Developing Feminist Conversation Analysis: A Response to Wowk

I have documented elsewhere (Kitzinger 2004) the intellectual history that led me—as an already established lesbian feminist scholar—to the decision to spend a sabbatical year at the University of California at Los Angeles, learning the foundations of conversation analysis from leaders in the field. The article Wowk critiques was my first attempt to formulate what might be involved in doing conversation analysis from a feminist perspective. Since then, I have published work in three distinct domains: feminist work that is not conversation analysis (e.g., Kitzinger and Willmott 2002; Kitzinger and Wilkinson 2006)—despite the fact that some of it analyses naturally-occurring talk-in-interaction (e.g., see Shaw and Kitzinger 2005 for a thematic analysis of calls to a helpline); conversation analytic work that is not feminist (e.g., Land and Kitzinger 2007a; Lerner and Kitzinger 2007a; Wilkinson and Kitzinger 2006)—despite the fact that some of these co-authors are feminists as well as conversation analysts; and feminist conversation analysis (e.g., Kitzinger 2005a, b) which embodies my commitments both to CA and to feminism. The claims made here about my work in feminist conversation analysis apply of course only to that subsection of my published work that I identify as feminist conversation analysis.

I am baffled by Maria Wowk’s (2007) claims (quoting Jeff Coulter and Harold Garfinkel as her sources) that social constructionists fail to appreciate the “objective facticity” attributed by social members to gender (as to other social constructions, p. 135). The term “social constructionism” derives (as she acknowledges, p. 136) from the work of Peter Berger and Thomas Luckmann. They say (and I have endorsed their claims in my own work, e.g., Kitzinger 1987):

The reality of everyday life is taken for granted as reality. It does not require additional verification over and beyond its simple presence. It is simply there, as self-evident and compelling facticity. I know that it is real.” (Berger and Luckmann 1966, p. 37)

There are many specific criticisms raised by Wowk which there is no space to address in this article—and in any case I take it that a list of detailed point-by-point rebuttals would be tiresome to the reader. I will occasionally footnote some such issues as they are relevantly invoked by my main text. Note here that Schegloff’s reference to conversational practices as the “infrastructure” of interaction has resonances with the ‘macro’/‘micro’ dichotomy which he discusses more comprehensively in “Between macro and micro” (Schegloff 1987b): note that (as do I) he uses these terms while acknowledging their “utter relativity and likely hopelessness” (Schegloff 1987b, p. 208). For a feminist discussion of the macro/micro dichotomy see Speer 2005, pp. 97–109. For a discussion of the different levels of analysis involved in a CA study “ranging from relatively ‘micro’ levels involving the design of individual turns and actions, to more ‘macro’ levels involving the encompassing system of speech exchange and its relationship to the broader institutional and sociopolitical environment” see Clayman and Heritage 2002, pp. 21–25.

“A person” here replaces Wowk’s formulation, “a member of hotel staff”—which she employed since it was such persons who featured in the comparison of exchanges made in the article she was critiquing. I take it that she means her statement to apply to persons in general, and not simply to hotel staff.

Wowk (2007, p. 146, ft. 23) incorrectly reports that I was “the interviewer in one of these sessions”. In fact (as I say in Kitzinger 2000, pp. 181–182) the ‘session’ to which she refers was a university seminar group run as part of a course on human sexualities and I was the teacher running the seminar group. There was no “interviewer.”

In the interests of space I have reproduced only a part of the data (for more context, see the longer data extract in Kitzinger 2000, pp. 182–183), although I have retained the same line numbering as in the original. (Note that a typographical error [an extraneous “=” at the original line 24] has been corrected.)

In the co-present data that lacks video I cannot be sure that there were not head nods or other non-vocal signals of recipiency. All the data extracts included in my collection are either from telephone conversations (where body behavioural cues cannot be used) or from video-recorded interaction in which I have inspected the data for non-vocal responses.

The remainder include: anticipatory completions of the “if” component (Lerner 1996), corrections to the “if” component, and a small set of instances of interjacent overlap (Jefferson 1986) analyzably prompted by affiliative interactional concerns (see Kitzinger 2008). These map closely on to Lerner’s (1996) findings and will be discussed in more detail in Kitzinger (in preparation).

See Rod Gardner (2001) for a discussion of the differential import of these various response tokens in talk-in-interaction. I develop analyses of these specifically in relation to if/then constructions in Kitzinger (in preparation). Also in Kitzinger (in preparation) I analyse and discuss the minority of those parenthetical sequences (six of the 28 parentheticals in my collection of 300 if/then TCUs) in which more than a response token is produced by a recipient.

My comments here are not uniquely a critique of Wowk. There is a substantial body of work on gender and language that treats the use of particular words (e.g. “boy,” “woman,” “lady” or “lad”) as offering adequate evidence of participants’ orientations to whatever those words are taken to denote (e.g. gender). Not only is this deeply problematic (since gendered linguistic terms are neither necessary nor sufficient to show participants’ orientation to gender, Kitzinger 2007a; Stockill and Kitzinger 2007; Wilkinson and Kitzinger 2007), but this discussion deflects attention away from what I take to be the core meaning of the phrase “participants’ orientation” for conversation analysis.

This formulation of “participants’ orientations” has led to a misunderstanding that occasionally surfaces in data sessions (sometimes accompanied by use of the phrase “next turn proof”), according to which a recipient’s understanding is definitive of its import and the utterance itself is taken to have no “objective” meaning. As Schegloff (1996a, fn 6, p. 173) says:

This view (and its attribution to conversation analysis) is mistaken on many counts, not least of which is its total subversion of the possibility of analytically specifiable ‘misunderstandings’ …. Care must be taken, then, to distinguish between, on the one hand, the professional analyst’s undertaking to establish the understanding of some utterance in some interaction—which should indeed seek to ground itself in the recipient’s displayed understanding, if possible, and on the other hand the recipient’s undertaking to understand the import of some utterance, which clearly cannot be so grounded, for that would presume its own outcome.

Contrary to Maria Wowk’s claims about EM/CA, conversation analysts do not eschew the term “subjective” nor necessarily reject cognitivism: see, for example, Emanuel Schegloff (1996a, p. 157) on “The defense of intersubjectivity,” in which he formulates a conversation analytic perspective on “socially shared cognition”; Robert Hopper (2005) writing as “a cognitive agnostic in conversation analysis”; and Paul Drew’s (2005, p. 180) discussion of “cognitive states manifest in the details of talk.” This is a topic of ongoing discussion within conversation analysis: see Hedwig te Molder and Jonathan Potter 2005 and the collection of articles on cognition in Discourse Studies 8(1) (including Kitzinger 2006c) for current debates.