Can Discourse Analysis Successfully Explain the Content of Media and Journalistic Practice?

It is argued here that textual analysis of media accounts requires the study of the social structures from which competing ideological explanations develop. A comparison of the methods of the Glasgow University Media Group with the work of Norman Fairclough and Teun van Dijk shows that discourse analysis which remains text-based has problems in its ability to show: (1) the origins of competing discourses and how they relate to different social interests, (2) the diversity of social accounts compared to what is present (and absent) in a specific text, (3) the impact of external factors such as professional media practice on the manner in which the discourses are represented, and (4) what the text actually means to different parts of the audience. There are other problems with "text only" analyses in relation to (1) the accuracy of representations, (2) the significance of texts to our own audience, and (3) the question of how rhetoric "belongs to" or is used by different social interests. To overcome these problems requires a method which analyses processes of production, content, reception and circulation of social meaning simultaneously.