by Chris Fowler
SETTLED AND UNSETTLED, HOUSE AND HOME
For much of the twentieth century Neolithic studies in Britain and Ireland took a sedentary farming lifestyle and the existence of houses for granted despite there being little evidence for either (Clark 1937). Early discoveries of Neolithic settlements such as Skara Brae, Orkney (Childe 1931) and Lough Gur, Co. Limerick (Ó Ríordáin 1954) were rare and apparently not representative. Surveys of Neolithic settlement increasingly conceded that ‘storage’ pits were just about the only surviving traces of occupation sites across much of Britain and Ireland (e.g. Field et al. 1964; McInnes 1971; Herity and Eogan 1977). Debate inevitably turned to whether the paucity of houses was simply an absence of evidence, or evidence for absence (Gibson 2003). Explanations for the ‘missing’ houses ranged from the utilization of building styles that left no trace to the loss of evidence in the plough zone (McInnes 1971; Kinnes 1985, 25; Gibson 2003, 137; Loveday 2006).
More recently an alternative narrative emerged: that ‘the earliest settlement was scattered and mobile’ (Whittle 1988, 85). The notion that Neolithic people moved around and between temporary settlements emerged as a key aspect of the rethinking of the Neolithic that took place in the 1980s. Contrary to traditional expectations it was argued that, ‘the majority of the population lived for most of the time in rather flimsy and temporary dwellings’ (Thomas 1996, 2). Within this framework, there was a tendency to emphasize the ‘unusual’ and non-domestic nature of the few buildings that had been found (e.g. Thomas 1996; 1999; Topping 1996): their rarity meant they were exceptional, not everyday. This in turn allowed a re-evaluation of the appropriateness of the concept of ‘house’. For a time, less loaded terms were preferred such as ‘building’ and ‘structure’ (e.g. Barclay et al. 2002). Concepts such as ‘home’ and ‘household’ were also critiqued (Thomas 1996, 5).
This view of Neolithic settlement was not universally accepted (cf. Cooney 1996; Barclay 2003). Developer-funded fieldwork in the 1990s and 2000s in Ireland and Scotland found more than 75 Neolithic buildings, permanent structures of an apparently domestic nature (Grogan 2002; Barclay 1996; 2003). In Irish Neolithic studies at least, the term ‘house’ has always been used, although not uncritically in recent decades. Cooney (1999, 52ff) argued that ‘house’ and ‘household’ are perfectly acceptable in a Neolithic context given the cross-cultural nature of these concepts. He argued that archaeologists shied away from calling buildings houses because of the universal nature of the mobility model. Nonetheless, it is also the case that the mobility model may still have validity in southern England (Thomas 1999; Pollard 1999; Bradley 2007), where evidence for ‘houses’ is limited, although they have even been found here in recent years (cf. Thomas 2010, 5). Therefore, there is now a greater acceptance that variable patterns of sedentism were evident in the Neolithic of Britain and Ireland, and this has been accompanied by a subtler approach to the classification of Neolithic structures. Such buildings are viewed somewhere between the poles of wholly domestic and wholly ritualistic which until recently had dominated discourse. For instance, the mundane routines of everyday life could be viewed as ritualized actions (Bradley 2005; Whittle 2003), a recurring theme throughout the remainder of this chapter.
HUT AND HALL
Buildings associated with the early centuries of the Neolithic are mostly wooden rectangular structures, ranging from long and thin, to almost square, and from a few metres across to over 20m in length. Although many of these have produced evidence for ‘domestic’ activity, such as hearths, cereals, pottery and so on, it is by no means certain they were all dwelling places, or permanently occupied. For the purposes of this chapter, it is assumed that any structures discussed potentially had some kind of domestic role.
The most coherent body of evidence for a rectangular house building tradition is found in Ireland (Grogan 1996; 2002) where some eighty such buildings are now known, some occurring in groups of two or three (Smyth 2006, 234; 2010, 4) (Fig. 17.1). Whilst these buildings share a roughly rectangular footprint, there is a great deal of variation in size, with lengths typically between 6m and 12m (Grogan 2002, 519). Often these buildings have one or more internal divisions splitting the building into ‘rooms’ and where found, entrance gaps tend to be on the short ‘gable’ ends or a corner. The buildings seem to have been constructed in a number of different styles, from planks set into a bedding trench to posts supplemented by wattle (Smyth 2006, 238). Little is known about roofing material and in nearly all instances floors have been lost.
FIG. 17.1. Interpretive plans of early Neolithic rectangular timber buildings in Britain and Ireland. A—Lismore Fields A, Derbyshire, England; B—Llandegai, Gwynedd, Wales; C—Chigborough Farm, Essex, England; D—Ballygalley 1, Antrim, Ireland; E—Ballyglass 1, Co. Mayo, Ireland; F—Lough Gur A, Limerick, Ireland; G—Raigmore, Highland, Scotland; H—Kinbeachie, Highland, Scotland (A–C, Darvill 1996, fig. 6.4; D–F, Grogan 1996, fig. 4.2; G, H, Barclay 2003, fig. 8.2).
Darvill (1996) compiled a group of 37 assorted early Neolithic buildings from England and Wales. These conform to a largely rectangular ground plan, although some are nearly square. They are generally small buildings, mostly 4–10m long and 3–7m wide (Darvill 1996, 88), with rare larger structures, such as Lismore Fields 1, Derbyshire, at 15m in length (Garton 1987). Again, the plough-truncated nature of these structures means internal divisions, entrances, and hearths are not always clear, although most buildings seem to have had at least one partition. Darvill (1996) was able to identify a few different construction styles for these buildings, mostly posts set into slot trenches, with evidence for a single central roof support at a few buildings. More recently, a few larger buildings have been found in southern England, such as White Horse Stone, Kent, measuring 18m by 8m and defined by ‘wall foundation gullies’ (Oxford Archaeological Unit 2000).
Early Neolithic buildings in mainland Scotland are rare, with settlement evidence dominated by scatters of pits, postholes, and slots (Barclay 2003; Brophy 2006). Barclay (2003) recorded only ten possible rectangular buildings in mainland Scotland, although some are little more than settings of truncated pits, such as Kinbeachie, Highland (Barclay et al. 2001). Since his review a possible post-built house, 15m by 6m, has been found at Laigh Newton, East Ayrshire (fig. 17.2). Most buildings had an association with cereals and Neolithic Carinated Bowl pottery, whilst some had hearths.
FIG. 17.2. Excavation of a possible Neolithic house at Laigh Newton, Ayrshire, located in the bottom right-hand corner of the trench (GUARD, aerial view taken by Edward Z. Smith, Hawkeye Photography).
These light wooden structures stand in stark contrast to the timber halls found in Scotland, the role of which is still open to debate. Four have been excavated: Balbridie and Warren Field, Aberdeenshire (Fairweather and Ralston 1993; Murray et al. 2009), Claish, Stirlingshire (Barclay et al. 2002), and Lockerbie Academy, Dumfries and Galloway (Kirby 2011). A few others have been recorded as cropmarks (Brophy 2007). These large buildings had a rectangular footprint, usually with slightly rounded ends, measuring in the range of 22–27m long and 8–12m wide. Walls of large oak posts (presumably with wattle screening) or posts set into bedding trenches, and a series of internal posts, supported a roof that may have been 6–8m high. All had a series of internal divisions and entrances at one or both gable-ends, and Warren Field and Lockerbie may have had open yards attached (Fig. 17.3). These ‘halls’ have been interpreted as farmhouses, cult houses, feasting halls, or communal buildings used for a variety of roles (Ralston and Fairweather 1993; Topping 1996; Barclay et al. 2002; Cross 2003; Brophy 2007; Murray et al. 2009).
FIG. 17.3. Plans of the four excavated early Neolithic timber halls in lowland Scotland (copyright Oxbow Books, J.C. Murray, H.K. Murray and S.M. Fraser). From top to bottom: Warren Field, Aberdeenshire (Murray et al. 2009); Claish, Stirling (Barclay et al. 2002); Balbridie, Aberdeenshire (Fairweather and Ralston 1993); Lockerbie Academy (courtesy of Gordon Barclay, Ian Ralston, Magnus Kirby and CFA Archaeolog
y Ltd).
Timber-framed rectangular buildings are found across much of Britain and Ireland, and appear largely to be an early Neolithic phenomenon. For the most part, they were built, used and destroyed in the earliest centuries of the Neolithic; rectangular buildings are rare in the late Neolithic. The houses in Ireland may have endured for only a few generations or at most centuries, spanning the period of roughly 3800–3600 BC (McSparron 2003; Smyth 2010). The timber halls in Scotland had a relatively short currency, perhaps also 3800–3600 BC (Marshall 2009), although dating is less clear for the smaller timber buildings (Brophy 2006). Darvill (1996) noted rectangular buildings in England and Wales are largely the preserve of the ‘early and middle Neolithic’ whilst a large rectangular timber building at Yarnton, Oxfordshire was dated to 3950–3640 BC (Hey 2001). Certainly, by 3300 BC, house plans had begun to change (Darvill 1996; Grogan 2002; Brophy 2006).
ROUNDER HOUSES
Late Neolithic houses tend to have been ‘ephemeral and generally smaller and circular structures’ (Bradley 2007, 94). Not all were ephemeral: a number of roundhouses, so-called ‘Grooved Ware’ houses, typically consisted of substantial four-post arrangements within a ring of smaller posts, such as at Wyke Down, Dorset, Trelystan, Powys and Leadketty, and Perth and Kinross. Examples of these structures have now been found across Britain and Ireland (Thomas 2010; Smyth 2010; Noble et al. 2012). The connection with Grooved Ware, and a wider change in ceremonial monument shape from rectangular to circular, suggests that alterations in house plan almost certainly reflect wider changes within society towards the end of the fourth millennium BC (cf. Cooney 1999, 66ff; Thomas 2010).
The ephemeral form of most late Neolithic houses makes them difficult to find and identify (Fig. 17.4). Only three sites with oval buildings have been recorded in mainland Scotland (Barclay 2003; Brophy 2006). At Cowie, Stirling, there were eight overlapping stake-defined structures no more than 4.4m across (Atkinson 2002). Grogan (2002) has identified some 50 late Neolithic round or oval buildings in Ireland; most of these seem to have been defined by stakes set into a shallow bedding trench, and they are small, typically 4–7m across (Atkinson 2002, 109–110) with central hearths. Darvill (1996) listed 67 late Neolithic buildings in England, Wales, and the Isle of Man, although this included various sites perhaps better interpreted as timber circles. (Indeed, it is possible that in the Neolithic some roundhouses were symbolically and visually associated with timber circles (Thomas 2010, 8–9)). Confirmed domestic structures in England were often small, with stake supports for the walls, which Darvill (1996, 93) argues would have supported animal skins or another form of weatherproofing. A remarkable group of small sub-rectangular houses was recently discovered at Durrington Walls, Wiltshire. The buildings had light timber-frames, wattle walls, plaster-floors, hearths and slots for wooden furniture (Parker Pearson et al. 2006). It seems likely that this ‘village’ of up to 300 houses is unique, and associated with work at nearby Stonehenge.
The internal layouts of the Durrington Walls houses bear an uncanny resemblance to the stone houses of Orkney. Although there are a few early Neolithic settlements, such as the Knap of Howar ‘farmstead’ (Ritchie 1983) and hints of earlier timber buildings, the majority of houses on Orkney are later in date, stone-built, and sub-rectangular, oval, or sub-circular in plan. ‘Villages’ at Skara Brae, Barnhouse, and Ness of Brodgar tell us a great deal about house architecture, economy, and community beliefs. The buildings typically had dry stone walls, with flagstone furniture and central hearths; were sometimes surrounded by (or dug into) midden; and most likely had pitched roofs supported by timbers and lined with straw, seaweed, turves, or thin flagstones. Buildings were adjoining and connected by low passages (Skara Brae) or more dispersed and clustered around a central yard (Barnhouse). Although most of these buildings appear to be largely domestic in nature, a range of ‘specialized’ ancillary buildings have also been noted, including workshops and exceptionally large communal buildings (e.g. Childe 1931; Richards 2005).
FIG. 17.4. Interpretive plans of various late Neolithic sub-circular buildings in Britain and Ireland. A, C—Lough Gur C1 and C2, Limerick, Ireland; B—SlieveBreagh 1, Meath, Ireland; D—Gwithian Phase 2, Cornwall, England; E—Trelystan A, Powys, Wales; F—Trelystan B, Wales; G—Hockwold-cum-Wilton, Norfolk, England; H—Beckton F111, Dumfries and Galloway, Scotland; I—Cowie, Stirling, Scotland. (A–C, Grogan 1996, fig. 4.3; D–G, Darvill 1996, fig. 6.10; H, I Barclay 2003, fig. 8.5).
There are many dozen possible Neolithic stone houses on Shetland, often associated with field banks (Calder 1956; Barclay 1996, 62ff). Dating is patchy, but three sub-circular to oval buildings at Scord of Brouster dated to the late Neolithic (Whittle et al. 1986). These houses had thick stone walls, recesses in the walls, and hearths, and at least one was preceded by a timber ‘flimsy transient shelter’ (Whittle et al. 1986., 133). There is similar untapped potential on the Western Isles (Barclay 1996, 70). Evidence about the character of settlements has been gained from excavations at Eilean Domhnuill, North Uist, a settlement on an islet occupied from as early as 3600 BC for a millennium (Armit 2003). A sequence of small stone-walled rectangular houses were constructed, perhaps only a few in use at a time, and it seems this location was used seasonally or periodically due to changing water levels (Armit 2003, 98).
There is a tendency for late Neolithic settlements to include a number of buildings, sometimes occurring in clusters, often with one building replacing the other rather than all being contemporary. This hints at groups returning to the same location, but not necessarily the same building, again and again. Grogan (2002, 524) has called oval timber structures ‘simple domestic dwellings’. These are at odds with the enduring stone buildings of Orkney and Shetland, although evidence suggests that even these ‘permanent’ buildings were being replaced and rebuilt through time. There seems to have been a concern with using, and reusing, the same places.
PITS AND MIDDENS
As noted above, the role of pits in identifying settlement locations has a long tradition in Neolithic studies (cf. Garrow, this volume). Initially, pits with Neolithic material were viewed as grain storage pits (or rubbish pits), and these often occurred in groups (e.g. Clark et al. 1960; Simpson and Coles 1990). As such, pit clusters were viewed as a proxy for settlement, with an assumption that associated buildings had been destroyed or remained undiscovered (e.g. Field et al. 1964, 367ff). However, Thomas (1999, 64ff) convincingly argued against the storage pit interpretation, using an alternative narrative developed in the 1980s, that of ‘structured deposition’. This does not mean they did not hint at domestic activity, as pits often contain the detritus of everyday life, from hearth sweepings and broken potsherds, to parts of quernstones. Pits may also be associated with pottery firing (Gibson 2002). The association of pits with a tent-like structure, cereals, and Carinated Bowl potsherds at Boghead, Moray (Burl 1984) suggested that early farmers had at least some temporary dwellings associated with pits.
Larger pit clusters are perhaps indicative of longer-term repeat visits, such as the 226 Neolithic pits excavated at Kilverstone, Norfolk (Garrow et al. 2005). These relatively small pits were arranged in several large groups and pits commonly overlapped one another. Most seem to have been opened and filled fairly quickly, and contained a wide range of material including pottery, burnt flint and stone, and carbonized organics; refits of material between small groups of pits were common. The excavators concluded that the pits indicated repeated visits over several centuries, with groups possibly occupying the area for a short period of time only. A similar pattern can be seen at Dubton Farm, Angus, where huge pits 2.7m across and 1.85m deep were dug to receive a range of deposits including potsherds and carbonized material (Cameron 2002); this location was re-used throughout much of the Neolithic.
Middens also hint at long-term patterns of deposition. Spreads of deposit within hollows at Eton Rowing Course, Buckinghamshire, consisted of thousands of potsherds, lithics, and fragments of animal bones
and quernstones (Allen et al. 2004, 85ff). The excavators noted nothing ‘exotic’ and these deposits were viewed as indicative of long-term dumping of domestic ‘rubbish’. No buildings were found in the vicinity. Spreads of occupation debris associated with multiple stake-holes suggest middens, pits, and temporary buildings/shelters were located together in Ireland in the middle Neolithic (Smyth 2010, 14ff). Pits and middens, then, are indicative of depositional activity at certain locations, and can offer an insight into the disposal of domestic and other materials, including patterns of leaving and returning to special locales (Pollard 1999).
CRAFT AND DESIGN
Buildings are almost always culturally significant in terms of use and social meaning (Samson 1990; Parker Pearson and Richards 1994a). Houses can form a sense of rootedness for communities, are related to family and kin, and are usually the medium for a range of routine activities. The materiality of buildings, their appearance, and layout can be meaningful in a range of ways, and it seems likely that this was the case in the Neolithic (Cooney 1999; Bradley 2005). Unfortunately, in Britain and Ireland, the evidence we have is largely plough-truncated and sub-surface. Most organic building materials are elusive archaeologically, whilst the height of walls, as well as roofing, are speculative. Inevitably, any discussion will draw disproportionately from stone buildings.
The act of entering—or looking—into these buildings would have been mediated (and controlled) via entrances, but we have little understanding of doorways in Neolithic houses other than their location. We have no sense of what ‘doors’ would have looked like, and it may be wooden buildings were sealed by a wattle screen or similar. The role of doorways as means to allow light into these buildings is a possibility: where ‘entrances’ have been identified, they are largely in the southern and eastern sectors (Topping 1996). Entering any building involved crossing a threshold, a liminal experience that may have been associated with rites of purification or a transformation (cf. Parker Pearson and Richards 1994b, 25). Entry into Balbridie and Claish timber halls would have been physically awkward, channeled through narrow and angled entrances (Barclay et al. 2002). Threshold slabs within entrance passages of Orcadian houses marked this change from outwith to within, and at Barnhouse one threshold may have involved crossing a burial cist (Richards 1990, 116; 2005).
