Puttings things into words. Ethnographic description and the silence of the social

This paper has been discussed with many colleagues in Germany, especially with Klaus Amann and Jörg Bergmann. The hard job of bringing its meaning into the English language was done by Michael Liegl (New York) and Roxana Preda (Edinburgh). Many thanks to all of them!

Not even in that seminal text which characterized ethnography as “thick description.” There, Clifford Geertz does not discuss the “description” part at all, but only the “thick” part of it. Ethnography, according to him, is less an issue of observation, than of interpretation (1973: 9). Thus, the ethnographer does not act as an author, but immediately as a reader, or rather a “literary critic” trying to “read a manuscript . . . written not in conventionalized graphs of sound but in transient examples of shaped behavior“ (1973: 10). This is a nice metaphor, but how should one read something that has not already been described? The problem of “putting into words” is covered up when cultural reality is assimilated prematurely into the textual universe of cultural studies. Compare Amann’s critique of Geertz’ notion of culture as text (1997).

An obvious alternative to grounding ethnography in participant observation, which can be found in the ethnography of communication, for instance, is to use participation only as a means of gaining access whereby the researcher is able to position himself or herself favorably in order to skim all kinds of data, and especially to get audiovisual recordings (Knoblauch, 2001). The recording situation in that case is not a field situation, with the task of verbalization of experiences, but an arrangement for the efficient production of data. Whether the brief visits in the field of a “quick and dirty ethnography” can still be called ethnography at all, is a terminological question. But if the goal is to follow long-term processes in a field at their own pace and if we are really interested in a reliable reconstruction of the participants’ patterns of relevance, it is indispensable to draw out the generation of data. In terms of terminology, I prefer a narrower notion, which takes the second part of the word “EthnoGraphy” more seriously, since the first part seems so questionable to me, that one might just as well speak of “praxeography” (Mol, 2002). (This cannot be discussed in more detail here).

For a distinction between these genres see Sanjek (1990); for a presentation of the whole process see the excellent textbook by Emerson, Fretz and Shaw (1995).

See also Bergmann (1993). Bergmann’s article is written from a conversation analysis perspective and refers primarily to audio recordings, as the most common practice of almost all forms of qualitative social research. By comparison, the use of the camera is a more specific practice, which is by no means more recent (Petermann, 1991), but which has not yet developed a methodology for visual data that could be compared with the one in conversation analysis. However, see Mohn (2002).

For example, according to the requirements of a good story, where “unimportant parts” have already been sorted out.

Bergmann, too complains that the sociological use of audio-recordings is not sufficiently concerned with the transformative quality of its data (1985: 317).

Linking up with the aspect of selectivity we may note a complementary relationship between recording and description: On the one hand, protocols can and must substitute technical recordings, which are useful auxiliaries and release researchers for observation, thus increasing their opportunities for description. On the necessity to consider selection in research not only as a problem of technical equipment but also as a problem of directing attention see Amann and Hirschauer (1997: 22).

Against any naïve-realistic expectations towards recording Bergmann (1985: 317f.) explicitly calls for epistemological “caution in dealing with audio-visual recordings” and he points to a “constructive aspect.” I will come back to these issues.

Here two kinds of differentiations have to be made, the first one regarding the form of recording: transcripts are primarily relevant for audio recordings, which make up the majority of all recordings. However, occasionally, we can also find recommendations for a “textual taming” of video-recordings through notation systems (Oevermann, 2000: 112ff.). Secondly, we find varying ways of dealing with recordings within empirical social research: Hermeneutical analyses or content-analytical interview studies are usually based on the wording in the transcript, whereas conversation analysis, which is more interested in the process of speaking, makes use of the sound- or video recordings during data analysis. Only by publication does the transcript displace the recording.

That is why it seems questionable when sequence analysis, (in conversation analysis and objective hermeneutics alike), in its prohibition of retrospection, invokes so strongly the standard of current participant-knowledge. This cannot count as a “gold standard” for a procedure which, by the selectivity of its medium, chronically stays below participants’ knowledge, while at the same time, by the exactness of its registering, leaves this knowledge so far behind.

Of course, this metaphor does not only refer to the everyday notion of “silence” – an interruption of speech – but to a blank in the kind of research that is restricted to verbal data; moreover, it refers to a mute challenge for description to “make something speak” that resists verbalization.

This problem finds written expression in the communicational privatism of fieldnotes: Since their first addressees are those who write them, they are usually illegible for others, incomprehensible and highly indexical; examples for this can be found in Sanjek (1990).

Since the 1990s however, this genre has also been taken up by a number of published “auto-ethnographies” (Ellis & Bochner, 1997).

Méadel and Rabeharisoa (2001) show that the development and selection of hundreds of descriptors is necessary – can orange juice taste “metallic” or “green”? – before it is possible for a “collective body” to have such corresponding experiences.

In ethnography, there are various ways of dealing with protocols that bring them closer to the status of an inviolable “original document” or to that of “literary subject matter,” which can be reworked for presentation purposes, or even rewritten. We could draw the following conceptual distinction: A subject-matter is made up of general literary themes and motifs which are given shape by individual design. Empirical materials, however, already have an idiosyncratic shape, and detailed structures of their own. They can be ordered and rearranged, but not created anew.

It is just this feature which makes description nearly useless for the reconstruction of conversation sequences.

This could also be put in another way: Field and text are not clearly demarcated zones; in many places they jut out in each other’s territory. Clifford (1990: 66) rightly objected to the concept of fieldnote in that it naturalizes a place protected from transformations within the (prospective–retrospective) temporality of the writing process.

Here we are dealing with the “literary” problem of ethnography in a more direct sense. Geertz already was very explicit in valuing its communicative performance over its aesthetic aspects: “A good interpretation of anything – a poem, a person, a history, a ritual, an institution, a society – takes us into the heart of that of which it is the interpretation. When it does not do that, but leads us instead somewhere else – into an admiration of its own elegance, of its author’s cleverness, or of the beauties of Euclidean order – it may have its intrinsic charms; but it is something else than what the task at hand calls for” (1973: 18). To drive this point home: a sociological description which draws the readers’ attention mainly to its aesthetic means of composition has failed.

This can also be formulated from the position of the author: The notion of observation is not only misleading because it implies a too strong separation between the subject and object of research, but also because “observation,” to the extent that it generates disciplinary knowledge only through writing processes, stands for a second membership: namely in the sociological discourse. Fieldnotes are an interface: as a local practice, they belong to the field, as a writing practice, they belong to academic discourse.

This knowing how (Ryle) has been the target of ethnomethodological “studies of work” (Garfinkel, 1986), which explicitly take professional “skills” as their object; for example, the embodied knowledge in human-machine-interaction.

It is obvious that there are fields of social practice where such participation competences can be acquired either only rudimentarily or at the cost of leaving the discipline (see the discussion in Lynch, 1993: 273–275). In other fields however, becoming a co-worker, that is, a strong hybridization of roles, can be an indispensible condition for getting access.

It was “non-verbal behavior” that in the history of visual sociology suggested itself as the “natural” object for the use of the camera (Scherer, Banse, & Wallbott, 2001). Like audio recording in regard to ephemerality, the camera cannot be outrun by description. On the other hand, it finds in the obvious its biggest problem. The camera intensifies the problem of visual self-evidence, just because it is an “optical type” oriented to the visual and hence relying on other people seeing “the same thing.” The fact that this doesn’t happen becomes immediately evident as soon as the need arises to speak sociologically about what has been shown. On the testing of strategies of using the camera in an alienating way to “watch people watching,” see Amann and Mohn (1998).

Here the comparison to photography suggests itself: this recording technology too manipulates the temporal structure of social processes. It turns them into stills and through this harsh way of decontextualization it allows for new forms of observation.

This is the serious aspect of the otherwise satirical classic “Body Ritual among the Nacirema,” portraying North-American bathroom habits in terms of religious ritual (Miner, 1956).

Outside the social sciences, language is of course transgressed in almost all the arts, especially in music; literature too has a “zone of the ineffable” (Fuchs, 1989: 163). Within sociology, the possibilities are more limited, but nevertheless, interesting: in “ethnodramatic” performances of research results (Mienczakowski, 2001) or in attempts to use visual media not only for the recording of “data,” but also as a semiotic extension of sociological communication, and as an analytically articulate form of showing (Amann & Mohn, 1998). If such a visual form of sociological communication were to become established, this might have two opposite effects for our topic: It would make the competence in description shrivel even more, and it would, once again, highlight the possibilities that are special to description.