'Always Momentary, Fluid and Flexible': Towards a Reflexive Excavation Methodology
Hodder, Ian, Antiquity
Although processual and postprocessual archaeologists conceive of 'data' in different ways (Patrik 1985), there has been little discussion of a postprocessual methodology (but see Carver 1989; Tilley 1989). This is understandable; any notion of a general methodology separate from the context of the production of knowledge could conflict with approaches which emphasize critique, interpretation and multivocality.
Most excavation record forms still separate description and interpretation in the way advocated by Barker (1977; 1982). Carver (1989: 669) argues that this tradition in Britain extends back to Pitt-Rivers, such that 'English excavators, particularly, believe that there ought to be a science of retrieving archaeological evidence which has nothing to do with the interpretations that are subsequently made'. The general emphasis on 'objective' recording found in field contexts in many parts of the world would suggest that the postprocessual debate has had little impact in this arena.
I wish to argue that there are two reasons for a reconsideration of archaeological 'data collection' techniques. The first - 'internal' or logical - concerns a contradiction that lies at the heart of the empiricist and objectivist approaches to 'data'. The second - 'external' - concerns the wider world within which archaeology today operates. I will illustrate the impact of this global context with reference to renewed work at Catalhoyuk (Hodder 1996).
Consider the following two statements from Joukowsky's A complete manual of field archaeology (1980: 218-19,175):
In addition to day-by-day notes, the square supervisor is responsible for a subjective interpretation of the meaning of his or her excavation. This subjective analysis is submitted at the conclusion of his/her work in a particular area and is physically kept separate from the objective facts, so that assumptions are kept distinct from field notes.
If the earth is from a sterile layer, it can be dumped, but if it comes from an occupation level, the earth should be carried to the screen, spread on it, and sifted so that no telltale signs will be overlooked.
The first statement argues that interpretation should be kept separate from objective fact: it should occur only after data have been collected. The second statement contradicts this by arguing that the methods used depend on prior interpretation. The excavator has to interpret a deposit in terms of whether it is an occupation level before screening (sieving) is used. How is one supposed to know whether a layer is sterile before it has been screened?
Interpretation occurs at many levels in archaeological research, and in the example just given, it cannot be confined to a higher level. How we excavate a site is generally determined by our prior interpretation of the site. The screening and point-proveniencing of all artefacts common on excavations of Palaeolithic cave sites are minimal on historic urban sites. Even within a site, decisions about whether to screen are frequently made on the basis of interpretation. A 'floor' context is excavated more intensively than one interpreted as 'fill', with 100% water-screening only being used in the 'floor' context. In such cases, whether an artefact exists at all within the archaeological view depends on interpretation. Microartefacts may only be recovered because of full water-screening of 'floor' contexts. The same artefact from a 'fill' would not be searched for or recovered; it might simply 'not exist'. The objective existence of an artefact as 'data' depends on the interpretation made prior to and during excavation. How can it be maintained that subjective data interpretation should only occur after objective data description and collection?
Archaeologists have typically dealt with the problem of needing to know what is being excavated before it is excavated by sampling different parts of deposits or sites in different ways, or by taking initial trial soundings (sondages) before full excavation. …