Some Thoughts on Ethnography

I recently read Robert Lowie’s The German People. It’s an ethnography of sorts of German culture, at least in the sense that it’s a “writing of culture” (more on this text as ethnography below). So far as I can tell, it’s a largely forgotten book, certainly much less widely read by anthropologists today than several other Lowie books, such as The Crow Indians, Primitive Religion, or Social Organization.

As I read The German People, I couldn’t help but to think of a more popular and more widely read ethnography, Ruth Benedict’s The Chrysanthemum and The Sword. The two books have some important things in common. Both were written and published during World War II, and both can be seen as attempts to understand “the enemy,” both for the war and the succeeding occupation. (Benedict’s research was specifically commissioned by the U.S. government for this purpose.) In the processing of making sense of the Germans and Japanese respectively, the two texts also no doubt offered an important humanizing of the two nationalities.

The two works have a very different “feel” in other ways. Neither is the conventional ethnography written on the basis of participant observation field work in the culture in question. Benedict’s work was largely based in extensive interviews with Japanese-Americans (so that there was the use of interview methods typical of ethnography, but without the often more prominent participant observation). Benedict’s work “feels” very much like a conventional ethnography, even if based on for the time an unconventional total methodology. Ironically, Lowie’s experience of German culture was much more direct than Benedict’s of Japanese culture (Lowie was not a participant observer there, but had had extensive first hand experience of the culture as a student earlier in the century), while his text has little of the stylistic “feel” of an ethnography at all, really fitting more into the genre of social history.

Fortuitously, while I was reading Lowie’s The German People, I was also reviewing the various essays in the mid-1980s text Writing Culture for an upcoming discussion with grad students in a seminar on culture theory.

In Mary Louise Pratt’s essay in the collection, “Fieldwork in Common Places,” she writes (p. 32; parenthetical note on “it” added):

“James Clifford speaks of it (the persistence of personal narrative alongside objectifying narrative) as ‘the discipline’s impossible attempt to fuse objective and subjective practices.’ Fieldwork produces a kind of authority that is anchored to a large extent in subjective, sensuous experience. One experiences the indigenous environment and lifeways for oneself, sees with one’s own eyes, even plays some roles, albeit contrived ones, in the daily life of the community. But the professional text to result from such an encounter is supposed to conform to the norms of a scientific discourse whose authority resides in the absolute effacement of the speaking and experiencing subject.”

This quality of the ethnography does several things.

1. It gives ethnography a distinctive feel. Those texts we call ethnography generally do present in some way the ambivalence between personal, subjective narrative and third person, objectifying narrative. The lack of this subjective and personal element is largely what makes The German People feel like it’s not an ethnography, even if it is “writing culture.”

2. The tension between these elements in ethnography is, I think, what is largely responsible for the long history of conscious experimentation with the form of ethnography – something that’s been going on far longer than the writers of Writing Culture tend to acknowledge (see Experimental Ethnography Old and New).

3. This tension is one of the things that makes ethnography continually interesting because it is continually problematic at its formal core – more on ethnography as interesting or boring in my following post.

4. As Clifford and Pratt are pointing out, it’s the inclusion of the personal narrative that grounds the authority of the ethnographic narrator who experienced the culture and gives credence to the objectified narrative. Without such rhetorically established authority, why should we trust the strange things we read about in so many well written ethnographies are true? (Of course, being aware of this source of authority, why trust what we read to be true and not simply an interesting account of something which may or may not correspond to anyone’s lived reality?)

5. The continual presence of personal narrative as grounding authority is the chief means through which field work as rite of passage (as discussed so well, or at least so nicely by Lévi-Strauss in Tristes Tropiques) is interjected into the text, creating what Susan Sontag called “The Anthropologist as Hero,” the ethnographer venturing out to where others dare not go and returning to bring us comprehension of the other. Contra the construction of anthropology as the “softest” of social sciences on the part of many other social scientists, we have here an image of anthropological ethnography as the most macho of social science endeavors, and one trafficking in an essentializing division of self and other. All of this is problematic (and silly), but I’d argue it’s still very much a part of the image and appeal of anthropology and ethnography. (In my own socialization into the discipline in a Ph.D. program in the 1990s, this was still part of how anthropologists thought of ethnography. My research along the U.S./Mexican border was suspect as ethnography, because my others might not have been other enough, and frankly because I’d be doing ethnography in places with running water and electricity, and that I could drive there, though ultimately, the fact that I’d be doing participant observation in some specific contexts that were sometimes actually potentially dangerous and always perceived as dangerous made it just acceptable. A friend, who did participant observation on human rights issues at the U.N. in Switzerland, never seemed to be able to shake people’s perceptions that he somehow wasn’t doing “real ethnography,” even if everyone agreed his work was “important.”)

Research, Teaching, and Music Performance

The other day I had a very nice conversation with a graduate student I work with. This particular student is just beginning field research for his thesis, a thesis which, in a nutshell, will address issues of booth rental and wage labor in hair salons, a topic that taps into debates in political economy going back at least to Ricardo and also rich with interesting ethnographic detail. This student, like a lot of, probably most, ethnographers is using a combination of participant observation and flexible, open-ended interviews.

He noted that he was pleased by how his first interviews had gone, also noting the highly flexible quality of the interviews, with interviewees often taking the conversation in interesting and unanticipated directions, but also that he felt confident in working in this highly flexible and even improvisatory setting because of a significant amount of preparation for his field work that he had engaged in along with me and other members of the committee.

I drew an analogy to certain aspects of teaching. Specifically, there is a performative quality to research methods like participant observation and flexible, open-ended interviewing that has something in common with the performative quality of some teaching, e.g. leading an effective class discussion. Effectively leading discussions requires preparation and organization – you have to know your stuff, but I find that the most effective discussions are true conversations that can often lead in unexpected directions. There is improvisation, but based on sufficient organization and preparation that I’m confident enough to set aside preset plans and follow an interesting lead. (This doesn’t mean that anything goes in class discussion – or open-ended interviewing – some comments are outside the domain of relevancy, are too tangential, and require reigning in, though it can sometimes be difficult to tell in the moment what is too tangential and what not.) Not all teaching works this way, though. Sometimes a thoroughly preplanned lecture is the best and most efficient way to communicate information to a class – there can always be room for questions and clarifications, but within a plan.

Then, another analogy struck me. Some research (in this case, participant observation and flexible interviewing strategies) and some teaching (e.g. leading class discussion) is analogous to jazz performance, while other research (e.g. more controlled interviewing or survey research) and other teaching (e.g. delivering a preplanned lecture) is more analogous to classical performance.

Jazz performance is highly improvisatory. When performed well, though, jazz is not chaos or noise, but based on thorough preparation and practice that allow a skilled musician to dispense with rigid adherence to formulae to play freely. The same is true with skillful performance of certain research and teaching strategies.

With some exceptions (typically highly delimited and occurring either in music from the baroque period or earlier or from very recent classical composition), classical performance is highly scripted rather than improvisatory. The musicians follow a definite score. Something like survey research tends to work similarly, with attention paid to following a scripted questionnaire and attempting to control as much about the research environment as possible so as to limit as far as possible the number of variables that might contribute to the production of the different question responses.

In both cases here, classical performance and survey research, though, even within the highly scripted context, there is nuance and interpretation to performance. Different performances of the same classical works can sound quite different based on subtle differences in interpretation and performance of the music’s details, producing highly different results. With something like survey research, there is an art to getting people to respond to questions, and doing so without either inhibiting or overly influencing respondents’ replies through the details of posture, facial expression, or a wide variety of vocal qualities. (As an aside, the film Kinsey presents several examples of such things to be avoided by interviewers in a formal research setting. In the film, we learn about Alfred Kinsey as a person via several scenes in which he trains students in interview techniques by having them interview him. It’s an innovative way of delivering exposition about the subject’s life in a biographical film without slipping into the clichés of biopics. Along the way, it’s the only movie I’ve ever encountered that seriously explores social science research methods.)

Discourse, Practice, and the Sewer

Like many anthropologists, I’ve long been interested in the relationship between patterns of discourse and practice, or in more simplified terms, between what we say and what we do. (“Saying” and “doing” is a more simplified presentation in that it potentially confuses the non-mutually exclusive nature of discourse and practice. Discourse and practice are not so much different human activities, as different modes of interpreting human activity – what we say, our discourse, is a key sort of practice, and our patterns of behavior constitute not just what we “do,” but also constitute a form of discourse – something recognized in phrases like “actions speak louder than words.”)

Recently, much of my own research has focused on public health issues, including especially the study of cultural models of drinking and related activities. I’ve also been interested in thinking about novel ways to do ethnography through utilization of a broader set of research methods than are typically used by anthropologists. (See my previous posts: Thinking Problem, Measurement and Interpretation, A Clarification on "Qualitative" and "Quantitative," Ethnographic Research Methods and Ethnographic Writing.)

On a topic such as this, it’s relatively easy to collect data on drinking discourse and on what people claim about their own drinking behavior – and in some cases observational data from participant observation style research can complement and inform such discursive data, but this often depends on the specific nature of the population being studied. For instance, along with other colleagues at my university, I’ve been specifically interested in students’ cultural models and actual drinking patterns. I could do participant observation on drinking behavior with this population, but I’m under no illusion that what I would see would come remotely close to “natural” behavior – while I strongly feel that participant observation is the best research method to use for some research purposes, here it’s simply not a very efficient or reliable way to go about studying the cultural context, and the precise relationship between what students claim about themselves and what they do when not being asked questions by pesky researchers is left murky.

I recently encountered an interesting news story on the website Medical News Today, Evidence in Sewage of Community-Wide Drug Abuse. Researchers analyzed raw sewage and were able to discern distinct community patterns of drub use. The research says nothing about any individual’s actions (and so, at least debatably doesn’t violate confidentiality or infringe upon informed consent rights – though that’s a debate I think would be worth having), but this style research could say quite a bit about aggregate patterns concerning any patterned behaviors which would result in chemical traces in urine. (I’m not sure what that range of patterned behaviors that could be studied this way would be, but it’s interesting to think about. On a related note, I remember reading a science news article from last January or so that talked about ecologists having detected elevated traces of cinnamon in Puget Sound as a result of people’s holiday baking.)

I certainly don’t expect ethnographers to dive into this style of research, but I do think it’s worth considering whether unconventional research methods such as this could be of use to the anthropological study of cultural patterns. In instances where behaviors result in chemical traces, such analysis could open an insightful window on that complex relationship between discourse and practice.