The Oxford Handbook of Neolithic Europe

by Chris Fowler

  The ‘pre-pit’ context

  In this section, I concentrate on one individual site, Kilverstone in Norfolk (Garrow et al. 2005, 2006), as this focus enables a fruitful depth and detail of understanding. The site produced a very large number of pits (226 in total), containing over 2000 sherds of Mildenhall pottery and over 13 000 flints. Importantly, detailed analysis of the conditions of artefacts within the pits shed light on the character of the place in which those artefacts had been ‘stored’ prior to deposition.

  The physical processes which had affected artefacts were similar across all assemblages. Substantial amounts of pottery and flint had been altered to varying degrees by burning and weathering (Beadsmoore 2006; Knight 2006). Similarly, hazelnuts and other seeds had only been preserved due to charring. Amongst the pottery, some sherds were burnt slightly, whilst others had been severely affected by heat. Similarly, most burnt flint pieces had probably been purposefully heated (in the process of cooking, to create temper for pottery, etc.), but some were only very lightly cracked, perhaps suggesting more incidental burning. Occasionally, burnt stones were also recovered, usually surrounded by charcoal and burnt flint. The pottery and flint had also weathered to different degrees: a number of flints had been severely frost-shattered before deposition, whilst many sherds’ surfaces had been eroded.

  Crucially, in the case of both flint and pottery, burnt or weathered pieces could be directly re-fitted with unburnt or unweathered pieces (Fig. 38.2; Knight 2006, 31). This implies that between the breakage of pots or the knapping of flint, and the deposition of material within the pits, there was a delay during which some artefacts were subjected to heat or left open to the elements. Refits between differentially weathered or burnt pieces suggest transformation after breakage, but prior to deposition. A post-breakage, pre-pit context for the material therefore existed, where broken pots, flint-knapping debris, food remains, and other materials accumulated. Widespread weathering on the material suggests that it had been left out in the open for some time, whilst burning on material in most deposits indicates fires or hearths nearby.

  FIG. 38.2. Burnt/unburnt and weathered/unweathered refitting sherds from Kilverstone, Norfolk (Garrow et al. 2005, fig. 9).

  (Photograph: M. Knight/D. Webb).

  The exact nature of these pre-pit accumulations is difficult to describe. It is possible to imagine an area where the ‘mess’ of broken pottery and flint debris, hearths and cooking, hazelnut shells, and un-needed burnt flint was perhaps contained within demarcated ‘zones’. The artefacts nearest the fire were burnt, whilst those further away were not; those that ended up underneath others were protected from the weather, whilst those on the surface were exposed. Sherds were trampled, and flints trodden into the soil, leading to ‘incomplete’ assemblages. The material links between pits suggest that these pre-pit contexts were relatively limited in size, and changed location each time the site was occupied (see below).

  The character of occupation

  Detailed examination of the material within pits also allows a better understanding of how ‘much’ occupation these sites, often viewed as fairly ephemeral, witnessed. Despite the fact that large amounts of material never made it into the pits (in some cases arguably because items were specifically selected for deposition), it is nevertheless possible to use those elements which did to provide a rough estimate of how many people visited a site for how long (see also Garrow 2010).

  In many cases, the association often made between pits and ephemeral settlement is understandable. It is indeed hard to view a couple of shallow scoops containing a few scraps of pottery and flint flakes as the archaeological signature of anything more than a short visit to that place. However, a number of sites, particularly in eastern England, appear to signify occupation of a rather more substantial kind: at Kilverstone, 226 pits were recovered (Garrow et al. 2006); at Hurst Fen there were at least 200 (Clark et al. 1960); at Broom Heath in Norfolk there were 67 (Wainwright 1972); and at Spong Hill in Norfolk there were 56 (Healy 1988).

  It is very difficult to link the number of pits on a site directly with the duration of occupation; people could well have dug large numbers of pits very quickly, especially as most sites are on soft sandy soils. However, the materials within those pits arguably do provide insights into the ‘amounts’ of occupation we are dealing with. On all sites with large numbers of pits in eastern England, there were large numbers of vessels. It is important to emphasize that all of these vessels were broken, and the vast majority were only partially represented (sometimes by very few sherds); clearly, substantial amounts of material never made it into the pits. At Hurst Fen, it was estimated that between 283 and 317 different pots were represented, at Broom Heath 272, at Kilverstone at least 150, and at Spong Hill 113 (Clark et al. 1960, 228; Wainwright 1972, appendix 3; Knight 2006; Healy 1988, Ch. 3). Interestingly, even on some sites with far fewer pits, the number of vessels represented was impressive. At Eaton Heath in Norfolk, for example, pieces of 26 different pots were found in just 12 pits, and at Hall Farm Reservoir in Norfolk a single pit contained the remnants of 12 vessels (Wainwright 1973; Birks 2000).

  As with the pits themselves, we must be cautious in reading the duration of occupation from the number of pots represented. Nevertheless, as there was no clear evidence that vessels had been deliberately smashed or brought specially to the sites for deposition, it does seem reasonable that the material within the pits resulted from occupation at these sites. I am usually reluctant to employ ethnographic analogy, as comparing practices of the archaeological past with those observed in the present usually ignores that we are dealing with entirely different contexts. However, remaining mindful not to make direct comparisons across such vast differences in context, I would nevertheless like to explore the rates at which pots are broken in contemporary (or near-contemporary) situations. These numbers are not ‘correct’, but provide us with a way of thinking through the quantities set out above.

  As part of his study of deposition on Iron Age settlements, Hill carried out a wide-ranging survey of ethnographic and ethnoarchaeological studies investigating the rates at which pots are broken (Hill 1995, 129–131). Based on evidence from a variety of contexts across the globe, he arrived at an average figure for the number of pots broken per ‘household’ per year of 5.4 (overall, numbers varied between 2.4 and 22). If we tentatively accept Hill’s average figure for the Neolithic in Britain, we might thus suggest that the c. 300 vessels at Hurst Fen were broken by the equivalent of one household occupying that place over 55 years, or five households over a decade. Similarly, the material at Kilverstone (151 pots) equates to one household occupying that place for 25 years, or five households for five years. Even the single pit at Hall Farm Reservoir, with 12 vessels, could be equivalent to two years’ occupation. On this basis, it can certainly be argued that the link between pits and ephemeral and highly mobile occupation has perhaps been overplayed (Garrow 2010). Whatever these numbers actually imply in a quantitative sense, it is clear that many sites—all but one of which produced no evidence for any structures—did witness substantial occupation(s).

  Pits as containers of significant ‘background’ information

  In addition to information about specific sites, it is possible to glean a great deal about issues of broader note to the Neolithic as a whole from the contents of pits. To illustrate this point, I will focus on evidence for the cultivation of domesticated crops and gathering of wild plants, again using eastern England as a case study (see also Garrow 2010).

  Over the past two decades, debate has raged as to whether cultivated cereals or gathered wild plants were the major contributor to Neolithic diet in Britain. On one side, it has been argued that a mixture of wild foods and cereals were used, but the former predominated throughout the Neolithic, whilst the latter played a more symbolic role (e.g. Richmond 1999; Thomas 1999). On the other side, it has been suggested that wild foods, particularly hazelnuts, are simply more visible due to preservation biases, and
that the Neolithic economy was actually almost entirely cereal-based (e.g. Rowley-Conwy 2000). Often, these arguments have been conducted with evidence from a rather selective and unusual sample of sites: causewayed enclosures and ‘freak’ assemblages such as those from the timber structures at Lismore Fields in Derbyshire, or Balbridie in Aberdeenshire (Thomas 1999, 25). By contrast, pit sites arguably allow something approaching a ‘background’ impression of the situation at the time.

  The archaeobotanical evidence from pits in eastern England strongly suggests that it is unwise to come down on either side of the wild/cultivated fence. Of the 33 early Neolithic pit sites in the region, 11 were sampled for charred plant remains. Nine of these produced wild plant remains, with hazelnut shells predominant but apple also in evidence. Seven of the 11 sites produced charred cereal remains (dominated by emmer and to a lesser extent barley). Two further sites, both excavated before flotation samples were taken routinely, produced secondary evidence for cereals in the form of seed impressions on pottery (Clark et al. 1960, 211; Wainwright 1972, 21). Broken down into individual feature-specific samples, the picture is slightly less well-balanced, but nevertheless comparable: 61 pits were sampled in total, 52 (85%) of which produced hazelnuts and 35 (57%) cereal grains.

  Considering that all these sites lie on sand or gravel subsoils (leading to extremely poor preservation conditions, particularly for charred seeds), and that for a variety of reasons hazelnut shells are much more likely to be preserved in the first place (Jones 2000), it can be argued that both wild and cultivated plant foods are actually well represented. It is not my intention, however, to perpetuate debates on the relative importance of wild or domestic foods. Given the problems and biases of initial preservation and post-depositional survival, it is almost impossible to assess this effectively. What seems most important is that on 64% of all sites sampled, and on 100% of sites where more than one sample was taken, both cereals and hazelnuts were found. Clearly, on the basis of this evidence, both formed an important part of people’s diets. Previous debates, which have argued for either wild or domesticated plant foods being dominant, might therefore be somewhat misplaced, in this region at least.

  Material dynamics and the temporality of occupation

  In this final section, I would like to illustrate the potential that material patterning across pit sites has for understanding another much-debated issue of wider significance: the permanence of settlement (e.g. Whittle 1997; Rowley-Conwy 2003). Again, I look firstly at Kilverstone, as it represents the most thorough analysis of material patterning on a pit site to date (Garrow et al. 2005). Although I will not dwell on other sites in detail here, similar analyses on a number of pit sites in eastern England produced extremely similar patterning (Garrow 2007b).

  Most pits at Kilverstone were grouped into spatially discrete clusters (Fig. 38.3), with some variability in the forms these clusters took. Whilst many were relatively well-defined (pits spaced closely together, and separated from the next cluster by several metres), others were more diffuse; there were also a few ‘isolated’ pits which did not clearly fall within any particular grouping. The clusters also varied considerably in the number of pits within them. Initially, clusters were defined purely through spatial proximity and were intended as an aid to interpretation, enabling finds to be analysed in relation to the layout of the site. However, it became clear that many of the pits grouped together on this basis were also connected through shared material assemblages (i.e. flint and pottery re-fits) (Beadsmoore 2006; Knight 2006).

  FIG. 38.3. Re-fits within Area E, Kilverstone, Norfolk—pottery in black, flint in grey (Garrow et al. 2005, fig. 10).

  The patterning of re-fits between pits was extremely revealing. Here, the same patterns were observed irrespective of whether pottery or flint was involved (unusually large sherd sizes ensured a very good potential to re-fit pottery). Within both assemblages, numerous re-fits were established (Fig. 38.3). Unsurprisingly, many involved pieces in the same pit. Interestingly, numerous others were found between pits in the same cluster. However, despite concerted attempts, no re-fits within the pottery, and only a single re-fit within the flint, were established between pits in different clusters.

  In summary, the processes behind deposition had ensured that parts of the same pots or of the same flint working sequences often ended up in separate pits within a cluster, but—with a single exception—never in separate clusters. These links existed within widely spaced, diffuse clusters and neatly defined, spatially discrete ones alike, and in clusters with small and with large numbers of pits. Whilst pits within spatially discrete clusters were closely linked, the different clusters were not. Ultimately, we argued that this had come about as a result of the temporality of occupation on the site: the clusters were materially separate because they were also temporally separate, produced during different occupations (Garrow et al. 2005). It proved possible to suggest a temporal interpretation as a result of assessing the material dynamics within and between the pits. Prior to this analysis, debates about settlement mobility in Neolithic Britain had been conducted largely in relation to the presence or absence of houses. As the evidence from Kilverstone shows, the absence of archaeologically visible domestic structures on a site does not necessarily mean short-term settlement there. The contents of pits, if analysed creatively, can help elucidate the character of Neolithic occupation.

  SUMMARY

  The fact that sites consisting only of pits exist in Britain is a relatively unusual aspect of the archaeological record. The concurrent absence of domestic dwellings across large parts of Britain has led to a situation in which archaeologists are forced to question what they mean by ‘settlement’, and to break apart the age-old links between farming/pottery on the one hand, and permanent occupation on the other. The fact that pits can be multi-faceted in terms of their original ‘uses’ has ensured that interpretations of these features have over the years been creative and fairly fluid. Recently, it has been generally accepted that pits may well have been dug purely to receive deposits, rather than having any prior ‘purpose’ as such. Arguably, this practice is the result of a changed ‘Neolithic’ attitude to the world, and between people and places.

  Over the past 25 years, the concept of ‘structured deposition’ has become central to most interpretations of Neolithic pits in Britain. The notion has proven so enduring that it has been applied to many different types of site, and indeed to other periods and regions (Garrow 2012b). However, the success of the concept has, in certain ways, also been its downfall. Structured deposition has shifted from being a means to an interpretive end to an interpretation in itself. This has effectively cut short interpretation especially of pit features. In this chapter, I have sought to demonstrate the potential of moving beyond the act of deposition itself, to look at what the material within pits can tell us about the dynamics of the lives and practices that material was caught up in prior to being deposited. As pits’ uses can be varied and sometimes vague, and as these features can seem small and insignificant next to impressive monuments, they have been easy to overlook. I hope to have provided more than enough evidence as to why it is vital to incorporate these undeniably small, but nevertheless materially rich and very rewarding features into our narratives of the Neolithic in Britain and beyond.

  ACKNOWLEDGEMENTS

  I would like to thank the Arts and Humanities Research Council and the Cambridge Archaeological Unit, who funded much of the research on which this chapter is based; Emma Beadsmoore and Mark Knight—the analysis and interpretation of Kilverstone was entirely a joint project; and Anwen Cooper and Daniela Hofmann for their comments on an earlier draft of this chapter.

  REFERENCES

  Allen, M. and Gardiner, J. 2002. A sense of time: cultural markers in the Mesolithic of southern England. In M. Wilson and B. David (eds), Inscribed landscapes: marking and making place, 139–153. Honolulu: University of Hawaii Press.

  Anderson-Whymark, H. and Thomas, J. (eds) 2012. Regional
perspectives on Neolithic pit deposition: beyond the mundane. Oxford: Oxbow.

  Beadsmoore, E. 2006. Flint. In D. Garrow, S. Lucy and D. Gibson Excavations at Kilverstone, Norfolk: an episodic landscape history, 53–69. Cambridge: East Anglian Archaeology 113.

  Birks, C. 2000. Report on an archaeological watching brief at Hall Farm Reservoir, Croxton. Norfolk Archaeological Unit: unpublished report 518.

  Bradley, R. 1993. Altering the earth. Edinburgh: Society of Antiquaries of Scotland.