2.1.1 Dialectics of Discourse

Discourse is, firstly, a semiotic practice. The key insight of CDA, however, is that discourse is also always a form of social practice (Fairclough Reference Fairclough1995). Defining discourse as social practice means viewing it as a mode of action that is a. socially situated and that b. has the ability to impact the world (Austin Reference Austin1962). This means that discourse is shaped by the situations, institutions and social structures in which it is embedded but at the same time helps to define and shape those situations, institutions and social structures. That is, discourse is both socially constituted and socially constitutive (Fairclough Reference Fairclough1995). Discourse constitutes social identities and relationships between groups of people as well as knowledge and understanding of social events. It follows that, in representing the world and positioning people within it in particular ways, discourse plays a significant part in the establishment and maintenance of power. By the same token, discourse is also capable of resisting and transforming existing power structures. This two-way relationship conceived between discourse and social contexts is described by Fairclough (Reference Fairclough1995) as dialectical. It entails an epistemological alignment with social constructionism (see Cap Reference Cap2019a), according to which knowledge is developed not separately within the individual but through coordination with others such that certain aspects of the world that are normally taken as immutable facts with a basis in physical reality in actual fact represent a shared set of assumptions which have developed historically within institutions and cultures.

It is increasingly argued that the dialectical relationship between texts and society is mediated by cognition (e.g. Chilton Reference Chilton, Wodak and Chilton2005a). The social actions that amount to or lead to power abuses, inequalities and other injustices are dependent on certain mental representations of the world, which are engendered through discourse. Approaches to CDA such as the discourse-historical approach and the socio-cognitive approach therefore draw on theories of social cognition (e.g. Moscovici Reference Moscovici2000) to provide the ‘missing link’ between texts and society and to model the mental representations which mediate in this relationship. Cognitive CDA draws on cognitive linguistics in this connection.

2.1.2 Dialogism

A concept that was developed originally in the context of literary criticism by Kristeva (Reference Kristeva and Moi1986), elaborating on Bakhtin’s (Reference Bakhtin, Holquist, Emerson and Holquist1981) notion of dialogism, but which forms an important part of CDA, is intertextuality. Texts are not produced or consumed independently of one another in a historical vacuum. Rather, texts are constituted by elements of other texts (Fairclough Reference Fairclough1992). Texts may contain explicit aspects of or more general allusions to past texts. They may similarly pre-empt and attempt to shape future texts. In this sense, as Kristeva (Reference Kristeva and Moi1986: 39) observes, intertextuality implies ‘the insertion of history (society) into a text and of this text into history’ (where texts are the primary artefacts that constitute history). Texts are therefore sites of ideological struggle which bear the imprint of competing discourses vying for dominance. From a semiotic perspective, past texts provide a resource for meaning-making in present texts with elements of the original text acquiring new meanings within the fresh context into which they are transplanted – a process usually referred to as recontextualisation.

A closely related concept developed in CDA and modelled on intertextuality is interdiscursivity (Fairclough Reference Fairclough1995). Like intertextuality, interdiscursivity highlights the historicity and hybridity of texts. However, where intertextuality refers to the appropriation of specific past texts within current ones, interdiscursivity refers to the borrowing of features associated with alternative genres (i.e. text-types) and discourses to override existing discourse practices and create new ones. This process is what gives rise to change in discourses and genres such as the marketisation of education discourse or the conversationalisation of the news (Fairclough Reference Fairclough1995). It follows from a dialectical view of discourse that such changes in discourse practices go hand in hand with changes in society, both in the sense of reflecting societal change but also facilitating or directly effecting it. Within Cognitive CDA, intertextuality and interdiscursivity are particularly important in connection with metaphor.

2.1.3 Power, Ideology and Legitimation

Power is an asymmetrical relationship between social actors which, following Weber (Reference Weber1978), can be defined in terms of the possibility that an individual has within a social relationship to accomplish their will, including in the face of resistance from others. Power can be exercised through acts of physical force and threats of violence. In democratic societies, however, power is maintained primarily by consent rather than coercion. A question addressed by Gramsci (Reference Gramsci, Hoare and Smith1971) in his theory of hegemony is why social actors should acquiesce to power structures which are against their own interests. For Gramsci, this is owing to ideological reproduction through the various institutions and practices of civil society (education, religion, family), which render the social, economic and political status quo as ‘common sense’. For CDA, it is discourse in particular that is central in this regard (Fairclough Reference Fairclough1989). As Reisigl and Wodak (Reference Reisigl, Wodak, Wodak and Meyer2016: 26) state, ‘power is legitimised or delegitimised in discourses’. CDA therefore shares Habermas’s view of language as ‘a medium of domination and social force [that] serves to legitimize relations of organized power’ (Reference Habermas1967: 259). Of course, Gramsci’s writings are reflective of a certain period in time. In contemporary society, it is the discourses of mass media institutions rather than the Church that are arguably most complicit in the inculcation of dominant ideologies.

Two important means by which discourse serves to maintain power and inequality, then, are ideology and its naturalisation and legitimation. The notion of ideology is somewhat slippery and has come to be defined and figure differently across models in the social sciences (see Eagleton Reference Eagleton1991; Thompson Reference Thompson1990). In CDA, ideology is normally defined in terms of the ‘perspective’ or ‘worldview’ that a text presents – one which usually serves the interests of established power. Ideologies are the social belief and value systems that are instilled in texts and reproduced through discourse to provide motivation and support for particular ways of acting and being in the world, including those that result in continued domination, discrimination and exploitation (Fairclough Reference Fairclough2003: 218). Van Dijk (Reference van Dijk1998) sees ideologies as residing in social cognition where they exist as ‘general systems of basic ideas shared by the members of a social group’ (van Dijk Reference van Dijk and van Dijk2011: 380). From a cognitive linguistic perspective, ideologies are conceptual systems consisting of a structured set of cognitive models representing ‘recurrent phenomena and their interpretation in culture and society’ (Dirven, Frank and Pütz Reference Dirven, Frank, Pütz, Dirven, Frank and Pütz2003a: 1–2). Discourse plays a central role in achieving hegemony where readers become habituated to dominant forms of representation which are then no longer treated as ideological but viewed instead as neutral or objective – a process Fairclough (Reference Fairclough1989) refers to as naturalisation. Fowler (Reference Fowler1991: 66) argues that ideology in discourse is inevitable. Linguistic (and other semiotic) representation is always necessarily representation from a specific ideological point of view (66). Through repeated patterns of instantiation, however, the worldviews espoused by dominant discourses lose their manifest ideological character so that the power relations enshrined within them come to be viewed as natural and expected rather than as having been socially engineered. CDA then involves the excavation of texts to identify the specific semiotic sites where ideology can be traced.

A second means by which discourse is able to maintain power and inequality is legitimation (Martin Rojo and van Dijk Reference Martin Rojo and van Dijk1997; Reyes Reference Reyes2011; van Leeuwen Reference van Leeuwen2007; van Leeuwen and Wodak Reference van Leeuwen and Wodak1999). Legitimation is the process whereby individuals or groups actively ‘seek to secure consent to their power from at least the most important among their subordinates’ (Beetham Reference Bevelas, Chovil, Lawrie and Wade1991: 3). Successful legitimation therefore results in the public’s endorsement or at least acceptance of expressions of power. Legitimation is described as ‘the principle goal of the political speaker’ (Cap Reference Cap2006: 7) whose discourse is aimed at persuading audiences that a course of action, state of affairs or policy is justified, necessary or morally incumbent. Legitimation normally occurs in contexts of controversial actions, accusations, doubts and critique (Martin Rojo and van Dijk Reference Martin Rojo and van Dijk1997: 528). In this sense, legitimating texts are dialogic and restorative. Through legitimation strategies, speakers seek accreditation or support for social behaviours that would otherwise lie outside the normative order. Legitimating strategies aim to show that such actions and relations are in fact consistent with the moral code of society; that is, fall within the system of laws, norms and values tacitly agreed upon by the majority of citizens (Martin Rojo and van Dijk Reference Martin Rojo and van Dijk1997: 528). To this end, as Martin Rojo and van Dijk observe:

Legitimation is closely linked with ideology where it is the worldviews promoted through discourse that make acts of violence, aggression, exclusion, discrimination and suppression appear more palatable and thus more acceptable.

2.1.4 Critique, Normativity and Resistance

CDA is not a neutral science but an explicitly critical project. To be critical in this sense means to go beyond the standard scientific tasks of observing, describing and explaining in objective terms to consider data within its social context, adopting an explicit socio-political stance and actively seeking social change through semiotic analysis. As Kress puts it:

This position is traceable to Habermas (e.g. Reference Habermas1971) and the Frankfurt School of Critical Theory (e.g. Adorno Reference Adorno1970; Horkheimer Reference Horkheimer and O’Connell1972). Continuing in this critical tradition, research in CDA aims to address perceived ‘social wrongs’, understood in broad terms as ‘aspects of social systems, forms or orders which are detrimental to human well-being, which could in principle be ameliorated if not eliminated, though perhaps only through major changes in these systems, forms or orders’ (Fairclough Reference Fairclough, Wodak and Meyer2016: 91). Examples include violence, discrimination, poverty, lack of freedom and other forms of structural inequality. Of course, critically addressing such issues assumes that they are wrong in principle and thus presupposes a certain normative stance on the part of the researcher.Footnote ¹ Here, CDA does not subscribe to a specific ethical framework but is broadly guided by a sense of justice based on ‘the unrestricted validity of human rights and by the awareness of suffering’ (Reisigl and Wodak Reference Reisigl and Wodak2001: 34).

Three types of critique are defined (Reisigl and Wodak Reference Reisigl and Wodak2001: 32–35). Text or discourse immanent critique highlights the logical inconsistencies, contradictions and paradoxes displayed by texts. Socio-diagnostic critique aims at demystifying the ideological, persuasive, propagandist or populist properties of texts which – from a certain normative standpoint – are perceived as problematic. Socio-diagnostic critique transcends the text-internal sphere to analyse texts within their wider social contexts, including intertextual, interactional, institutional and politico-historical contexts. This requires analysts to make use of their own background knowledge and interpretations as socially situated researchers, which in turn demands reflexivity on the part of researchers. From this perspective, the social situatedness of the analyst is not an impediment to research but an imperative for it. Critical textual analysis requires knowledge of the social contexts in which texts are produced and interpreted since these affect meaning. Finally, prospective critique aims at directly changing and improving the communication practices of public institutions in the future by making recommendations and drawing up guidelines etc. The form of critique most commonly practiced in Cognitive CDA is socio-diagnostic critique. This is often best achieved through comparative forms of analysis which can reveal sites of difference between texts that imply or can be explained by competing ideological positions.

In so far as the goals of socio-diagnostic critique are intended to be illuminatory – ‘to root out a particular kind of delusion’ (Reisigl and Wodak Reference Reisigl, Wodak, Wodak and Meyer2016: 7) – there is a danger of reducing readers to passive receptors of knowledge who otherwise lack the capacity to make for themselves critical judgements about texts. Of course, this is not the case and readers are perfectly capable of conducting their own form of critique and of engaging in resistant forms of reading (Browse Reference Browse2018), although there are linguistic and contextual circumstances in which resistant reading is inhibited or made more difficult as, for example, when no counter-narratives are available, when textual input misdirects cognitive processing (Maillat and Oswald Reference Maillat, Oswald and Hart2011) or when discursive steps are taken to satisfy the epistemic safeguards of sceptical readers (Hart Reference Hart2011a). The aim of CDA in any case is to go beyond the kind of critique that ordinary people are capable of to offer something, based on the theoretical models, research tools, time and other resources uniquely at our disposal, that is additionally insightful and that has a firm footing in the academic field of linguistics. In this regard, as Fairclough, Mulderrig and Wodak (Reference Fairclough, Mulderrig, Wodak and van Dijk2011: 374) state, CDA is ‘not just a replication of everyday critique: it can draw upon social theories and theories of language, and methodologies for language analysis, which are not generally available, and has resources for systematic and in-depth investigations which go beyond ordinary experience’. In the case of Cognitive CDA, theories of language and methods of language analysis are provided by cognitive linguistics.

Similarly, there are occasions when the ideology or persuasiveness of a text is immediately apparent and readers are likely to be left with a very clear impression of its character and orientation but lacking relevant linguistic expertise are unable to account for this impression in any precise or scientific way. The role of CDA is then not so much one of diagnosis but of exposition. CDA is able to elucidate, in precise terms, the particular linguistic or other semiotic properties of texts that are responsible for the impressions readers form. In cognitive CDA, this means providing cognitive details of how specific textual features contribute to meaning construction.

2.2.1 Generalisation and Conceptualisation (Non-autonomy of Language)

In contrast with generative linguistics, cognitive linguistics rejects the modularity of mind hypothesis (Fodor Reference Fodor1983) and does not see language as an autonomous cognitive faculty operating on a uniquely dedicated set of principles and processes. Rather, the principles and processes that operate in language are essentially the same as those found to function in other areas of cognition like memory, perception, imagination and reason (Croft and Cruse Reference Croft and Cruse2004). That is, language is dependent on the same underlying systems and processes that support other areas of mental experience, many of which are conceptual and imagistic in nature. For example, just as we necessarily assume a particular viewpoint in memory, perception and imagination, so viewpoint figures extensively in language. Accordingly, the structures and processes of language are not propositional or algorithmic in form but are likewise conceptual and imagistic. They therefore have analogues in other areas of cognitive experience and can be described appropriately in the same terms.

This position gives rise to what Lakoff (Reference Lakoff1990: 40) calls the cognitive commitment of cognitive linguistics: that a plausible account of language should be informed by and fit with what is already known about the way the mind and brain work from other disciplines of cognitive science. Hence, cognitive linguistics makes frequent appeal to models in cognitive psychology, as well as neuroscience and to some extent artificial intelligence, to help shed light on the mechanisms involved in human language.

From this perspective, language is seen as a set of cues for the co-construction of meaning achieved through conceptualisation. Conceptualisation is a rich and imaginative process in which language connects with background knowledge and general cognitive processes to yield fully modal, dynamic mental representations. It is vast and kaleidoscopic with language providing only a guide to an extensive array of conceptual activity. As Fauconnier states:

Conceptual processes discernible in language include categorisation, metaphor, image-schematic patterning, figure-ground segregation, selective attention, viewpoint and perspective (Croft and Cruse Reference Croft and Cruse2004). To the extent that these conceptual processes produce specific construals they are sometimes referred to as construal operations (Croft and Cruse Reference Croft and Cruse2004).

2.2.2 Continuity of Grammar and Lexicon

Cognitive linguistics extends its non-autonomous view internally so that language itself is not treated as modular. That is, language is not seen as being made up of distinct components representing grammar, phonology, the lexicon and so on. Rather, since the same conceptual processes are able to account for meaningful distinctions at different rank levels of linguistic structure, it follows that cognitive linguistics should take a ‘minimalist’ approach to linguistic description (Langacker Reference Langacker2008).

Such a minimalist approach is encapsulated in the symbolic thesis. According to Langacker (Reference Langacker2008), the only elements that need necessarily be postulated in an account of language are (a) semantic structures, (b) phonological structures and (c) symbolic structures linking semantic structures with phonological structures. On this account, a language is characterised as an inventory of symbolic assemblies or constructions in which forms and meanings are conventionally associated with one another.

Semantic structures take the form of conceptualisations which are exploited by language to provide meaning to linguistic expressions (Langacker Reference Langacker2008: 15). This accords with the widely held view of the lexicon that has its roots in the Saussurian notion of the linguistic sign. The radical claim of cognitive linguistics, however, is that grammar is principally no different and can be accounted for without the need to posit any additional architecture. Lexicon, morphology and syntax ‘form a continuum fully reducible to assemblies of symbolic structures’ (Langacker Reference Langacker2008: 15). Grammatical structures are therefore represented as whole units, alongside lexical units, in a system of symbolic assemblies, where they are paired with semantic structures in the form of conceptualisations. From this perspective, grammatical forms are themselves inherently meaningful by virtue of the conceptualisations with which they are conventionally associated. As Langacker (Reference Langacker1987: 1–2) puts it, ‘grammatical structures do not constitute an autonomous formal system or level of representation’ but instead are ‘inherently symbolic, providing structuring and conventional symbolization of conceptual content’. Thus, in discourse, the lexical, inflectional and grammatical structures presented all contribute meaning to the overall conceptualisation evoked.

Importantly, the meanings expressed by both lexical and grammatical units represent schematisations over experience. The difference between lexical and grammatical units resides in the degree of schematicity inhering in the semantic structures of which they are indexical and in the symbolic complexity of those structures (with more specific and simple symbolic assemblies tending to be characterised as lexical and more schematic and complex ones tending to be characterised as grammatical). Otherwise, grammar and lexicon are made of the same stuff. Similarly, no principled distinction is to be drawn between literal and figurative language where conventional metaphoric expressions index semantic structures in the form of conceptual metaphors (Lakoff and Johnson Reference Lakoff and Johnson2003). Such an account of language is attractive for it adheres to principles of conceptual unification and theoretical austerity, which are properties considered desirable of scientific theories in general (Langacker Reference Langacker2008: 14–15).

Linguistic structure emerges through usage events whereby recurrent form-meaning pairings, by process of abstraction, become conventionalised inside the system of symbolic assemblies. Cognitive linguistics therefore presents a usage-based account of linguistic structure in which form follows function. Crucially for multimodality, form extends beyond what is traditionally understood as linguistic so that representations included under the rubric of phonological structures include any form of expression that is capable of fulfilling a symbolising role (Langacker Reference Langacker2008: 15). This includes gestures and images.

2.2.3 Construal

The notion of construal is integral to cognitive linguistics and refers to our ‘manifest ability to conceive and portray the same situation in alternate ways’ (Langacker Reference Langacker2008: 43). Talmy (Reference Talmy2000: 14) describes this cognitive capacity as the principle of conceptual alternativity stating that ‘the same ideational complex can be represented in terms of alternative conceptualisations’. A speaker can therefore select one or other of such conceptualisations to represent the ideational complex they currently wish to communicate (Reference Talmy2000: 14).

The role of construal in language is most obvious in cases where more than one linguistic formulation is truth-conditionally licensed to describe the same situation. Classic examples are deixis and the active/passive distinction. For instance, the sentences the people came into the building and the people went into the building are truth-conditionally synonymous. However, they are meaningfully different in that they construe the scene from different vantage points – from inside versus outside the building respectively. Similarly, sentences like the dog attacked the person and the person was attacked by the dog are not semantically equivalent to one another, contrasting only in their surface form, but differ meaningfully in the relative attention they confer on the participants involved. While deixis and voice distinctions provide clear examples, construal is ubiquitous in language and upon inspection is evident in connection with nearly every aspect of every sentence, from its lexical content to its morphological inflections and its syntactic organisation.

Many facets of construal are analogous to visual perception where ‘in viewing a scene, what we actually see depends on how closely we examine it, what we choose to look at, which elements we pay most attention to, and where we view it from’ (Langacker Reference Langacker2008: 55), as well as whether we observe it directly or indirectly via some refracting lens or medium.Footnote ² Corresponding conceptual parameters through which construal is effected include specificity, distribution of attention, viewpoint and metaphor.

The claim in cognitive linguistics is that ‘every symbolic structure embodies conventional images and thus imposes a certain construal on the situation it codes’ (Langacker Reference Langacker1991: 294). Indeed, language generally eschews synonymy and therefore functionally it is precisely because of their conceptual import – the competing construals they impose – that alternate linguistic expressions are available to code the same situation (Reference Langacker1991: 295). It follows that the lexical and grammatical (including morphological) units presented in discourse serve to guide readers, directing them and their attention in different ways, to arrive at a particular construal or experience of the events described. From a critical standpoint, particular construals may be seen as motivated by, and as serving to maintain, specific ideologies or worldviews which, in turn, legitimise social actions and relations that may be discriminatory or otherwise difficult to justify.

2.2.4 Experiential Grounding

Through construal, then, language encodes experience. But language is also grounded in experience, which is to say that many of the meanings attached to linguistic expressions are rooted in our experience of the world. Two sources of experience from which the meanings of linguistic expressions are derived are experiences we have of our bodies and the physical environment we inhabit and experiences we have as members of a given culture or society.

Traditional conceptions treat the mind and body as independent of one another. In viewing meaning as having a basis in bodily experience, however, cognitive linguistics rejects Cartesian dualism and subscribes instead to an embodied view of language consistent with embodied cognitive science more generally (e.g. Gibbs Reference Gibbs2005). The principle of embodied cognition maintains that features of the mind, including language, are shaped significantly by the body and the environment in which it functions. Cognitive processes are therefore constrained by the kind of bodies we have and the experiences they afford of the world around us. This perspective gives rise to the epistemological position of experiential realism (Lakoff Reference Lakoff1987). According to this view, there is a world that exists objectively outside the body but our only access to it is via our bodies which, owing to their unique forms and position within a particular ecological niche, necessarily guide and constrain our experience of that world.

With respect to language, many of the meanings attached to linguistic expressions are rooted in embodied experience, defined by the nature of our bodies and our interactions with and observations over the physical environment (Johnson Reference Johnson1987). Such meanings take the form of image schemas, which include container, support, source-path-goal, various force-dynamic schemas and many others. Image schemas develop pre-linguistically as distillations over recurrent patterns of experience but come to figure in language as the meaningful basis of linguistic expressions (Reference Johnson1987). A word like enter, for example, involves both the container schema and the source-path-goal schema, while the meaning of a transitive clause is provided by an action-chain schema. Theories that treat language as embodied provide a solution to the symbol grounding problem (Harnad Reference Harnad1990), which arises when abstract, amodal symbols are defined only in terms of their relationship with other similar symbols (Glenberg and Robertson Reference Glenberg and Robertson2000).

Embodied concepts of this kind are extended to provide structure to other mental and social domains of experience via mappings in conceptual metaphors (Lakoff and Johnson Reference Lakoff and Johnson2003). For example, the source-path-goal schema is invoked whenever we talk about ‘making progress’, ‘moving forward’, ‘taking steps’, or ‘being on the right path’. When applied figuratively such expressions realise a primary metaphor purposeful activity is movement toward a destination in which the source-path-goal schema inherently features. Primary metaphors arise where concepts are correlated in embodied experience (Grady Reference Grady1997). Through inheritance hierarchies (Lakoff Reference Lakoff and Ortony1993), many conceptual metaphors that do not necessarily have an obvious experiential basis can in fact be traced back to primary metaphors. Thus, when we talk about politics as a journey, as in ‘the road to Brexit’, the conceptual metaphor expressed elaborates the primary metaphor purposeful activity is movement toward a destination. In this sense, significant portions of language and the conceptual system that underpins it ultimately have a grounding in embodied experience. This is not to say that political metaphors are inevitable or that alternate construals, based on other primary metaphors, are not possible. Equally, not all political metaphors have an experiential grounding. Political metaphors, whether elaborations of primary metaphors or not, reflect the ideological frameworks of the institutions in which they are produced and/or are guided by the rhetorical intentions of the individual text-producer.

The second kind of experience that language draws upon to provide meaning to linguistic expressions is experience cultivated (including vicariously via texts) as members of a given culture or society. Language has access to vast repositories of encyclopaedic knowledge in the form of frames. Frames are structured bundles of encyclopaedic information relating to areas of socio-cultural experience (Fillmore Reference Fillmore1982, Reference Fillmore1985). Examples include a commercial transaction frame, a marriage frame and a journey frame. The argument advanced in cognitive linguistics is that lexical meaning extends beyond the concept immediately evoked by a word to include the entire network of knowledge with which that concept is associated. Thus, for example, the meaning of road is appreciable only as a function of the wider journey frame in which the concept road figures. The range of knowledge accessed by an expression on a particular occasion of use is not fixed but is determined by contextual factors and is accessed preferentially to diminishing degrees of salience. A further continuity is therefore to be found between aspects of meaning traditionally conceived as semantic and aspects traditionally treated as pragmatic. The distinction hinges on what is considered to be linguistic knowledge versus what is considered to be extralinguistic or world knowledge. However, while orthodox theories locate the boundary at slightly different points, cognitive linguistics makes no sharp distinction. To all intents and purposes, linguistic knowledge is world knowledge. Any boundary imposed to extricate pragmatic meaning from semantic meaning is therefore imposed artifactually with semantics and pragmatics instead forming a continuum. In discourse, the frames accessed by a linguistic expression, including as part of a conceptual metaphor, contribute to construal.

2.2.5 Simulations

Another sense in which language may be described as embodied relates to simulations. A branch of cognitive linguistics known as simulation semantics (see Bergen Reference Bergen, Gonzalez-Marquez, Mittelberg, Coulson and Spivey2007 or Matlock and Winter Reference Matlock, Winter, Heine and Narrog2015 for overviews) argues that language is grounded in perceptual and sensorimotor experience so that understanding certain kinds of language involves running a perception-based or action-based simulation of the situation described (Barsalou Reference Barsalou1999; Zwaan and Madden Reference Zwaan, Madden, Pecher and Zwaan2005). By simulation, it is meant a mental experience ‘resembling perceptual or motor experience occurring in the absence of the relevant external stimuli’ (Bergen, Matlock and Narayan Reference Bergen, Gonzalez-Marquez, Mittelberg, Coulson and Spivey2007: 735). Simulations, in other words, provide an experience close to seeing and acting without actually seeing or acting. From this perspective, meanings are not just schematic, as is sometimes suggested by Langackerian diagrams which capture their ‘essence’, but at a higher level of conceptualisation possess ‘life’.

Such simulations are grounded in the co-occurrence of linguistic units and specific perceptual and motor experiences. They are rich in detail and include information not made explicit in the utterance but derived instead from background knowledge and past experience, such as the size, shape, orientation, weight, colour, smell, texture of an object or the relative speed, force, direction etc. of an action. Simulations are performed with simulated situatedness whereby the language comprehender is an immersed experiencer (whether agent or observer) in the situation described (Zwaan Reference Zwaan and Ross2004). In other words, simulation involves a re-enactment of the referential scene as though one were there themselves directly perceiving the objects and relations described or performing the actions designated.

Evidence for simulations comes from priming studies which show that subjects are faster in responding to an image when the image matches the image implied by a sentence (Stanfield and Zwaan Reference Stanfield and Zwaan2001; Zwaan, Stanfield and Yaxley Reference Zwaan, Stanfield and Yaxley2002) or are faster in performing an action when the action matches the action implied by a sentence (e.g. Glenberg and Kaschak Reference Glenberg and Kaschak2002), as well as eye-tracking studies which show that eye-movements ‘follow’ the direction of successive scenes described in an unfolding story (Spivey et al. Reference Spivey, Richardson, Tyler, Young, Gleitman and Joshi2000). Neuroimaging studies further show that understanding language depends on neurophysiological structures of the brain involved in actual action and perception (Hauk, Johnsrude and Pulvermüller Reference Hauk, Johnsrude and Pulvermüller2004). Evidence that understanding metaphorical language also involves simulation comes from similar studies which show that subjects are quicker to understand a metaphorical phrase like ‘grasp an idea’ if they have previously performed a grasping action (Wilson and Gibbs Reference Wilson and Gibbs2007) and eye-tracking studies which show eye-movements in response to fictive motion sentences but not in response to static spatial descriptions (Matlock and Richardson Reference Matlock and Richardson2004). Neuroimaging studies also show that areas of the brain associated with aspects of physical experience, such as somatosensory experience, are activated by metaphors like ‘had a rough day’, which draw on those areas of experience (Lacey, Stilla and Sathian Reference Lacey, Stilla and Sathian2012). Neuroimaging studies in particular provide robust evidence for the cognitive reality of conceptual metaphors.

2.2.6 Multimodality

Communication is inherently and extensively multimodal. The modes and resources that human beings exploit for communicative purposes include typography, intonation and prosody, proxemics, gesture, gaze, facial expressions, bodily stance, sound, music, images and material artefacts including objects available within the immediate communicative environment. As Turner (Reference Turner and Dancygier2017: 97) states, human beings use ‘all these resources as part of performances meant to prompt audiences to construct meaning’. Modes and resources are deployed simultaneously as part of semiotically complex performances where they interact with one another in different ways – complementing, clarifying, contradicting – to direct conceptualisation.

Cognitive linguistics has multimodality built into it where meaning is characterised as embodied, analogue and modal. Cognitive linguistics is therefore appositely placed to incorporate multimodal forms of communication within its analyses, allowing parallels to be drawn between language and other semiotic modes like image and gesture which more manifestly exhibit the kind of modal properties ascribed to language. Indeed, the theoretical apparatuses of cognitive linguistics are pre-configured to be accommodating of multimodality. For example, the range of forms that can feature in a usage event and which cognitive grammar allows to become conventionalised as part of the phonological structure in a symbolic assembly ‘includes the full phonetic detail of an utterance, as well as any other kinds of signals, such as gestures and body language (conceivably even pheromones)’ (Langacker Reference Langacker2008: 457). Langacker’s parenthetical reference to hormone-like secretions is not necessarily intended to be taken literally but to make the case for a view of language that is broader in semiotic scope and encompassing of multimodality.

The semantic properties that provide meaning to linguistic expressions are often of a spatial nature, motivated by the fundamental role of space in embodied experience (Langacker Reference Langacker2008; Talmy Reference Talmy2000). It is therefore not surprising to find expressive modes which make use of two- or three-dimensional space, such as image and gesture, reflecting the spatial discriminations made by the language usages they accompany. Indeed, where this is observed to be the case, it provides strong evidence for the spatialised conceptions ascribed to linguistic forms (Kok and Cienki Reference Kok and Cienki2016). Similarly, observing metaphors in modes besides language breaks us out of the potentially circular reasoning whereby conceptual metaphors are adduced on the back of verbal metaphorical expressions which, in turn, provide evidence for conceptual metaphors (Cienki Reference Cienki and Pinar Sanz2015: 200–201). The occurrence of metaphors across multiple modes is best explained by a common underlying factor in the form of a conceptual metaphor.

To date, multimodal research in cognitive linguistics has focussed most prominently on co-speech gestures (see Cienki Reference Cienki2016), to the extent that attention to gesture has become mainstream in the field (Turner Reference Turner and Dancygier2017: 95). However, cognitive linguistics continues to expand its multimodality program to include more of the modes that humans exploit in creating meanings for one another, whether in pursuit of art or politics (Dancygier and Sweetser Reference Dancygier and Sweetser2012; Forceville and Urios-Aparisi Reference Forceville and Urios-Aparisi2009; Pinar Sanz Reference Pinar Sanz2015).

2.4.1 Cognitive Grammar

Cognitive grammar is a theory of grammar associated primarily with the work of Langacker (e.g. Reference Langacker1987, Reference Langacker1991, Reference Langacker2008). However, I include under the same banner the model of conceptual semantics developed by Talmy (Reference Talmy2000) since both frameworks are concerned with the relationship between grammar and conceptualisation and overlap to a considerable extent. For example, both theories treat grammatical structures as having a basis in bodily experience to present semantic characterisations of grammar. Image schemas, attention and perspective figure in both frameworks as the conceptual bases of grammatical distinctions and thus as the conceptual parameters along which construals may vary. Both frameworks therefore maintain the centrality of schematic structure, arguing that the structural specifications of linguistic forms are represented conceptually by ‘abstract, idealized, and often virtually geometric delineations in particular relationships with each other’ (Talmy Reference Talmy2000: 13). And both frameworks include the representation of such schematic structure in diagrams whose relevant parts are labelled (Reference Talmy2000: 13).Footnote ⁴ As a consequence, there is often a duplication of terminologies with essentially the same phenomena receiving different designations in each model. For example, so far as I can see, the notion of profiling within the architecture of cognitive grammar, whereby a portion of an experiential complex is selected for initial attention, is synonymous with windowing of attention as defined within conceptual semantics. In the proceeding analyses, I am somewhat promiscuous with respect to these two models drawing on one or other of them, or some combination of the two, as led by the data. For example, since the representation of motion events has been extensively analysed in conceptual semantics, it is appropriate to draw predominantly on conceptual semantics in analysing representations of migration. Conversely, where the action-chain is a central feature of cognitive grammar, cognitive grammar is most pertinent in analysing representations of inherently ‘transactive’ events such as violent encounters between the police and protesters. In so far as cognitive grammar and conceptual semantics share common conceptions of language, such promiscuity and pliant use of interchangeable or ‘translatable’ terms does not pose a problem. I make extensive of cognitive approaches to grammar throughout the book to analyse strategies of structural configuration, identification and positioning.

2.4.2 Frames and Projections

The theory of frame semantics developed by Fillmore (Reference Fillmore1982, Reference Fillmore1985) relates lexical meaning to encyclopaedic knowledge. It describes frames as the organised bundles of conceptual knowledge which stand as pre-requisites to understanding the meanings of related words. The FrameNet project (Ruppenhofer et al. Reference Ruppenhofer, Ellsworth, Petruck, Johnson, Baker and Scheffczyk2016) compiles an index of the frames that receive linguistic expression in English. In discourse, frames are activated by lexical units denoting their elements to bring relevant background knowledge and associated affect to bear in the current conceptualisation.

Frames enter into relationships with other frames including relations of metaphorical frame projection. The ability to draw connections between different areas of experience, establishing correspondences between conceptual elements and projecting whole regions of conceptual structure, is fundamental to human language, the expanse of our imagination and our capacity for complex reasoning (Lakoff and Johnson Reference Lakoff and Johnson2003; Fauconnier and Turner Reference Fauconnier and Turner2002). The role of frames in metaphorically structuring other frames is described in most detail in conceptual metaphor theory (Lakoff and Johnson Reference Lakoff and Johnson2003). Conceptual metaphor theory is concerned with the relatively stable set of frame associations that are evidenced by systematic and conventionalised uses of metaphor and which account for much of the conceptual system. Conceptual metaphor theory is applied in Cognitive CDA to account for the metaphorical frame projections found in political discourses where metaphor is recognised as a key index of ideology and a fundamental instrument in the legitimation of social action. Conceptual metaphor theory is applied mainly in relation to framing strategies.

2.4.3 Worlds and Spaces

Mental spaces, described in mental spaces theory (Fauconnier Reference Fauconnier1994, Reference Fauconnier1997), are the temporary conceptual pockets that we open up and move through in the course of discourse to maintain reference, resolve ambiguities, and entertain negation and counterfactuality, among other things. As discourse unfolds, new spaces are created to represent the different times, places, texts, minds and scenarios described. Linguistic expressions including prepositional and adverbial phrases, markers of tense and modality, and conditional conjunctions act as space builders. Links are then forged between elements in different spaces to create a complex lattice of interwoven spaces which hearers are steered through in the construction of meaning, continuously assuming new viewpoints and shifting attention within the unfolding structure.

A particular kind of mental space described in discourse space theory (Chilton Reference Chilton2004, Reference Chilton2014) is a three-dimensional geometric space in which the ‘world’ espoused by the text gets mapped out according to values for space, time and modality, relative to a deictic point of reference. The ‘coordinates’ of actors, actions and events referenced explicitly or implicitly in the text are assigned by linguistic cues including pronouns, prepositional and adverbial phrases, tense and modal markers etc. or else are based on background assumptions and value positions presupposed by the text. Crucially, such worlds do not reflect objective realities but are a matter of construal. They therefore constitute ideologically invested ontologies or worldviews which readers are invited to share in. I appeal to the notions of worlds and spaces primarily in connection with positioning strategies.

2.4.4 Introspection and Empirical Methods

The standard method of investigation in Cognitive CDA remains qualitative and introspective, drawing on a combination of linguistic theory, one’s own knowledge as a language user, and an appreciation of the social and historical contexts in which texts are necessarily embedded to critically interpret the meanings evoked by a text, where meaning encompasses the full range of effects that a text potentially achieves from the conceptual to the ideological, emotional and actional. Such an approach does not, as critics of CDA in general have argued (e.g. Widdowson Reference Widdowson2004), lead automatically to overly subjective or biased interpretations. Rather, as in any scientific method, critical interpretation must be conducted with rigour, appealing to theoretical models that are plausible and well founded and applying them consistently and accurately. Likewise, introspection, on which critical interpretation relies, can be justified in much the same way as the methods found in any form of scientific inquiry. As Talmy (Reference Talmy2000: 5) argues, in any science the researcher must go to where the relevant data is to be found. Thus, in archaeology, one must survey and excavate the earth. Similarly, if one’s area of investigation is linguistic meaning, then one must go to where meaning is located and since meaning is located in conscious experience, this consists of introspection. However, the results from critical interpretation should, wherever possible, be correlated with those resulting from other methodologies – a principle usually referred to in CDA as triangulation (Wodak Reference Wodak, Tracey, Illie and Sandel2015b). This may involve collecting and analysing semiotic data from multiple modes and genres, comparing the interpretations of multiple readers, consulting large corpora, or conducting experimental research to test the effects that texts have on audiences. All these methods of triangulation are available to Cognitive CDA. However, the two that I draw on most extensively are intersemiotic comparison, where convergence between modes provides evidence of both entrenched conceptualisations and their modal (visuo-spatial) properties, and experimental investigation.