Within structuralist-descriptive linguistics the development of a disourse perspective entailed two reorientations: one towards prioritizing the study of actual language use and the interpretative challenges posed by variation, often making use of large electronic corpora for studying the distribution of particular structures and uses. One can notice here developments which "opened up" enquiry, e.g. by insisting on the use of complete texts as input units for, say, grammatical analysis, or by selecting specifically textual or discursive phenomena as objects of study (e.g. developing new insights into the structural functioning of cohesion devices or - in spoken language use - discourse markers). At the same time, one can detect a tendency to "contain" the interest in the discursive. A second reorientation therefore pertained to the identification of the "discursive" as a specific area of linguistic meaning (e.g. as delineated from the "semantic").


The study of stylistic variation and registers is based on the observation that language variation depends not only on the social and geographic origin, position and trajectories of the speakers (traditionally the concern of variationist sociolinguistics). It also varies according to the activity in which one is engaged in (e.g. writing up a research article, addressing parliament, giving a sermon, etc.). Diatypic variation of this kind was initially grasped through the notion of style (Crystal & Davy 1969 CRYSTAL, David & DAVY Derek, 1969. Investigating English Style. London: Longman.  ) or, within the systemic-functional framework, register (Halliday et al. 1964 HALLIDAY, Michael A.K., McINTOSH, Angus & STREVENS, Peter, 1964. The Linguistic Sciences and Language Teaching. London: Longman. , Halliday 1978 HALLIDAY, Michael A.K., 1978. Language as a Social Semiotic. London: Edward Arnold.  , Halliday 1985). Such a perspective on diatypic variation has more recently been re-affirmed in Conrad & Biber (2009 BIBER, Douglas & Susan CONRAD, 2009. Register, Genre and Style. Cambridge: CUP. ), who continue to stress the systematic nature of language variation which goes together with particular occasions of language use and locate distinctions in point, purpose and focus in the analysis of register, style and genre (Conrad & Biber 2009: 15-24). While recognizing that the terms have been used in many different ways, it is suggested that register analysis is oriented to frequencies and regularities in the occurrence of particular lexico-grammatical features which serve particular communicative functions as can be gauged from an analysis of a corpus of text excerpts which are drawn from a sphere of activity. Unlike Crystal & Davy (1969), their use of the term style echoes more literary uses, as its analysis is more a matter of distributional prevalences and (salient) preferences within a particular register or text type, often for aesthetic purposes and often associated with a single author. In subtle contrast, genre prioritizes features of textual organization which are conventionally associated with complete texts; they are part of an expected format, but are not necessarily functionally motivated. Its scope typically includes specialized expressions, forms of rhetorical organization and aspects of formatting.

In these models of diatypic variation, language use within a particular register is seen as a functional effect of situational variables. Crystal & Davy talk about "dimensions of situational constraint" which inform frequencies of occurrence at the level of language use, whereas the Hallidayan framework views register as a configuration of the semantic resources which members of a culture associate with a situation type and which correspond to a configuration of features in the context of situation (Halliday 1978: 111 HALLIDAY, Michael A.K., 1978. Language as a Social Semiotic. London: Edward Arnold.  ). Both planes of analysis are seen as organised metafunctionally. First. Configurations of semantic resources count as probabilities, that is, as likely choices among lexicogrammatical options at the textual, interpersonal and ideational levels of meaning. Register is thus seen as having a meaning potential. Second. The same metafunctional diversity is echoed in the analysis of situational context, with corresponding distinctions between mode, field of activity and tenor (cf. Halliday 1995 HALLIDAY, Michael A.K., 1985. An Introduction to  Functional Grammar. London: Edward Arnold. )

  • mode: what role is language playing? - this includes channel (e.g. written, spoken, spoken-to-be-written, written-to-be-cited) but also rhetorical mode (e.g. persuasive, rhetorical, expository, didactic, etc.)
  • field of activity: what is actually taking place - which includes as relevant distinctions particular domains of social activity such as science, religion, law, etc.
  • tenor: who is taking part? - this includes the social role relationships which obtain between the language users in a particular situation (e.g. teacher-pupil, preacher-congregation, parent-child).

More recent theoretical explorations within systemic-functional linguistics (Martin 1992 MARTIN James R., 1992. English Text.. Amsterdam: John Benjamins.  , Martin & Rose 2002 MARTIN, James R. & ROSE, David, 2002. Working with Discourse. Meaning beyond the Clause. London: Continuum.  ) have extended the model with two additional, higher planes of analysis: genre and ideology. Genre here is not a formal category, but it is used to refer to types of texts which, within a culture, enact types of social context and are recognised for doing so. Genre is seen as a goal-oriented social process (genres are used to get things done); genres are staged (there are steps involved to reach a goal); and, genres are realised through register: they can be defined as a set of register patterns. One underlying motivation has been that in a given culture, not all combinations of field, tenor and mode variables occur. Examples of genres in a Western culture include greetings, service encounters, casual converations, telephone enquiries, lectures, jokes, book reviews, research reports, etc. The introduction of the still-higher plane of ideology, finally, is motivated by an observation borrowed from the work of the British sociologist of education, Basil Bernstein: meaning potentials are not evenly distributed across participants in a given social-cultural space (e.g. Bernstein 1969 BERNSTEIN, Basil, 1969. A Sociolinguistic Approach to Sociolization: with Some Reference to Educability (vol. 1). London: Chaucer Publishing Company. ). The plane of ideology thus has to do with the system of coding orientations which position speakers/listeners in such a way that options in genre, register and language are made selectively available (with divisions along lines of class, gender, ethnicity and generation). Social power then depends on the range of options which are available to a user, the extent to which these can be used for purposes of control, submission or negotiation, and the degree to which these options can be taken up in order to transform the context which makes them available. "Interpreted in these terms, all texts manifest, construe, renovate and symbolically realise ideology, just as they do language, register and genre." (Martin 1992:581 MARTIN James R., 1992. English Text.. Amsterdam: John Benjamins.  ). The introduction of the plane of ideology aims to account for a broader societal dialectic of difference, systemic inertia and evolution. In addition, it are the "tensions produced by the unequal distribution of meaning potential that forces a culture to change" (Martin 1992:575 MARTIN James R., 1992. English Text.. Amsterdam: John Benjamins.  ).

Certainly in the Crystal & Davy and the Conrad & Biber-type of register analysis, language use is rather exclusively seen as an effect of contextual variables - a trait shared with a lot of work in variationist sociolinguistics. This leaves little room for a view in which language use also contributes to the creation of context (e.g. becoming a police suspect usually begins with being addressed in an interrogative, sincerity-doubting mode). Compare with post-structuralist theory, conversation analysis, linguistic anthropology. More recent versions of the systemic functional linguistic model (e.g. Martin 1992 MARTIN James R., 1992. English Text.. Amsterdam: John Benjamins.  ), however, have come with a fairly explicit recognition of the context-creating and transforming capacities of textual resources in a programme oriented to combatting inequality. Viewed from a yet different angle - that of a project of understanding language use as "social practice", it has been argued that such a two-directional view still comes with a prevailing "textualism": social practice is "bundled together and reduced to the 'context' of language and the focus is on how language internalises them, in a one-sided way which gives no account of how they [social practices] internalise language or how language constitutes part of the 'context' for them [...] The apparatus also pushes the analyst to the side of the system [...] the analysis of texts is overwhelmingly an account of what choices the text makes from the potential of the system, of the text as an instantiation of the system" (Fairclough & Chouliaraki 1999:141 FAIRCLOUGH, Norman & CHOULIARAKI, Lilie, 1999. Discourse in Late Modernity. Rethinking Critical Discourse Analysis. Edinburgh: Edinburgh University Press. )


Early text linguists concentrated on the development of various paradigms for the study of how sentences interconnect. They have drawn attention to the various linguistic devices that can be used to ensure that a text "hangs together" (cf. the concept of textual cohesion). Such devices include the use of articles, lexical repetition and personal pronouns to refer back to entities mentioned earlier in a text and the use of linking words to establish a particular logical relationship of, say, contrast, concession or addition between two or more sentences in a text. Other text linguistic themes include:

  • establishing criteria or standards of textuality: De Beaugrande & Dressler (1981 DE BEAUGRANDE, Alain & Wolfgang DRESSLER, 1981. Introduction to Text Linguistics. London: Longman. ) list: formal cohesion, thematic coherence, intentionality, acceptability, informativity, situationality and intertextuality.
  • developing a typology of text types (esp. written text types). The most commonly known classification is that typological variation can be reduced to 5 functional types: argumentative texts, narrative texts, descriptive texts, expository texts and instructive texts (see Wehrlich 1992 WEHRLICH, Egon, 1976. A Text Grammar of English. Heidelberg: Quelle & Meyer. ). In some versions of this theory, the 5 types tend to be viewed as textualisation-strategies. It is not uncommon for a single text to incorporate parts which fall under different functional headings (for instance, a novel may consist of descriptive, narrative and argumentative episodes; a newspaper editorial is likely to contain narrative and argumentative parts).
  • the study of how sentences functionally interrelate within particular rhetorical schemata (e.g. types of textual sequencing such as top-down and bottom-up methods of proceeding; an example of the former is a sequence consisting of a general claim > a specific application > listing arguments > giving examples; an example of a bottom-up way of proceeding is: an example > analysis > next example > analysis > a conclusion).Toulmin's (1954 TOULMIN, Stephen, 1958. The Uses of Argument. Cambridge: CUP. ) model of argument as containing the three central elements of 'claim', 'evidence' and 'warrant' and the optional components of 'backing', 'rebuttal' and 'qualifier' serves as an early precursor of such research, be it though one with normative underpinnings. For Toulmin, a good argument can succeed because it is achieved through processes of justification which correspond to a particular textual-functional layout of argumentation.
  • understanding the cognitive processing of textual units in a theory of discourse comprehension (e.g. Kintsch & Van Dijk 1978 KINTSCH, Walter & VAN DIJK, Teun, 1978. Toward a model of text comprehension and production. Psychol. Rev. 85:363-94. , Kintsch 1988 KINTSCH, Walter, 1988. The role of knowledge in discourse comprehension construction-integration model. Psychological Review, 95, 163-182. ).


Pragmatics, as a sub-discipline of linguistics, can be said to thematise the relationships between language use and the language user in a situational context (cf. the adjective "pragmatic" refers to the capacity of a social actor to adjust to situational circumstances). Initially, pragmatics was mainly bracketed by analytical philosophy, as among the first themes that were developed were indeed speech act theory and the study of principles of information exchange. Since, however, a number of thematic strands have been added, with a certain amount of import from sociology:

  • The study of presuppositions. The pragmatic interest in the implicit meaning dimensions of language use has been extended to include meanings which are logically entailed by the use of a particular structure. Presuppositions are implicit meanings which are subsumed by a particular wording in the sense that the interpretation of the latter is conditional upon the tacit acceptance of these implicit meanings (cf. pre-supposition = "an assumption that comes before"). For instance, a sentence such as "The Cold War has ended" presupposes that the existence of the entities it refers to, in this case the "Cold War". The study of presuppositions therefore often concentrates on meaning dimensions which are "taken for granted" in an utterance or a text and hence this area of pragmatic research offers an instrument which is well-suited for examining the links between language and ideology (CLICK HERE for an example analysis).

  • Face and politeness phenomena:: The pragmatic interest in the communication of indirect speech acts, in particular, as well as the interest in the social-relational aspects of and situational constraints on information exchange, more generally, are at the basis of an interest in face and politeness phenomena. One entrance to the study of politeness phenomena can indeed be built around the observation that language users often depart from the conditions of optimal information exchange because a failure to do so would result in an amount of lost face. For instance, a "white lie" can be described as a linguistic strategy through which a speaker intentionally and covertly violates the maxim of quality so as to "spare the feelings" of the person s/he addresses or in order to save his/her own face. It is on the basis of observations like the above that some pragmaticists have proposed to complement Grice's cooperative principle and its four maxims of information exchange with a politeness principle and attendant maxims (CLICK HERE for a schematic overview). A politeness perspective can also be detected in the an analysis of many indirect speech acts. For instance, the use of an indirectly formulated request such as (son to dad) are you using the car tonight? counts as a face-respecting strategy, among other reasons, because it leaves room for the interlocutor to refuse by saying sorry, it is already been taken (rather than the more face-threatening you may not use it). In this version of politeness, speaker and hearer face are simultaneously attended to.

    By far the most influential theory of politeness phenomena is that of P. Brown and S. Levinson,  Their theory is based on a particular interpretation of E. Goffman's writings on the role of "face" in social interaction (Brown & Levinson 1987:63):

      BROWN, Penelope & LEVINSON, Stephen, 1987. Politeness. Some universals in language use. Cambridge: Cambridge University Press. Our notion of 'face' is derived from that of Goffman and from the English folk term, which ties up face notions of being embarrassed or humiliated, or 'losing face'. Thus face is something that is emotionally invested, and that can be lost, maintained or enhanced, and must be constantly attended to in interaction. In general, people cooperate (and assume each other's cooperation) in maintaining face in interaction, such cooperation being based on the mutual vulnerability of face.

    According to Brown & Levinson, one can subsequently distinguish between two types of face wants: positive face and negative face. Positive face refers to the desire to be appreciated as a social person. Negative face refers to the desire to see one's action unimpeded by others. Corresponding to these two face-types, language communities develop strategies to attend to positive and negative face wants. These strategies are referred to as positive and negative politeness strategies. With particular reference to negative face wants, Brown & Levinson have developed the concept of a face threatening act to refer to verbal acts which intrinsically threaten face and may therefore require face-redressive action (Click  for a schematic overview of available options). According to Brown & Levinson, there is a direct correlation between the amount of face work speakers put in and particular situational variables: (a) power, (b) social distance and (c) the gravity of the imposition (cf. a request to borrow someone's car usually involves more face-work than a request to use that person's pencil).

    Brown & Levinson predominantly see face wants in individualistic terms. Their speaker is a rational model person, who, when interacting, adopts rational goals of which she is conscious. The underlying assumption is that the behaviour of interactants displays a sensitivity towards a satisfaction of mutual face wants. In contrast, one may stress the situational diversification of systems of politeness as well as their conventional nature. See, for instance, Bourdieu who sees politeness in terms of conventions which reflect the determinate nature of power relations in a social space. Subcription to these conventions counts as an act of political concession. Compare also with critiques of speech act theory.

    Brown & Levinson are preoccupied with "losing face", but there is hardly an equivalent discussion of "gaining face". This choice of metaphor has been criticised as ethnocentric.

    The relevance of "face" in interactional analysis can be extended beyond Brown & Levinson's particular utterance-oriented interpretation of it. Suggestions for this can be found in Goffman's own work. In addition, one can think here of situations where speakers enter into confrontations with institutions in order to (re)claim certain entitlements. In terms of scope, this takes us beyond a pre-occupation with the "local" face-related dimensions of individual utterances towards a more "global" analysis of the face work dimensions of complete exchange sequences or encounters, especially disputes (see Sarangi & Slembrouck 1997 SARANGI, Srikant & SLEMBROUCK, Stefaan, 1997, Confrontational asymmetries in institutional discourse: a socio-pragmatic view of information exchange and face management )

  • The study of reference is essentially a pragmatic theme. The focus is on how speakers establish various types of linkage between their utterances and elements in a situational context (e.g. objects, persons, etc.). One central question is the functioning of deictic elements, sometimes called shifters (i.e. lexical items such as "I, you, here, now, there, tomorrow, etc." whose referential meaning shifts with every new speaker or occasion of use). Within a linguistic anthropological strand of enquiry, deixis is viewed as a linguistic phenomenon which fundamentally challenges the view that language would be a self-contained, autonomous system. The presence of deictic elements ties up an utterance with contextually variable factors and such can even be argued to affect the meaning of other lexical items in the co-textual vicinity (see Duranti & Goodwin (1992b:43-4 DURANTI, Alessandro & GOODWIN, Charles, 1992b. 'Editors' introduction to W. Hanks, ) for a lucid argumentation to this effect). See also indexicality.

[continue to contributions from LINGUISTIC ANTHROPOLOGY]