Critical discourse analysis: how to do critical discourse analysis – a framework for analysis

For key information relating to this material, including copyright details and how to reference this work, visit the introductory page Critical discourse analysis

The criteria for the Frame (Hyatt 2003, 2005), to be applied to texts, genres and discourses are outlined in Figure 1 below:

The criteria have been ordered for pedagogical purposes and to allow analysts to move from the more micro elements of lexico grammar, through discourse semantics, register and genre. This allows the mapping of texts onto the notions of language, and the extra-linguistic levels of context and ideology. Figure 2 (below) diagrammatically represents the ways that these criteria relate to linguistic and extra linguistic elements of systemic functional linguistics (SFL) theory (Hyatt 2007, Eggins 1994).

Analysts could and should supplement these criteria according to their contexts, the context of the text(s) under examination and the needs of the research project.

(1) Pronouns – Participant Choices

This aspect of the Frame considers the way in which pronouns may be used in the text, whether they are inclusive (our, us, we, etc.) or exclusive (they, their, them, he, she, it, you, your etc.). It also considers how the reader and other participants are positioned as allies or in-group members with the author, thus assuming shared knowledge, beliefs and values, or how readers and other participants are marginalised as ‘outsiders’ with different beliefs and agendas. Pronouns are central to the way individuals and groups are named and so are always political in the way they inscribe power relations.

(2) Activisation / Passivisation

Transformations of active constructions into passive forms can be motivated by the desire to elide agency and therefore systematically background responsibility for actions in some instances or to foreground responsibility in others. The manipulation of agency transparency serves to construct a world of various responsibilities, and power, e.g. ‘The present perfect is used to …’. By removing the agent, the use of a particular grammatical form is given an unquestionable, universal function, in spite of its context of use and the political dimensions raised here. Such an analysis is almost always absent from textbooks and grammar reference books using such definitions.
It is important to note that to assume that such a basic transitivity shift as passivisation or activisation would lead to a complete shift in the understanding of the reader would be an over-simplification and patronising to the reader. However, as noted earlier, the construction is effected through a layering of strata of representations and the claim for relevance of this aspect of the Frame is as one of these myriad strata.

(3) Time – Tense and Aspect

This relates to the way in which tense and aspect are used to construct ‘understanding’ about events. For example, the use of the present simple tense constructs an event as reality or fact; the use of the present perfect simple constructs a past event as being of relevance at the moment; the past simple tense can represent a past event as no longer being important or relevant. The effect of tense choices can be demonstrated by converting the past simple tenses to present perfect and vice versa and noting the different semantic effects.

It is therefore important to understand that choices made in terms of tense and aspect are not merely concerned with the time frame of an action or process but also impact clearly on the representation of that action or process as true, relevant or significant.

(4) Adjectives/Adverbs/Nouns/Verbal Processes

The use of loaded, dramatic, and stereotyping adjectives, adverbs and nouns are central to the construction of an event or a person, whether or not that construction is evaluating its object positively or negatively. Also the use of non-hedged adverbs, such as surely, obviously, clearly and so on, position a contention as being incontrovertible ‘fact’. The use here of overgeneralisation and overstatement is worthy of note. All-inclusive expressions (all, every, none, no-one, always never etc.) are rarely accurate, but can be used to construct a generalising, stereotyping or over-simplifying evaluation. Other comment adjuncts expressing the authors attitude to the whole proposition, such as ‘constantly’, ‘totally’, ‘entirely’, ‘absolutely’, ‘wholly’, ‘utterly’, etc. fulfil the same purpose.

The concept of evaluation is useful here. Hunston and Thompson (2000: 5) define evaluation as ‘the broad cover term for the expression of the speaker or writer’s attitude or stance towards, viewpoint on, or feelings about the entities or propositions that he or she is talking about’. Evaluation can further be divided into two main categories, inscribed and evoked (Martin 2000). In the inscribed category the evaluation is carried by a specific lexical item, overtly displaying the attitudinal judgement of the text producer e.g. excellent, terrible, etc.

In addition to inscribed evaluation, it is also important to consider what Martin terms evoked evaluation. This type of evaluation uses superficially neutral ideational choices but which have the potential to evoke judgmental responses, in those who share a particular set of ideological values. These evoked evaluations, in themselves do not denote the text producer’s attitude to the content overtly, but leave the value judgement to the reader/listener. However, they are mechanisms through which evaluation is covertly constructed. For example in tourism texts (Cunha de Freitas 2000) the terms natural and sunny operate at a experiential level yet do help to construct a positive image and in food promotional text terms such as natural, and organic operate in a similar way. Negative evaluation can also be constructed by terms such as suspected asylum seeker. Such mechanisms can be seen as powerful devices in a hegemonic view of language construction in the role they play in projecting a notion of ‘common sense’.

(5) Metaphor – literal and grammatical

Metaphor is more than just a literary device – it plays a fundamental part in the way people represent social reality. The use of metaphor is central in the way it positions what is described and the reader’s relationship to this. This is starkly seen in the description of individuals or the personification of entities, e.g. Saddam Hussein is a ‘monster’, Margaret Thatcher was the ‘Iron Lady’ etc. It is also significant to realise that the metaphor and its alternative congruent or literal form do not express exactly the same meaning – indeed the purpose of metaphor is functional in that it serves to construe a differently foregrounded meaning than its alternatives. Metaphors are neither better nor worse than their congruent counterparts – they are simply performing different functions.

It is significant to note that metaphors need not only be lexical but can be grammatical as well (Halliday 1985: 319-345), whereby the meaning is expressed ‘through a lexico-grammatical form which originally evolved to express a different kind of meaning.’ (Thompson 1996: 165).

One clear example of grammatical metaphor is nominalization, or presenting as a noun or noun phrase something that could be presented with other parts of speech, e.g. her understanding as opposed to what she understood. This has the effect of making a text more ‘lexically dense’, a feature commonly noted with ‘written’ texts. Characteristic of this are more ‘packed’ texts, texts that are more information heavy, can make these texts appear more prestigious, academic, and serious. It can construct an argument as significant and well thought through. Ivani č (1997: 267) notes that through the process of nominalisation ‘…writers identify themselves with those who engage in such knowledge compacting, objectifying and capturing practices’ and so can represent themselves as ‘intellectual’ or those who use ‘reasoned thought’.

(6) Presupposition / Implication

Presuppositions help to represent constructions as convincing realities and there are a number of lexico-grammatical means by which this can be achieved:

• the use of negative questions and tags which presuppose a certain answer – isn’t it the case that…?, wouldn’t it be fair to say that…?, you’re in even more trouble, aren’t you?;
• the use of factive verbs, adjectives and adverbs, verbs that presuppose their grammatical complements, adjectives and adverbs that describe entities and processes they presuppose, and therefore represent them as facts – we now know…, we realize…, we discovered that…, you forget that… I believe that…, as you will be aware…, odd…, obvious…, previously… and so on. Factive verbs have been noted in Hoey (2000) as a form of embedded evaluation;
• the use of change of state verbs which presuppose the factuality of a previous state – when did you stop beating your wife?, their policy on Europe has changed…, this school has improved…; transform, turn into, become, and so on;
• the use of invalid causal links presupposing that if one fact is true then the next is also true – ‘90% of my class passed FCE this year, 80% of my class passed last year, therefore my teaching is getting better…’;
• rhetorical questions, which pre-suppose the answer implied by the questioner in open questions – Is it not reasonable to ask the PM such questions? – or in the case of closed (wh-) questions provide the questioner with the opportunity to answer their own question, the question they have framed and therefore presuppose the self-response as ‘true’ – What did they do to British manufacturing industry? They destroyed it, that’s what.

(7) Medium

The conversationalising of a text is a form of interdiscursivity, which goes beyond the ways in which texts borrow from, steal from and interpenetrate each other, to the ways in which genres and discourses do this. Examples of interdiscursivity can be seen in the way in which the discourse of business has penetrated the discourse of higher education (Fairclough 1993), with the perception of students being addressed more explicitly as customers and the attendant implications of this managerialist discourse – value for money and accountability being positively associated with this change, and the changing perception of teachers as being in need of scrutiny (Smyth 1995, Hargreaves 1994) being the negative aspect. In the same way the presentation of advertising copy in a conversational style serves to imply a close social relationship between the copywriter and the reader, which does not exist. This ‘masquerade’ (Hyatt 1999) of friendship, a shared communication with a trusted confidant, an individual projected as someone you can believe in, who wouldn’t lie to you, who has your best interests at heart, can predispose the text receiver to believe what the text producer is communicating.

(8) Audience

Central to the notion of language as a social semiotic is the idea that language is utilized for some form of communication, and therefore a party or parties at whom communication is aimed, in other words, the audience. Any analysis would therefore be inadequate if it did not focus some attention on who is perceived as being the audience, and how they are projected in terms of social distance – relationship to and familiarity with the text producer – and status. In light of the fact that there is no way that the author can know exactly who the audience is, the notion of audience can be read as an idealised, projected construction. In this idealisation and projection, clues can be found as to the ideological presuppositions of the text producers.

(9) Visual Images

Significant recent work in these visual and multi-modal areas has been conducted by Kress and van Leeuwen (2001: 4), particularly looking at “the four domains of practice in which meanings are dominantly made”.

“We call these strata to show a relation to Hallidayan functional linguistics, for reasons of the potential compatibility of description of different modes. We do not however see strata as being hierarchically ordered…Our four strata are discourse, design, production and distribution.”

In this theory of multimodal communication, discourses are seen as socially constructed knowledges of reality, designs are the uses of semiotic resources to realise discourses, production refers to the ‘organisation of the expression, to the actual material articulation of the semiotic event’ (2001: 6) and distribution as the facilitation of the pragmatic functions of preservation and distribution.

Historically, the association of the camera recording ‘a set image’ and as such being associated with ‘truth’ and ‘objectivity’, has impacted on the way visual images are read. Despite the potential for the manipulation of images, and the potential for displaying an image with a constructed impression of its contextual setting, visual images do play a powerful role in the construction of truth and reality. In this respect there are clear relationships with notions of hegemony in presenting a picture of ‘this is how it is’. As Fairclough notes (1995:7) images have primacy over words.

(10) Age, Class, Disability, Gender, Race/Ethnicity, and Sexuality Issues

Within a text it can be revealing to note any comment regarding individuals who may be projected as less socially valued, as a result of these issues, in order to legitimize the assertions of those who hold power, or to identify any pejorative or stereotyping presentation or labelling of such people as being a ‘normal’, naturalised and commonly-shared viewpoint.

Cole (2001) has noted the impact that labelling has had in the area of educational inclusion, noting Ballard’s (1995) argument that the language of Special Educational Needs and in particular the term ‘special’ ensures continuing segregation, and well as Corbett’s (1996) use of the term ‘bad mouthing’ to represent the type of labelling which lays the blame for barriers to inclusion on individual ‘deficit’ rather than systemic failures, such as the cultures, practices and policies of educational institutions.

(11) Reference to other texts, genres, discourses and individuals

One consistent way it which texts from all genres seek to establish the legitimacy of their claims, their common-sense assumptions and their world views is through reference to other texts, genres, discourses and individuals. Fairclough (1992) offers the terms interdiscursivity (or constitutive intertextuality) for the wider appropriation of styles, genres and the ideological assumptions underpinning discursive practice. Interdiscursivity operates on a more macro level than intertextuality and refers to the diverse ways in which genres and discourses interpenetrate each other, as exemplified previously with the examples of the co-penetration of the discourses of advertising, science and medicine, and the discourses of academia and consumerism (Fairclough 1993). Intertextuality is perhaps better viewed as the identifiable (either clearly or more indistinctly) borrowings from other texts. Quotation from, citation of and reference to other texts are lucid examples, whereas the use of phrasing, style and metaphor originating in other texts may be more opaque, yet equally revealing.

The impact of intertextuality, where used as a technique for particular construction, representation and projection of preferred meanings, can be to support reinforce and legitimize the argument of the writer. Careful selection and editing of ‘borrowed’ texts, and the utilization of other genres and discourses can achieve required evaluation, yet reference to other texts, directly through quotation or indirectly, retains projected links to ‘reality’ and, hence, claims for the truth-value of the assertion. Key figures are often used as their status is used to imply a legitimising respectability and again support the claim to the truth content of the writer’s assertions. (c.f. the way academic writing uses quotation and citation of key research literature.)

Return to introductory page Critical discourse analysis