After moving, having pneumonia, sustaining an incredibly painful injury to the ribs (from coughing), and being without the internet for two weeks (from moving), I'm back.

Moving on in discourse analysis theories, I will present my thoughts on Conversation Analysis:

Defining Characteristics of Conversation Analysis
Harvey Sacks founded Conversation Analysis in the mid 1960’s, as a means to investigate social order as it is revealed in everyday conversation. Sacks believed that there is “Order at all points.” Anywhere we look at human activity, we will find evidence of social order. Certainly the work in conversational sequencing, adjacency pairing, preference, and repair seems to bear this out. Each of these aspects of conversation are preformed systematically; they are rule governed. In demonstrating that people follow rules in their conversations, Sacks proved that social order is evident in conversation. It is believed that the theory can shed light on how social order works.

There are four basic principles of CA:
· Talk in interaction is systematically organized and deeply ordered
· The production of talk in interaction is methodic
· The analysis of talk in interaction should be based on naturally occurring data
· Analysis should not initially be constrained by prior theoretical assumptions

Sacks did not believe in the existing methodologies for linguistics or sociology, and his principles demonstrate this divergence. The last two are clearly reactions to existing theories that focus on contrived data sets or allow the theories to drive the interpretation of the data. Sacks wants us to look at real data with an open mind, and see what we can discover.

Sacks chooses to view humans as “free agents,” intentionally accomplishing actions in conversation. This too is in conflict with preceding theories in sociology, which prefer to look at people as “cultural dopes” carrying out social order unconsciously.


Strengths and Weaknesses of Conversation Analysis
One of the key differences between Conversation Analysis and its predecessors is in methodology. While it borrows heavily from the work of Erving Goffman and from Harold Garfinkel’s ethnomethodology, CA’s methodology is far more rigorous and systematic.

There is a simple, cookie cutter method that can be applied to virtually any conversational data set, even by analysts with minimal background in the theory. It is very easy to get information from your data set, and the more you look at it, the more you get out of it. The methodology allows the data to inform the theory, to show us things that we didn’t already know.

CA is also very open about publishing data. Major studies in the field are always accompanied by transcripts, which allow the reader to look at the same data as the analyst and draw their own conclusions. Of course the idea is that anyone should be able to look at the same data and draw the same conclusions.

One particularly thorny issue with CA is that it attempts to get at motivations of conversational participants, while divorcing the data of its broader context. It seeks to get at something that is very human, while at the same time dehumanizing the participants. This conflict is not resolved in the theory.

Another and perhaps more significant problem with CA is that there is no way to validate the analysis. CA does not perform post interviews like IS. All data interpretation is up to the analyst, and while the analyst’s personal background should not affect the interpretation, it can be easily demonstrated that it does. It is not the case that any analyst, looking at the same data will draw the same conclusions. This casts a reasonable doubt on any conclusions. Even an entire conversation can be misinterpreted without the right background information.

Despite CA’s attempts to be scientifically rigorous and systematic, the fact that all data interpretation is literally made up by the analyst puts the entire methodology at risk. This is not to say that CA has not produced anything of value, but overall it is too flexible to be convincing.