Critics of discourse say that it is not real - events and objects are real, but not thoughts or ideas.

van Dijk has tried to deal with this from the perspective of cognition (van Dijk, 2001, pp. 108-112). In particular he proposes that we form cognitive models of both the contexts we are in and of the events that occur.

However, in my work, I have created an external representation of the context. I have created a 60 webpage set of data, summarising the history of food production, the food production companies, environmental agencies, cultural issues around food (such as recipes, advertising) and official bodies. Although a survey like this is always incomplete (I missed out a page from a typical supermarket, for example), it can give an idea of the context in which the 'players' operate. In this case the survey was done with the aim of looking at Government policies, and this may have influenced the way I gathered the data. Potentially, further data could be included until all interested parties agreed that this was a valid representation of the context of food and farming in the UK. Admittedly, each of us reads this data and modifies some sort of internal mental map (cognitive context), and none of us think in an identical way. However, I think it is possible to reach some sort of external consensus context as a working basis for future discussions.

Returning to the cognitive approach, it is clear from van Dijk's theories that CDA can only succeed if discourse structures can be related to internal structures of local and global context. (though I claim that an external consensus context is possible, the remainder of this essay focusses on theoretical approaches to mental models in CDA)

After commenting on mental models, ideologies, situations, aims, and social groups and institutions, van Dijk states that these need to be defined in explicit theories, and he provides the first fragments of such theories. He defines two main types of model - context and event models. He has systematised his work in a very thorough way, and it provides very interesting background material, though I have not used it directly here.

Wodak also gives several instances where mental models are operating (Wodak, 2001, pp. 64-5)

	If we take politicians as a specific (and non-homogeneous) elite, then they are best seen as
	Shapers of specific public opinions
	Seismographs that react to the atmospheric anticipation of changes in public opinion, and to the changing interests of specific social groups and affected parties
	It is likely that they fulfil these functions through the use of mental models, plus some special antennae
	The relationships between media, politics and 'the people' are very complex, and we cannot clearly say who influences who, and how these influences operate. Again, this entire process seems to involve mental models and sensitive antennae.

Mental models provide a range of useful tools which might allow us to imagine the mental models of the authors of the text.

Social structures (and social models)

Fairclough has described how social practices are networked together, constituting a social order, and dominance and hegemony operate in these social orders. He uses a short text by Tony Blair as an example. From this, Fairclough draws an initial conclusion that the economic is colonising the non-economic fields (art, environment .. ), and there is a new dominance by the new capitalism.

He points out that globalisation has a very long history, and is more accurately seen as 'new relations between scales' rather than a change from 'national to global'. Capitalism, globalisation, 'the economic', art, etc. can all also be looked at as social structures.

Language and semiosis play a major role in the restructuring of capitalism, and in particular, the idea of the new 'knowledge-based economy' :

	knowledge is produced, circulates and is consumed as discourses.
	The discourses are operationalised as new ways of acting and interacting (including new genres)
	The discourses are inculcated as new ways of being, new identities, new styles
	For example, there are new ways of organising and managing companies
	The restructuring and rescaling of capitalism is partly a semiotic process - restructuring and rescaling of orders of discourse
	This involves new structural and scalar relations between genres, discourses and styles
	Much of these processes also relate to the language used in the restructuring - and perhaps neo-liberalism is part of this too.

Fairclough concludes that discourses are diverse representations of social life which are inherently positioned - different social actors see and represent social life in different ways, with different discourses. For example, people have different social lives at work and at home. (Fairclough, 2001, p. 123-7)

Scollon also adds (mostly adapted by me) (pp. 139-141) :-

	Social problems in our contemporary world are inextricably linked to texts ….. Social problems are couched in public and private discourses that shape the definition of these problems as well as inhibiting social change.
	Our actions are frequently accompanied by language and, conversely, much of what we say is accompanied by action
	The programme of CDA is founded in the idea that the analysis of discourse opens a window on social problems because social problems are largely constituted in discourse
	Nevertheless, it remains problematical to this programme to establish the links between discourses and social actions
	Also, there is sometimes an almost unbridgeable gap between the discourse and the people whose social actions are involved, e.g. with HIV, the 'official' discourse is almost entirely disconnected from the drug users and others who suffer from HIV, or whose behaviour means they are likely to get infected

Once again, we have a picture of discourse that clearly affects social structures and systems, but the way that it does this cannot be directly ascertained.

Please contact me if you are interested in using my services.
(0044)(0) 1372-749803