Critical Discourse Studies

ISSN: 1740-5904 (Print) 1740-5912 (Online) Journal homepage: http://www.tandfonline.com/loi/rcds20

Towards critical multimodal discourse analysis: a

response to Ledin and Machin

John A. Bateman

To cite this article: John A. Bateman (2018): Towards critical multimodal discourse analysis: a
response to Ledin and Machin, Critical Discourse Studies, DOI: 10.1080/17405904.2018.1550430

To link to this article: https://doi.org/10.1080/17405904.2018.1550430

Published online: 24 Nov 2018.

Submit your article to this journal

Article views: 2

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rcds20
CRITICAL DISCOURSE STUDIES
https://doi.org/10.1080/17405904.2018.1550430

Towards critical multimodal discourse analysis: a response to

Ledin and Machin
John A. Bateman
Faculty of Linguistics and Literary Sciences, Bremen University, Bremen, Germany

ABSTRACT ARTICLE HISTORY

Ledin and Machin’s critique of the use of some current approaches Received 8 November 2018
to multimodality for the purposes of critical discourse studies raises Accepted 16 November 2018
some important methodological concerns that need to be
KEYWORDS
addressed. However, both the particular position they develop as Multimodality; empirical
well as some of the key points they raise are themselves studies; systemic-functional
problematic. In this response, I argue that Ledin and Machin linguistics; critical discourse
misconstrue some significant aspects of the body of theory they analysis
critique and, as a consequence, offer a potentially misleading view
of the relevant state of the art in multimodality. I suggest that
pursuing closer alignments with theoretically more nuanced
accounts of multimodality would allow the account that Ledin
and Machin propose to be revised in ways better suited to the
task of conducting empirically-driven multimodal critical discourse
studies.

1. Introduction: a need for critique

There is nowadays a growing awareness, across many fields and contexts of application,
that communicative events and their media are often, if not always, constituted by
highly diversified forms of expression – verbal, visual, acoustic, spatial, diagrammatic,
architectural, gestural and many more. Addressing such forms and, most importantly,
their combinations, is the main concern of ‘multimodality’ (cf. Bateman, Wildfeuer, & Hiip-
pala, 2017, pp. 22–73; Jewitt, Bezemer, & O’Halloran, 2016, pp. 2–3). An increasing number
of studies demonstrate that multimodal communication is just as capable of distorting or
misrepresenting power relationships and enacting social inequalities as verbal language is
now accepted to be. It is then logical that Wodak and Meyer should explicitly signpost
‘non-verbal (semiotic, multimodal, visual) aspects of interaction and communication:
gesture, images, film, the internet and multimedia’ as legitimate targets for critical analysis
(Wodak & Meyer, 2015, p. 2).
The authors of the article being responded to here, Per Ledin and David Machin, have
long been engaged in extending the field of critical discourse studies (CDS) in this direc-
tion. In their position paper, ‘Doing critical discourse studies with multimodality’ (Ledin &
Machin, 2018b), they reflect on certain tendencies exhibited by current multimodality
research that they see as compromising the value of that research for critical discourse

CONTACT John A. Bateman [email protected]

© 2018 Informa UK Limited, trading as Taylor & Francis Group
2 J. A. BATEMAN

studies. In particular, they argue that those inadequacies stem directly from the theoretical
principles assumed in the kinds of multimodality research they target. Their discussion
consequently turns on some rather specific technical details underpinning that research.
This argumentative strategy is not entirely unproblematic. First, for readers who are
unfamiliar with the details of the theory critiqued, Ledin and Machin’s discussion enacts
a position of power with respect to its characterisation of that theory that is, arguably,
unwarranted. In certain central areas, the theoretical claims Ledin and Machin make are
at best contentious and, as a consequence, could well mislead critical discourse analysts
concerning the state of the art in multimodality research. Second, for readers who are fam-
iliar with the theory, misconstruals in Ledin and Machin’s presentation open up sufficient
‘wiggle room’ for those readers to evade important issues that Ledin and Machin do raise.
This is then doubly unfortunate. Several practices common in the kind of multimodality
research that Ledin and Machin criticise are certainly in need either of revision or, at
the very least, of focused critical reflection. However, the manner in which Ledin and
Machin reach their criticisms rests on, I will argue, some misreadings of the core principles
of the theory they target and, as a consequence, their final proposals are inherently limited
and fail to make contact with the current state of the art in multimodality research.
In this response there is insufficient space to engage with all of Ledin and Machin’s
theoretical points at the level of detail required for showing why they cannot be accepted
at face value. Instead, therefore, I will attempt to extract the main line of their argument
pinpointing some key places where misconstruals have undesirable consequences for the
research proposed. This will lead to a revision of Ledin and Machin’s methodology that is
more aligned with certain constructs of a broader, and more powerful, account of
multimodality.

2. The argument
Ledin and Machin’s linking of their argument to a particular body of theory requires at the
outset that at least the core principles of that theory be identified. The theoretical orien-
tation they discuss is that of ‘systemic-functional’ multimodal social semiotics or, more suc-
cinctly, ‘SFL-based multimodality’, a style of multimodality research drawing most of its
basic tenets from Michael A.K. Halliday’s systemic-functional linguistics (SFL: cf. Bartlett
& O’Grady, 2017). Although there are many other approaches to multimodality –
Bateman et al. (2017) offer a broader introduction – in the context of CDS Ledin and
Machin’s focus on SFL-inspired approaches is in many respects justifiable. SFL and CDS
have had a close relationship for a considerable time (cf., e.g. Kress, 1983) and many
researchers engaging in SFL-based multimodality consequently see themselves as
offering social critiques of (multimodal) discourse practices as well (cf. Djonov & Zhao,
2013; Zhao, Djonov, Björkvall, & Boeriis, 2018).
The essence of Ledin and Machin’s argument is that SFL-based approaches to multi-
modality address their objects of analysis with a generic interpretative scheme regardless
of what those objects may be and which ‘modalities’ they might employ. This scheme rests
on three principles. First, it is possible to characterise material regularities in terms of
‘systems of choice’ (called system networks), that are intended to capture the semiotic pos-
sibilities that some form of expression offers for making meanings. Second, SFL uses the
choices made from these networks to derive contextualised meanings for any object
 CRITICAL DISCOURSE STUDIES 3

analysed: Ledin and Machin call this ‘projecting’ context or semantics from the text. And
third, the meaning made within any form of expression is factored along three distinct
dimensions of generalised signification, called metafunctions, consisting of the ‘ideational’,
characterising how forms of expression offer ways of structuring the world, the ‘interper-
sonal’, characterising how forms of expression enact social relationships and evaluations,
and the ‘textual’, characterising how messages are organised to form coherent communi-
cative wholes.
Ledin and Machin see analyses performed according to these principles as multiply pro-
blematic. Most importantly, they consider it untenable that general descriptions of forms
of expression (represented as systems of choice with specific features) should be assumed
to apply to any object of analysis exhibiting those forms of expression. This is taken to be a
case of ‘one-size fits all’ ‘grand theorising’ which ignores context and situations of use
because the classificatory features applied are not themselves variable. Ledin and
Machin argue this to be an inappropriate stance for CDS because, in CDS, the social con-
texts involved are always the main target of analysis and it is, therefore, these contexts that
must be looked to for determining ‘relevant’ features of description. Thus, applying some
general schemes of description for, for example, pictorial content regardless of whether
that content appears as a press photograph, an oil painting, a piece of advertising, etc.
is considered in danger of missing socially significant patterning specific to particular con-
texts of use.
To address this, they propose that analysis must begin from what they term the ‘canons
of use’ of any artefacts or performances being analysed. Such canons of use are to be
drawn from the relevant literature and communities of practice involved in producing
and consuming any object under investigation and are claimed to offer a more appropri-
ate analytic tool than the generalised categories found in SFL-based multimodal analysis.
Rather than generalised features which may well not be ‘contextually relevant’ in some
particular case, canons of use allow identification of ‘the contextually relevant choices’
to be considered. In support of this, Ledin and Machin consider a press photograph
that I will return to below, noting that ‘much of the meaning of a photograph cannot
be established at the level of the contents of the image itself’. Now, while this is no
doubt true, it is also unsurprising: the ‘coherence’ even of verbal texts is not generally
seen as a text-internal property (cf. de Beaugrande & Dressler, 1981) and there is little
reason to expect that this would be the case for communicative artefacts such as photo-
graphs either. Further inference steps are always necessary for interpretation. Here, there-
fore, it will be useful to pick out more specifically just why Ledin and Machin consider this
important.
An answer lies in their critique of the second general principle of SFL-based accounts
listed above: the idea that SFL-based analyses ‘project’ contextualised interpretations
directly from the systemic choices used for describing ‘form’. This can indeed be highly
problematic, although I will argue below that the problems lie deeper than Ledin and
Machin suggest. In SFL multimodal analyses, ‘projection’ of this kind tends to be built
directly into the descriptive apparatus by virtue of labels indicating direct relationships
between observables and context. For example, Kress and van Leeuwen (2006 [1996],
pp. 209–210) label information presented in the top portion of a page as ‘ideal’ in contrast
to information presented in the bottom portion of a page, which is labelled ‘real’; similarly,
a pictorial representation that is looking up at a person visually (e.g. by virtue of camera
4 J. A. BATEMAN

angle) is often labelled as placing that person in a position of relative power (Kress & van
Leeuwen, 2006 [1996], pp. 148–149). In both cases, a directly observable material regularity
is being labelled in terms that suggest contextual social interpretations. As Ledin and
Machin point out, it is unlikely that such socially-specific readings of material regularities
will always apply and, although Kress and van Leeuwen provide many suggestive
examples, their categorisations have rarely been subjected to empirical validation. Many
SFL analyses nevertheless adopt them as if constituting established ‘fact’, rendering
those analyses, as Ledin and Machin state, largely circular (cf., also Thomas, 2014).
Ledin and Machin’s proposed solution to analysis drawing on canons of use neverthe-
less shares a critical weakness with most current SFL-based multimodality approaches:
there is still no systematic treatment of how descriptions of regularities in form can be
related to the situated interpretation of those regularities. For this, stronger models of mul-
timodal semiotics are required. I will support this claim here by considering a recontextua-
lisation of another of the critiques that Ledin and Machin bring against SFL-based
approaches, that concerning the assumed general applicability of the three metafunctions.
Gunther Kress, one of the original proponents of SFL-based multimodality, has taken this
so far as to propose that the existence of metafunctional diversification is actually defini-
tional for a semiotic mode (Kress, 2014): a form of expression – be that colour, layout, or
verbal language – is to be considered a semiotic mode only when it performs the three
kinds of communicative work that the metafunctions characterise. Kress motivates this
position by considering all social communicative tasks to necessarily engage with repre-
senting the world and enacting interpersonal social relationships. Since semiotic resources
develop in the service of such tasks, they will also consequently be approachable in terms
of metafunctional diversification. Most SFL-inspired multimodal analyses then apply this
principle to the internal organisations of the system networks of choice they develop
for describing semiotic modes as well.
This position conflates two sources of potential evidence. On the one hand, since Kress
characterises social situations in terms of metafunctionally-distinguished configurations,
discussion of the presence or organisation of such metafunctions should draw on
models of context and situation. On the other hand, the use of metafunctions within
descriptions of languages is quite different: originating in Halliday’s work on English
and other languages, the metafunctions emerged as an elegant refactoring of the internal
organisation of grammar. When viewed in terms of networks of interconnected choices, it
was suggested that there were substantially more interconnections within functional areas
of a grammar concerned with individual metafunctions than across functional areas con-
cerned with distinct metafunctions. Since then, similarly motivated distinctions have been
argued for semantics as well. Discussion and motivation for metafunctional organisations
in this sense is consequently best drawn from considerations of grammatical, or more gen-
erally formal, organisations of the realisations of semiotic resources.
Combining Kress’s claim that all communicative artefacts and performances can be
characterised metafunctionally with the SFL assumption that semiotic descriptions of
those artefacts and performances will similarly show metafunctionally diversified internal
organisation consequently conflates two quite different ‘levels’ of semiotic abstraction.
While, on the one hand, it may well be possible to interpret communicative artefacts or
performances in terms of the consequences that they have for structuring experience,
enacting interpersonal relationships, and organising messages (i.e. the three
 CRITICAL DISCOURSE STUDIES 5

metafunctions), this is by no means a guarantee that this will also offer an appropriate
‘internal’ description of the formal properties of those artefacts or performances. A particu-
lar choice of layout design, for example, may well give rise in a particular context of use and
for a particular community of users to an interpersonal evaluation of ‘cold, alienating for-
mality’ or, even, as Ledin and Machin (2016) suggest, a neoliberal management style,
but these are situated discourse interpretations of material patterns. Such discourse
interpretations may be further classified as performing work compatible with that
assumed by the distinct metafunctions, but cannot be attributed directly to the material
distinctions themselves. Although it may turn out to be the case that the internal semiotic
organisation of certain kinds of artefacts and performances transparently reflects this
structuring – thus supporting a metafunctional division throughout – this is always and
essentially an empirical issue. To assume that this is the case as a methodological principle
is premature. Moreover, binding together interpretation and form in this way would in any
case be detrimental to the flexible functioning of a form of expression. Probably as a con-
sequence of this, in the empirical studies with which I have been concerned for several
years, such tight binding appears almost never to occur. One consequence of this has
been the development of the rather more detailed semiotic account I sketch in the
section following.

3. A semiotically adequate multimodality

Ledin and Machin’s basic proposal for avoiding the difficulties raised by SFL-based multi-
modality is to assume that canons of use can reveal ‘relevant’ distinctions for analysis.
However, the process of moving from canons of use to analytic categories appears pro-
grammatic at best. If such analyses are to be carried out systematically, far more methodo-
logical guidance will be needed. For this, Ledin and Machin propose that CDS-relevant
analyses of multimodal phenomena can best be performed by beginning from what
they characterise as the ‘affordances’ of the objects of analysis, rather than with the
pre-determined coding schemes found in SFL-based multimodality (cf. Ledin & Machin,
2018b; Machin, 2016). The notion of affordance, originating in the ‘ecological’ approach
to perception of Gibson (1977), is widespread in approaches to multimodality and is gen-
erally adopted in order to direct attention back to the ‘materiality’ of any objects of study.
Although paying close attention to the materiality of what one is analysing is certainly to
be recommended (cf. Bateman et al., 2017, pp. 102–103), Ledin and Machin’s account faces
several difficulties that show that they are still assuming a semiotic model that conflates
important theoretical distinctions similar to those they identify as problematic in SFL-
based approaches.
One of the reasons they give for using affordances is to achieve what they term an
‘expression-to-content’ directionality in their analyses: this is the sense of the ‘from meta-
functions to materiality’ in their title. However, since they wish to do this in a way that
brings out what they see as the semiotically relevant distinctions for a given object of
analysis, rather than by applying pre-determined ‘grammars’ of forms of expression,
they are led to combine materiality and canons of use as well: i.e. the affordances of
objects of analysis are seen as already being structured through and through with their
cultural use. It is no doubt the case that the perception of materiality cannot be entirely
divorced from the social practices and experiences of a perceiver, but characterising
6 J. A. BATEMAN

material distinctions in ways that already include the uses made of those material distinc-
tions is circular in precisely the same manner as SFL’s labelling of technical features in con-
textual terms or importing metafunctional diversification into form descriptions.
For specific instances of analysis, this may (perhaps) be an effective short-term strategy,
but there is little support for systematicity or general results. The claims Ledin and Machin
make concerning affordances are primarily informal restatements of positions discussed in
existing literature, which may itself well need problematising in precisely the manner that
Ledin and Machin do for SFL-based multimodality. Some of the consequences of Ledin
and Machin’s position can be seen in an earlier statement of Machin concerning how
analysis of photographs might be pursued more effectively from a CDS perspective:
In multimodal work, it is possible to find photography treated as a mode. But there are reasons
why an affordance-driven approach might be better, avoiding the notion of mode altogether.
… The affordance of the photograph is that it claims to represent unmediated reality. … This
idea of the ideological trick of the photograph has led to a particular trajectory of critical work
on the photograph in visual studies and in media and cultural studies. This provides a number
of valuable insights for a multimodal approach to images … (Machin, 2016, p. 328)

However, avoiding ‘the use of mode altogether’ in effect reduces the analysis to pre-
cisely those standard debates of indexicality and truth in photographic representations
that currently stand in crisis due to radical changes in production technologies on the
one hand and in media competence of consumers on the other. The potential value of
a contribution anchored in multimodality is drastically reduced.
In contrast, reinstating a more sophisticated notion of semiotic mode can move the dis-
cussion onto a different level. The model of semiotic modes presented in Bateman (2016),
for example, formally distinguishes distinct levels of contributions to multimodal meaning-
making in ways that not only subsume the kinds of issues Ledin and Machin appropriately
point out as important, but which also maintain a far stronger methodological position for
practical analysis (cf. Bateman & Wildfeuer, 2014). The press photograph that Ledin and
Machin discuss showing a young girl standing near two soldiers in Sarajevo offers a
good example of how this can play out.
First, media competence plays a role in helping to identify (i.e. to abductively assume
for the purposes of further analysis) the medium that is relevant: i.e. press photographs.
Media have socioculturally established purposes and functions, distribution and reception
practices, and so on. Much of this information is to be drawn from studies of media land-
scapes as Ledin and Machin suggest. More formally, a range of genres with conventiona-
lised realisations will be associated with the medium, thereby covering many of the
important facets intended by Ledin and Machin’s ‘canons of use’. In addition, however,
media are defined multimodally in terms of the semiotic modes that they make available.
Each medium is seen as a socioculturally stabilised configuration of such semiotic modes.
Semiotic modes are then defined formally as empirically established bundles of materials,
classification networks of material regularities (the technical features of the semiotic
mode) and discourse semantics that abductively propose contextualised interpretations
for any material regularities present. Semiotic modes are then the crucial mediators by
which materiality becomes imbued with social meanings (and vice versa). They are also
the locus of focused empirical investigation, since formulating their internal details rests
solely on the regularities that can be observed in their use. In contrast to SFL-based
 CRITICAL DISCOURSE STUDIES 7

multimodality as critiqued by Ledin and Machin, however, no metafunctional organisation

is assumed prior to obtaining appropriate empirical support.
This is compatible with Ledin and Machin’s aims, but begins to offer a more rigorous
theoretical and practical position to conduct the empirical research that is necessary.
Ledin and Machin’s discussion is quite inconclusive in this regard. For example, they
suggest that their position is both quite different to the SFL-based approach and is to
be preferred:
If, rather than coming at the photograph from linguistic processes and syntax, we first describe
what we see it comes out slightly differently. We might take look at the photograph and bodily
and facial expression and pose as starting points. We might observe that the girl takes a
restricted pose and stands still with the hands behind her back. Her eyes are open with a
steady gaze and the mouth shut, with the lips slightly going down. (Ledin & Machin, 2018a,
p. 10)

They argue that this is clearly a better strategy than applying pre-set ‘grammatical’
categories of the kind proposed by Kress and van Leeuwen (2006 [1996]) and, as
noted above, widespread in SFL approaches. But just what the ‘affordances’ of photo-
graphs are is by no means so clear cut. One of the visual categories of Kress and van
Leeuwen rejected by Ledin and Machin rests on the recognition of ‘gaze vectors’ in
an image. Since it is well known from a considerable number of perception studies
that people pay close attention to gaze vectors, ruling such information out on principle
as not being ‘afforded’ is hardly convincing. Similar criticisms hold for Ledin and Machin’s
claim that photographs cannot show movement: sufficiently dynamically-charged static
images are also well known to trigger neural processing of motion and action. The entire
question of ‘affordances’ therefore needs a far more precise treatment than Ledin and
Machin’s account assumes.
Semiotic modes as defined above take us further. One of the semiotic modes involved
in the medium of press photographs is that of visually iconic pictorial representation. This
recruits many of the mechanisms of natural scene perception, including motion proces-
sing, facial expression and emotion processing, gaze tracking, spatial processing, etc. Evi-
dence for the information potentially made available by the operation of this mode is then
a matter of empirical perceptual psychology. Another semiotic mode involved in this
medium is that for framed representations in general, whereby the position of an
assumed ‘observer’ of the depicted scene, the framing of the material shown, etc. are
semiotic choices. Further modes available may include colouring, lighting, depth of
focus, and so on: all of which can be subjected to empirical investigation to bring out
the distinctions constituting the mode and conventionalised within particular commu-
nities of use. Crucially, however, these ‘choices’ cannot be labelled in socio-situational
terms because they have not yet been subjected to discourse interpretation. It is only in
the operation of discourse semantics, mediating between the formal distinctions and
their contexts of use, that interpretations are produced abductively as hypotheses
raising overall discourse coherence (Bateman & Wildfeuer, 2014). Thus, while it is the
assumption of a particular semiotic mode that licenses identification of particular material
regularities as being available for meaning-bearing, what those meanings are is always a
matter for the discourse semantics of a mode. It cannot be ‘read into’ the classifications of
material distinctions directly.
8 J. A. BATEMAN

Canons of use then become directly and empirically accessible by exploring just which
semiotic choices are taken up in a medium for performing particular communicative pur-
poses. Variations in the options taken up may then ‘realise’, or correlate with, differing
communities of practice. Thus, the particular options of framing, angle, motif, colour,
and so on found within press photography (determined empirically by examining larger
data sets) are what then provide the basis necessary for systematic critical analysis.
Without this information, analysis is limited to suggestive description, which may be
sufficient for Ledin and Machin’s aims but is not supportive of more systematic empirical
study capable of fine-grained work relating materials and socio-political questions. Ledin
and Machin suggest that ‘canons of use come with established meanings and patterns’
(Ledin & Machin, 2018a, p. 5), but then the challenge is to uncover those established mean-
ings and patterns since, as critical analysis reveals, they are often far from obvious. It
cannot be guaranteed that simply ‘asking’ producers (or the literature) will unveil what
is structurally buried in practice.
In essence, what Ledin and Machin have done is replace richly organised sets of predic-
tions concerning potentially pertinent material regularities within specific semiotic modes
with a more intuitive collection of everyday interpretative categories that make little dis-
tinction between modes, media and genres (cf. Bateman, 2017). The problem with the SFL-
based approach in terms of codifed descriptions of semiotic resources is not then that they
are system networks, but that they are, as Ledin and Machin correctly state, far too rarely
subjected to the kind of empirical probing that would be necessary before adopting them
as analytic tools. Approaching issues of multimodality with a stronger commitment to the
empirical basis of any analyses performed is then a more promising strategy for future
research than, as Ledin and Machin appear to want to suggest, avoiding such analysis
altogether. Whereas they conclude: ‘we cannot use a model that seeks context from
within the text’ (Ledin & Machin, 2018a, p. 14), I have argued that their position goes
too far in the opposite direction, making it difficult to account for the contributions of
texts as anything other than a reflection of context.
The overall frame for analysis set out here consequently identifies many of the particu-
lar points that Ledin and Machin wish to accommodate, while also placing them in particu-
lar relationships to one another so that systematic fine-grained analysis is supported. This
possibility is still largely missing in Ledin and Machin’s proposal, precisely because of that
account’s lack of sufficiently robust semiotic foundations.

Disclosure statement
No potential conflict of interest was reported by the author.

Notes on contributor
John A. Bateman (PhD in Artificial Intelligence, University of Edinburgh) has been a professor of
applied linguistics in the Faculty of Linguistics and Literary Sciences at the University of Bremen
since 1999. His main fields of research range over computational linguistics (particularly natural
language generation, discourse and dialogue), formal ontology, and the theory and practice of multi-
modality. He has published widely in all of these areas, including monographs on text generation,
multimodality and genre, film, text and image, and an introduction to multimodality as a new dis-
cipline (2017, de Gruyter, with Janina Wildfeuer and Tuomo Hiippala). Recent work focuses
 CRITICAL DISCOURSE STUDIES 9

specifically on the semiotic foundations of multimodality and the use of empirical methods for their
investigation, combining interdisciplinary studies drawing on eyetracking, brain-imaging and corpus
studies.

References
Bartlett, T., & O’Grady, G. (Eds.). (2017). Handbook of systemic-functional linguistics. London: Routledge.
Bateman, J. A. (2016). Methodological and theoretical issues for the empirical investigation of multi-
modality. In N.-M. Klug & H. Stöckl (Eds.), Handbuch Sprache im multimodalen Kontext (Handbooks
of linguistics and communication Science (HSK) 7) (pp. 36–74). Berlin: de Gruyter Mouton.
Bateman, J. A. (2017). Triangulating transmediality: A multimodal semiotic framework relating media,
modes and genres. Discourse, Context & Media, 20(Supplement C), 160–174. doi:10.1016/j.dcm.
2017.06.009
Bateman, J. A., Wildfeuer, J., & Hiippala, T. (2017). Multimodality – foundations, research and analysis. A
problem-Oriented introduction. Berlin: Mouton de Gruyter.
Bateman, J. A., & Wildfeuer, J. (2014). A multimodal discourse theory of visual narrative. Journal of
Pragmatics, 74, 180–208. doi:10.1016/j.pragma.2014.10.001
de Beaugrande, R., & Dressler, W. U. (1981). Introduction to text linguistics. London: Longman.
Djonov, E., & Zhao, S. (Eds.). (2013). Critical multimodal studies of popular discourse. New York:
Routledge.
Gibson, J. J. (1977). The theory of affordances. In R. Shaw & J. Bransford (Eds.), Perceiving, acting, and
knowing: Toward and ecological psychology (pp. 62–82). Hillsdale, NJ: Erlbaum.
Jewitt, C., Bezemer, J., & O’Halloran, K. (2016). Introducing multimodality. London: Routledge.
Kress, G. (1983). Linguistic and ideological transformations in news reporting. In H. Davis & P. Walton
(Eds.), Language, image, media (pp. 120–138). Oxford: Basil Blackwell.
Kress, G. (2014). What is mode? In Carey Jewitt (Ed.), The Routledge handbook of multimodal analysis,
2nd edn, 60–75. London: Routledge.
Kress, G., & van Leeuwen, T. (2006 [1996]). Reading images: The grammar of visual design. London:
Routledge.
Ledin, P., & Machin, D. (2016). A discourse-design approach to multimodality: The visual communi-
cation of neoliberal management discourse. Social Semiotics, 26(1), 1–18.
Ledin, P., & Machin, D. (2018a). Doing critical discourse studies with multimodality: From metafunc-
tions to materiality. Critical Discourse Studies, 1–17. doi:17405904.2018.1468789
Ledin, P., & Machin, D. (2018b). Doing visual analysis: From theory to practice. London: Sage.
Machin, D. (2016). The need for a social and affordance-driven multimodal critical discourse studies.
Discourse & Society, 27(3), 322–334.
Thomas, M. (2014). Evidence and circularity in multimodal discourse analysis. Visual Communication,
13(2), 163–189. doi:10.1177/1470357213516725
Wodak, Ruth & Michael Meyer. (2015). Critical discourse analysis: History, agenda, theory and meth-
odology. In Ruth Wodak & Michael Meyer (Eds.), Methods of critical discourse analysis, 3rd ed., 1–33.
London: SAGE Publishers.
Zhao, S., Djonov, E., Björkvall, A., & Boeriis, M. (2018). Advancing multimodal and critical discourse
studies. Interdisciplinary research inspired by Theo van Leeuwen’s social semiotics. London:
Routledge.