary purpose was teaching the language as well as
intellectual inquiry in the philosophy of language. Pānini’s Astādhyāyī, a
_ __
masterpiece of linguistic endeavour in the history of language study, was
motivated by both pedagogical purpose as well as his inquiry in the phil-
osophy of language. As the tradition of language study developed, however,
researchers began to ask more interesting questions in order to understand
language diversity, and genetic relationship among languages was a natural
extension of the intellectual inquiry into the philosophy of language that
led to comparisons of structures of different languages. This later came to
be known as language typology (see Section 29.3). Language typology was
thus introduced as a method to investigate genetic relationships among
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 769
languages and group them under different stocks on the basis of their
genetic origin and linguistic structures. Greenberg’s (1963) classification
of language stocks on the basis of the order of major constituents of the
clause, traditionally known as Subject, Object, and Verb, was based on this
principle. (With typological developments in SFL, SFL researchers have
observed that these traditional categories have inconsistencies at several
levels. SFL typologists therefore have begun to use the order of Subject,
Complement, and Predicator to categorize language (e.g. see Matthiessen
2004).2)
As can be expected, the findings of typological study help us to under-
stand the amazing cross-linguistic diversity of the world, and that can be
understood by no other means than describing, comparing, and analyzing
languages. For example, while the first-person pronouns in most, if not all,
languages of the world do not show gender and the first-person singular
pronouns generally lack the marking for honorification, we find interesting
variations in the second- and the third-person pronominal forms, as a
comparison of the pronominal systems of English (a Germanic language)
and Bajjika (an Eastern Indo-Aryan language (Kashyap 2014)) demonstrates,
even though both these languages have developed from the same root (i.e.
Indo-European). Bajjika has developed a four-level honorification system
(high-honorific, mid-honorific, plain honorific, and non-honorific) to refer
to addressees of different social statuses (Kashyap 2012; Kashyap and Yap
2017). English, in comparison, is limited to the use of only one second-
person pronoun you, regardless of the addressee’s social status.
This kind of linguistic diversity helps us to make sense of different
sociolinguistic situations in the world and often provides insights into the
range of linguistic and meaning-making systems in world languages. For
example, for a speaker of English, or even an expert of English with little
exposure to how different languages behave and how they develop struc-
tures to express meanings, it will be hard to make sense of the morpho-
logical complexity of the following example from Mundari (a Munda
language within the Austroasiatic family) given in (1) and whether it is a
word or a sentence.
(1) Omamtanain
om-am-tan-a-in
give-2sg-prs-1sg
‘I give you.’
2
The traditional categories Subject, Object, and Verb are reinterpreted and renamed in some new twentieth-century
theories of linguistics. In SFL, while the category Subject retains its name, although with a fresh interpretation, Object is
reanalyzed as Complement, and Verb (at the clause rank) is reinterpreted as Predicator (see Halliday 1994; Halliday
and Matthiessen 2014; Matthiessen et al. 2010). In SFL there is no such term as Object. The category ‘verb’ refers to a
unit in word classes; a corresponding unit above word class is ‘verbal group’, which has a verb as the Head, e.g. walk in
is walking and target in has been targeted.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
770 ABHISHEK KUMAR KASHYAP
Even more difficult would be making sense of why in the Bajjika con-
struction in (2) the addressee (shown by bold) is encoded in the verbal form
if the addressee is not a functional constituent or referent in the clause:
The hierarchical social order of the Bajjika community demands that the
speaker acknowledges the presence of the addressee even if the addressee is
not a functional constituent in the clause, and therefore the language has
developed a system in the agreement paradigm (for details, see Kashyap
2012; and Kashyap and Yap 2017).
Example (3) demonstrates how languages develop ways to manifest dif-
ferent sociocultural phenomena and how a good understanding of linguis-
tic diversity helps to grapple with the linguistic and semantic variations.
This example, from Maithili, shows the complexity of agreement marking.
In this example, three referents are marked in its paradigm of verbal
agreement.
3
The interlinear glosses in this example contain the modification by the author on the basis of his knowledge of Maithili.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 771
As more and more languages are described and cross-linguistic data are
compared and typologized, we encounter unexpected linguistic facts that
demonstrate that some ‘exotic’ languages have unique ways of encoding
the same information. In several cases, we also find surprising similarities
that stimulate us to ask whether they had any historical relationship. The
kind of addressee-oriented agreement markers shown in Indo-Aryan Bajjika
and Maithili above have also been found in Basque languages (see example (12)),
although no genetic relationship has yet been established between the Indo-
Aryan and the Basque languages. Rather, Basque is classified as a language
isolate, which means that as yet we have no clue to its origin.
Language typology has a great deal of similarities with two related lin-
guistic sub-disciplines, contrastive linguistics and comparative historical
linguistics, although they each have a unique focus of attention. While
language typology often focuses on many languages, and preferably larger
language samples on which firmer generalizations can be made and that
can enrich our theoretical understanding of linguistic diversity and lan-
guage evolution, contrastive studies have traditionally focused on the study
of two languages. Moreover, while comparison of linguistic structures and
related phenomena is at the core of both language typology and contrastive
linguistics, they have a different preference of inclination: in typological
explorations equal weight is given to both similarities and differences,
while the primary focus of contrastive linguistics is on differences of
linguistic structures, and linguistic similarities take a back seat.
Whereas language typology seeks to explore cross-linguistic variations
and chart the language diversity that exists across languages, language
contrast, which is sometimes considered a branch of language typology,
aims to contribute to applied areas of linguistics such as language teach-
ing, error analysis, and translation. Language contrast developed for peda-
gogical reasons, to improve foreign language teaching, from Charles Fries’
(1945) contention that in foreign language teaching ‘the most effective
materials are those that are based upon a scientific description of the
language to be learned, carefully compared with a parallel description of
the native language of the learner’ (Fries 1945:9). His colleague Robert
Lado later advanced Fries’ suggestion and proposed the ‘Contrastive
Hypothesis’ (Lado 1957), which suggested that ‘the second-language
learner’s language was shaped solely by transfer from the native lan-
guage’ (Tarone 2006:134). Lado emphasized that the comparison of the
learner’s target language with the learner’s first language would accur-
ately predict the learner’s difficulty of learning a second/foreign language,
offer better insights for learnability, and help to improve language teach-
ing. It is worthwhile noting that the typological anchoring of SFL also
developed through the pedagogical concerns of Michael Halliday (e.g. see
Halliday 2007), and contrastive analysis was an area of interest particu-
larly in the regions where English had the status of a second or foreign
language (e.g. see Prakasam 1970).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
772 ABHISHEK KUMAR KASHYAP
4
Although some works claim to be theory-free, ‘there is no such thing as atheoretical description’ (Dryer 2006:207; also
see Matthiessen and Nesbitt 1996) and the same is true with respect to typology, given that description is a
prerequisite of typological explorations.
5
Some scholars confusingly suggest that SFL is a sub-discipline of linguistics, which is clearly inaccurate. SFL is not a sub-
discipline; it is a theory of language, a comprehensive method of language study, or a ‘metalanguage’ (Matthiessen
2007), which is applied to study various subjects within a range of linguistic sub-disciplines, including (but not limited
to) language description, language typology, discourse analysis (e.g. Butt et al. 2004), language teaching/learning (e.g.
Rose and Martin 2012; Dreyfus et al. 2016; Hood 2016), World Englishes (e.g. Halliday 2003; Kashyap 2014),
stylistics (e.g. Halliday 1971), and verbal art (e.g. Hasan 1989; Miller and Turci 2007; Butt 2009). See Mwinlaaru and
Xuan (2016) for a detailed review and references of language description using SFL as the theoretical framework.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 773
6
Sir William Jones was the founder of this society. The society later was renamed as the Asiatic Society of Bengal and was
a primary forum for scholarly debate and publication.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
774 ABHISHEK KUMAR KASHYAP
philologer could examine them all three, without believing them to have
sprung from some common source, which, perhaps, no longer exists; there is
a similar reason, though not quite so forcible, for supposing that both the
Gothick [Gothic] and the Celtick [Celtic], though blended with a very differ-
ent idiom, had the same origin with the Sanscrit, and the old Persian might
be added to the same family, if this were the place for discussing any
question concerning the antiquities of Persia.
(Jones 1798:422–3; quoted in Teignmouth 1804:388; Pachori 1993:175)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 775
7
Friedrich von Schlegel’s elder brother, August von Schlegel, was an equally influential typologist.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
776 ABHISHEK KUMAR KASHYAP
development of what is the field of ‘language typology’ today. The field was
greatly advanced by the works of Sir William Jones, Friedrich von Schlegel,
Franz Bopp, Georg von der Gabelentz, Sir George Grierson, and many other
philologists in the eighteenth and nineteenth centuries, and Franz Boas,
Sapir, Bloomfield, and other linguists in the early years of the twentieth
century, all under the rubric of general linguistics. Linguistic structure took
precedence over meaning in these approaches, except in new approaches
such as SFL that are primarily concerned with meaning and the analysis of
morphological and syntactic forms as a means of understanding how
meaning is made in natural languages, as we will see below.
In the second half of the twentieth century, we can trace the emergence
of two significant lines of approaches to language typology – functional and
formal. The functional typology included several distinct theoretical frame-
works and was in a sense led by Greenberg’s functional typology, until the
emergence of new approaches such as SFL (Section 29.4.2), West-Coast
Functional, and Dik’s Functional Grammar (Butler 2003).
Each of the functional and formal approaches emerged on the scene as a
reaction to different approaches. Generative grammar emerged as a reac-
tion against the behaviourist psychologists’ anti-universalist view that dis-
approved of the existence of innate and universal mental ability for
language learning and postulated that ‘linguistic competence is acquired
through learning of stimulus-response pattern’ (Croft 2017). The generative
linguists (or cognitive linguists, to put it more appropriately) led by
Chomsky, interestingly, assumed all languages to be ‘English-like but with
different sound systems and vocabularies’ (Evans and Levinson 2009:429),
and proposed the well-known Universal Grammar, which lacked the empir-
ical base of the kind of Greenbergian or Hallidayan functionalism.
Greenberg’s functional typology emerged in response to anthropological
relativism, which postulated that languages of the world vary arbitrarily, as
we see in this famous quote of Martin Joos: ‘Languages could differ from
each other without limit and in unpredictable ways’ (Joos 1957:96). Green-
berg, as we see in his research and subsequent publications (e.g. Greenberg
1978), advocated for more systematic sampling of a greater number of
languages, which ‘reveals not only range of variation but constraints on
that variation’. He strongly believed that those linguistic constraints would
‘demonstrate that languages do not vary infinitely, and the constraints
represent language universals’ (Croft 2017). Thus, although the approaches
of both Chomsky’s generative theory and Greenberg’s functional typology
had a fundamental theoretical juxtaposition, they laid a great emphasis on
language universals for different reasons.
While cognitive linguists of the generative tradition and Greenbergian
typologists emphasized language universals, later research by functional
typologists painted a strikingly different picture. Languages of the world
‘differ so fundamentally from one another at every level of description
(sound, grammar, lexicon, meaning) that it is very hard to find any single
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 777
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
778 ABHISHEK KUMAR KASHYAP
order universals). The typological research from the early 1980s, however,
shifted the focus from language evolution and genetic classification to
language diversity, and typological researchers began to focus more on
how different meanings are expressed in different languages, thus guiding
the focus on a combination of factors, e.g. morphosyntax and lexicogram-
mar, as well as the semantic and pragmatic inferences drawn from lexico-
grammatical patterns in language.
In the last forty years, language typology has seen tremendous advance-
ment, especially from functional perspectives, with the emergence of dis-
tinct functional theorists who conduct research in distinct ways and as an
inevitable consequence have opened up numerous theoretical frontiers.
The functional theories of linguistics that emerged in the second half of
the past century and have significantly contributed to typological research
include Systemic Functional Linguistics (SFL), also referred to as Systemic
Function Grammar (SFG), first developed by Michael Halliday (e.g. Halliday
1994; Halliday and Matthiessen 2014) and further expanded by other sys-
temic functional linguists (on the development of SFL, see Matthiessen
2005), West-Coast functionalism (e.g. Givón 1995), Simon Dik’s Functional
Grammar (e.g. Dik 1997a, 1997b), and Role Reference Grammar (Van Valin
and LaPolla 1997; Van Valin 2005). As a description and comparison of
these theoretical approaches is not within the scope of this chapter,
I narrow down my focus to SFL’s contribution to language typology to be
consistent with the aims and scope of this handbook. An appreciative and
comprehensive comparison of the functional theories can be found in
Butler (2003). For a snapshot of key functional theories readers are advised
to consult Butler (2005, 2006).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 779
Language has evolved to satisfy human needs; and the way it is organized is
functional with respect to these needs.
b. sofā par.
sofa loc
‘On the sofa.’ Hindi
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
780 ABHISHEK KUMAR KASHYAP
8
Greenberg’s (1963) classic paper on word order universals in language ‘contrasts sharply with most other work at the
time in assuming a set of descriptive notions that are to a large extent simply those of traditional grammar’ (Dryer
2006: 210).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 781
Teruya et al. (2007) analyze six languages (Òkó, Spanish, French, Danish,
Thai, and Japanese), and find that all those languages ‘operate with the
prosodic mode of expression in the realization of options in the interper-
sonal system of mood, but the languages vary with respect to whether
these prosodies are manifested sequentially, segmentally or intonationally’
(Teruya et al. 2007:914). In English (a Germanic language), for example, the
polar interrogative and declarative moods are distinguished by the
‘the relative sequence of Subject and Finite’ (Teruya et al. 2007), as in the
examples of English in (7) and (8).
b. indicative: declarative
thǝ:1 maj3 daj3 ju:2 kruə1thep3:
she neg asp:pfv live Bangkok
‘She does not live in Bangkok.’
Thai (Teruya et al. 2007:901)
(9) a. indicative: polar interrogative
tú te acuerd- as de aquel señor?
2-sg 2.ref.sg remember 2.sg.prs.ind. of that man
‘Do you remember that man?’
Spanish (Teruya et al. 2007:892)
b. indicative: declarative
María se acord- ó de nosotros
Mary 3.ref. sg remember 3.sg.pst.ind of us
‘Mary remembered us.’
Spanish (Teruya et al. 2007:893)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
782 ABHISHEK KUMAR KASHYAP
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 783
In examples (11) and (12), we can see that the addressee is not a constituent
in these clauses, and yet these examples host suffixes for a second-person
referent to reflect the presence of the addressee. The English translation
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
784 ABHISHEK KUMAR KASHYAP
here shows the ideational meaning, but the interpersonal meaning encoded
by the allocutive suffixes for reference to the addressee’s social status is
missing because these suffixes have no role in the experiential segment of
the clause. In Bajjika, however, the meaning implied here is purely inter-
personal and indicates that the speaker is mindful of the presence and
social status of the addressee. Furthermore, such unfamiliar structural
and semantic differences also have relevance for inter- and cross-cultural
communications.
In the current scenario of globalization, when people more frequently
travel across territorial borders for various reasons such as business, immi-
gration, education, tourism, and diplomacy, language typology has a bigger
role to strengthen one’s intercultural and cross-cultural communication
skills. Communication styles significantly vary across cultures, and how
successfully one communicates largely depends on one’s social, linguistic,
and cultural backgrounds and training. For example, it is quite common to
express gratitude and apology by the use of, for instance, please and sorry, in
English culture, while Indian speech communities do not favour overt
verbalization of gratitude (Apte 1974) because most languages spoken in
India have developed verbal systems that register different strands of
politeness including gratitude and that are integrated into the speakers’
communicative styles (see Hasan’s 1984 work on Urdu). Moreover, in most
Indian communities, requests and expressions of gratitude are overtly
verbalized for strangers or those who have a weak level of solidarity with
the speaker but rarely among family members, close relatives, and friends.
In today’s multicultural globalized societies where colleagues and co-
workers of different linguistic and cultural backgrounds interact, the socio-
linguistic knowledge gained from language typology can help to empower
people with a better communication style and develop sensitivity toward
one’s colleagues and co-workers’ languages and cultures, and can facilitate
the creation of materials for training programmes in this area.
Researchers in language learning/teaching and acquisition have long
been interested in typological research, with a firm belief that a grasp of
linguistic diversity and how languages vary with respect to linguistic
structure and related meaning will help us to better understand the
cognitive processes of language learning by children as well as adults.
The evolution of contrastive linguistics was propelled by the impact of
typological differences in language learning, as we noted earlier. And the
famous Sapir-Whorf hypothesis reflecting this typology-acquisition/
learning duet is well known; according to the Sapir-Whorf hypothesis,
the structure of a language influences the cognitive processes of the
speaker of that language (see Whorf 1956; Wardaugh 1970:123). Cognitive
scientists have noted that typologically different languages significantly
vary with respect to which linguistic items children acquire first and
which items they acquire later (Berman and Slobin 1994; Slobin and
Bowerman 2007; Berman 2014).
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 785
29.6 Conclusion
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
786 ABHISHEK KUMAR KASHYAP
References
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 787
Bickel, B., W. Bisang, and Y. P. Yādva. 1999. Face vs. Empathy: The Social
Foundation of Maithili Verb Agreement. Linguistics 37(3): 481–518.
Bowerman, M. 2010. Linguistic Typology and First Language Acquisition. In
J. J. Song, ed., The Oxford Handbook of Linguistic Typology. Oxford: Oxford
University Press. 591–617.
Butler, C. S. 2003. Structure and Function: A Guide to Three Major Structural-
Functional Theories, Part 1: Approaches to the Simplex Clause. Amsterdam:
John Benjamins.
Butler, C. S. 2005. Functional Approaches to Language. In C. S. Butler, M. de
los Ángeles Gómez-González, and S. M. Doval-Suárez, eds., The Dynamics of
Language Use: Functional and Contrastive Perspectives. Amsterdam: John Ben-
jamins. 3–17.
Butler, C. S. 2006. Functionalist Theories of Language. In K. Brown, ed.,
Encyclopedia of Language and Linguistics. 2nd ed. Amsterdam: Elsevier.
696–704.
Butt, D. G. 2009. Technique and Empire in the Poetry of Ee Tiang Hong.
World Englishes 28(3): 365–93.
Butt, D. G., A. Lukin, and C. M. I. M. Matthiessen. 2004. Grammar: The First
Covert Operation of War. Discourse and Society 15(2–3): 267–90.
Caffarel, A., J. R. Martin, and C. M. I. M. Matthiessen, eds. 2004. Language
Typology: A Functional Perspective. Amsterdam: John Benjamins.
Cannon, G. 1990. The Life and Mind of Oriental Jones: Sir William Jones, the Father
of Modern Linguistics. Cambridge: Cambridge University Press.
Chen, L. and J. Guo. 2009. Motion Events in Chinese Novels: Evidence for an
Equipollently-framed Language. Journal of Pragmatics 41(9): 1749–66.
Collins English Dictionary. 2017. Collins English Dictionary. New York: Har-
perCollins Publishers. Available online at: www.collinsdictionary.com/
dictionary/english. (Last accessed 25/09/2017.)
Croft, W. 2017. Typology and Universals. In M. Aronoff and J. Rees-Miller,
eds., The Handbook of Linguistics. 2nd ed. Oxford: Wiley Blackwell. 39–56.
Daneš, F., ed. 1974. Papers on Functional Sentence Perspective. The Hague:
Academia.
Dik, S. C. 1997a. The Theory of Functional Grammar, Part 1: The Structure of the
Clause. 2nd ed. Edited by K. Hengeveld. Berlin: Mouton de Gruyter.
Dik, S. C. 1997b. The Theory of Functional Grammar, Part 2: Complex and Derived
Constructions. Edited by K. Hengeveld. Berlin: Mouton de Gruyter.
Dreyfus, S., S. Humphrey, A. Mahboob, and J. R. Martin. 2016. Genre Peda-
gogy in Higher Education: The SLATE Project. Basingstoke: Palgrave
Macmillan.
Dryer, M. 2006. Descriptive Theories, Explanatory Theories, and Basic Lin-
guistic Theory. In F. Ameka, A. Dench, and N. Evans, eds., Catching
Language: Issues in Grammar Writing. Berlin: Mouton de Gruyter. 207–34.
Evans, N. and S. C. Levinson. 2009. The Myth of Language Universals:
Language Diversity and Its Importance for Cognitive Science. Behavioural
and Brain Sciences 32: 429–92.
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
788 ABHISHEK KUMAR KASHYAP
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 789
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
790 ABHISHEK KUMAR KASHYAP
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Language Typology 791
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
792 ABHISHEK KUMAR KASHYAP
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:25:25, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.031
Index
abstract vs concrete 59, 98–9, 273–4, 289, 322, amplification 301, 340, 348, 350 (see also message
452, 528, 562–3, 565, 568, 572, 578, 623, semantics)
628, 631, 639–40, 645 anaphora, anaphoric reference (see under
abstraction 3, 14, 22, 26, 30, 46, 70, 77–80, 85–7, cohesion)
152, 156, 162, 205, 358, 384, 437–8, 441, ancillary 294, 754, 759 (see also Mode (contextual))
459, 488, 548, 565, 570–1, 575, 621, 636, annotation 347, 564, 568–9, 577–80, 758 (see
671, 741, 762 also corpora/corpus, automated tagging)
academic discourse 390, 397 aphasia 6, 589, 598–9, 601, 603–4
accent 644 appliable linguistics 1, 51–2, 92, 160, 276, 306–7,
and juncture 180 (see also juncture) 333, 657, 709
as in pronunciation 185 Application Programming Interfaces (APIs) 729
as in phonology appraisal and educational discourse 396–8
word accent 182 appraisal and legal discourse 399, 471, 480
acquired brain injury (ABI) 601 appraisal and media discourse 398–9
action literacy 542–3 (see also literacy) appraisal and multimodal discourse 400–1
action research 301–2, 537–8, 545, 551 appraisal (system of ) 4–5, 81, 301, 334, 360,
Action, Reflection, Contact (ARC) 152, 476, 494, 372, 375–6, 378, 382–4, 401, 592, 600, 660
503, 513–14, 543 (see also graduation , judgement ,
Activation (in context) 148–9, 151, 164, 221, 578 attitude , affect , appreciation ,
active voice 104, 337, 429 engagement )
activity sequence 216, 359 appreciation 360–1, 384–5, 388, 397 (see also
Actor 39, 94–6, 98–9, 337, 415, 465–6, 567 (see appraisal (system of ))
also lexicogrammar) apprize 343, 345–6, 349, 504–5, 658 (see also
addressee (Addressee) 41, 218, 235, 267, 344, semantic networks)
507, 570–2, 769–70 (see also Performer) Arabic 545, 564, 624, 626, 628
adjective group 130, 138, 246–7 architecture (of language) 3, 35, 60, 86, 151, 173,
Adjunct, adjunction 40, 42, 110, 113–15, 193, 260, 267, 274–5, 291, 302, 383, 549, 621,
240, 242, 244, 251, 253, 346, 349 623, 633, 638, 643–4, 740
circumstantial Adjunct 42 artificial intelligence 27, 562, 565, 570, 579, 643
comment Adjunct 378, 383 Asperger Syndrome 593
conjunctive adjunct 115, 314, 707 assumptive 344–7, 658 (see also semantic
Confirmation Seeker 238 networks)
adverbial group 122, 127, 131, 138, 246–7, 383 articulate, articulation, articulator 493–4, 622
Ælfric’s Homilies 413–14 (see also symbolic articulation, double
affect 39, 360, 378, 384, 386–7, 704, 717, 728, articulation)
754–5, 759 (see also appraisal (system of )) attention deficit hyperactivity disorder (ADHD)
Affected (as in lexicogrammar) 241, 250–1 589, 596 (see also neurodevelopmental
affiliation 5, 361, 399, 402–4, 655, 661, 725–7, disorders)
732 (see also ambient affiliation) attitude 5, 81, 301, 360–1, 384–6, 388–90,
affordance 361, 715, 728 392–4, 396–8, 401, 403, 545, 551, 726 (see
Agent 39, 241, 250–1, 498, 573 also appraisal (system of ))
Alzheimer’s Disease 6, 589, 598 audience 388, 402, 512, 528–9, 727
ambient affiliation 399, 726 (see also affiliation) autism spectrum disorder (ASD) 592
ambiguity 2, 160, 214, 601, 668 automated tagging 569
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
794 INDEX
autonomous wheelchair 572 (see also dialogue clause complex 59–60, 73, 75, 78, 82–3, 267,
systems) 301, 314, 321–2, 360–1, 365–6, 393, 415,
auxiliary verbs 40, 107, 110, 126–7, 135, 244, 445
252–3 CLIL 529
clinical linguistics 6, 277, 587–90, 608, 610
Bajjika 769 clinician–patient interaction 654 (see also
Basque 641, 771 doctor–patient talk)
belief system 234–5, 514 closed class (items) 315
Bhojpuri 770 COBUILD corpus 653
Bihar 774, 785 code 45, 463–4, 633, 656, 675 (see also
boundary marker 94, 175–6 Legitimation Code Theory (LCT))
elaborated 294, 463–4, 498
Cardiff model of language, Cardiff Grammar 4, restricted 294, 463
29, 58, 230–55, 260–1, 266, 268, 270, 275, cognitive development 497, 526
474 cognitive models 266, 278–9
Carrier 102, 250–1, 287, 445, 745 Cognitive Linguistics 232, 476–7
casual conversation 149, 180, 298, 305–6, 383, cohesion 4, 41, 60, 80, 186, 288–91, 300, 311–28,
403 (see also pragmatic conversation) 334, 358, 360–1, 377, 379, 417, 467, 525–6,
category (grammatical, linguistic) 13, 46, 78, 591–3, 595, 599, 654, 656, 746
102–3, 119, 206, 210, 231, 243, 266, 268, anaphora, anaphoric reference 314, 318–19,
279, 294, 301, 316–18, 320–1, 336, 384–7, 359, 361, 364, 378
422, 474, 521, 630, 756 cataphoric reference 319, 361
Channel 630, 638 (see under Mode) cohesive conjunction 4, 80, 83, 115, 311–28,
Chaucer 414–15, 623, 631 334, 358–60, 370–1, 378, 457
Child language 4–6, 21, 52, 63, 92, 286, 293, 396, cohesive devices 19, 291, 312–18, 323, 326–7,
489–94, 502, 504, 651 (see also 358, 599
protolanguage) cohesive relations 60
Childhood-onset fluency disorder 606 (see collocation 17, 20, 145, 275, 317–18, 360, 591
stuttering) comparative reference 315, 361, 365, 370
children’s story 179–80, 188, 190 ellipsis 114, 288, 312–13, 359–60, 366–7, 591,
Chinese 22, 29, 49, 232, 399, 402, 564, 653, 668 593, 599, 655
Choice 7, 17, 19, 23, 35–6, 44, 92, 116, 148–9, endophoric reference 314–15, 361, 593, 598,
151, 156, 158, 160, 163, 173, 177–83, 185, 755
187, 189–90, 195, 197, 220, 224, 238, exophoric reference 314–15, 361, 591, 593,
273–5, 279, 340, 343, 346, 348, 350–1, 602
375, 403, 433–4, 436, 439–41, 462, 464, expectancy relations 317–18 (see also lexical
467, 473, 479–80, 520, 562, 566, 643–4, cohesion under cohesion)
671 demonstrative reference 315
Circumstance (as in grammatical unit) 38, 47, lexical chain 326–8, 360, 363, 605 (see also
93, 104, 113, 131–2, 240–1, 267, 360, cohesive chain)
367, 371, 435, 442, 445, 504, 516, 600, lexical cohesion 80, 288, 312–13, 316–18, 359,
745 363, 378, 525, 591 (see also lexical
circumstantial Adjunct (see under Adjunct) organization under cohesion)
Circumstantial Role (CR) 241, 251 lexical reference, lexical identity 312, 316
circumvenience 287, 296, 303, 305 lexical organization 312–13, 316–18 (see also
class 38, 121–3, 130, 243, 318, 471, 496, 623, lexical cohesion under cohesion)
630, 634 (see also rank scale) personal reference 315
classification 158, 211, 213, 250, 295, 312, 322, reference 314–15, 361, 593, 598 (see also
363, 564, 569, 578, 640, 768 anaphora, cataphoric reference, lexical
classification 340, 348, 350 (see also message reference, personal reference)
semantics) substitution 288, 312, 315, 360, 591
classroom 295–6, 348–9, 377, 392, 395, 515, 520, taxonomic relations 317–18, 363 (see also
668 lexical cohesion under cohesion)
clause 3, 39, 41, 59, 64, 73, 92–116, 118, 120, visual 451
127, 129–30, 137, 177, 188, 193, 236, cohesion in clinical situations (see also Alzheimer’s
241–5, 251–5, 267, 290, 359, 365, 378–9, Disease (AD), aphasia, Asperger Syndrome (AS),
384, 393, 415–17, 429, 435–6, 438, 445, autism spectrum disorder (ASD), schizophrenia,
474, 500, 552, 567–8, 574, 580, 591, 595–7, traumatic brain injury (TBI))
600, 603, 607, 623, 628, 630, 632, 637, 769 cohesive chain 318–21, 602, 605
(see also syntactic unit) cohesive harmony 4, 290–1, 317, 320–1, 334,
co-ordinated clause 187 602, 605, 660, 663 (see also traumatic brain
declarative clause 42, 107, 187, 253, 368, 595, injury (TBI) and cohesion in clinical situations)
780 colligation 17, 20
embedded clause 42, 707 collocation (see under cohesion)
major 106, 339 communicating body 26–7
minor 106, 339 communication 8, 16, 173, 265, 336, 348, 361,
projecting clause, projecting verb 64, 83, 399, 411, 423–4, 434, 436, 587–9, 591–4,
393–4, 435, 501, 506 596, 598, 601, 603, 605, 608, 610
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
Index 795
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
796 INDEX
dimensions 38, 58, 72, 83, 86–7, 174, 218, Ender (E) 253 (see also Starter (St))
267, 569, 592, 594, 623, 719, 722, 729, endophoric reference (see under cohesion)
741 engagement 301, 360, 373, 375, 383–4, 390,
discourse 205, 225, 285–307, 488, 500, 503, 505, 394–6, 398–9, 401, 755 (see also appraisal
507–8, 522, 526–7, 529, 537–50, 601, (system of ))
662–3 English for Specific Purposes 525
academic discourse 397 episode 438, 442, 447, 458, 660 (see also
as social practice 467 multimodal/multimodality)
discourse marker 128, 314 epistemic 108, 423, 600 (see modality)
discourse semantics 4, 78, 80, 162, 300–3, 334, ergative 39, 412, 429, 495, 641, 786
358–9, 362, 375, 378–9, 382, 384, 393, 402, Ethnicity 726
438, 442, 651, 660, 762 etiology 609
electronic discourse 718, 729 evolution 45–7, 339, 387, 411, 427, 622, 641,
Field of discourse 751 (see under Field) 771–2
HIV discourse 7, 656 (see also clinician–patient evolutionary systems 28
interaction) evolutionary thinking 50, 627
media discourse 383, 398, 651, 664–5, 667, evolutionary theory 47, 645
715, 717, 727 exchange 16, 20, 36, 40, 295, 360, 435, 572–3,
medical discourse 7, 663 754
metadiscourse 370–1, 376, 378, 663, 742 exchange structure 81, 297, 360, 594, 601, 727
Mode of discourse 468, 743 (see under Mode) Exegesis 760
multimodal discourse 5, 333, 400–1 existential process, existential clause 104, 435,
spoken and written modes of 597 704
Tenor of discourse (see under Tenor) exophoric reference 591, 593, 602 (see under
discourse analysis 4–5, 283, 285–7, 298, 333, cohesion)
348, 436, 608, 664, 773, 778 expansion 59, 75, 82–3, 318, 360, 401, 415,
discourse semantic systems 5, 358–9, 375–6, 439–40, 459
378–9, 384, 445 (see also discourse semantics experiential (meta)function 38–9, 82, 93–105,
under discourse) 108, 116, 240, 249, 267–8, 301, 340, 436,
discursive practice 467 (see also discourse) 745
display stratum 438 (see also multimodal/ exploring 158, 218 (see also field of activity)
multimodality) exponence 20, 244, 246, 254
distance 389–90 (see also graduation) , expounding 158–9, 218, 625, 664 (see also field of
doing 39, 95, 98–9, 154, 158, 218, 267, 435, 537 activity)
(see also field of activity) extra-linguistic environment 62, 140, 164
doctor–patient talk 348 (see also clinician–patient extravagance 71–2, 75–6
interaction) evaluation 81–2, 360, 373, 375, 378,
double articulation (as symbolic articulation) 47, 382–4, 386–8, 393–4, 396, 400–4, 592,
63, 690, 701 704
drift (in grammatical metaphor) 429, 522, 621
(see also grammatical metaphor) Facebook 7, 715, 729 (see also social networking
dynamic context 224–5 sites)
factorial explanation 373, 375
echolalic responses 594 Field 48, 85, 147, 150, 152, 154, 157–8, 161,
eco-social environment 62–3, 66, 85, 87 163–4, 186, 218, 221–2, 291, 303, 359, 361,
education 29, 537, 783 363, 365, 367–8, 379, 383, 392, 396, 402,
primary 6, 51, 544–7 404, 415, 426, 436–7, 457, 468–9, 475, 480,
secondary 6, 51, 547–8 518, 539 (see also context of situation)
tertiary 6, 397, 537, 550 field of activity 159, 223, 537 (see also registerial
educational context 163, 327, 396 cartography)
educational discourse 396 (see also appraisal and figure 82, 359–60, 365, 367, 391, 438, 442, 446,
educational discourse) 458 (see also multimodal/multimodality)
educational linguistics 52, 277, 383 filling 242–4, 246, 254
elaborated code (see under code) film 313–16, 320–1, 577, 754
element (of syntax) 82–3, 105, 111, 113, 123–7, Finite 40–1, 106–7, 113, 126–7, 289, 366–7,
130, 133–4, 176, 188, 238, 241–55, 313, 412–13, 418, 422, 629, 704, 781
319, 338, 347, 376, 413, 416, 441, 443, 445, First-order context 164 (see also context of
451, 458, 745 situation, second-order context)
element (of intonation) 181, 183, 185, 290 Firthian influence 11–30, 728
ellipsis 591, 593, 599, 655, 752–3 (see under Flickr 577
cohesion) Focus
embedded clause (see under clause) information focus (system of ) (see under
emergency medicine 655, 657 (see also medical information)
discourse) in graduation network 5, 389, 391 (see also
emoticon 719, 727 graduation)
enabling 158, 218, 565 (see also field of activity) foot (in system of intonation) 176–7, 185, 196,
end of life care 658 (see also clinician–patient 199, 379
interaction, medical discourse) foot composition 177–8, 182
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
Index 797
Ictus, silent ictus 175, 177, 182–3, 185 hashtag 399, 717 (see also metadata, social
ictus state 177–8 tagging)
Remiss 178, 182–3 health/healthcare 348, 397–8, 404, 651 (see also
tonic foot 183 medical discourse, sexual health discourse)
foregrounding 41, 64, 600, 691–2, 695–6, 706, hedging 412, 427
709 heteroglossia 394–5, 397 (see also monoglossia)
foreign language teaching 4, 771 (see also second high autonomous professional (HAP) 464 (see
language acquisition, second language also code, low autonomous professional (LAP))
development/learning/teaching) high-functioning autism (HFA) 593
formalist models 260, 262, 264–5, 691 (see also Hindi 777
Russian formalist approaches) historical linguistics 49, 410, 771–2, 774 (see also
fractality 75, 77 old English)
French 411, 421, 427–8, 526, 564, 653 history (language of ) 46, 372, 397–8, 402, 428, 772
functional models 35–52, 230–1, 259–60, 262, history texts 163, 428, 548
264–5, 268, 664, 780 HIV discourse 656 (see under discourse)
functional syntax 4, 230–55 homophora 361, 365
functional-cognitive space 4, 259 human language 12, 628
functionality 35, 493, 574 (see also systematicity) human–wheelchair interaction (see autonomous
wheelchair)
gender 21, 719, 722, 753, 769 humour 423, 656, 719 (see also patient-initiated
generic stage 23, 217, 750–1, 753 humour)
generic structure 134, 150, 216–17, 219, 224–5, hybridity 627
248, 373, 601 hybrid voices 304
generic structure potential (GSP) 154, 223, Hypertext 441, 443, 755, 760
291–3, 601
genetic counselling 663, 674 (see also Ictus, silent ictus (see under foot)
clinician–patient interaction, medical ideational 35, 38–9, 42, 49, 65, 82–3, 149, 187,
discourse) 225, 267, 274, 287, 291, 339, 359, 382–3,
genitive cluster 248–9 393, 403, 435, 443–6, 448–51, 454–5,
genre agnation 160, 163 457–8, 604, 725, 783
genre-based literacy programs 362, 375 (see also ideational function in clinical linguistics 517 (see
genre pedagogy) aphasia)
genre (as text type) 160, 215, 217–22, 265, 293, ideology 20, 50, 398–9, 760
298, 300–2, 375–7, 518, 539–40, 588, 675, identity (social) 6, 35, 146, 398, 402
752 identity chain 319, 364, 366, 717–20, 722, 724,
genre (as extra-linguistic stratum) 161–2, 728 (see also cohesive chain)
216–17, 221, 358, 379, 382, 433, 436–7 identification 301, 334, 339, 358–9, 361, 365,
genre pedagogy 375, 545 370, 378, 588 (see also discourse semantics)
gestalt theory 439 IELTS 550
gesture 400, 434, 452, 575, 725 image–text combination 441, 443 (see also
Given, Givenness 41–2, 111–12, 185, 188, 267, intersemiosis, intersemiotic complementarity,
415, 417 (see also New) multimodal/multimodality)
given information 41, 594 (see Given, Givenness) imperative 40–1, 60, 64, 105–7, 187, 326, 453–4
globalization 784 indeterminacy 45, 69, 160, 268
Goal 336–7, 429, 445, 552 indicative 106, 345–6, 745, 781, 786
grammatical 39, 95–7, 114, 546 individuation 21, 715, 721–2
communicative 186, 296 Indo-Aryan languages 770
goal orientation 157, 301, 751 Indo-Iranian languages 773
graduation 5, 301, 340, 360, 384, 386, 388–93, ineffable 13, 26, 625
399–401 (see also appraisal (system of )) infographic 433, 450, 459
grammaticalization 110, 272, 412, 425, 427 Information 177, 181, 187, 189, 297, 343, 361,
grammars (types of ) 358, 740 367, 378, 436, 442, 745, 780 (see also Given,
notional 207, 212 New)
formal 15, 212, 236, 242, 248 information distribution (system of ) 174,
functional 206, 208–9, 212, 230, 232–3, 237, 181, 189, 267, 745
242, 312, 322, 776 information flow 42, 234, 346, 361, 417, 436,
grammatical intricacy 596, 607 451, 453, 746
grammatical metaphor 622–3, 741 (see under information focus (system of ) 41, 171, 181,
metaphor) 185, 189, 196, 240–1, 438, 442
grammatical structure 60, 74, 79, 122–8, 242–3, information giver 236, 241, 252
274, 292, 320, 379, 630, 779 information group 177
grammatics 37, 546 information grouping 177
graphology/graphological systems 38, 67, 148, information prominence 177, 368, 445–6
162, 174, 302, 363, 384, 438, 701 information status 240, 425
Gricean maxims 16 information structure 41, 190, 193, 222, 279,
group 3, 174–7, 179, 183, 185, 187, 267, 359, 361, 415, 746
365, 390–1, 393, 414–15, 420, 428, 438, information unit 41, 111, 175–6, 181, 185,
454, 707, 780 (see also phrase, rank scale) 188, 193, 197
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
798 INDEX
information thinking 627 Japanese 232, 399, 401, 524, 653, 781
Instagram 7, 577, 723, 729 (see also social journalistic discourse 4, 348, 350–1
networking sites) judgement 360, 373, 385, 387, 397–8, 623, 633,
instance, instantiation 35, 66, 70–2, 145, 149–51, 704 (see also appraisal (system of ))
162, 177, 231–2, 268, 270, 287, 338, 384–5, juncture 180 (see also accent)
388, 390, 393, 402–3, 519–20, 644, 703,
721, 730 key 187, 194 (intonational system of, choices in)
instantiation effect 44 KPML 563–4 (see also natural language generation)
instantiation hierarchy 162
institutional linguistics 277, 652 L2 6, 512–28, 530 (see second language acquisition,
institutionalization 152 second language development/learning/
intelligent wheelchair 573 (see autonomous teaching)
wheelchair, dialogue systems) language-based knowledge construction 526
interpreting 7, 661–2, 742–3, 758–9 (see also language-based linguistics 3, 12
translation) language change 45–8, 271, 286, 569, 762
intensity 389, 391–2, 626 language learning 4, 52, 58, 396, 487, 509, 512,
intercultural communication 755, 762, 767 (see 515, 519, 521, 523–4, 530, 537, 538, 541,
also cross-cultural communication) 546, 550, 553, 784 (see also child language,
interpersonal analyses in clinical linguistics (see education, educational linguistics, ontogenesis,
aphasia, traumatic brain injury (TBI)) second language acquisition, second language
interpersonal (meaning, metafunction) 12, 20, development/learning/teaching)
40, 48, 77, 80–2, 104–10, 116, 186, 225, language typology 402, 776, 778, 782, 785 (see
240, 244, 267–8, 287, 291, 301, 349, 358, under typology)
360, 379, 382–4, 393–4, 397, 400–3, 412, langue 1–2, 44–5, 152, 641
427–8, 435–6, 442–4, 453–4, 458, 517, latent patterns/patterning 28
630, 662, 675, 716–17, 724 Legitimation Code Theory (LGT) 464
interpersonal interactions 600–1 (see Alzheimer’s levels of analysis 2, 12, 19, 24, 173, 205, 700
Disease) lexical density 190
interpersonal metaphor (see under metaphor) lexical cohesion 525, 591 (see under cohesion)
interpersonal prosody 393 lexical metaphor (see metaphor)
interpersonal resonance 403 lexicogrammar 37, 92–116, 162, 174–5, 181,
intersemiosis 433, 440, 447, 455–6, 663, 674 (see 230, 235–6, 273–4, 289, 303, 324, 333, 339,
also multimodal/multimodality, 341, 351, 382, 384, 393, 438, 487, 570, 580,
resemiotization, semiotic) 632, 637, 702, 744, 756, 780
intersemiotic complementarity 440, 444 (see linguistic criticism 694 (see also literary criticism,
also intersemiotic relations, multimodal/ stylistics)
multimodality, resemiotization, semiotic) literacy 349, 542, 657, 665, 675 (see also
intersemiotic complementarity framework (see recognition literacy, action literacy, reflection
intersemiotic complementarity) literacy, learning to read/write)
intersemiotic relations 439, 444, 446–7, 450, 452 literary language, literary text 49, 218, 692, 700
(see also intersemiotic complementarity, literary theory 621, 691
multimodal/multimodality, resemiotization, literature 7, 690, 692–5, 698 (see also poem, poets/
semiotic) poetry, verbal art)
intersemiotic downranking 449 living language 410–11
intersemiotic package 450 logical metafunction 82, 122–5, 127, 133–4, 136,
interstratal relationship 148, 156, 220 (see also 149, 187, 267, 301, 340
intrastratal relationship) logical metaphor (see under metaphor)
intersubjectivity 27, 145, 490, 724 logogenesis 23, 152, 286, 709
interventionist medical linguistics 653 (see also London School 14, 145
medical discourse) low autonomous professional (LAP) 294–5, 494
intonation (in clinical linguistics) (see also (see also code, high autonomous professional
Asperger Syndrome, high-functioning autism (HAP))
(HFA))
intonation (features of, theory of ) 3, 25, 40–1, macrofunction 494–5 (see also microfunction, child
253, 297, 360, 438, 593–5 (see also tone, language)
tonality, tonicity) macrophenomenon 101–2
intonation (and literacy, learning to read) 655 Maithili 770
intonation systems 178–9 (see also intonational markedness 41–2, 44, 104, 368, 429, 595, 762
systems, tone, tonality, tonicity) marked Theme (see under THEME)
intonation unit 190 (see also tone group, tone) Marxism 49–51, 475
intonational systems 173, 177 (see also intonation material situational setting (MSS) 22, 152, 295
systems) mathematical discourse 425, 547
intrastratal relationship 148 (see also interstratal mathetic function 495 (see also macrofunction,
relationship) microfunction)
item (word, morpheme, or punctuation) 127, matrix 742
144, 242, 244–5, 290, 311, 313, 317–18, experiential matrix 17, 26
591, 605, 623, 772, 784–5 matrix RU 294
iterative contextualization 739–40 maxims (Gricean) (see Gricean maxims)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
Index 799
meaning (definition of in SFL) 56–8, 143, 173, Mood 187, 267, 289–90, 567, 780–1 (see also
207–9, 232, 236–8, 240, 244, 254–5, 266, mood (system of ), lexicogrammar)
303, 384, 390, 520–1, 595, 601, 604 Declarative Mood 40–1, 105–7, 267, 289, 781
meaning potential 35–7, 43–5, 58, 68–70, 85, Imperative Mood 105–6, 267, 289, 453–4
148–9, 162, 242, 274, 289, 292, 333, 338, Interrogative Mood 105–7, 267, 289, 781
384, 398, 400, 434, 458, 621, 623, 639, 654, mood (system of ) 38, 40, 60, 64, 66, 80, 105–8,
693 149, 240–1, 255, 347 (see also Mood,
media discourse (see under discourse) (see also lexicogrammar)
radio broadcasts/commentary) mother–child interactions 4, 6, 80, 293, 334,
medical consultation 654 (see also clinician–patient 338, 341–8, 396, 544 (see also child
interaction, doctor–patient talk) language, primary carer)
medical discourse (see under discourse) mother tongue 487, 507
medical emergency teams 666 (see also surgical motion verbs 782–3
teams, medical discourse) move (in discourse) 81, 153, 160, 290, 296–9,
medical interpreting 661 360, 575
melody (in speech) 175–6 multifunctional 3, 250, 494, 744
meme 716, 727, 728 multilingual text production 7, 746, 749–50 (see
mental health 659–60, 664, 675 (see also health, also interpreting, translation)
medical discourse) multiliteracies 546
message 4, 290–1, 500 (see also message analysis Multimodal Analysis Image 441
and message semantics) Multimodal Analysis Lab 575
message analysis 500–1 (see also message multimodal corpora 578–9
semantics) multimodal/multimodality 293, 307, 319, 327,
message semantics 335, 338–9, 341, 348, 350, 400, 430, 433, 575
666 SF-MDA 433–40, 444, 458
metadata 715, 716, 725 (see also social networking SF-MDA Framework 439, 576–7
sites, social metadata, social tagging) multimodal semiosis (see multimodal/
metafunction 400, 403, 433, 435–6, 440, 503, multimodality)
517–18, 567, 587, 654, 658, 725 multiperspectivism 71–5
metafunctional shift 742 multiscalar model of context 164
metalanguage 13, 641, 700 multisemiotic 459, 541, 547–8 (see multimodal/
metaphor 621, 633, 669 multimodality)
conceptual metaphor theory 476 multi-stratal 174, 502, 644, 656, 660
grammatical metaphor 6, 46–7, 63–6, 72–4, multivariate structure 119, 124, 133–5 (see also
213–14, 266, 276, 371, 378, 397, 521, univariate structure)
622–3, 628, 637, 645, 741, 747 mutual expectancy 17–22
interpersonal metaphor 64–5, 360
lexical metaphor 46, 64–5, 81, 392, 401 narrative, narrating 43, 760
logical metaphor 360 native speaker, non-native speaker 180, 327, 642
metaphorization (in translation) 622–3, 747 natural language generation 235, 324, 562–3 (see
metaredundancy 3, 66–9, 72, 75, 86, 88 also KPML, Penman project)
metarepresentation 501 Negator (N) 244, 253
metaphenomenon 101–2 negotiation 81, 289, 301, 383–4, 654, 666 (see
fact 101 also discourse semantics)
idea 101 Neurocognitive disorders 598
metastability 66–9, 88, 149 Neurodegenerative disease 598 (see Alzheimer’s
microblogging 716, 724 (see also social networking Disease)
sites) neurodevelopmental disorders 592
microfunction 490 (see also macrofunction, New 111–12, 185, 188, 193, 267 (see also Given,
protolanguage) Theme)
microplanner 235 Contrastively New, Contrastive New (CN) 240–1
Middle English 410, 425, 623 Culmination of New 185, 188
modalization battery 600 (see also Alzheimer’s Hyper-New 361, 369, 394
Disease) macroNew 445, 448, 457
modality 40, 66, 80, 108–10, 267, 435–6, 654, New1 186
693, 745 New2 186
modalization 108–10, 391, 423, 592, 600 Unmarked New (UN) (unmarked newness)
modulation 108–9, 389, 423 188, 240
Mode (contextual) 17, 147, 149, 163, 165, 190, New Criticism (literary) 692, 695 (see also literary
219, 303, 399, 402, 468–70, 475, 480, 518, criticism)
539, 627, 633, 754, 758, 779 New information 445, 451, 454 (see New)
channel 17, 186, 436, 630, 638, 754, 759 Newton (Isaac) 415, 426, 624, 628, 631
medium 85, 399, 429, 573, 668, 754 nominal group 101, 103, 122–3, 129–30, 135–6,
role of language 217, 219, 469, 515, 620, 651, 177, 208, 238, 244–6, 248, 251, 292, 415,
754 454, 622, 637 (see also Epithet, Deictic,
rhetorical mode 218–20, 222, 291 Modifier)
Modifiers 124, 135, 246, 420, 428 Thing vs Head 125, 236–8, 245–6, 292, 420,
monoglossia 394 (see also heteroglossia) 452, 629
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
800 INDEX
nominalization 46, 266, 413–14, 417, 634, 636, phonetics 25, 174, 333, 571, 587–8, 604, 607
660 phonological paragraph 177, 190 (see also
non-discreteness of language 268 paraphone)
non-literary stylistics (see under stylistics) phonological rank scale 121, 172, 176
non-verbal action 144 (see also verbal action) phonological realization 37, 187
nursing 658 (see also medical discourse) phonological system 40, 378
phonological units (in English) 176
observation 642, 644 phonology 162, 171–2, 174–7, 179, 207, 273,
Òkó 782 333, 341, 359–60, 378, 520, 589, 604, 607
Old English 411–12, 427, 430 (see also segmental phonology)
oncology 656 (see also clinician–patient interaction, photograph (analysis of ) 346, 411, 426, 433, 436,
medical discourse) 441, 444–6, 575–6 (see also multimodal/
onset (in phonology) multimodality)
onset syllable 190 phrase 118–19, 121–3, 129, 131, 133, 135–6, 138,
rise–fall hook onset 192 140, 379, 393, 420, 438, 564, 567, 600, 655
fall–rise hook onset 192 (see also group, rank scale)
onset (in clinical linguistics) 595 phylogenesis 45, 152, 286
child onset fluency disorder 606 (see also picture books 349, 400, 546
stuttering) pitch 173, 182, 187, 189, 191–2, 196–8, 438
ontogenesis 63, 152, 286, 396, 487, 721 (see also pitch contour 174, 181, 183, 185, 191–2, 194–5,
child language, protolanguage) 197–9
open context 338, 351 pitch range 191–2, 196, 198
operator 127, 236–8, 244, 252, 254, 412–13 Pitjantjatjara 780
oral-literate continuum 525 planner 234–5, 238 (see also microplanner)
orthography 730 discourse planner 235
over-lexicalization 466 sentence planner 233, 236
poem 691, 702
paradigm 14, 26, 339, 610, 622 poets, poetry 625, 707
paradigmatic 18, 35, 43–5, 58, 156, 181–2, 223, polarity 240–1, 268, 345–6, 385, 507, 704
274, 278, 318, 351, 478, 520, 566, 568 (see polysystem/polysystemic 11–26, 172
also syntagmatic) population thinking 627, 640
paragraph, paragraphing 177, 190, 359, 363, posture 17, 19, 26, 400, 493, 642 (see also
375, 444, 751–2, 755 communicating body)
paralinguistic 726, 732 power (in language) 298, 305, 341, 348, 360, 436,
parallelism 629, 754 (see also foregrounding) 462, 467, 472, 543, 548, 753 (see also
grammatical parallelism 695 critical discourse analysis (CDA), critical
pervasive parallelism 691–2, 706 linguistics)
paraphone 177, 190 PRAAT 191
paraphrase 347, 742–3, 747 Practical Criticism (literary) 692 (see also literary
parole 1–2, 44–5, 152 criticism)
parser/parsing (in computational linguistics) 565 pragma 11–18
(see also computational complexity) Pragmatics 15, 23, 265, 274, 289, 473, 503,
automated parsing 569 587
Stanford Dependency Parser 567 pragmatic conversation 306 (see also casual
systemic parsing 567, 570 conversation)
Universal Dependency 564, 567 Prague School linguists 692
Participant Role (PR) 147, 214, 235, 241, 249–51, Predicator 106, 244, 412–13, 502, 707,
253, 435, 442 769
patient-initiated humour 656 (see also humour) prefaced 506–7 (see also semantic networks)
passive voice 104, 465, 525 prehension 18, 23
pattern 14, 161, 180, 189, 216, 220, 223, 265, prepositional group 123, 128, 132, 246,
275, 292–3, 297, 299, 305–6, 339, 479, 543, 251
550, 577, 590, 592, 600, 607, 610, 620, 627, prepositional phrase 119, 126, 129, 136, 138,
644, 654–5 244, 414, 521, 629–30
pedagogical discourse 348 pre-text context 165 (see also context of situation)
Penman project 231, 564 pretonic 184–5, 190–2, 194, 196–9 (see also
Performer 235 (see also addressee) tonic)
periodicity 86, 334, 358, 361, 371, 378 (see also pretonic prominence (see under prominence)
discourse semantics) pre-translational text analysis 750–1
permeability 155–6 preventive medicine 668 (see also medical
personal reference (see under cohesion) discourse)
pervasive parallelism (see under parallelism) primary carer 499 (see also child language)
phasal shift 224 primary progressive aphasias 609
phasal analysis 224–5, 334, 594 principal components analysis (PCA) 20
phase (of discourse), phase analysis 14, 224–5, probability 435, 591
298, 304 probability theory 568
phoneme 25, 176, 179, 379, 638, 761 probes 95–7
phonetic expectancies 19 process test 95, 99, 251–3
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
Index 801
process type 212–14, 238, 240, 250–1, 316, 422, realization rules or statements 56, 179, 209, 223,
435, 445, 579, 600, 604, 703 236–44, 255, 273–4
behavioural 104, 435, 447, 604, 703 same pass rules 238
existential 104, 422 preselection 238, 241
material 212–14, 238, 240, 250–1, 316, 415, realizational systems 644, 675
419–20, 422, 435, 445, 463, 495, 498, 553, recognition literacy 542–3 (see literacy)
626, 704 recommending 158, 218, 664 (see also field of
relational 102–4, 267, 422, 435, 449, 604–5, activity)
637, 703 reconstrue 39, 527
attributive 102, 445, 457 recoverability 41, 94, 106, 240
identifying 103, 435, 454, 457, 637 recreating 158, 218 (see also field of activity)
mental 99–102, 238, 378, 383, 386, 422, 516, reference chain 314, 327, 361 (see also cohesion)
604, 704 reflection literacy 542–3, 546–7 (see also action
verbal 104, 209–11, 214, 422–4, 478, 522–3, literacy, recognition literacy and literacy)
542, 546, 604, 703 register 29, 45, 71, 83–5, 145–6, 180,
progressive message 339 (see also message 214, 216–17, 219–24, 235, 265,
semantics) 290–1, 300, 303, 358, 379, 382, 426,
projecting verb (see under clause) 436, 518, 525–6, 538, 540, 542, 547, 552,
projection 39, 59–60, 75, 83, 160, 209–10, 394–5, 579, 666 (see also situation, context of
415, 507, 516, 661, 780 situation)
prominence 445–6, 696 (see also salience, register variation 17, 19, 98, 181, 190, 225,
tonicity) 296, 411, 457, 509, 552, 627, 659, 668,
tonic prominence 41, 178–9, 181–2, 185 742–3
textual prominence 193, 394 registerial cartography 157, 160, 552
prominent 177, 448, 458, 655 (see also Relevance theory 474, 749
prominence, information prominence) relevancy 145, 147, 149, 152, 164
prosodic analysis 24, 172 Remiss (see underfoot)
prosodic composition 180, 185 reporting 158, 218, 390, 394, 433, 444, 664–5,
prosody 372–3, 393–4, 465, 608 (see also semantic 675 (see also field of activity)
prosody) resemiotization 433, 439–40, 719 (see also
Proto-Indo-European 641 intersemiosis, multimodal/multimodality,
protolanguage 396, 489–94, 502, 504 (see also semiotic)
child language) resonance 403
protolinguistic sign 489–90, 492–3 (see interpersonal resonance 403
protolanguage) fractal resonance 75–6
protoscience 620 restricted code (see under code)
punctuation 179, 597 restricted language 16–19, 26 (see also register)
punctuative message 501–2 (see also message Rheme 42, 111–16, 188, 267, 294–5, 313, 413,
semantics) 415, 417, 425, 445, 524, 527 (see also
theme, lexicogrammar, New)
quality group 246–7 rhetorical mode (see mode)
quantity group 248–9 rhetorical relations 324–8 (see also Rhetorical
Structure Theory (RST))
radio broadcasts/commentary 180, 189 (see also nucleus–satellite relations 299–300, 325
media discourse) multi-nuclear relations 299, 326
Range 129 Rhetorical Structure Theory (RST) 4, 29, 83, 153,
prepositional phrase as 126, 130 159, 298–300, 324, 334
rank scale 38, 82, 120, 123, 131, 133, 137, 176–7, Rhetorical Unit (RU) 153, 155, 293–6, 334
242, 338, 706 rhythmic unit, unit of rhythm 177 (see also foot,
phonological 121, 177–8, 182, 341, 347, 438, rhythm, syllable)
495, 599, 655 rhythmic patterns 180
lexicogrammatical 120, 122, 125, 129–30, 132, rhythmic properties 27
135–6, 155, 336–7, 345, 378, 393, 491–2, role enactment 86, 340 (see also message
500, 502–5, 515, 517–18, 522, 539, 569, semantics)
572, 592, 596, 604, 668 role of language (see under Mode, context of
ranking clause 359, 367, 501 situation)
rankshift 38, 124, 134, 136, 138, 140, 426, 629 Rolland 573 (see autonomous wheelchair, dialogue
reactance (grammatical) 206, 208, 212 (see also systems)
realization)
reaction (in appreciation network) 378, 388 salience 185, 425
reading 473, 541, 604 (see appreciation) salient syllable (see under syllable) (see also
reading aloud 179–80, 189 Tonicity)
reading comprehension 540 Sanscrit/Sanskrit 773
reading to learn pedagogy 540 scale-and-category linguistics 119, 740
realization 37, 47, 56, 67–8, 148–9, 151, 155–6, scale of delicacy 145
161–2, 174–5, 187, 207, 219–20, 222, 226, schizophrenia 590
232, 248, 254, 299, 304, 401, 502, 644, 662, schizophrenic discourse/speech 591, 608
752 (see also realization rules) (analysis of, features of)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
802 INDEX
science 383, 418, 475, 477, 479, 519, 524, 542, social activity 148, 154, 286, 514, 516, 751, 753,
545, 547–8, 551, 621 758
register 419, 475, 477, 479, 551, 579, 666 social class 21, 342
experimental science (discourse of ) 415, 628 social distance 165, 344, 347, 594, 659, 754
scientific developments 418, 420–1, 631 social media 399, 664, 715–16, 718–19, 722
scientific discourse 46, 52, 74, 414, 416, 428, social network 727
546, 548, 620, 623, 633 social networking sites 715
scientific English 414–15, 638 social processes 13, 158, 161, 277, 437, 440, 539,
scientific explanation 622, 631 731
scientific research article 418 social purpose 7, 292, 518, 539, 550, 627
science of meaning 15, 49, 644 social semiotic(s) 266, 269, 398, 401, 436, 440,
scientific themes 626 462, 474, 479, 487–8, 498, 514, 517–18,
scope 389–90 (see also graduation) 537–8, 542, 548, 550–1, 576, 720
Scope (grammatical element) 96–7, 248 language as 2, 488, 498, 514, 517–18, 537–8,
second language acquisition (SLA) 6, 512 542, 548, 550–1
second language development/learning/teaching theory of language 1–2, 233, 266, 279, 286,
349, 512 462, 474, 479, 487
second-order context 47, 164 (see also context of social metadata 716–17 (see also metadata)
situation, first-order context) social tagging 715, 717 (see also metadata)
segmental phonology 179 sound wave 754, 759
selfies 723 (see also identity) source text 744–7, 760–1, 782 (see also target text,
semantic cluster 19, 508 translation)
semantic component 59, 75 Spanish 113, 401, 645
semantic domain 388, 468, 470, 770 Specific Language Impairment (SLI) 593
transcategorical semantic domain 75, 82 speech function 267, 296, 334, 347, 360, 442,
semantic drift 595, 660 453, 594 (see also metaphor of mood)
semantic networks 44, 60, 77, 236, 333–41, 343, speech sound disorder 596
348–51, 502–8, 670 (see also message statements of meaning 12, 18
semantics) stratification 3, 56–8, 60–6, 220, 519–21, 632,
semantic prosody 13, 20, 28, 81 657, 721, 741
semantic stratum 44, 56, 58–63, 222, 289, 291, streaming 718 (see also social media, social
500, 550 networking sites)
semantic variation 21, 29, 334–5, 341–2, 347–8, stroke 589, 601 (see also acquired brain injury
464, 480, 496–7, 501, 503–5, 508–9, 538, (ABI))
544, 657, 770 structuralist theory 58, 171–2, 209
semantics 55–88, 251, 254, 265, 273–4, 289, 300, stuttering 606 (see also onset (in clinical linguistics))
333–4, 340–1, 351, 358, 473, 494, 508, stylistics 276, 400, 622, 690
520–2, 524, 542, 545, 741, 744 British stylistics 690
as discourse-structural meaning 60, 508 non-literary stylistics 693
as higher-level systemic meaning 59, 158, 174, structural stylistics 691
494, 520–2, 524, 542, 545 Systemic Socio-Semantic Stylistics 690
as topological meaning 60, 473 Subject 4, 41–2, 106–7, 113–15, 193, 238, 241,
semiotic 36, 285, 398, 462, 473–4, 478, 488, 244, 251, 253–4, 289, 316, 368, 412–13,
493–4, 508–9, 515–17, 520, 525–6, 529, 421, 429, 472, 539, 567, 642
537–8, 552, 726 supervenience 287, 305, 307
artefact 458, 729 Support Vector Machine (SVM) 577
mediation 497, 516, 544, 552 suprasegmental 171
processes 158, 223, 439, 474, 540, 547, 664 (see surgical teams 666 (see also medical emergency
also social processes) teams)
resources 37, 426, 433–4, 436, 450, 458–9, syllable 25, 176, 178–9, 181–3, 185, 193, 368, 379
540–1, 547, 576–7, 722, 727 (see also intonation)
systems 36–7, 49, 63, 67, 85–6, 153, 160–1, salient syllable 176–7, 182, 185
216, 221, 233, 271, 287, 293, 402–3, 434, weak syllable 177, 182
437, 495, 542, 636, 701, 706, 742 syllable-initial consonant (see under consonant)
semogenic 36, 634, 638 syllable margin 179 (see also syllable-initial
semogenic power 47, 68, 628 consonant, syllable-final consonant)
serial contextualization 18–19 symbolic articulation 7, 622, 695, 701–2, 707 (see
sexual health 669 (see also health, medical double articulation)
discourse) synchronic linguistics 410
SF-MDA 576–7 (see under multimodal/ syntagmatic 43, 58, 72, 177, 182–3, 186, 274,
multimodality) 318, 440, 479, 508, 520, 566, 706 (see also
signalling 323–4 paradigmatic)
silent ictus (see under foot) syntax 464, 487, 489, 523, 642–3, 777, 780
situation 468, 470, 472, 475, 503, 518, 571, 671, system network 43–4, 58, 69, 85, 94, 156–7, 163,
754 (see context of situation) 177–9, 182–3, 220, 231–2, 235–44, 254, 266,
situation type 146, 155, 518 (see context of situation) 268, 274, 279, 297–8, 334, 336, 338, 351,
SLA 512–16, 520, 523, 527–31 (see second 384, 395, 400, 434, 441, 492, 503, 569, 590,
language acquisition) 658 (see also contextualization system networks)
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
Index 803
systematicity 493 (see also functionality) tone unit 174–8, 181, 183–4, 189, 594
Systemic phonology 171, 177 (see also phonology) tonic 368 (see also tonic syllable, tonicity)
tonic foot (see under foot, see also tonicity)
tag phrases 107 tonic prominence (see under prominence, see also
Tagalog 777 tonicity)
target text 742 (see also source text, translation) tonic syllable (see under tonicity)
taxis 19, 38, 82, 435, 703 tonicity (system of, choices in) 174, 178–81,
Thai 781 186, 189, 368 (see also pre-tonic, pre-tonic
teacher education/teacher training 529, 537, prominence)
539, 550–1 unmarked 368
teaching–learning cycle 539 tonic syllable 368
Telugu 780 topology 375 (see also typology)
Tenor 48, 84–5, 147, 161, 164, 217–20, 225, 291, trace 164, 420, 440
293, 303, 360, 379, 426, 436–7, 443, 453, transdisciplinary 417, 462, 473, 488, 497, 499,
468, 643, 746, 751, 753 (see also context of 514, 528, 550, 644
situation) transformational generative grammar 29, 776
text analysis 359, 362, 379, 463–4, 541, 751 (see (see TG grammar)
also pre-translational text analysis) transformative/creative system 97–8
text-based transitive vs intransitive 95, 412, 425, 429
curriculum 540 Transitivity 95, 211, 213–14, 231, 240, 244,
pedagogy 541 249–51, 254, 255, 267, 317, 322, 429, 464,
syllabus 540 467, 567, 600 (see also lexicogrammar)
teaching 552 Transitivity analysis in clinical linguistics 702
text synthesis 362 (see Alzheimer’s Disease)
TextTiling 327 translation 276 (see also interpreting)
textual metafunction 358, 413, 436, 517, 593, equivalence 744–6
624, 744 hermeneutics of 760
textual function (in clinical linguistics) 503, 595 literal 744
texture 41, 60, 80, 154, 163, 288–9, 291, 293, machine translation 327, 561, 741
311–12, 319, 326, 334, 384, 437, 440, 577, process 748
638 quality assessment 741, 755
thematic structure 40–1, 222, 411, 415, 421, 425, translator 675, 749
429, 436, 746 traumatic brain injury (TBI) 601 (see also acquired
thematic progression 42–3, 328, 417–18, 668, brain injury (ABI))
752 trinocular perspective 62–3, 158, 207, 348, 518,
theme 41–3, 66, 111–16, 149, 188, 240–1, 538, 552
255, 267, 294, 346, 361, 367–9, 371, Twitter 7, 399, 576, 716 (see also social networking
394, 413, 415, 417–18, 420–2, 425, sites)
436, 442, 445, 448, 451, 458 (system of, tweeting 361
choices of see also lexicogrammar, type 151, 243, 567
New, Rheme) ‘typical actual’ 15–16, 18, 30
adjunct Theme 421–2 Typology 375, 402 (see also topology)
marked, unmarked Theme 41, 113–15, 240, Formal approach to language typology 776
595–6, 745 Functional approach to language typology 776
and New, and Information 42, 79 History of language typology 775–82
experiential Theme 42 Language typology 269, 402, 776
hyper-Theme 43, 361, 369, 371, 393, 445, 448, Typology of genres 160, 375 (see also genre)
453–4, 527, 549 Typology of grammatical metaphors 634 (see
interpersonal Theme 42 also grammatical metaphor)
macro-Theme 83, 361, 371, 373, 375, 445, 448, Typology of registers 157 (see also register)
453, 457
topical Theme 445, 448 UAM Corpus Tool 564, 758 (see also corpus)
textual Theme 42, 313 unit 566, 630 (see rank scale)
threat (semantic network category) 336–7 unit of analysis 4, 325, 338, 474
token 445, 449, 567 (see also cohesive harmony) universal grammar 12, 15, 27, 642–3, 776
central token 290, 321 univariate structure 119, 133–6, 140 (see also
relevant token 290, 321 multivariate structure)
peripheral token 290, 321 unmarked New (UN) (see under New)
tonality (system of, choices in) 178–81, 186–7, unmarked stress 593
189 unmarked Theme (see under THEME)
tone (system) 174, 178, 180–1, 184, 186–7, 194,
345, 347, 379, 428 valeur 14, 19, 25, 56–9, 69, 641, 721
compound tone 181, 183–4, 191, 193–4 valuation (in appreciation network) 373, 388
primary tone 181, 184, 191, 655 (see appreciation)
secondary tone 195, 197 verbal action 144, 210, 660
simple tone 193 verbal art 7, 46–7, 628, 633, 639, 690–1, 698, 701
tone group 111–12, 121, 174, 176, 178–9, verbal group 18, 123, 126–7, 130, 135–6, 138,
182–3, 185, 187, 190, 193, 593 244–5, 359, 391, 629, 637, 769, 772
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032
804 INDEX
verbal science 7, 46, 621, 623, 638, 645 webpage 433, 435–7, 439–57
vernacular 620–2, 641 Weibo 715 (see also social networking sites)
via-text context 165 (see also context of situation) WHO 433, 435, 437, 439–41, 454
Vietnamese 401 word phonology 179
visual display 444, 447, 452, 457, 547 work 479 (see also multimodal/multimodality)
voice 104, 465, 524 (see also active voice, passive writing 361, 365, 376, 520, 604, 754
voice) written language 521, 525–6, 539, 544,
voice quality 172, 180, 182, 198–9 575–6
Downloaded from https://www.cambridge.org/core. Nottingham Trent University, on 20 Aug 2019 at 13:21:23, subject to the Cambridge Core terms of use, available
at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781316337936.032