A Specific Problem? The Detection, Protection and Exploration of Romano-British Cremation Cemeteries through Competitive Tendering
Introduction The context of this paper was Pete Wilson’s provocatively titled ‘Field archaeologists don’t think…’ session at TRAC 2006. This was a fascinating discussion that quickly deconstructed an often reified dichotomy: that there are two types of archaeologist, the (ivory tower-based) ‘thinkers’, far removed from the ‘real world’, and the ‘dirt archaeologists’, who in fact pride themselves on their involvement with the ‘real world’, wherein there is little time to think, let alone to do all the digging required by the specification. Anyone who so narrowly defines themselves (in either way) surely restricts both their experience of, and their potential contribution to archaeology. Any archaeologist might be expected to have an interest in the archaeological process as well as the evidence it produces, in order to comprehend the dynamic relationships between data collection and interpretation. Self-styled ‘pure theorists’ or ‘purely practical’ archaeologists would seem to be missing an integral part of what it means to be an ‘archaeologist’, and both conditions are arguably just the types of estrangement for which we might consider the term ‘alienation’ (see below). In my contribution to the TRAC session, I wanted to draw attention to what might be seen as a particular way in which both ‘field’ and ‘research’ archaeologists currently may not ‘think enough’, in relation to a very specific archaeological example.
‘Using a machine to find Roman burials’ The use of mechanical excavators to clear surface and deeper deposits from sites is a commonplace in developer funded archaeology, and entirely understandable in most cases. Of course many sites, including Roman period cemeteries, have already been severely truncated anyway as a result of the post-depositional processes, including activity such as agriculture. The use of a machine to swiftly remove mixed deposits in such cases is perfectly justified (provided that adequate evaluation has demonstrated such conditions), allowing time and resources to be focussed on the more archaeologically significant (relatively undisturbed) features and deposits. Yet, however skilled the machine watcher, it is increasingly clear that certain types of archaeological feature are more adversely treated by the adoption of machining strategies than others: for example any relatively ephemeral features that lie at the ‘interface’ between perceived ‘overburden’ and ‘archaeology’. Exactly where this divide exists is of course an on site interpretation in itself, and a decision necessarily taken once machining has already started (and in response to subjective perceptions).
The detection, protection, and exploration of Romano-British cremation cemeteries is an excellent example of the type of archaeology where such considerations are increasingly apparent. Research into this area during the last decade or so has considerably expanded our understanding and expectations of the potential complexity and fragility of these sites. Using such evidence, archaeologists would now not only hope to compare burial contents, but reconstruct and compare much more of the ‘funerary sequence’ leading to, and otherwise related to burials (see Pearce 1997; 1998; 1999; 2002; Fitzpatrick 2000; McKinley 2000; Weekes 2005).
Surely no archaeologist (however they style themselves) would either promote, or simply not care about the unrecorded destruction of this sort of archaeological evidence? Why then is the question of machine use not more hotly debated in these and other circumstances where that use is detrimental? Are there ‘pre-understandings’ (Hodder 1999) at work here, governing and producing a routine methodology? Is it perhaps (for example) because Roman period cemetery sites are not considered as ‘urban’ (being outside the Roman pomerium?), and can therefore be investigated via a more ‘rural’ strategy? Surely an ‘urban’ site would be far more likely to be subjected to a single context excavation and planning (or something like it) these days? Certain cemetery horizons might easily be complicated enough that only a context by context evaluation would be considered sufficient for their analysis. Or would some archaeologists consider such interests to be limited to a few specialists (in comfortable ‘Ivory Towers’?), and overly optimistic, with little application in the practical ‘real world’; this would be a world where burials are located (usually already ‘plough damaged’) once extraneous material has been removed, revealing their contents. Historically (hopefully) this attitude was simply a matter of not recognising the significance of features other than burials for reconstructing funerary activity, and privileging the latter above wider aspects of the ‘funerary sequence’ (Table 1).
Table 1:Simple outline of the funerary sequence model and evidence types. Funerary sequence Archaeological contexts Types of evidence Pyre (and pre-pyre?) Pyre features Other features containing pyre material Burials Structures, pyre deposits (nature of deposits as well as spatial arrangement) Pyre deposits, feature type, micromorphology etc. Cremated bone and other burnt material Initial deposition Burials Combined objects, spatial features, modification of objects etc. Further actions (continued focus for ritual, further deposits, disturbance etc.) Burials Other contexts Upper contexts, including ‘backfills’, markers, superstructures etc. Buried soil horizons, structured deposits of material other than human bone etc.
Using a machine simply to locate burials (or rather the contents of burials) in this way, and ignoring other evidence, can be compared with other antiquarian excavation methods, such as ‘wall chasing’ to locate buildings at the expense of stratigraphy. Clearly the parts of the funerary sequence that are most at risk would be the early and later stages. Such data are often highly complex and typically ephemeral, and much of this evidence is easily destroyed by hand excavation, let alone mechanically.
The reconstruction of cremation techniques and pyre-side rituals (considering uniformity and diversity of treatment in separate instances) requires a meticulous scanning of fragile material that has already suffered from the damage and selectivity from the cremation itself, the deposition of the material, and the post-depositional processes. What is more, particularly in relation to pyre sites, we need to see this material in stratified association if we can really hope to understand it. What we might call the ‘secondary rites’, involving the potential revisiting of burial sites, further deposition, and even the removal of materials, are surely even more difficult to reconstruct. This is an area that deserves far more attention, but until it informs methodologies (and more specifically is considered in the drafting of specifications for archaeological works in the developer funded system and elsewhere), we are unlikely to have any evidence to consider. In a machine excavation focussed on the hunt for burials, the upper contexts of burials, the buried horizons into which burial pits were originally cut, and other super-structural features (unless of a durable material and recognised by the excavator in time) that might provide clues into this area, will often have been removed to the spoil heap before a proper examination has been carried out.
More generally we might say that in the wrong hands, and in poor conditions for example, the machining away of large amounts of material is a method that will automatically focus attention on the solid objects that form the contents of burials, at the expense of their context, or of other less solid or noticeable contexts. It seems appropriate here to invoke Hodder’s statement that ‘interpretation is not something which happens after description. It cannot be marginalized or ghettoized in ivory towers’ (1999: 103); but the idea that ‘it happens at the trowel’s edge’ is perhaps in itself a little optimistic if considered terms of the developer funded archaeological process. In this environment, more often than not, it happens on the blade of the ditching bucket (!).
Interestingly, an archaeological site that was all but destroyed in the process of machining out foundation trenches in the 1980s provides excellent evidence for the types of contexts we would be missing if we were to routinely level such sites using a machine. The site of new warden-assisted housing at Cranmer House, Canterbury was considered at that time to be beyond Canterbury’s Area of Archaeological Importance (Bennett 1987: 56–58), and was therefore not protected as the trenching started. As a consequence, of a total of 53 ‘burials’ reported as having been found within the trenching footprint, only seven were actually recorded in situ, while the rest were ‘rescued’ by workmen and archaeologists (who had voluntarily maintained a watching brief) from the machine’s ditching bucket (as can be seen occurring in Fig. 1).
Despite its destructive qualities, the very nature of this excavation actually preserved information that could so easily have been lost had an ‘open area’ strategy been adopted, maintaining in section evidence of the Roman period buried soil horizon into which the burial pits and other cemetery features had been cut (Figs. 1–3), and onto which deposits such as scatters of broken pottery vessels and human bone (probably cremated, given the acidic soil conditions) had apparently been deposited (Bennett 1987: 68). Moreover, certain burial pits preserved the possible evidence of robbed out superstructures or markers, such as in burials 28 (Fig. 2) and 45 (Fig. 3).
Figure 1: Archaeologists and the machine at Cranmer House, Canterbury in 1982, from Frere et al 1987 (Plate VIII).
Figure 2: Cemetery horizon and potential machine levels for burials 11 and 28 at Cranmer House, Canterbury (after Bennett 1987: 69, Fig. 22). Not to scale.
Figure 3: Cemetery horizon and hypothetical machine levels for amphora burials 46 (left) and 45 (right) at Cranmer House, Canterbury (after Bennett 1987:69, Fig. 22). Not to scale.
Figure 4: Cemetery horizon and hypothetical machine level for ‘burials’ 9 and 47 at Cranmer House, Canterbury (after Bennett 1987:69, Fig. 22). Not to scale.
The archaeologists who attended the site even noticed in other cases what appeared to be preserved mounds covering particular burials (as with burial 9 in Fig. 4, above). All such evidence clearly has a special significance for the latter phases of the funerary sequence, yet how many of these features would have survived machine levelling of the site? Evidence such as this should make us reconsider the implications of the machining practice of removing successive spits of soil in order to ‘find the right level’ (a faulty premise in itself, at least in the case of cemetery excavation, unless we expect all cemeteries to be on level ground, and all deposits therein to conform to an overall level). In order to discuss this further, Figs. 3 and 4 are composites of original Cranmer House section drawings, with burials placed side by side in a purely hypothetical manner.
While these do not represent the actual spatial relationships between amphora burials 46 and 45 (Fig. 3), or burials 9 and 47 (Fig. 4) respectively, this hypothetical view clarifies further the destructive potential of machine levelling in order merely to identify burials by finding their contents. The deeper level of burial 46 in relation to surrounding features would probably protect the feature if a machining level had already been established prior to its discovery; it would most likely be subjected to hand excavation and recording following the ‘strip and map’ phase. But what if this was the first burial to be ‘discovered’? Would the excavator, seeking to characterise the feature more clearly through over-cutting, necessarily stop the machine before grazing the top of the amphora?
On the other hand, (and still hypothetically speaking) even where ‘cemetery level’ and ‘machine level’ appear to be practically identical (as in Fig. 4), the fact that a preserved mound protrudes above the expected level (burial 9) would seal its fate, with the loss of what may have been a wooden cover for the burial, as well as other potentially significant material. The loose or bagged cremation deposit in ‘burial’ 49, without a durable protective container, stands little or no chance of surviving its discovery; the mineralised bone and carbon of this mixed deposit would probably first become evident to the machine watcher as a dark, flecked smear scraped across the adjacent natural clay. Even relatively durable objects forming the contents of burials are bound to suffer damage if we use a machine to ‘find’ them (even though experienced digger drivers can often amaze with their mastery of the machine to perform delicate tasks), such as removal and destruction of lids for cremation containers (ceramic or otherwise) as well as pot rims themselves, not to mention the explosion into powder of glass vessels under severe compression, the scattering of hobnails for example.
Above all then, although the use of the machine is obviously quicker, this is not in itself an archaeological consideration: archaeologists would surely favour the preservation of as much data as possible, at least by record, and enough time to do that job. Why then might some archaeologists at least continue to use a machine in such circumstances? A return to Marx’s concept of ‘alienation’ might offer a possible explanation for such a passive acceptance of priorities that are, from an idealistic point of view, not archaeological.
Alienation? According to some, the use of early Marxist language and concepts has become quite unfashionable, and not just in archaeology. Insoll (2004: 53) has recently stated, for example, that:
‘Today, however, Marxist archaeological approaches, as with Marxism in general, are of little consequence and part of the history of our discipline…’.
We can be sure that many neo-Marxist theorists would be equally as unconvinced as Insoll by unsophisticated projections of deterministic Marxist models onto the material culture of past societies (or deterministic models of any kind). Perhaps more seriously, many others would argue that important ideas derived from Marx are increasingly relevant in a wider sense today, particularly in an increasingly global economy where workers sell their labour in what seems to be increasingly a buyers market. Even so, sociologists (mirroring Marx himself) have apparently come to regard the term as somewhat outmoded; if the following declaration from a widely available introduction to sociology is to be heeded, that term’s final demise was immanent over a decade ago:
‘The term has been used less often in recent sociology. Many modern Marxists believe that Marx abandoned alienation in his mature work in favour of exploitation and they see little point in preserving the concept. Most non-Marxist sociologists find that it has become too indeterminate to be useful’ (Abercrombie, Hill and Turner 1994: 14–15).
But it seems to me that this could be a very useful term for distilling the essence of what was being discussed in much of the ‘Field archaeologists don’t think…’ session at TRAC 2006. In the context of the current ‘relations of production’ of archaeological data via the development process, perhaps there is a real danger that as archaeology becomes a commodity it comes to have less ‘value’ in other respects, in particular those qualities which, for Marx, marked out human labour as a creative enactment, refinement and development of ‘species being’ (Marx 1963 [in his ‘Economic and Philosophical Manuscripts’, (1844)]: 120–144).
Marx’s application of Feuerbach’s term is clearly a complex formulation for which there is little space for discussion here (see Mulhall 1998 for a detailed exposition). However, for the purposes of this paper, it could perhaps be applied thus: archaeologists, being so estranged in the workplace from whatever aesthetic ‘drive’ or ‘vocation’ that made them want to be archaeologists in the first place, can come to view the work as merely a chore, with only a limited personal input or outcome, whose methodology and even analytical and interpretive components are governed purely by the perceived external necessity to ‘get the job done’ (in exchange for wages), as quickly and as cheaply as possible, in order to compete in the marketplace. In a wider sense, the experience of this condition could be concisely summarised by Eagleton: ‘… they are no longer able to recognise themselves in the world that they have created.’ (1998: 28).
In this (constantly constructed) ‘real world’ archaeology can be seen as a service like any other in a market economy: its quantity and quality driven by market forces. From such a Marxist angle, the time and money saving machining strategy would surely be seen as embedded in the ‘objectified’ PPG16 process. This system for archaeological investigation, perhaps rightly considered by many non-archaeologists (and an increasing number of archaeologists ‘going native’?) as being simply part of ‘the development process’ it serves, can be viewed as incapable of what Hodder might call a ‘reflexive method’ (Hodder 1999), and therefore unlikely to engender a reconsideration of the use of a machine in circumstances where it is inappropriate. Actually, from this perspective, the process would rather seem to constitute a division of labour that potentially stifles the channels of communication and feedback that could promote group discussion, reflection, and multivocality.
This lack of communication could be seen as existing between archaeological units involved in ‘the process’ (in terms of competitive tendering), and also within them (in terms of individualised, competitive and hierarchical occupational frameworks). We might also spare a thought for university based archaeologists and their respective departments here, who must compete through the Research Assessment Exercise, or its equivalents. In fact, as I suggested in my introduction, assertions of self-limiting identifications with or ‘loyalties’ to any given ‘type’ of archaeology might also be seen as symptoms of alienation resulting from externally imposed divisions of labour.
Returning to PPG16, many would argue (and I have concurred above) that the system works perfectly well for archaeology in general (on the premise of a totally generalised research question: ‘sample and characterise whatever is there prior to its irrevocable loss’), and some would go further to say that, as archaeologists, we should simply be grateful that this system exists at all. But the capacity of the system as it stands to respond to the more specific research questions thrown up by certain types of archaeological evidence (often those that are most likely to be damaged by such a generalised process and its routine methods) should be questioned and discussed further. Is our acceptance of the system and its wholesale methodologies a function of a type of alienation? Could it perhaps be an archaeologist alienated by an objectified process who simply accepts the damage caused by the use of the machine in the specific circumstances of Romano-British cemeteries (and of other particular types of archaeological data), almost as if there were no human agency in the decision to employ that method? In a more hegemonic sense, perhaps we should say that it is archaeologists themselves who ‘structurate’ their working practices, and who have, in some cases at least, come to see the use of the machine as ‘common sense’, whatever the archaeology.
Conclusion If we believe in alienation, it is perhaps understandable that an individualised archaeologist might feel a sense of powerlessness to adopt an appropriate (and more time an money consuming) excavation strategy in a hierarchical context where developers and colleagues accept and expect a more destructive method: in other words, where he or she feels alienated from their work by social and economic forces beyond their control. If on the other hand we ‘believe’ in market forces (i.e. as a suitable mechanism for archaeological research), we should remember more often perhaps that a machine is still frequently used to retrieve data from archaeological contexts precisely because a machine removes more material more quickly. It is a more competitive method to include in a proposed scheme of works. Who could compete in the developer funded system if they tendered a more data-sensitive archaeological strategy?
Acknowledgements Thanks to Andrew Gardner, the TRAC editorial team and an anonymous referee for their insightful comments on this paper. Any views expressed or questions raised remain my own.