John Robb

Steps to an archaeology of agency

John Robb^*

Paper presented at Agency Workshop, University College London, November 2001

Premises

Why agency? Well, why not? It might seem that agency theory is inherently and intuitively a good thing, and little reflection is needed before proclaiming the latest theoretical agenda. But before we rush headlong into the new millennium, we need a bit of serious reflection on some basic questions:

what is agency anyway?
what do we want it to do for us archaeologically? And
how are we going to devise methodologies for using the concept?

Even the first of these has received relatively little soul-searching among archaeologists, and the second is necessary if we want to consider our own particular needs and ambitions rather than simply picking out a new wardrobe from the anthropology fashion store. The final question, devising methodologies, has been neglected, partly (one suspects) due a bias against methodology as a journeyman task involving a descent from the theoretical stratosphere. I would argue, to the contrary, that in many ways methodology is theory, and I will tack back and forth between them.

I started out writing this paper as an overview of archaeological methodologies for agency, mostly for my own use. However, I quickly realized that virtually all archaeological interpretation could be related to agency in some way, with the possible exception of some of the more bizarre things emanating from the American Southwest. Moreover, as Hodder (2001: 8) points out, archaeologists with widely divergent theoretical momentums are thinking about similar problems in parallel ways but not hearing each other. What was needed was not a literature review but some set of systematizing concepts, and these had to derive from agency theory itself. Here I'll try a rather inelegant roughing out of ideas about agency under some very general rubrics. This is very much a last minute first draft, and I'll be genuinely grateful for any comments, benign or otherwise.

What is agency?

`What is agency?' is not a simple question. In one review of it, Marcia-Anne Dobres and I counted a dozen definitions in use in archaeology (Dobres and Robb 2000: 9). Our most useful strategy here is to specify both a set of minimal necessary characteristics and the latitude within which the question is best left vague.

Agency is often associated with `putting people back into the past', and most definitions would center around the relationship between the constitution of the actor, in terms of cultural and psychological structures, and behavior. This relationship is usually considered dialectical: structures shape actions, and these actions reproduce structures. One key area where theories of agency disagree is in how tight the relationship between structure and action is. Agency is the socially significant quality of action involved in social reproduction. Given this view, one important province of agency studies focuses on the emergence of experience in action. Another important line of study has concentrated on the relationship between the dialectic of agency and the material resources and external conditions of social reproduction.

We now need to turn briefly to the question of how to imagine actors. Any agency theory implies agents; in fact, any interpretation of the past at all does, which is for me one of the most pressing reasons why we should formally theorize these imaginary beings. Most agency theories fail to do justice to the complexity of human life by proposing more or less unidimensional views of the actor. They show people pursuing a political ambition, or experiencing a landscape, or acting in a gendered way, or crafting a stone tool knowledgeably. But these tend to be narrow band broadcasts from Radio Agency. There is little attention to how people do all of these things at once and what they have to do with each other.

All serious theories of agency imply analysis of action at several levels. The core constituting the actor has to be a set of deeply embedded cultural and psychological structures. It is currently somewhat unfashionable to see the actor in terms of enduring cultural structures rather than fluid meanings; this is partly in reaction to structuralist schemes making humans prisoners of rigid cognitive edifices. Indeed, one risks being accused of essentialism for broaching the matter. But failing to consider entrenched dispositions reduces experience to an ephemeral flickering on a screen. It denies humans' commitment to a particular cultural world. By deeply embedded cultural and psychological structures, I mean the basic values and oppositions which shape our thinking, the terms of identity and personhood which make us who we are, and the emotional currencies we live through. These are the basic tools of thought and vocabularies of experience. This is the world of the implicit, unquestioned and unquestionable values which Bourdieu calls `doxa'. Doxic tools of thought are not `blueprints' for being, to use a genetic metaphor, or systematic deep structures to use a linguistic one; they are full of ambiguities, potential contradictions, and slippages---the raw material for creativity, conflict and heterodoxy as well as orthodoxy.

Doxic tools of thought cannot be deployed and exercised directly; they require a second level, that of genres of action. Genres of action specify situations and arenas of social performance, their boundaries, rules and strategies, what kinds of people can be involved and how they have to behave. We might deconstruct the fox hunt into a structural morality play about class, privilege and identity, but this does not mean that these values and identities could be expressed or experienced nakedly without the hunt. Social life includes very many nested genres of action. Some are narrowly defined social games, others are broadly encompassing projects of identity. They may develop conflicts to explore or divert attention from contradictions at a more structural level. Genres may also have prescribed `truth conditions', conditions which have to be fulfilled for a given performance to be considered a valid act.

Finally, within each genre of activity, people command knowledge, skills, techniques, and strategies. These are usually context-specific and discursively understood and taught. What makes a good pig sacrifice, or fox hunt? How can a ritual utterance be striking without breaching the limits of sanctity? This is the level of maneuver, tactic, practice, and performance. As with the relationship between doxa and genre, the relationship between genre and action is subtle. A genre of action is a potential reality, not a reality. As Barrett notes (2001: 153), `agents recognize the coming into being of their own existence in an engagement with the remaking of the world itself.' Knowledge and actions are things we exercise, and often the exercise is more important than the script; otherwise we reduce a game to its rules.

One might propose lots of potential rubrics here, but I would like to try calling this level of things dramas. The concept of drama has not been widely used in anthropology since Victor Turner's attempt to interpret historical events as dramas. Within archaeology the metaphor has been productively raised by phenomologists discussing movement, the senses and performative aspects of culture. As a systematizing concept, drama has a number of useful features. It incorporates intentionality, emotional engagement and experience, while recognizing the structuring power of contexts. It recognizes social psychological insights into the belief and emotional commitment created by enactment. It provides a way to understand the relationship between structure and agency which does not reduce either to the other. Indeed, dramas incorporate contingencies, doubt, uncertainty, and risk. `Drama' also makes active many things which are often considered passive or mechanical. Knowledge is an example: knowing is active, and reproducing knowledge is a social act (Barth in press). Practical skill is another, laden as it is with meanings of identity, efficacy and the production of certainty.

Archaeologies of agency

This is far too rapid a sketch of the depths of agency, but I want to turn now to methodologies for understanding dramas, genres, and doxas archaeologically.

The archaeology of drama

Mention agency to most archaeologists and you quickly raise the issue of intentions. Intention is inherent to drama. Analysis of ambitious political actors and their stratagems is especially an American obsession. It would be superfluous, if entertaining, to critique it here. The archaeology of intentional strategy is a powerful but very narrow tool. Methodologically, it focuses how political actors control symbols of power and economic resources.

A second major tradition is the archaeology of experience. This is far too involved to summarize in a sentence or two, especially in the presence of one of its founders. Suffice it to say that the stress is on the emergence of experience through the agent's engagement with the material world. A number of very important methodologies have been inspired by this focus. We might mention above all the interpretation of architecture and landscapes as cultural constructions defined through human history and experienced by people moving through them. Discussion of the body and the senses calls attention to the creation of actors through presentation, clothing, bodily modification and consumption.

A third line of investigation here is the analysis of technology, knowledge, and agency, especially as developed by Dobres and others studying the chaine operatoire. This work makes clear the expressive and self-creative dimensions of carrying out even the most mundane tasks.

We might supplement these existing lines of thought with some general tactics. Performance, after all, requires a specific context and occasionwi With a close enough reading, archaeologists can monitor both. In artifact studies, minor variations in execution are usually ignored as normal `noise' but by examining them within their micro-settings we might understand the nuances of expression. We must also consider more closely the world of knowledge ancient people worked in. For instance, a landscape, artifact or history must be understood in terms of layers of local knowledge differently situated people would have had about it. Several other lines of investigation arise from close attention to context. To take just one example, people often play with a sense of time through conscious anachronism, intentional difference from the past, or transplanting meanings to different temporal contexts.

The archaeology of genre

Every drama presupposes a genre, a set of beliefs, practices, social relations, things and places which together define categorically what is being done. As such, a number of archaeological openings are immediately apparent. Iconography may suggest polysemic key symbols and their situational interpretations in a particular genre. Symbols and environment may suggest the subjective experience participants are expected to have. Recurrent artifact assemblages may give us an idea of the practices taking place in a given context and of the social relations involved.

Many of these things, of course, have already been cited as clues to dramas, and it's clear that the same material, looked at in differing scales, may inform us about normative expectations for a genre and its actualization as a particular statement in a particular context. There is more to genres than this, however. Archaeologically, we often study activities; however, with archaeological genres we probably need to think in terms of projects, long term undertakings requiring planning, assembly and management of material and social resources. Projects require considerable social and material entanglement---which is why their potential for self-aggrandizement is almost always diffused in a web of debts, obligations, and qualifications.

Genres of action are important because they are places where doxic belief becomes mobilized in the service of concrete practice. Archaeologically, what we see as cultural change---the rise of a new burial rite, the spread of an assemblage, a dramatic shift in economic production---is often moments of genre formation, where an existing variant practice is proclaimed as a new orthodoxy, often with a new uniformity of practice and material culture, a rearrangement of social relations and a new elaboration of ancient symbols.

The archaeology of doxa

Archaeology affords at least three avenues to the archaeologies of doxic belief. The first is what we might call the archaeology of structure. In the crudest terms, where doxic belief has a logic, this emerges in relations between symbols, often across the boundaries between genres of action. Interestingly, much of the most valuable and interesting work inspired by post-structuralism in archaeology actually uses structural analysis as an encapsulated methodology to define points of meaning. This may a necessary methodological reflection of social reality, in which people need fixed reference points, even if only as fulcrums for opposition. We need to understand the generative logic people used to decorate a pot to understand what it means when such a logic is rejected, transported to a different context, or overturned. Methodologically, we might approach this through (for example) the recurrence of key symbols across divergent times and contexts, the elucidation of generative grammars of artifact form, and analysis of structural oppositions.

A second line of work is the archaeology of the body. If we follow the line of argument which derives cultural knowledge from bodily practices, tracing the logic of the body is one key to the production of cultural order, and not only via gender. Without reviewing this canon extensively, it is worth noting both the variety of material available for our analyses---iconography, representations of people, traces of clothing, bodily modification, architectural analyses of movement, gestural analyses of artifact use, studies of food production and consumption, evidence of means of altering bodily experience through narcotics, and so on.

Finally, one should consider what I call the `doxic dirt' argument. Paradoxically, archaeologists have not thought seriously about the meaning of garbage. But garbage is what differentiates the real streets of London from a stage set put up yesterday out of wood and canvas. Without any discursive meaning, it provides a history. The remains we live among preclude other possibilities and commits us to a specific history. Because they are trivially unimportant yet absolute, they lend the social world a cumulative effect of unbreakable normality. There is much here yet to be theorized.

Gulliver's travails: The archaeology of social reproduction

After this rapid tour of methodologies, let's return to the question of social reproduction. As the careful listener will note, essentially, I am advocating the analysis of a given archaeological situation on multiple levels, from the performance to the script to the structures of ideas and language buried within them. Here let me make one final point very clearly. This strategy is not merely to learn as much as possible from the material at our disposal. Nor is it naïve and hopeful eclecticism. Rather, it reflects where the locus of social reproduction is. We can only accomplish an intention through the use of dozens of tools, structures, social relations and contexts. This is most readily apparent when these things are contradictory. You have absolute freedom of speech in an academic paper provided you don't incite people to behave unacademically or run over time. You can feast your neighbors and impress them with your power, but you are dependent upon them to come and be impressed. You can vote politicians out of office, but this only validates the electoral system which creates them. You can hold academic seminars on class inequality which help perpetuate the idea that only people with tickets to middle class childhood educations can ultimately utter authoritative statements. And so on. The point is that to rebel, or even to perform, in one aspect of a drama requires you to conform rigidly in all the other aspects of semantics, material provision, and social relationships. Social reproduction binds actors inextricably in an arbitrary cultural order. We wind up like Gulliver, tied down by the Lilliputians by a hundred threads; struggling in one direction binds the threads more tightly in others, and ultimately only a major wrench or rupture will change things. And only analysis on all levels at once will reveal this network of threads.

Where do we go from here?

In this paper I've tried to provide an overview of what I think agency is, and how we can devise methodologies for studying it archaeologically. So, in conclusion ...where do we go from here? I will close with just a couple of ideas about how we can use agency theory.

First, and somewhat paradoxically given the above, we should use agency theory tactically. Agency is not a thing; it's a quality of many things and the relationships between them. We rarely have the materials or opportunity to explore all aspects of social reproduction. We should instead recognize the diversity of things under the rubric of agency and define which is of interest in a particular interpretation, what methods are best for getting at it and how it might have related to things beyond our immediate scope. We can't just say we are `doing agency' and leave it at that; we inevitably end up substituting the part for the whole.

My other concluding point concerns generalization. Clearly we are beyond worrying about lawlike generalizations. Probably everyone here would agree that our goal is to create historically specific stories about particular pasts. At the same time, I believe some degree of generalization is useful; it is in fact implied by our applying terms such as `agency' to different cultures and histories. Here our goal should probably be not to develop theoretical rules or even arguments, but rather lexicons. We need terms for common patterns or trajectories, for example, for things like the formation of new genres of action, the mobilization of the routine as a symbolic resource in new political circumstances, and so on. Developing as diverse an archaeological lexicon as possible may help us escape from the paradoxical situation where prehistoric societies reconstructed in a given theoretical tradition all look remarkably alike. After all, having a common language does not mean you can only tell one story.

*: Department of Archaeology, University of Cambridge, Downing Street, Cambridge CB2 3DZ United Kingdom; jer39@cam.ac.uk

Department of Archaeology