The Haunting Question of Textual Agency: Derrida and Garfinkel on Iterability and Eventfulness

Abstract

In this article, I propose to broaden and deepen the concept of agency by focusing on what is called textual agency, the capacity of texts to do things in specific settings, and from that analysis, I draw implications for the way one conceives of how interaction works. This concept is drawn from complementary elements in Derrida's (1988, 1994) and Garfinkel's (1967, 2002) work in which deconstruction shows that for order to be found in situ, for order to be endogenously produced, as ethnomethodology contends, it is also the product of a variety of sources beyond the scope of people's ethnomethods. People's ethnomethods are crucial and intractable aspects of this order, but they do not exhaust the variety of its sources.

Notes

¹Even though Lynch and Bogen (1991) were quite critical of the way Denzin (1990) claimed to deconstruct Garfinkel's (1967) portrait of Agnes the transsexual, Lynch and Bogen (1991) never dismissed Derrida's approach as irrelevant (see also Lynch & Bogen, 1996). As Lynch and Bogen (1996) wrote

Perhaps we can begin by asserting that “truly interesting” postconventionalist writings never attempt to disown or disavow their own situated intelligibility by insinuating that there can be an alternative to local, reflexive, and rhetorical practices. Such writings find themselves always and already situated within an order of moves in some language game. They take such a locale as an unavoidable starting point for mining (and undermining) the conventional surround in interest of displaying possibilities of parapraxis. (p. 270)

In this regard, see also Lynch's (1993) insightful comparison between Garfinkel and Sacks's (1969/1986) position regarding the question of structure and what Derrida (1970) called the “structurality of structure.” I explicitly address this comparison in the third section of this article.

²Even in this case, neither Derrida (1970) nor Searle (1977) have used naturally occurring interactions to make their respective points.

³As Garfinkel (2002) noted, ethnomethodology “deliberately misreads Gurwitch's theory of the coherence of objects so as to provide for the concertedly achieved coherence of sociology's organizational things” (p. 177). See also Lynch (1993), especially what he writes on “phenomenology and protoethnomethodology” (pp. 117–158).

⁴Note that Derrida (1973) was not only playing with words here (although he does like to do that!). For him, this type of aporia is absolutely unsolvable. For instance, it is the iterable character of the forms one perceives that makes perception possible, but it is also this very iterability that prevents one from ever perceiving the pure presence of something.

⁵Regarding the question of Gestalt psychology, which is a recurring reference in ethnomethodology (Garfinkel, 2002; Garfinkel & Livingston, 2003; Heritage, 1984; Livingston, 1995, 1999, 2006), one could say that Derrida's (reflection confirms, in some respects, the idea that perception is always contaminated by representation. For instance, when one looks at a Rubin vase, which can be seen either as two silhouetted heads or as a chalice, one realizes that “immediate perception consists of totalities, wholes, or ‘gestalts’,” whereas “perception is not determined by physical sense data” (Livingston, 2006, p. 414). Perceiving, as demonstrated by Livingston's (1999, 2006) analyses of mathematicians' and checker players' skills, is something that has to be learned, acquired, and cultivated, whether it is to be able to see a sequence of possible moves in a checker game or to see a mathematical proof in a given figure.

⁶As Garfinkel (2002) wrote

EM [ethnomethodology] is not in the business of interpreting signs. It is not an interpretive enterprise. Enacted local practices are not texts which symbolize “meanings” or events. They are in detail identical with themselves, and not representative of something else. The witnessably recurrent details of ordinary everyday practices are constitutive of their own reality. They are studied in their unmediated details and not as signed enterprises. (p. 97)

⁷In this respect, Lynch (1993) wrote

It would be misleading to treat ethnomethodology's interest in indexicality as a basis for a general skeptical position. Once it is agreed that all utterances and activities are indexical, then it no longer makes sense to suppose that a system of context-free and standardized meanings can apply to all occasions of natural language use. Less obviously, however, it no longer makes sense to treat the unrealizable possibility of such a context-free system as a general backdrop for analyzing situated practices. … [W]hat becomes prominent is not that all expressions are indexical but that members manage to make adequate sense and adequate reference with the linguistic and other devices at hand. The question for ethnomethodology is, How do they do that?” [italics added] (p. 22)

⁸One could say that Garfinkel has done what Derrida overlooked, that is, to examine how people observably and accountably manage to go about their business not only despite but also by means of the fluidity of the tools they have to work with. In other words, Garfinkel has pursued the empirical side of the question of what happens rather than, as Derrida, the theoretical side of the complexities and vagaries of the tools people work with. I thank Robert E. Sanders for bringing this point to my attention.

⁹Incidentally, ethnomethodologists address this idea of an infinite context by calling it the “etcetera problem” (see especially Lynch, 1993, pp. 28–30).

¹⁰For the sake of the demonstration, I imagine that A is a male lawyer and that B is a female witness, although nothing in this transcript allows me to make such inferences with complete certitude.

¹¹For instance, we could imagine situations in which some members of the jury would never verbally problematize a specific trait that they recognize in the client (e.g., cowardice), and yet this trait would be extremely consequential in the way they assess the whole situation. In other words, if something is not problematized in interaction, that is not a reason for it to not matter, practically speaking.

¹²This spectrality is, I think, implicitly acknowledged by Livingston (1987) when he wrote, commenting what happens in a freeway traffic, “The traffic flow production cohort's practices and the order-productive character of those practices—the phenomena of practical action and reasoning—are the ghosts in the machine [italics added] that make it work” (p. 30).

¹³The parallel with ethnomethodology and conversation analysis is, again, interesting given that Sacks (1984) did not hesitate to speak in terms of “machinery” to account for the “technology of conversation” (cf. Silverman, 1998, p. 65).

¹⁴As Sharrock and Button (1991) pointed out

Ethnomethodology's respecification is … to treat the solution to “the problem of social order” as completely internal to [the sites of everyday activities]. It conceives social settings as self-organising and for just that reason has no further need for the received concepts of “social actor” and “social structure.” (p. 141)

If various forms of agency are taken into account to refer to such self-organizing phenomena, I completely agree with this position. Note that Lynch (1993) seemed extremely aware of the danger of such forms of reductionism (see, especially, pp. 147–149), which makes me think that the multiagency approach I propose is not at all in contradiction with ethnomethodology's tenets.

¹⁵To be sure, this coherence may present doubling or mixed messages, but not entirely chaotic or arbitrary messages. Texts certainly “do things,” but the spectral agency they present is not identical with the author agency. Texts give rise to ghosts who operate at varying angles of variance with author agency.

¹⁶Here, I point out that not all ethnomethodologists operate this form of reduction. Speaking of a videotape showing a typist's hands typewriting a document, Lynch (1993) wrote

On the videotape, the typed page can be seen as the product of a course of work, but when the page is read as a disengaged text, its coherent semiotic features implicate a different order of “authorship.” The completed sentences stand as document of a coherent set of “ideas,” “intentions,” “grammatical competencies,” and so forth, which no longer display the local history of production documented by the videotape. (p. 290)

Lynch (1993) even quoted Derrida's (1977a) Limited Inc:

As Derrida points out …, the “orphaned text” is far from unintelligible. Although it would be absurd to figure that one would need to observe the writer “in the act” of writing in order to understand the written text, Garfinkel's point is that a distinctive order of intelligibility—and one that is definitely part of “writing”—is opened up by inspecting the lived work of writing. (p. 290)

¹⁷Commenting on Garfinkel's (1967) famous study of psychiatric files, Sharrock and Button (1991) wrote

Ethnomethodology … assumes that (forgive the paradoxical sound of this) the specifics of readers are the essentials, that any actual reader will have specific relevances, purposes, procedures for reading, stocks of knowledge at hand, and that these will be integral to any determination of sense that the party will make. (p. 150)

I contend that the specifics of readers are indeed important, but that they are not the essentials to the extent that they are no more essential than the texts they are reading. It is precisely to fight against this prioritization of readers that the idea of textual agency is put forward.

¹⁸Speaking of the driver's environment in freeway traffic, Lynch (1993) wrote “This concrete environment is also a graphic ‘text,’ as it is composed of a grid of lines and intersections along with inscribed notations and directional signs, all of which encompass [italics added] and inform [italics added] an actively developing field of other vehicles” (p. 156). It is precisely these types of contributions (encompassing, informing) that the concept of textual agency attempts to grasp explicitly.

¹⁹Although I do not have enough space here to develop this point further, it is noteworthy that Derrida (2005) used the “as if” figure a lot in several of his texts, especially his latest ones (see Derrida, 2005).