2. Lithic Scatters, a Methodology

Many methods have been used to assess the technological composition and dating of lithic assemblages. The methodology adopted here combines the study of raw materials, technology–typology, artefact condition and metric analysis (cf. Bond 2007a). This may be viewed as a 'whole assemblage' approach (Gardiner and Shennan 1985, 66). This approach combines the typological study of tools (retouched forms), with waste and metric analysis. Other approaches centre on dating ploughzone assemblages only by the presence and absence of chronologically diagnostic tools (Schofield 1991). Changes in Holocene flaking technology, the by-product of tool production, can also be used to demonstrate a broad date range for an assemblage. From the early Mesolithic to the Neolithic and the Bronze Age, there was a gradual change from blade to flake dominated industries of different shape and size (Ford et al. 1984; Pitts 1978; Pitts and Jacobi 1979). Therefore, individual lithic artefacts, both tools and waste (cores, flakes and blades) can be assigned to broadly chronologically significant groups. These groupings enable the quantification of that change in flaking technology through the analysis of assemblage composition. All lithic artefacts recovered from the ploughsoil can be assigned to a broad data quality and period group. Cores, waste and tools (retouched forms) recovered from the disturbed context of the ploughsoil, can be grouped by comparing the size and shape of flakes and blades, considering negative flake scar patterns and condition (patination and rolling of the artefacts can also in some case indicate the age). In central Somerset, as noted by others (Brown 1986; Norman 2003), the presence of a patina on lithics recovered from the ploughsoil appears to correlate with the presence of a Mesolithic/earlier Neolithic blade or narrow flake technology. The type of core and negative flake scar pattern on core faces can also be used to attribute broad dates. Tools, through their shape, size and type of retouch, can also indicate chronology. These assemblages, in turn, can then be compared to excavated and radiocarbon-dated lithic assemblages from the region, so to cross-check relative dating. For example, the excavated and radiocarbon-dated lithic assemblages from the Sweet Track, provides comparison for the earlier Neolithic (Brown 1986) or Charterhouse Warren Farm Swallet for the later Neolithic/Beaker period (Levitan et al. 1988).

In this way, a range of hierarchical groupings of lithic artefacts can be constructed, from well dated technologically, typologically and chronologically discrete groups, to less well understood groupings of artefacts. Hence, after preliminary analysis identifying lithic artefacts and recording attributes, those artefacts can be assigned to broad groupings that constitute data quality and period filters (Table 1). This is an interpretative device, but enables the relative phasing of a lithic scatter assemblage. Importantly, rather than only being able to attribute dates to tools, often representing less than c.5% of any assemblage, the total lithic scatter is attributed a range of periods.

'Whole assemblage' approaches towards lithic scatters have been applied across southern England. For example, mixed-period lithic scatters were investigated in this way, from the East Berkshire survey, Berkshire Downs (Ford 1987a; Richards 1978), Oxfordshire (Ford 1987b), and East Hampshire Survey (Gardiner and Shennan 1985), to the Stonehenge environs (Richards 1990). Other examples include Healy's (1991) analysis of the lithics recovered as part of the Fenland Project Survey - the Wissey embayment, Norfolk and Waddington's survey of the Milfield Basin, Northumberland (1999).

Lithic artefacts from these surveys have been grouped in different ways. Groupings of artefacts based on data quality and period filters, vary in number from four to nine in Norfolk (Healy 1991, 65). Variables that affect the ability to group lithics are: the relative presence or absence of chronologically diagnostic tools; the level of completeness of artefacts, reflecting the impact of ploughing on a lithic scatter (limiting the potential for metric analysis); the presence or absence of certain lithic technology; and the condition-preservation of artefacts. Some analysts may choose to split or lump assemblages; inferences may also be limited by the few comparative well-published excavated and radiocarbon-dated assemblages available regionally. Importantly, the actual artefact patterning reflected in local and regional lithic scatters must also reflect long-term occupation histories across a region; the 'nuts and bolts' of prehistoric settlement patterns.

In Somerset, fifteen separate but complementary lithic data filters have been applied (see Table 1). Lithic artefacts were attributed to groups of undated material (liming flints), questionable prehistoric material, prehistoric material (but otherwise unclassified), modern (gun-flint) and other more chronologically sensitive or data quality groupings. Hence, the earliest broadly dated grouping is the 'Mesolithic', whereas this is succeeded in quality and period resolution by the 'early Mesolithic' grouping. The latest prehistoric grouping is broadly termed 'indeterminate middle Bronze Age or later Bronze Age' (Table 1). Although Iron Age lithics have been argued to be present in a small group of excavated assemblages regionally (Smith 1981; Young and Humphrey 1999, 237-238), these industries were not identified in this analysis using recently published criteria (Young and Humphrey 1999). Curation of older lithic industries, 'recycling', was observed locally as indicative of Iron Age working (Smith 1981). But, during analysis this was not recorded as unusually high at any one lithic scatter. It has also been argued that such lithic industries were used solely in the domestic sphere (Young and Humphrey 1999, 240). Thus, it may be that Iron Age lithic industries did not leave a similar signature in the ploughsoil as with earlier lithic industries. It can also be acknowledged that there were few artefacts assigned to the indeterminate middle to later Bronze Age. These groupings may or may not mask later industries. Further work is needed to clarify the later Bronze Age/Iron Age social change in the role of knapping and its outcome in the landscape (Edmonds 1995, 187-188). This approach has been applied successfully on a systematic fieldwalked assemblage from the Shapwick Project, a parish on the northern slopes of the Polden Hills adjacent to the Somerset Levels and Moors and across the study area (Bond 2004a; 2004b; 2005; 2006; 2007a; 2007b).

PREVIOUS NEXT CONTENTS HOME