I knew Teun A. van Dijk had started to work on context when I read his academic biography in 2005, and I have been expecting his new publications ever since. But I never expected his study to appear in the present form of “two independent, but closely related, monographs” (vii). To study context in relation to discourse is huge work involving all of the aspects of social situations on the one hand and all of the variable structures of language on the other. It is even more demanding if one attempts to develop a multidisciplinary theory of context that can be applied to the investigation of production and comprehension of discourse. Van Dijk has successfully done that, with Discourse and context focusing on developing the theory while discussing context studies in linguistics, sociolinguistics, and cognitive psychology, and Society and discourse extending the theory to social psychology, sociology, and anthropology. This joint review will show where this theory comes from, what it is, and how it is extended and applied.
The study of context forms an essential part in linguistic (as well as philosophical and anthropological) inquiry that pays special attention to parole or performance (Goodwin & Duranti Reference Goodwin, Duranti, Duranti and Goodwin1992), and much of the literature about context in this field (and also other fields like social psychology and sociology) is reviewed, discussed, and analyzed in the present two volumes. One aim of reviewing these previous studies of context is to identify their remaining problems, and on many pages, in fact, we see van Dijk's thorough and thought-provoking analysis of the identified problems.
In Discourse and context, for example, Ch. 2, “Context and language,” is devoted to assessing the work of Systemic Functional Linguistics (SFL) in relation to context. Works of various authors, such as Bronislaw Manlinowski, John Firth, M. A. K. Halliday, Jim Martin, Suzanne Eggins, Micheal Gregory, and Philipp Wegener, are discussed, and many concepts of context, such as context of culture, context of situation, register, and genre are reviewed. Van Dijk identifies that the three register variables of field, tenor, and mode are rather arbitrary, and the mapping of these variables on the three metafunctions of language and on the language structures controlled by them are also arbitrary, incomplete, and confused (54). In Ch. 3, “Context and cognition,” van Dijk reviews studies of “context-effects” in cognitive psychology and comes to the conclusion that they bear no overall theory of context as a specific mental construct influencing discourse production and comprehension. Ch. 4, “Context and discourse,” assesses the work of sociolinguistics on context, analyzing two of its problems: (i) the direct correlation of gender, class, and age with the pronunciation of the speaker, and (ii) the influence of context on language (e.g. pronunciation, lexicalization) rather than on discourse (e.g. topic choice, turn-taking, and the strategies of persuasion and manipulation).
Similarly, Society and discourse also discusses problems of previous context studies, but shifts its focus to sociology, conversational analysis, and linguistic anthropology. In Ch. 3, “Context, situation and society,” the sociological studies of Max Weber, Alfred Schutz, Erving Goffman, and many others are reviewed. Although some of the studies are consistent with van Dijk's study and in one way or another provide him food for thought, many others lack sufficient attention to properties of context. For example, “social situation,” an essential notion in context study, is studied less than notions such as “interaction” in modern sociology (95). Van Dijk has a ambivalence in this chapter about conversation analysis (CA). The classical, pure CA, defended by Emanuel Schegloff, is considered a context-free, autonomous approach, scrutinizing the mechanisms of talk-in-interaction itself. The same is true with the CA-oriented studies of institutional interactions that also focus on the details of actual talk, leaving context properties (e.g. goals, roles, identities, and power) unattended in the analysis of talk and text. In Ch. 4, “Context and culture,” van Dijk reviews in great detail context studies by scholars such as Dell Hymes, John J. Gumperz, Stephen Levinson, Alessandro Duranti, and William Hanks, but only to show that these anthropological studies of context lack a cognitive dimension.
To van Dijk, however, not all previous context studies are problematic. Some have in fact inspired him or at least have been consistent with his study. The profound analysis of those studies in psychology, sociology, and social psychology, for example, even enables van Dijk to find room for developing his new theory, which, in sharp contrast to previous context studies, is systematically formulated and significant in a number of ways. For example, it first emphasizes the nondeterministic feature of the relation between social situation and discourse; it also highlights that such a relation is by its definition personally and situationally variable. The main thesis of the theory is that it is the definition, interpretation, representation, or construction of participants of the their social situation, in terms of subjective context models, that influences how they speak, read, listen, and understand. In other words, societal or situational structures can only affect discourse through the mediation or the interface of the mental representations of language users. Based on this, a fundamentally different concept of context emerges, in which contexts are not kinds of objective condition or direct cause, but rather subjective constructs designed and ongoingly updated in interaction by participants as members of groups and communities.
Principles, key concepts, and tenets of this new theory are stated and restated in the two introductory chapters of the two volumes and summarized in the two conclusions (Ch. 6 in both). Detailed explorations are discussed in Ch. 3 (“Context and cognition”) in Discourse and context and extended throughout several chapters in Society and discourse. In the elaboration, van Dijk's theory derives an important concept—context model—from “mental models” in cognitive psychology. A mental model is a subjective and socially based construct of participants about the relevant properties (for them) of a social situation. To apply the concept of mental model to language communication, in which language users are involved in processing discourse as well as in dynamically constructing their subjective analysis of the communicative situation, context model is devised for the crucial interface between mental models of events and mental models of the discourses about such events. Thus, context models organize the way our discourse is strategically structured and adapted to the whole communication situation in the same way that more general mental models organize how we adapt our action to the social situation and environment. According to van Dijk, the concept of context models is very helpful in explaining properties of discourse that can not be explained by theories that assume that discourse is directly controlled by social situations. For example, context models can explain how news articles about the same event in different newspapers will always be different when written by different journalists, and why a journalist cannot possibly tell the same story twice in different circumstances, and for newspapers, with different constraints on reporting. Context models also explain, for instance, how readers retell what they have read in the newspaper or seen on TV in subsequent conversations (72). Van Dijk lists as many as nine properties of context models, in addition to eleven more general context models that can explain what they do (71–72).
In Ch. 2, “Context and social change,” of Society and discourse, the cognitive theory of mental models is extended in the discussion of social psychology. In cognitive psychology, contexts are defined as mental models in episodic memory, but in the perspective of social psychology these models are not only forms of personal experience or subjective interpretations of communicative situations, but are also constituted in ongoing interaction with other participants, embodying many aspects shared with other language users, such as their social identities or group memberships. In this chapter such social psychological notions as social situations, social actors, and social beliefs are discussed in great detail, and van Dijk extends and elaborates his theory to the fullest extent as a sociocognitive approach. In his discussion of social beliefs, for example, notions of attitudes and ideologies are introduced and taken as general mental representations shared by the member of social groups (81). These attitudes or ideologies, like context models, control the formation of the representation of the event we talk about, and also the representation of the current communicative event, that is, the social situation, including the setting (time and space) and speakers (social identities). Thus, in the perspective of social cognition, the concept of context models is extended from cognitive features of various kinds of personal (and socially shared) knowledge to social features that involve dimensions of social beliefs (e.g. ideologies) and social relations (e.g. power and identities of social actors).
The extension of context models is also seen in Ch. 3 “Context, situation and society,” and Ch. 4 “Context and culture,” in Society and discourse. Following the definition of context models as personal interpretations of situations as well as socially shared understandings of time, place, participants and their roles, identities, and relations of power, Ch. 3 examines in more detail the societal basis of such categorizations, as well as their institutional and organizational embedding, adding more to the sociological dimensions of such a theory of context. Van Dijk elaborates, for example, on the relations between context and interaction, the institutional embedding of social actors, interactions and situations, and, more generally, the influence of social structure on the definition of the situation known as context models. It is then understood that once language users are taken as social members who are able to model social structures, such structures are linked to talk and text.
In the same way that Ch. 3 enables the theory of context to account not only for the face-to-face situations, but also for the macro structures of the society, Ch. 4 discusses how the theory accounts for the cultural specificity and variability of the ways cultural members define communicative situations as contexts. In a discussion of various anthropological studies of context, van Dijk shows that, though there is no such an equivalent of the very Western notion of context in Chinese, Japanese, Arabic, Turkish, Hungarian, Thai, and many native American cultures, the fundamental structures and categories of context models are not that different. Thus participant relations, social identities (e.g. power, status, kinship, gender, and age) and social beliefs (e.g. attitudes and ideologies) might be very general for human interaction, if not universal.
In addition to the analysis of previous context studies and the formulation of the sociocognitive theory of context models, van Dijk applies his theoretical framework to a contextual analysis of political discourse, namely, a speech by the former British Prime Minister Tony Blair in a parliamentary debate about the Iraq war. Case studies have very important roles in van Dijk's work; the whole present study begins with a case study (see Ch. 1, “Towards a theory of context” in Context and discourse) and concludes with another. This case study (see Ch. 5, “Context and politics” in Society and discourse) is more extensive than others from van Dijk's work (1998, 2002) and features the following four points.
First, it involves most of the categories or properties of context models. Besides the obvious category of setting (current date, parliament, Britain), the analysis displays to a great extent the context modeling of the speaker's combined identities as a current speaker, a prime minister (PM), a member of parliament (MP), a leader of the government, and someone who favors the war against Iraq. Second, much of the political points of the speech are seen to be inferred from the context models of the participants rather than explicitly expressed. Third, properties of context models are controlled by various parliamentary rules, but these rules need to be applied to the current situation by each MP's context models. Fourth, the modeling of power relations between participants is also fully analyzed, as the usual arrogance of Tony Blair is seen in the analysis of his reply to “The main parties?” interruption by Liberal Democrats.
The contextual analysis of the debate, though fragmented, confirms the theoretical claims in the study summarized by van Dijk (245–46). It illustrates the process of context modeling both in the direction of text and talk as constituents of their context and in the opposite direction of text and talk constituting their context. In other words, the case study illustrates the ongoing process of how social context influences speech and vise versa.
In sum, it seems that the present study applies to two principles of critical discourse analysis (CDA). First, CDA is multidisciplinary, and this study brings discourse and political science together. In the case of the parliamentary debate, PM and MPs not only engage in grammatical and meaningful talk, and not only follow the rules and strategies of interaction, but also and especially engage in political action, such as defending and attacking policies. This is seen in the context analysis followed by the theory of context models, and definitely not, say, in classical conversational analysis. Second, CDA sees that discourse and society are mediated, and this study brings about a sociocognitive interface in between. In the context analysis of the debate, the subjective, on-going construct of the speaker's identity greatly contributes to the use of language forms, such as choice of topics and style. Thus, the new theory of context models cracks the shell of the sociocognitive interface that links forms of language use to their social and communicative situations. In short, this theory is very useful for understanding how discourse is socially shaping while socially shaped.
