There was a time when the phrase discourse analysis (DA) meant something that was fairly clear and conceptually coherent.It dates from the early to mid 70s. It was part of the dissatisfaction with a linguistic tradition that dealt with nothing above phrases and sentences, a tradition that had no appreciation for "discourse" as a profoundly social venue of social action. Austin reminds us that we do things with words. An early work of this kind was by Sinclaire and Coulthard's (1975) Towards an Analysis of Discourse, Oxford University Press. (I encourage my students to get and read the preface and introduction of this text). These two sections provide the reader with a profound historical account of this development.

But, that was then, and this is now. And what 'DA' means now is very difficult to say. It has become a generic phrase. Of all people, Foucault was hugely responsible for popularizing the phrase [epistemic discourses, historical discourses, etc.]. It has become a covering phrase for 'programs' that otherwise share very very little in common. I recommend reading Foucault's The Archeology of Knowledge (1969), Routledge.

When working on my dissertation, I examined what I called the Janusian* figure of the Zimbabwean president,Robert Mugabe for his presence in both antagonistic positions in a communication battle in a political setting – as first the rebel clandestine broadcaster, and then as the sanctioned oppressor of communication. In my study, I examined message propagation in an information restrictive environment. It paid attention to this complex communication process by analyzing the discourse of messages propagated by sanctioned and clandestine radio stations.

This kind of DA encompasses, among other approaches, the idea of turn taking, the looks of transcript, ways of finding structure,and the clash of information cognitive authorities.(Read Patrick Wilson's (1983)Second-hand knowledge: An inquiry into cognitive authority.

There are other researchers who work with a different form of literature, and make use of other terms: Conversation analysis,conversational analysis or sequential analysis (Thanks to my friend, Dr. Doug Macbeth, Ohio State University). According to Macbeth, sequential analysis is fundamentally a sociological program. It treat conversation as the primordial site of "language use." It has strong affiliations to "natural language study."

In an e-mail message to this author, Macbeth argued that sequential analysis treats talk (conversations) as social action, and the achievements of talk as the achievements of common understanding and thus worlds in common.

But natural conversation is not what I examined in my dissertation research. My research is about radio messages. Is that a form of 'discourse'. Well, yes.

People have said that how we "dress" is a form of discourse too. Is it analyzable? Of course. But sequential analysis would not have too much to say about it. Sequential analysis looks at "talk-in-interaction." It looks at sequential structures, how we "go on" in conversation.

Analyzing a billboard, for example, is quite different from analyzing a moving image document. There, we must say something about its 'content'. Sequential analysis speak of what people say too -the content- but as social action. That is, what we 'mean' is inseparable from 'how we say it', e.g., the difference between a joke and an insult. Thus, the practice of talking is inseparable from what gets said. How we interpret and react to a STOP sign on a highway is different if we saw it on our neighbor's backyard.

PS: Visit University of North Texas library for an electronic version of D.N.Wachanga's (2007)Sanctioned and controlled message propagation in a restrictive information environment: The small world of clandestine radio broadcasting.
___________________________________________________
*Janus was a Roman god who guarded the doors – both the entry and the exit – of Forum Romanum, the idyllic center for the Roman People. Janus was therefore portrayed as double-faced, looking back and forward, the beginning and the end.