Research in Critical Discourse Studies - Website Teun A. van Dijk

What do we mean by 'Discourse Analysis'?

Many papers submitted to this journal are rejected because they do not, or insufficiently, engage in what we call 'discourse analysis', a requirement that is the first on which all articles are being evaluated.

We are aware that there are many ways of "doing" discourse analysis (conversation analysis, narrative analysis, etc.), and that also different scholars may have different opinions about what is or is not (good, adequate or simply acceptable) DA.

But still, there are some criteria most discourse analysts will recognize, so that they can tell apart discourse analysis from other kinds of analysis (like content analysis, or social analysis), or indeed from no analysis at all.

One such criterion is the special attention paid to STRUCTURES of any kind. That is, couched in terms of some THEORY, analysis may focus on --for instance-- structures of expression (sounds, image, movement, etc., including those of words, word order or sentence structure), on the one hand, and structures of meaning and (inter)action, on the other.

Thus, structures of meaning may involve such diverse ones as overall topics and their organization in text or talk, local patterns of coherence between propositions or the functions of propositions in a sequence, as well as implication or entailment, presupposition, vagueness, allusions, more or less detailed descriptions, the ways acts or actors are being described, and so on.

Similar observations hold for the analysis of action and interaction in talk, for instance in terms of turn taking, interruptions, hesitations, pauses, or the overall organization of a conversation (beginnings/endings, conventional categories that appear in a specific type of talk, such as greeting and leave taking at the beginning and end of a conversation, or formulas being used when opening or closing a session or meeting, but also typical categories of telling a story, and so on). Indeed, many other forms of (inter)action may thus be identified in discourse, such as promises and threats, agreements and disagreements, mitigating and exaggerating, attacking others or defending oneself, and so on.

Such attention to 'structure', 'form', 'organization', 'order', or 'patterns', is characteristic of virtually all contemporary approaches to discourse or conversation analysis. Some of these approaches are very sophisticated and detailed, and may be very technical -- as is the case of much work on the grammatical structures of sentences and sequences of sentences in discourse, or studies of narrative or conversational organization. Contributions to this journal should of course be aware of the current literature about the  different types of structural patterns text or talk may exhibit.

Note that such a "structural" analysis need not be limited to "fixed" or "abstract" structures, but may also focus on the more 'dynamic' aspects of discourse organization, such as the mental, interactional or social STRATEGIES participants engage in. Thus, we may analyze the abstract structures of a story or news report, but also moves and strategies of credibility enhancement, persuasion, impression formation, derogation, legitimation, and so on. And each of such more global strategies that may characterize a discourse as a whole, may again be analyzed in smaller, functional components, that is, in terms of moves, as we also know from a game of chess. For instance, a journalist may locally enhance the credibility of a news report by recurring to the semantic moves of mentioning numbers or statistics or quoting credible sources.

And finally, especially in a more psychological perspective, an analysis may not only focus on structures or strategies but also on PROCESSES, such as those of production and comprehension of discourse, the activation of knowledge or opinions during such processing, the way discourse or its meanings are represented in memory, or how mental models of events are formed or activated during production or comprehension. Such a process analysis may very well be combined with an analysis of structures or strategies. Indeed, processes involve structures or strategies of mental representations.

In other words, typical of discourse analysis is an explicit, systematic account of structures, strategies or processes of text or talk in terms of theoretical notions developed in any branch of the field.

In practice this will usually mean attention for structures that are not trivial or irrelevant in interaction and communication or that are so obvious that any native speaker may observe them. A scholarly article observing only such 'obvious' properties of discourse loses much of its point of providing new insights. This is admittedly a very subjective criterion, but all scholarly journals and publishers use it in judging a paper or book. That is, descriptions of structure, strategy or process should also be at least somewhat interesting, new or original.

This also means that merely summarizing, paraphrasing or repeating (fragments of) talk or text, something any reader may also do, is NOT a form of discourse analysis for this journal.

The same is true for merely commenting ABOUT (the contents of) a fragment of discourse without any regard for structural or dynamic properties, even when such comments may well be relevant in a social perspective.

The majority of papers submitted to this journal are rejected for this reason: They do not go beyond repeating, paraphrasing, summarizing or (merely) commenting upon a fragment of text or talk.

In other words, all comments on fragments of text/talk should be framed in terms of theoretically based categories of structure or strategy, and hence presuppose knowledge of classical theories as well as new developments in the field.

For examples of such analyses the best recommendation is to carefully read some other papers contributed to this journal. There are also many introductions to discourse/conversation analysis which discuss the kinds of structures and strategies referred to here, such as Teun A. van Dijk (Ed.), Discourse studies. 2 vols. (London, Sage, 1997).

Authors who want to contribute a paper and are not familiar with contemporary discourse analysis, are recommended to have their paper read first by a local discourse analyst. Another option is to invite a discourse analyst to be the co-author of a paper.


Especially (but not exclusively)  for authors in discursive psychology the following article of the Loughborough Group  is relevant:  Discourse Analysis means doing Analyisis, by Charles Antaki, Michael Billig, Derek Edwards and Jonathan Potter.