Creating communities new advances in Central European Neolithic research Edited by Daniela Hofmann and Penny Bickle Published by Oxbow Books, Oxford © Oxbow Books and the individual authors, 2008 ISBN 978- A CIP record for this book is available from the British Library This book is available direct from Oxbow Books and The David Brown Book Company PO Box 511, Oakville, CT 06779, USA Phone:860-945-9329; Fax: 860-945-9468 or from our website www.oxbowbooks.com Cover image: Printed in Great Britain by Contents Introduction: researching across borders Penny Bickle and Daniela Hofmann 1 Diverging trajectories? Forager-farmer interaction in the southern part of the Lower Rhine area and the applicability of contact models Luc Amkreutz, Bart Vanmontfort and Leo Verhart 11 Frontier setlements of the LBK in central Belgium Marc Lodewjckx, with Corrie Bakels 32 The extreme eastern periphery of the Linearbandkeramik: the landscape and geographical contexts Olga Larina 49 Setlement history of the Linear Band Potery culture in Kuyavia Joanna Pyzel 70 The exchange of LBK adze blades in central Europe: an example for economic investigations in archaeology Brita Ramminger 79 Setlement history, land use and social networks of early Neolithic communities in western Germany Erich Claßen 94 First relections on the exploitation of animals in Villeneuve-Saint-Germain society at the end of the early Neolithic in the Paris Basin (France) Lisandre Bedault 110 Scene by the brook: early Neolithic landscape perspectives in the Paris Basin Penny Bickle 131 Mobility in a sedentary society: insights from isotope analysis of LBK human and animal teeth Corina Knipper 141 New aspects and models for Bandkeramik setlement research Oliver Rück 158 4 Contents A monumental prestige patchwork Joachim Pechtl 185 The LBK setlement with pit enclosure at Herxheim near Landau (Palatinate) Andrea Zeeb-Lanz, Rose-Marie Arbogast, Fabian Haack, Miriam Haidle, Christian Jeunesse, Jörg Orschiedt, Dirk Schimmelpfennig and Samuel van Willigen 199 Cemetery and setlement burial in the Lower Bavarian LBK Daniela Hofmann 216 Bone temper in early Neolithic vessels from southern Poland. Examinations using Scanning Microscopy Anna Rauba-Bukowska 231 The people who lived in longhouses: what’s the big idea? Alasdair Whitle 245 Contributors Luc Amkreutz National Museum of Antiquities (Rjksmuseum van Oudheden) Rapenburg 28 NL – 2311 EW Leiden Rose-Marie Arbogast Institut für Prähistorische und Naturwissenschatliche Archäologie (IPNA) University of Basel Spalenring 145 CH – 4055 Basel Corrie Bakels Faculty of Archaeology, University of Leiden Postbus 9515 NL – 2300 RA Leiden Lisandre Bedault Université de Paris 1 Panthéon-Sorbonne CNRS UMR 7041, ArScAn, Protohistoire Européenne Maison de l’Archéologie et de l’Ethnologie 21, allée de l’Université FR – 92023 Nanterre Cedex Penny Bickle School of History and Archaeology Cardif University Humanities Building, Colum Road UK – Cardif CF 10 3EU Erich Claßen Bayerisches Landesamt für Denkmalplege Referat B1 Oberbayern/München Dienststelle Ingolstadt Unterer Graben 37 DE – 85049 Ingolstadt Corina Knipper Eberhard-Karls-Universität Tübingen Institut für Ur-und Frühgeschichte und Archäologie des Mitelalters Naturwissenschatliche Archäologie Rümelinstr. 23 DE – 72070 Tübingen Olga Larina Institutul Patrimoniului Cultural Str. Banulescu Bodoni 35 MD-2012 Chisinau Marc Lodewjckx Dept. of Archaeology, University of Leuven Postbus 33 BE – 3000 Leuven Fabian Haack Generaldirektion Kulturelles Erbe Rheinland-Pfalz Direktion Archäologie – Speyer Kleine Pfafengasse 10 DE – 67346 Speyer Miriam Haidle Intitut für Ur- und Frühgeschichte und Archäologie des Mitelalters Abt. Ältere Urgeschichte – Quartärökologie Schloss, Burgsteige 11 DE – 72072 Tübingen Daniela Hofmann School of History and Archaeology Cardif University Humanities Building, Colum Road UK – Cardif CF 10 3EU Christian Jeunesse Institut d’Antiquités Nationales Université Marc Bloch – Strasbourg II 9, place de l’Université FR – 67084 Strasbourg Cedex Jörg Orschiedt Archäologisches Institut Universität Hamburg Edmund-Siemers-Allee 1, Flügel West DE – 20146 Hamburg Joachim Pechtl Institut für Ur- und Frühgeschichte Ruprecht-Karls-Universität Heidelberg Marstallhof 4 DE – 69117 Heidelberg 6 Joanna Pyzel Institute of Archaeology and Ethnology Polish Academy of Sciences PL – Poznan Brita Ramminger Archäologisches Institut Universität Hamburg Edmund-Siemers-Allee 1, Flügel West DE – 20146 Hamburg Anna Rauba-Bukowska AGH University of Sciences and Technology PL – 30-059 Krakow al. Mickiewicza 30 Oliver Rück Lahnstr. 31 DE – 51105 Köln Dirk Schimmelpfennig Institut für Ur- und Frühgeschichte Universität zu Köln DE – 50923 Köln Contributors Bart Vanmontfort Faculty of Archaeology University of Leiden Reuvensplaats 3–4 NL – 2311 BE Leiden Samuel van Willigen Swiss National Museum Department of Archaeology Museumstraße 2 CH – 8023 Zürich Leo Verhart Faculty of Archaeology University of Leiden Reuvensplaats 3–4 NL – 2311 BE Leiden Andrea Zeeb-Lanz Generaldirektion Kulturelles Erbe Rheinland-Pfalz Direktion Archäologie – Speyer Kleine Pfafengasse 10 DE – 67346 Speyer New aspects and models for Bandkeramik setlement research Oliver Rück Translation by Daniela Hofmann This contribution is a summary of parts of my doctoral dissertation, published in German in 2007, which critically reviewed LBK architecture and setlement structure on the basis of the Weisweiler 111 site on the Aldenhovener Plate, Rhineland, and other setlements. The interested reader is referred there for further detail and a fuller exposition and illustration of the points made here. This article aims to briely introduce some new models for Bandkeramik setlement, including the secondary enlargement of buildings (additive building system), the row setlement model and the possible reconstruction of houses with a raised dwelling platform. The later seems to have been necessary on most sites in order to compensate for the hill slope. At irst, I introduce the idea that LBK buildings were not static and may have been enlarged through time. The most commonly used typology of LBK houses is based on a scheme by Waterbolk and Modderman (1958/59), who classify LBK buildings according to the observed combination of characteristically disposed features, i.e. parts of a building’s plan which are recognisably distinct and which they identify as northwest, central and southeast parts. They could show that: 1. the central part can occur as a structure in itself (Kleinbau; type 3) 2. the central part can be combined with a northwest part (Bau; type 2) 3. the central part can occur with both a northwest and a southeast part (Großbau; type 1). However, this does not necessarily mean that all parts must have been built at the same time, an assumption which often underlies interpretations on house construction. Yet if we accept that structures could be secondarily enlarged, there are also consequences for the potential durability of the houses and hence for setlement layout and organisation. The second half of the paper hence critiques the widespread Hofplatz (or ‘ward’) model developed for the LBK in the Rhineland, which stipulates short-lived houses located far from their contemporary neighbours in autonomous wards. Instead, I propose a more planned, linear setlement layout. Finally, I ofer some alternative reconstructions of LBK buildings based on the idea of a raised dwelling platform supported by posts. Building-speciic observations When analysing house plans from Bandkeramik setlements, it becomes apparent that the majority of buildings exhibits morphological peculiarities. These either concern single elements of the buildings (postholes, wall trenches) or the structure of whole parts of the house (e.g. the northwest or southeast part). Numerous plans also stand out due to their speciic constellation of house parts (e.g. combination of central and southeast part) and the accompanying loam pits. The sheer number of instances of such peculiarities suggests a departure from the traditional, rigid LBK building scheme. The different characteristics of the Weisweiler house plans show that the degree of correspondence to a normative building scheme is reduced to a minimum. The site furthermore conirms the observation that building morphology changed over the course of the LBK (e.g. Modderman 1977). New ways of construction were experimented with and known ones were varied; inter-regional contacts probably also played a part here. It seems evident that the spectrum of architectural possibilities and fashions New aspects and models for Bandkeramik setlement research from which the builders of new houses could choose increased throughout the LBK1. Overall, this creates the impression that, beginning from the mid LBK and even more strongly in the late and latest LBK, the architectural uniformity and the normative trends in building construction within a setlement decreased. According to Coudart, too, the degree of architectural standardisation is smallest “at the time at which the Bandkeramik had reached its greatest extent” (Coudart 1998, 238). However, this seems to be partially opposed to the emergence of ‘ceramic’ regional groupings in the latest LBK2. This opposition could suggest that the diferent kinds of material culture groupings, houses and ceramic vessels, were bearers of symbolism with diferent intended messages. Enlargement of buildings When analysing the buildings from the Weisweiler 111 setlement, located on the Aldenhovener Plate in the Rhineland, it became apparent that in three buildings (No. 2, 3 and 4) the orientation of the southeast part diverged from the orientation of the longitudinal axis of the northwest and central parts. In plan, this appears as a southeast part at a marked angle relative to the rest of the house (Figure 1). Other observations also hint at the possibility that the southeast part could have been constructed ater the central and northwest sections of a building had already been in use for some time. The later addition of a southeast part is also postulated for buildings 10 and 17. In the following, I will introduce examples from Weisweiler 111 and other setlements to suggest the enlargement of existing buildings during the LBK. The southeast part of house 4 deviates from the length axis of the central part by 6° (Figure 2b)3. An additional post was noted at the transition from the central to the southeast part; it was probably necessary for the enlargement of the structure. The northwest part, too, showed numerous posts which deviated from the normal building scheme and which could hint at later refurbishments or repairs. Figure 2a shows the hypothetical appearance of the house before its suggested enlargement. Having discussed a possible secondary enlargement with the example of building 4, similar processes can also be supposed for houses 2, 3, 10 and 17 (Figure 1). For building 2, the potery recovered from loam pits could be used – in addition to the plan – to argue for a secondary expansion: once the potery from the pit diagonally in front of the southeastern end of the structure is dated by correspondence analysis, the 159 material overall appears younger than the potery from loam pit 3574. Other Bandkeramik settlements also feature buildings which were probably expanded and whose southeast parts diverge from the orientation of their central and northwest parts. In the following, I will use mainly structures from Elsloo and Stein in Dutch Limburg as examples, but further instances can be cited from Ulm-Eggingen (Baden-Würtemberg) and Mold (Lower Austria). Figure 3a shows the hypothetical appearance of house 88 at Elsloo before the enlargement, and Figure 3b its actually observed plan. The southeast part, potentially added at a later date, and two pits accompanying it are marked in red. In addition, two posts in the northwest part, also shown in red, may indicate further rebuilds or repairs. Overall, the observations already made for the Weisweiler buildings can be repeated here. In addition, the very short distance between the two post rows just northwest of the area with double posts is particularly noteworthy. What is more, the posts in these two rows are arranged ofset to each other. This could be because the building enlargement necessitated an additional post row. In order to be able to move the beams necessary for an enlargement past already standing posts, any newly erected posts in this row (shown grey in Figure 3b) had to be ofset. This may be the reason why the longitudinal post rows of the southeast part are not in a straight line with those in the central and northwest parts. It is hardly possible to ind a clearer example of a secondary enlargement of a building by the addition of a southeast part. Houses 3 and 76 are further instances at Elsloo. As house 4 at Weisweiler 111, these buildings also feature loam pits which lank the southeast part only. Summary of indicators for building expansion and/or repair5 1. loam pits dug along the northwest and/or central part which end at the transition to the southeast part 2. loam pits accompanying only the southeast part (and mostly of the same length as the later) 3. a different orientation of the southeast part respective to the northwest/centre 4. ‘additional’ posts or post row at the transition between central and southeast parts 5. ‘additional’ posts in the northwest and/or central parts which suggest repairs or rebuilds. This kind of additional posts seems to be absent from southeast parts. 160 Oliver Rück Figure 1: Plan of Bandkeramik features at Weisweiler 111. Doted lines are suggested reconstructions in areas disturbed by younger features. New aspects and models for Bandkeramik setlement research 161 a N 0 5m a) N b 682 645 0 5m b) Figure 2: House 4 at Weisweiler 111. a) Hypothetical appearance before enlargement. b) as excavated. The southeast part, ofset by 6° from the rest of the house, a loam pit and some of the posts in the central part are marked in dark grey; they may be connected to the episode of enlargement and possible repair. The indicators listed above can appear singly or combined. This begs the question of how far these observations reflect an underlying regularity and whether they can be paralleled for buildings from other sites. A irst look at several setlement plans shows that this is indeed the case and that indicators of this kind can be found in large parts of the LBK distribution; a comprehensive gathering and statistical analysis of the data must, however, follow elsewhere. These indicators for a secondary expansion of buildings may also be related to the observation that at Weisweiler 111, no type 3 buildings (Kleinbau) were revealed. Houses classed as type 2 (Bau) can equally not always be unequivocally identiied. 50% of all houses excavated at Weisweiler 111 can be identiied Figure 3: House 88 at Elsoo, Netherlands. a) Hypothetical appearance before enlargement. b) as excavated. Modderman 1970, plate 36. Scale: 1:200. as type 1 (Großbau), a further 35% are not well enough preserved to allow a decision of whether they are type 1 or type 2 buildings. Some of these, too, can probably be numbered among the type 1 structures (e.g. house 10, possibly also house 17). The remaining 15% cannot be addressed in any detail. At this point, it must be mentioned that at the Langweiler 8 setlement, also on the Aldenhovener Plate, only three of the 108 buildings could be identiied as Kleinbauten, and even these are not entirely certain (von Brandt 1988). At the LBK setlement of Kückhoven in the Rhineland, where 85 house plans could be reconstructed, there were no type 3 buildings at all (Lehmann 2004). How can we explain the small number of Kleinbauten at so many sites? 162 Oliver Rück Figure 4: Schematic model of possible enlargements for type 2 and 3 LBK buildings (additive building system). The model of house enlargement For Waterbolk and Modderman (1958/59), the central part of a house is the fundamental element of construction present in every building type. Other combinations (double central part, or a central and southeast without a northwest part) have only been observed very rarely in the entire LBK area (e.g. Modderman 1988). This scheme, based on different construction modules, seems predestined to be understood as a kind of building-kit system (additive building system; see Figure 4). This kind of outlook would mean that houses could well have been expanded in the course of their use life. Certainly not all buildings have been modiied according to the model presented here, but it probably applies to a certain percentage of houses. With this in mind, some of the Großbauten on sites where this building type is uncommonly frequent and the occasionally small number of Kleinbauten could be explained6. Kleinbauten, as well as some houses originally planned as Bauten, could thus have been enlarged with a southeast part some time ater their initial construction. Alongside the addition of a southeast part, it is possible that houses were later enlarged with a northwest part (as also suggested by Bradley 2001), or that the northwest part was extended, as can for instance be suggested for house 5 at Ulm-Eggingen. These examples support the possibility of the secondary enlargement of buildings from a morphological point of view7. Setlement geographers (e.g. Lienau 2000) have repeatedly pointed out reasons for structural changes to buildings in the course of their use, i.e. also including changes in family or social structure. “If a farmhouse at the point of its construction can be understood as a facility meeting a certain purpose, its functions and actual requirements oten drit apart with the increasing age of the farm. […] The building’s substance oten renders a radical structural adaptation to changing operational conditions too expensive and diicult. Hence, an adaptation generally New aspects and models for Bandkeramik setlement research happens in stages, and only rarely radically through demolition and rebuilding at the same or a diferent location […]” (Lienau 2000, 56–57). The suggestion that houses could be enlarged, repaired or altered in the course of their use life has a fundamental efect on the plausibility of a setlement model irst formulated about 30 years ago (Kuper et al. 1977). Now known as Hofplatzmodell (ward model), it atempted to explain the structure and development of an LBK setlement (e.g. Boelicke et al. 1988). The model divides a setlement into several spatial areas, the so-called Hofplätze or wards8. The model implies that no two buildings could have stood on a ward at the same time. In addition, it assumes that each house was in use for roughly one generation, i.e. on average for about 25 years (Stehli 1989). If we assume the secondary enlargement or repair of buildings, this would irst of all point to the fact that houses were constructed with a longer span of use in mind and were lived in for a considerable time – probably for much longer than the postulated 25 years. This would mean a much larger number of contemporary houses and hence more inhabitants in a given settlement. Consequently, to retain the Hofplatzmodell, one would either have to increase the number of wards or the number of contemporary houses per ward. However, this would also mean that the minimum distance between contemporary houses, generally quoted as 25–50 m, would have to be reduced. As neither of these options can be reconciled with the Hofplatzmodell, the alternative is to suggest a new model altogether. Setlement-speciic observations At the Weisweiler 111 site, there are no overlapping LBK house plans9. The buildings at this site show a regular arrangement – they are grouped in several spatial clusters and lie alongside each other with a roughly parallel orientation (Figure 1). The setlement plan thus appears structured and intentionally shaped and is well suited for considerations of setlement structure and development. At smaller sites or those with few overlapping house plans, the formation of the setlement can be more easily traced than at large and long-lasting ones such as Köln-Lindenthal, Bylany or Langweiler 8. The observable characteristics of such an intensively built-on site are the outcome of several decades or centuries of setlement activity in the same location. These constant activities can result in unclear site plans with numerous intercuting houses, as is the 163 case at the sites mentioned above. Where the “original regularity of a facility is crippled beyond recognition by later building works […]” (Lienau 2000, 66), it seems less suitable for the construction of models. Using Weisweiler 111 as an example, I will outline structural principles which are observable on the majority of Bandkeramik setlement plans (for instance at Cuiry-lès-Chaudardes, France; the Dutch sites of Elsloo and Geleen-Janskamperveld; at Frimmersdorf 141, UlmEggingen and Regensburg-Harting in Germany; and at Füzesabony-Gubakút, Hungary) and point towards a setlement structure which is fundamentally diferent to the Hofplatzmodell (Boelicke et al. 1988). The following paragraphs outline the observations which led to this conclusion in more detail. The feature-free area in front of the southeast end A closer look at the plans of LBK setlements, especially at those where house plans only overlap to a small extent, shows that the area immediately in front of the southeastern or southern end of a house is generally devoid of features. This observation can be replicated for most sites. At Weisweiler 111, it can be made for eight or ten buildings (Figure 1), but this peculiarity is also evident at other setlements10. It is supposed that this area was somehow connected to the house. One possibility is the existence of an unpreserved terrace raised above ground level (Figure 20)11. Posts which occur as extensions of a building’s outer walls or of the interior longitudinal post rows could hint at such a porch or extension. At Weisweiler 111, traces of such posts were observed for houses 3 and 4; at Cuiry-lès-Chaudardes for buildings 45, 89, 90, 225, 245, 280, 320, 360, 380, 390, 400, 425, 500 and 530 (Figure 5, highlighted by arrows). Due to their frequent occurrence, these single or grouped postholes can no longer be regarded as coincidental. As far as the durability of LBK houses is concerned, it seems interesting that even at setlements with a greater density of buildings and several parallel house plans with aligned gable ends, the areas to the southeast of the structures are devoid of features. This makes it likely that more of the houses were in use simultaneously. If this had not been the case, this empty area could have been used for digging pits or for the construction of new buildings. This is especially clear at the sites of Elsloo (Figure 6) and Ulm-Eggingen, where the areas in front of the houses have remained devoid of features in spite of the frequent intercuting of house plans. 164 Oliver Rück Figure 5: Plan of Cuiry-lès-Chaudardes, northeast France. Areas in front of the buildings which are virtually devoid of features are highlighted in grey. Arrows denote posts or groups of posts which are aligned with the house walls or longitudinal post rows in the interior. Ater Coudart 1998, 136, ig. 130. Even between the diferent rows of houses, the area immediately to the southeast of the structures rarely yields any features and is only overlaid by later houses in very densely built-over areas of the site. Aligned gable ends and parallel houses The analysis of numerous site plans shows that most LBK setlements in central Europe have one factor in common: within the site, groups or rows of parallel houses with aligned gable ends can be identiied. The distance between individual structures varies between one or two building widths. The number of houses arranged in this way seems to vary with the size and extent of the site. Thus, at Straubing-Lerchenhaid in Bavaria, there are at least two pairs of parallel buildings with aligned gable ends (Brink-Kloke 1992, 10; ig. 1.8). In the southern part of Ulm-Eggingen, three houses (buildings 6, 4 and 8) lie parallel to each other and their southeastern gable ends are located on a shared baseline or alignment. The largest number of parallel buildings with aligned gable ends was observed at Cuiry-lès-Chaudardes in northeast France (Figure 7). Weisweiler 111 shows a similar arrangement of houses. New aspects and models for Bandkeramik setlement research 165 Figure 6: Plan of Elsoo, area west of the Koolweg. Areas in front of the buildings which are virtually devoid of features are highlighted in grey. Ater Modderman 1985, Beilage 2. Here, buildings 2, 4, 5 and 9 lie next to each other and feature aligned gable ends. Before its suggested extension, the gable end of house 3 would also be aligned with that of the other buildings (Figure 1). At setlements with large numbers of overlapping house plans, the original setlement structure, or initial form, has oten been greatly modiied. Nevertheless, parallel buildings with aligned gable ends can also be recognised over larger distances within such sites. Although the numerous instances of intercutting houses hide the structured arrangement of buildings, if such later houses are metaphorically blinded out the original setlement plan can oten be iltered out. This is for instance the case at Elsloo (Figure 8), Langweiler 8 and Regensburg-Harting (Becker and Braasch 1984). On the basis of setlement geographical investigations of historical and recent setlement forms, it can be suggested that parallel buildings with aligned gable ends are contemporary to each other (Lienau 2000). As an example, one could cite central European rural setlements, where farms are strung out along the main street. Cities, newly developed areas or terraced housing could be added here. Ethnographically documented setlements in southeast Asia, for instance those of the Batak on Sumatra (Guidoni 1976) or of the Toraja on Sulawesi (Fraser 1968) also illustrate the contemporaneity of neighbouring buildings (Figure 14). The lake-side setlements of the Alpine foreland can 166 Oliver Rück Figure 7: Cuiry-lès-Chaudardes, showing linear arrangement of roughly parallel houses. Although there is no type 1 building, each row seems to have a house which stands out due to its size (length and/or width). Ater Coudart 1998, 136, ig. 130. be quoted as an archaeological example for regularly placed contemporary houses. In all, these observations suggest that on LBK setlements, too, a large part of parallel buildings with aligned gable ends was in use simultaneously12. Setlement structure – the arrangement of houses in rows “The form of a setlement results from the shape of the setlement plan and the density of buildings. The setlement plan is the result of the arrangement of the houses and/or yards which are combined into a setlement unit and their relationship to the streets (paths) and squares” (Lienau 2000, 64). Can we also recognise a speciic setlement form for LBK sites? The analysis of several setlement plans shows that houses were arranged in rows. Within a row of houses (or setlement row), neighbouring houses are not or only marginally ofset longitudinally. The observed height of such a row lies between one or two house lengths (Figures 7–14)13. The number of rows probably depends New aspects and models for Bandkeramik setlement research 167 Figure 8: Elsloo. Houses of the same colour seem to form rows. Note the few overlaps, which appear mostly between narrow sides of structures. This suggests that houses in the same row could have existed simultaneously. Also note that the ive type 1 houses belong to ive diferent rows. Light grey houses cannot be atributed unequivocally. Ater Modderman 1985, Beilage 2. on the size and duration of the setlement. Equally, not all houses in a row need to have existed at the same time. There is, however, the possibility of coexistence, especially where no overlaps of houses are observed. The appearance of a row of houses probably changed over time: new buildings were added, old ones decayed or were built over. A setlement plan with houses laid out in rows can be reconstructed for most sites – even where the original setlement structure was hidden by high levels of setlement activity and overlaps. Until now, LBK sites have not been systematically examined for a row-based setlement structure. This can be explained by the predominance of the Hofplatz model and the fact that, for the Rhineland at least, alternative interpretations did not seem necessary. In the course of inter-regional investigations on the LBK, Coudart (1993, 128) reached the conclusion that setlements could generally be divided into ive spatial units. These units could be linear or irregular. Although her Figures show a row-based structure for several setlements (Coudart 1993, 129, ig. 15) she does not interpret them in this way. In all probability, she was inluenced by 168 Oliver Rück Figure 9: Füzesabony-Gubakút, northeast Hungary. Part of area excavated in advance of motorway construction (M-3). Note the parallel arrangement of houses and their aligned gable ends, as well as the aligned loam pits. Domboróczki 2001, 197, ig. 4. the Hofplatz model14, as well as assuming only ive contemporary longhouses for large LBK sites. Domboróczki (2001) described a linear arrangement of buildings on Alföld-LBK sites (Figure 9). Based on his original observation that large pits on the northeast Hungarian site of Füzesabony-Gubakút were arranged 8–10 m apart in rows, Domboróczki’s analysis of the setlement features reached the following conclusions: “The most important result of the excavations at Füzesabony-Gubakút was the recognition of the setlement structure of the ALPC [Alföld Linear Potery Culture]. The setlement consisted of four parallel rows, with two rows situated along both sides of a one-time riverbed. The setlement rows were composed of houses and refuse pits” (Domboróczki 2001, 202). The illustrations of several setlement plans presented here were the starting point for the description of the Bandkeramik village as a continuously growing New aspects and models for Bandkeramik setlement research setlement in which several parallel rows of houses were in use at the same time (Figures 5–13)15. On the reconstruction of Bandkeramik buildings A truthful reconstruction of LBK houses is a diicult task, and one which remains unresolved in detail. This could change through further inds such as the well from Kückhoven or the discovery of similarly well-preserved house walls, roofs or walking horizons. For now, we can only base our tentative conclusions of the aboveground elements of the house on detailed observation and documentation of the dug features and their embeddedness into the site’s topography16. The houses thus reconstructed will only roughly relect prehistoric reality. Most architectural characteristics, such as interior subdivisions and features, roof construction, windows, entrances and the question whether the building had more than one loor are as impossible to determine as the details of a probably highly developed decoration and colour scheme. For the past 70 years, there have been suggestions on the reconstruction of LBK buildings from structural, economic, technological and static points of view (Butler and Haberey 1936; Paret 1946; Meyer-Christian 1976; Startin 1978; Lüning 1980; Masuch and Ziessow 1983; von Brandt 1988; Luley 1992). Ever since Paret’s (1946) criticism of Butler’s theory of pit houses and raised granaries (Buttler and Haberey 1936), the patterned postholes of an LBK building are used to reconstruct the same, ground-level building all across central and western Europe17. In spite of the large number of archaeological inds – more than 2000 investigated LBK house plans — it cannot be proven whether ground-level houses even existed in the Bandkeramik. Part of the reason is that the former walking horizon was destroyed by erosion, resulting in most information being lost. The reconstruction accepted at present is hence based on hypotheses which, in the course of research into the Neolithic, assumed paradigmatic traits18. On the topography of LBK sites However, observations on the topography of LBK sites (Rück 2004) and investigations on climate at the time (Schmidt et al. 2004) make the interpretation of a ground-level living space seem doubtful. As early as 1972, Sielmann summarised the topographic location of LBK sites as follows: the setlements were 169 generally located on the top or the upper third of rises sloping down to watercourses or on the edge of loess-covered river terraces (Sielmann 1972). New information from LBK sites conirms these general paterns and forcefully demonstrates that steep slopes were chosen as setlement sites. Table 1 shows the topographic situation on slopes of 14 LBK sites, but could be extended to almost all setlements in southern and western Germany, as the following examples show: mapping more than 100 LBK sites in the area of the Nördlinger Ries, Bavaria, shows setlement concentrations on the slopes at the edge of the Ries and along river courses. The Ries itself, i.e. the fertile loess soils in the plain itself, in contrast, remained largely devoid of setlement (Zeeb-Lanz 2003, 296–98, maps 1 and 2). Similarly, setlements along the Merzbach valley on the Aldenhovener Platte are also concentrated on rises or valley slopes descending towards the Merzbach stream (Lüning and Stehli 1994, IX). Parts of the Langweiler 9 setlement lay on a slope which dropped three metres over a distance of 75m (from 134 to 131 m above sea level) in the area of houses 8 to 11 (Lüning 1982, 24, ig. 10). A comparable situation exists in Poland. “The Danubian I population nearly always colonized the lower part of the valley slopes. The setlements were situated at their edges immediately above the inundated terrace covered with silts of various rotation” (Kruk 1973, 250). Plateaus or plains adjacent to the higher ground were not setled. The setlements at Geleen and Sitard can stand as examples for the Dutch Limburg. Waterbolk and Modderman (1958/59, 36) describe the topography of Sitard as follows: “Globally speaking, the contours run northsouth. Before modern house construction, the largest height diference was 4.75 m. However, this Figure increases further if we consider the level at which the Bandkeramik features were found; here, it is 5.5 m”. Soil scientiic investigations conirm that erosion processes led to erosion in post-LBK times and contributed to a levelling of the ground (Schalich 1977). Kuper et al. (1975, 17) note: “Intensive mapping of soil types in the Merzbach area showed that this landscape was much more structured 6000 years ago and featured much greater height diferences than today. Thus, the Neolithic Merzbach valley lay up to 4 m below the modern-day ground level, while the hill slopes have to be imagined as augmented by the mass of the since eroded soil”. This observation can be applied to other loess areas. For the Weterau region in Hesse, Thiemeyer (1988) established a latening of 170 Oliver Rück Figure 10: Geleen-Janskamperveld, Netherlands. Houses of the same colour seem to form rows. The boundary between the orange and purple rows corresponds with a ditch. Type 1 houses are spread in diferent rows. Ater Koojmans et al. 2003, 376, ig. 2. 47 46 45 48 49 N 42a 44 43 41a 40 50 51 42 54 53 41 55 52 39 57 59 56 58 28 38 36 26 27 25 24 34 33 35 23 22 31 32 30 29 13 12 1 20 21 4 11 5 2 15 17 9 14 19 18 10 8 6 3 16 7 0 25m Figure 11: Geleen-Janskamperveld. Dashed lines are the wards reconstructed by Koojmans et al. (2003). The houses which the authors have dated to ceramic phase 3 are shown in black. As seen in igure 10, they form a line. Ater Koojmans et al. (2003, 387, ig.8). New aspects and models for Bandkeramik setlement research 98 99 77 78 2 4 70 65 1 41 67 71 54 91 68 55 56 42 3 79 6 73 81 59 95 57 110 61 80 49 53 48 44 63 76 60 108 46 86 5 82 58 83 43 103 102 75 69 64 101 100 85 74 171 62 52 47 107 51 40 45 38 39 31 7 50 37 36 32 94 26 29 27 25 33 35 21 22 2 1 24 34 8 66 10 8 14 87 11 5 17 15 6 4 9 19 18 88 3 30 28 23 7 9 13 10 20 12 11 16 12 Langweiler 8 Wards dated building undated building 0 50m N Figure 12: Langweiler 8, Rhineland. Diachronic structure of wards 1–12, illustrating the Hofplatzmodell. The suggested ward boundaries seem artiicial. Ater Stehli 1994, 87, ig. 1. slopes of 50%, i.e. a prehistoric gradient of 15% is only 3–6% today. The gradient at the sites listed in table 1 also falls within this range. LBK climate This begs the question of why slope locations were preferred. According to new dendro-climatological investigations, above-average rainfall could be reconstructed for the LBK (Schmidt et al. 2004). Alongside results from dendrochronology (Figure 15), the following observations support the idea of a moist and warm climate during the LBK: 1. The formation of large lime deposits, for instance in Stutgart-Bad Cannstat, Baden-Würtemberg (Wagner 1995, 24–25) and Witislingen, Bavaria (Stirn 1964, 76; Seitz 1990, 20f). 2. The occurrence of the pond turtle19 in central and northern Europe (Willms 2003). 3. Kreuz (2007) discusses the dominance of twograined einkorn over emmer in the LBK as a possible climatic indicator. For several reasons, einkorn is the worse choice. Its only advantage over emmer is that plants remain upright in heavy rainfall and do not lie down, as all other cereal species, thereby avoiding substantial harvest losses. Interestingly, from the Flomborn phase onwards, which according to Schmidt et al (2004) correlates with a rise in precipitation, an einkorn species which produces two grains per spikelet appears (Kreuz and Boenke 2003, 233f). 172 Oliver Rück Langweiler 8 Rows dated building undated building N 0 50 m Figure 13: Langweiler 8 as a row setlement. Rows seem to correspond well with contour lines. Over time, houses were added, decayed or were built over; this dynamic situation resulted in shits within rows and of the rows themselves. Nevertheless, most houses fall within the suggested row boundaries. Ater Stehli 1994, 87, ig. 1. Climatic factors seem to have played a large role in the selection of locations for setlements and have led people to build on slopes and to avoid areas at risk of looding. Sielmann (1971, 101) noted that the areas in the top third of a rise or at a terrace edge are the zone characterised by the greatest soil aridity. This shows that lat, badly drained areas on (high-lying) plateaux were generally too wet for setlement. The measurements summarised in Table 1 show that height diferences of a metre or more over 20 – 40 m long houses were frequent. If LBK houses had been directly constructed on the ground surface, the living loor would also be sloping. This can almost certainly be excluded. There are no known ethnographic parallels for dwellings with a sloping loor. If we hence assume a level loor, a possibility for reconstruction is to see the LBK house as raised partially or entirely above ground by posts. Building houses with an elevated dwelling platform would not have presented a great diiculty to Neolithic builders. Since the excavation of the LBK well from Erkelenz-Kückhoven in 1990, we have detailed knowledge of the highly developed carpentry skills of the time (Weiner 1995). On the basis of the knowhow revealed there, it is certainly appropriate to base further relections on the existence of well-developed house constructions20. New aspects and models for Bandkeramik setlement research 173 Table 1. Slope gradient for 14 LBK sites. Calculations are based on site plans and topographical maps. Figures luctuate between 1.8 and 8.7%. Most LBK houses in the area are oriented NW–SE; the few exceptions are not considered here. All sites are located no more than 400 m from a watercourse. Note that for Ulm-Eggingen, the former ground surface has been radically altered due to gravel extraction (Kind 1989, 23); it is likely to have been steeper originally. Site Gradient (in %) 1.0 m hight difference over a distance of 6.3 – 8.7 15.9 - 11.5 m S Hienheim 4.8 20.8 m SE Modderman 1977, plate 3 Utzwingen 5.0 20.0 m SE Rück 2001, 18 Wittislingen 2.8 35.7 m SE Rück 1999, 8 Ulm-Eggingen 2.3 43.4 m SE Kind 1989, 23 and Beilage 1 Diemarden 4.8 20.8 m SE Posselt and Saile 2003, 312, fig. 3 Hempler 3.0 33.3 m SSE Schade-Lindig and Schwitalla 2003, 352 Köln-Lindenthal 2.5 40.0 m SSE (north ring) Langweiler 8 2.6 (near house 31) – 3.3 (near houses 1-9) 38.5 – 30.3 m SE Stehli 1994, 87 Langweiler 9 4.7 (near houses 8-11) 21.3m SE Stehli 1994, 95 Langweiler 2 1.9 (near houses 8-11, 13, 14) – 3.6 (near houses 1-4) 27.7 m SE Stehli 1994, 90 Geleen 1.8 (minimum; data imprecise) 55.6 m SE Waterbolk 1958/59, 123, plate XVIII ca. 2.5 40.0 m SE figures in Ilett 1982, 25–7 3.7 27.0 m S contours in Lenneis 2004, 383, fig. 4 Sallmannsberg Cuiry-lèsChaudardes Mold Archaeological indings Archaeological indings, too, support the position of LBK sites on slopes, as the following example of two buildings shows. Figure 16 illustrates the 35.8 m long house 12 from Ulm-Eggingen, oriented southeastnorthwest. The plan is irst of all notable for the stepped proile, decreasing towards the southeast, of the two ca. 12.5 m long wall trench arms (Figure 17). In the western arm, the base of the northernmost post pipe (143/10) reaches an absolute height of 532.97 m above sea level. In contrast, the base of the southernmost post pipe (20/30) in the same wall trench arm is 0.40 m lower at 532.57 m above sea level. The situation is similar in the eastern arm of the wall trench. But the heights do not only decrease in the northwest part of the house. Over the entire length of the structure, the absolute heights of post bases decrease from the northwest to Direction of Reference slope Brink-Kloke 1992, 9 Buttler and Haberey 1936, plate 3 the southeast (Kind 1989, 47). These diferent depths are not an isolated example; the same patern is repeated in further structures, such as houses 8 and 10 at UlmEggingen (Kind 1989), and houses S02, S07, S08, S11 and S13 at Landshut-Sallmannsberg (Brink-Kloke 1992). At the later setlement, houses S02 and S04 have a stepped wall trench (Brink-Kloke 1992). Several buildings from Hienheim (such as houses 8, 17 and 29) show a comparable patern (Modderman 1977), if one takes into account the ground level, let out of the schematic section drawings. In the second example, house 1 from Mold in Lower Austria, the former slope is very clearly visible (Figure 18). The height measurements given on the plan (Lenneis 2004, 383, ig. 4) drop from 292.60 m above sea level in the preserved portion of the north part to 291.20 m above sea level at the southern gable end – a height diference of 1.40 m21. As the postholes are of a 174 Oliver Rück Figure 14: The Sa’dan setlement of the Toraja in cental Sulawesi, Indonesia. Botom: schematic layout. Fraser 1968, ig. 38 and 39. Figure 15: Reconstruction of rainfall paterns between 5600-4600 BC. Ater an extremely dry period at 5360 BC (1), precipitation progressicely increased (phases a-c). A renewed dry episode (2) coincides with the end of the LBK. Period (3) is based on data from Kaster, Kreis Bergheim, Nordrhein-Westfalen. Points B1 and B2 denote the dendrochronological dates for the Erkelenz-Kückhoven wells. The overlay diagram shows the chronological distribution of 154 LBK houses from the Aldenhovener Plate dated to ceramic phases. Setlement seems to correlate well with increasing rainfall. From Schmidt et al. 2004, 304, ig. 1. New aspects and models for Bandkeramik setlement research 175 Figure 16: House 12, Ulm-Eggingen, Baden-Würtemberg. Kind 1989, 49, ig. 25. Figure 17: Ulm-Eggingen, house 12. Section through the western (top) and eastern (botom) arm of the wall trench, showing stepped proile. Kind 1989, 49, ig. 26. roughly equal depth throughout the entire building, their bases in the southern end are consequently also 1.40 m deeper in absolute terms. The archaeological examples collected here give the impression that the dominant slope was made used of in the course of house construction. As shown, in order to give suicient stability to all posts within the house, their botom ends had to be dug to a deeper level further down the slope, which resulted in stepped wall trenches and increasing depth measurements (Figures 17 and 18). At irst, one could suppose that the posts in the southeastern or southern ends of the house were generally dug in deeper, but the indings from well- preserved buildings contradict this. In such cases, it is generally the posts in the northwest part which are the deepest, or the northwest and southeast posts reach the same depth below ground level. At Ulm-Eggingen, this is for instance the case with houses 2, 12 and 20 (Kind 1989); at Hienheim, similar observations were made for houses 2, 5, 8 and 31 (Modderman 1977). The model of a dwelling platform Bandkeramik houses on southeast facing slopes could be envisaged as follows: where a northwest part was present, it rested on the ground surface. In the central 176 Oliver Rück and southeast parts, depending on the slope gradient and the length of the house, the distance of a putative platform to the ground level would progressively increase (Figures 19 and 20). The central parts of LBK houses oten contain posts which difer in diameter and depth from those in the northwest and southeast; they are larger and more deeply buried, as for instance in houses 14 and 17 at Hienheim (Modderman 1970, 25, 28), houses L02 and L05 at Lerchenhaid (Brink-Kloke 1992, 54, 58) or houses 3 and 18 at Weisweiler 111 (Rück 2007). In this part of the house, such posts generally seem to have been suicient to carry the weight of the platform and the roof. However, the situation is different in the southeast part, where doubled post holes are found. Here, the height difference between the ground and the platform was so large that a reinforcement was necessary for static reasons. A second, higher level (storage space) at this point of the house is also possible (Modderman 1970, 110). The presence of ive internal post rows with double or triple post holes, as for instance observed in the Großbau at Mold, shows that the slope was integrated into the structure (Figure 18). This house is also notable for the presence of double and triple posts in the centre; here, there was already a height diference of 0.80 m. Precipitation seems to have markedly increased in the course of the LBK and reached a maximum towards its end (Schmidt et al. 2004). In this case, the wall trench and the wall it supported may have had a protective function: to prevent surface water draining down slope from entering the northwest part and the space beneath the house. At the same time, the existence of a sturdy northwest part would have diminished the risk of the other house posts being destabilised by water. It is notable that towards the end of the LBK – and towards the end of the marked ‘wet phase c’ (Schmidt et al. 2004) – the wall trenches of houses were additionally strengthened. Archaeologically, this is manifest in the appearance of so-called protuberances (e.g. Modderman 1977, 28; Brink-Kloke 1992, 28, 58)22, i.e. additional posts on the outer side of the wall trench, which may have provided increased stability23. Butler and Haberey (1936) suggested a further possibility for the reconstruction of an LBK building by proposing a platform entirely separated from the ground surface. Kleinbauten, too, could have been entirely supported by posts in this way (Figure 20, building in the background). The height above ground level of the dwelling platform can only be guessed at, but the space beneath a platform could only be used in a meaningful way if it was a metre high or more. The advantages of an elevated construction are clear: ground humidity is reduced and a storage space for wood or tools is gained. In addition, freely roaming animals are prevented from entering the structure and pests, too, are more easily kept at bay. There are several possibilities for reconstructing an LBK house. The varied examples of southeast Asia provide vivid parallels (e.g. Young 1974; Condominas 1974; Rousseau 1974; Stirn and van Ham 2000; Hasenbichler and Hanreich 2004). Relections on the durability of Bandkeramik houses On the basis of his analysis of the excavations at Elsoo and Stein, Modderman (1970) assumed a use-life of 25 years for LBK buildings. Stehli reached the same conclusion in 1989. 14C analyses allowed LBK setlement in the Rhineland to be dated to 5300–4950 cal B.C. (Stehli 1989). Hence, the archaeologically established 14 setlement phases of the Aldenhovener Plate lasted a total of 350 years, resulting in an average length of 25 years per phase. These 25–year setlement phases have since been referred to as house generations (Hausgenerationen) and equated with the use-life of the buildings themselves (Stehli 1989). This model has hardly been questioned since and has become well-established in LBK research. One exception is a recent article which atempts to evaluate the durability of LBK structures on the basis of indings from dendrochronology, wood biology, archaeology and historical building studies (Schmidt et al. 2005). The relections and data collected there support the idea that prehistoric farmers and stockherders planned a use-life for their houses which went beyond the otquoted 25 years. The indings are briely summarised in the following: 1. A ield experiment on the durability of wood in diferent soil types, carried out in Great Britain, showed that 5 by 5 cm oak posts possess an average durability of 27 years (Purslow 1976; Purslow and Williams 1978; Smith and Orsler 1996). If this is transferred to LBK house posts, with their diameter of between 20 and 30 cm (e.g. Stieren 1951; Modderman 1970), we can calculate a use-life of ca. 100–160 years, as the durability of a wooden construction element in the ground is proportional to its diameter (Smith and Orsler 1996). 2. LBK buildings reconstructed in several open-air museums (e.g. Asparn an der Zaya, Austria, in 1970, or Oerlinghausen, Germany, in 1980) show New aspects and models for Bandkeramik setlement research 177 Figure 18: Mold, Lower Austria. House 1. Contour lines with hights above sea level (grey) show a diference of 1.40 m between northern and southern parts of the house, although post holes were equally well preserved throughout. Lenneis 2004, 383, ig. 4. 178 Oliver Rück Figure 19: Sketch of possible reconstruction of a house built on a slope. Drawing by R. Mauss. Figure 20: Garo longhouse, West Garo hills, India. Stirn and van Ham 2000, 62. no traces of decay on their wooden construction elements ater several decades. 3. Economic considerations could support the idea that people atempted to use houses for as long as possible. The efort of building a Bandkeramik house (Startin 1978) is disproportionately larger than that for repairing it or adding parts. Thus, from the point of view of cost-beneit analyses, maintenance and enlargement should always be the preferred options (Lienau 2000). 4. There are now numerous known examples for repair/maintenance and for enlargements (additive building system). 5. The chronological diference between overlapping house plans was calculated for some Aldenhovener Plate sites (Table 2). The average time between a irst and second build on the same spot is around 85 years (Schmidt et al. 2005). There seems to be a trend towards a ‘long’ use-life for LBK buildings (as a rough guess between 75 and 100 years). During his analysis of Elsloo and Stein, Modderman (1970) found only few instances of overlap. On the basis of datable potery from pits near the houses, he suggested a clear chronological diference between the irst and second builds. If the data on the durability of oak in the soil given in Schmidt et al. (2005) is summarised, an average of 100 – 125 years results. The lower Figure of 100 years New aspects and models for Bandkeramik setlement research 179 is chosen as the basis for further calculations. Here, it should be pointed out that the postulated use-life of 25 years (e.g. Stehli 1989; Zimmermann 2003) is itself based on an arithmetical model. The Hofplatzmodell itself provides far fewer arguments for its idea of short-lived houses and the resulting setlement structure than have been listed here in support for a 100–year use-life. Number of inhabitants in Bandkeramik buildings Figure 21: Southeast and central part of an LBK house upon excavation. The scale in the foreground is 2 m long, showing the impressive dimensions of the building. Kuper et al. 1975, frontispiece. The part of an LBK house shown in Figure 21 makes the dimensions of these structures clear. Table 3 provides numerical data on the average area of a house. The efort for constructing such a building is enormous and cannot be handled by only a few people. Where LBK houses are reconstructed, for instance, professional organisations such as the German Technisches Hilfswerk are often asked for support, as they provide the necessary personnel and modern machinery. With this in mind, it is hard to comprehend why the ‘magic number’ of ten inhabitants per house is so rarely exceeded in LBK research. Lüning and Stehli (1989) for instance assume only an average of 6.25 inhabitants per house. Zimmermann (2003) suggests seven and refers to ethnographic observations, but without specifying these more clearly. However, it is precisely ethnographic and historical data which furnish a completely diferent picture. Wherever the longhouse is the primary dwelling unit, there is a much higher Table 2. Time diference between irst and second builds in the same locations at the Aldenhovener Plate sites Langweiler 8, 2 and 16. The duration between rebuilds varies between 25 and 150 years, but most second builds take place 75-100 years ater the irst construction. The gaps of 25 or 50 years suggest that gaps in the setlement structure were closed quickly. Although the table is only based on 11 overlaps, a trend towards a diferent use-life for houses (roughly between 75 and 100 years) is evident. Data from Boelicke et al. (1988), Stehli (1994) and Münch (1999); table from Schmidt et al. (2005, 158, tab. 4). Site LW 8 LW 8 LW 8 LW 8 LW 8 LW 8 LW 8 LW 8 LW 2 LW 2 LW 16 House number older / younger 58 / 80 8 / 87 5 / 108 68 / 70 68 / 57 91 / 70 9 / 87 40 / 39 3/4 15 / 16 2/3 Phase older house VI X V IV IV I XIII II X IX IV Phase younger house X XIV VII VII VIII VII XIV VII XIII XII VI phase difference 4 4 2 3 4 6 1 5 3 3 2 time difference in years 100 100 50 75 100 150 25 125 75 75 50 180 Oliver Rück Table 3. Comparison of Langweiler 8 with 13 sites on the eastern Aldenhovener Plate. Calculations for Langweiler 8 are based on von Brandt (1988). Note that using only the 25 complete house plans at the site (von Brandt 1988, 179, 199) would result in an average area of 140 m2 per house. Thus, the total built-over area may be up to 30% greater than calculated here. The other 13 setlements are treated together; the lower average house area of 87 m2 is due to diferent preservation and recovery conditions. Table from Schmidt et al. (2005, 158, tab. 3). Site Langweiler 8 Langweiler 2 Langweiler 9 Langweiler 16 Niedermerz 4 Laurenzberg 7 Laurenzberg 8 Aldenhoven 3 Lamersdorf 2 Weisweiler 6 Weisweiler 17 Weisweiler 29 Weisweiler 110 Weisweiler 111 Lohn 3 Number of houses 108 Total area of all houses [m²] 11000 Average area per house [m²] 102 230 20000 87 number of inhabitants than is generally postulated for LBK houses. This is as true for Iroquois longhouses with their division into several family compartments (e.g. Warrick 1984) as for the longhouse societies of southeast Asia (e.g. Guidoni 1976; Hirschberg 1988). Meier-Arendt (1979) also assumes a larger number of inhabitants per house. “Given their size, one should assume that these Bandkeramik houses were not only the dwellings of a family or rather clan” (Meier-Arendt 1979, 62). Following these estimates, one could suggest that 30 people or more lived under one roof. Conclusion In sum, in spite of the large number of LBK houses so far excavated, we can still not be certain of the appearance and use of these structures, nor of the way in which they were arranged relative to each other. To some extent, this is due to the over-reliance on very few models, mostly developed for sites in the Rhineland and subsequently uncritically accepted as facts. While immensely fruitful as heuristic devices, they have created a sense of security and familiarity regarding LBK buildings and setlement structure, and this has hindered the search for alternative explanations of our data. The most prominent example is the Hofplatzmodell, hugely inluential in the reconstruction of LBK social structure as a whole, which postulates a short use-life of houses and a large distance between contemporary buildings. However, the criticism also applies to the reconstruction of the above-ground features of houses, where alternative possibilities have so far been litle discussed. The present article has outlined several ways in which these assumptions can be challenged. It is argued on the basis of house repairs and the qualities of sturdy construction timbers that houses may have stood for as long as 100 years, which has profound implications for the structure of setlements and the number of inhabitants at any given time. On several sites, it has been possible to suggest an ordered setlement layout of rows of houses with parallel gable ends. In this model, contemporary houses are relatively close and the traditional idea of independent wards is strongly relativised. In addition, the pronounced slopes on which LBK houses were built, probably to respond to climatic factors, should be taken into account in reconstruction atempts. One possibility is the presence of a raised living platform in parts or all of the building. New aspects and models for Bandkeramik setlement research Notes 1 Architectural variation is clearly relected in the diversity of house plans. 2 A diferentiation of regional groups on the basis of ceramic decoration was already carried out at the beginning of the last century (e.g. Lehner 1912; Koehl 1914; Bremer 1925). 3 The northwest part of this building is also incompletely preserved. Seen in its entirety, the structure shows a remarkable patern of post setings. It seems slightly bent and the northwest part appears to have been trapezoidal. One of the post rows in the central part consists of only two posts, while an unusual post seting also occurs in the southeast. 4 The pit is situated diagonally right in front of the southeast end of the structure and seems to reference the building. Usually, pits associated with the southeast part are at the long sides of a house. For unexplained reasons, in house 2 the interior double posts and the position of the pit have been rotated or ofset by 90°. 5 To facilitate understanding, the indications and structures are here discussed for buildings oriented northwest/ southeast, but they can analogously be transferred to houses with other orientations, e.g. north-south. 6 Socio-economic and social factors are postulated as the reason for diferent sizes and types of structure (Lienau 2000), but they require further discussion. 7 From socio-political and socio-economic points of view, the practical execution of a house enlargement would also be a simple means to react to changes. 8 In the German literature, the term Wohnplatz is used alongside Hofplatz. 9 There is one possible case of overlap. Houses 4 and 5 may partially overly house 19. Due to the bad preservation of house 19, no certain conclusions can be drawn; but given the overall situation, an overlap seems unlikely. 10 Whether consciously or not, this is probably the reason why on the plan for Cuiry-lès-Chaudardes (Figure 7) the house numbers were added parallel to the southeast end of the house. 11 Quite apart from the architectural elaboration, the southeastern or southern end of a house would generally be a favourable place for activities due to its greater exposure to sunlight. 12 A possible rule-of-thumb is that on large LBK sites, typologically distinct buildings are oten ofset, while morphologically similar houses are frequently parallel Bibliography Barthel, H.-J. and Cot, J. 1977. Eine Sumpfschildkröte aus der neolithischen Station Erfurt-Gispersleben. Ausgrabungen und Funde 22, 170–73. Becker, H. and Braasch, O. 1984. Plan einer Siedlung aus der Jungsteinzeit bei Harting, Stadt Regensburg, Oberpfalz. Das Archäologische Jahr in Bayern 1983, 27–30. Boelicke, U., v. Brandt, D., Lüning, J., Stehli, P. and Zimmermann, 13 14 15 16 17 18 19 20 21 22 23 181 and have aligned gable ends. This is very clear at Straubing-Lerchenhaid (Brink-Kloke 1992). This Figure must be interpreted in relation to the average house length at a given site. In metres, depending on the setlement, a row of houses could have been between 25 and 50 m wide. Langweiler 2 and 9, with ive wards each, are among the setlement plans reproduced in Coudart (1993, 129, ig. 15). The term ‘village’ is here used in a general sense to refer to a setlement of several houses and characterised by agricultural social and economic structures. The prerequisite is that sites are carefully and completely excavated, a situation that is becoming ever rarer due to restricted funds. House reconstructions based on ethnographic parallels are so far rare in the archaeological literature (but see Butler and Haberey 1936; Coudart 1998). “[...] what I have since called ‘paradigms’. These I take to be universally recognized scientiic achievements that for a time provide model problems and solutions to a community of practitioners“ (Kuhn 1996, X). Bones of the pond turtle (emys orbicularis) have been retrieved from the following sites (ater Willms 2003, 190): Eilsleben (Döhle 1994), Erfurt-Gispersleben (Barthel and Cot 1977), Straubing-Lerchenhaid (Ziegler 1985/86), Päigen (Stork 1993) and Brześć Kujawski and Strzelce (Bogucki 1982). The natural habitat of the pond turtle requires the constant presence of water and suiciently high summer temperatures for several months, see htp:// www.swissherp.org/Swissreptiles/Emys_orbicularis. html. (accessed 09.03.2008). Given the usual dimensions of the posts of an LBK house (0.25–0.40 m; Stieren 1951, 66, ig. 3; Modderman 1970, plate 23), a multiplicity of reconstructions is possible on static grounds. This Figure was probably even greater in Neolithic times, as the house most likely continued further to the north (Lenneis 2004, 383, ig. 4). Houses with strengthened wall trenches irst appear at the end of the LBK (Modderman 1970, 109). This kind of wall trench construction becomes more frequent in the Middle Neolithic (Großgartach, Stichbandkeramik, Rössen; see Hampel 1989, 81, ig. 67). The statics of load-bearing walls and their stability are discussed in Rapoport (1969, 124, ig. 5.27). A. (eds). 1988. Der bandkeramische Siedlungsplatz Langweiler 8, Gemeinde Aldenhoven, Kr. Düren. Beiträge zur neolithischen Besiedlung der Aldenhovener Plate III. Köln: Rheinland Verlag. Bogucki, P. 1982. Early Neolithic subsistence and setlement in the Polish lowlands. Oxford: BAR. Bradley, R. 2001. Orientations and origins: a symbolic dimension to the longhouse in Neolithic Europe. Antiquity 182 Oliver Rück 75, 50–56. Brandt, D.v. 1988. Häuser. In U. Boelicke, D. v. Brandt, J. Lüning, P. Stehli and A. Zimmermann (eds), Der bandkeramische Siedlungsplatz Langweiler 8, Gemeinde Aldenhoven, Kr. Düren. Beiträge zur Besiedlung der Aldenhovener Plate III, 36–289. Köln: Rheinland Verlag. Bremer, W. 1925. Flomborner Typus. In M. Ebert (ed.), Reallexikon der Vorgeschichte 3, 388–90. Berlin: de Gruyter. Brink-Kloke, H. 1992. Drei Siedlungen der Linienbandkeramik in Niederbayern. Buch am Erlbach: Marie Leidorf. Butler, W. and Haberey, W. 1936. Die bandkeramische Ansiedlung bei Köln-Lindenthal. Mainz: de Gruyter. Condominas, G. 1974. Proto-Indochinesen (Montagnards) in Vietnam und Kambodscha. In E. Evans-Pritchard (ed.), Bild der Völker. Malaiischer Archipel mit Philippinen, Hinterindien. Brockhaus Völkerkunde 6, 152–67. Wiesbaden: Brockhaus. Coudart, A. 1993. De ľusage de ľarchitecture domestique et de ľanthropologie sociale dans ľapproche des sociétés néolithiques: ľexemple du Néolithique danubien. In Le Néolithique du nord-est de la France et des régions limitrophes. Actes du XIIIe colloque interrégional sur le Néolithique 1986, 115–35. Paris: Maison des Sciences de l’Homme. Coudart, A. 1998. Architecture et société néolithique. L’unité et la variance de la maison danubienne. Paris: Maison des Sciences de l’Homme. Döhle, H.-J. 1994. Die linearbandkeramischen Tierknochen von Eilsleben, Bördenkreis. Ein Beitrag zur neolithischen Haustierhaltung und Jagd in Miteleuropa. Halle: Landesamt für Archäologische Denkmalplege Sachsen-Anhalt. Domboróczki, L. 2001. The excavation at FüzesabonyGubakút. Preliminary report. In R. Kertész and J. Makkay (eds), From the Mesolithic to the Neolithic. Proceedings of the International Archaeological Conference held in the Damjanich Museum of Szolnok, September 22–27, 1996, 193–214. Budapest: Archaeolingua. Fraser, D. 1968. Village planning in the primitive world. New York: Studio Vista. Guidoni, E. 1976. Architektur der primitiven Kulturen. Stutgart: Belser. Hampel, A. 1989. Die Hausentwicklung im Mitelneolithikum Zentraleuropas. Bonn: Habelt. Hasenbichler, I. and Hanreich, T. 2004. Pfahlbauten. Papua Neu Guinea. <http://atelier-tms.at/PNG> (accessed 09.03.2008). Hirschberg, W. (ed.). 1988. Neues Wörterbuch der Völkerkunde. Berlin: Reimer. Ilet, M. 1982. The Neolithic of north-eastern France. In C. Scarre (ed.), Ancient France. Neolithic societies and their landscapes, 6–33. Edinburgh: Edinburgh University Press. Kind, C.-J. 1989. Ulm-Eggingen: Die Ausgrabungen 1982–1985 in der bandkeramischen Siedlung und der mitelalterlichen Wüstung. Stutgart: Theiss. Koehl, K. 1914. Ältere und jüngere Spiralmäanderkeramik. Mannus 6, 53f. Koojmans, L., van de Velde, P. and Kamermans, H. 2003. The early Bandkeramik setlement of Geleen-Janskamperveld: its intrasite structure and dynamics. In J. Eckert, U. Eisenhauer and A. Zimmermann (eds), Archäologische Perspektiven; Analysen und Interpretationen im Wandel. Festschrit für Jens Lüning zum 65. Geburtstag, 373–97. Rahden/Westf.: Marie Leidorf. Kreuz, A. 2007. Archaeobotanical considerations on the beginning of agriculture north of the Alps. In S. Colledge, J. Conolly and S. Shennan (eds), In honour of Gordon Hillman. New perspectives on the origins and spread of farming in southwest Asia and Europe. London: University College London Press. Kreuz, A. and Boenke, N. 2003. Zweikörniges Einkorn zur Zeit der Bandkeramik: Archäologisches Indiz oder Laune der Natur? In J. Eckert, U. Eisenhauer and A. Zimmermann (eds), Archäologische Perspektiven; Analysen und Interpretationen im Wandel. Festschrit für Jens Lüning zum 65. Geburtstag, 233–41. Rahden/Westf.: Marie Leidorf. Kruk, J. 1973. Studia osadnicze nad neolitem wyzyn lessowych. Wrocław: Ossolineum. Kuhn, T. S. 1996 [1962]. The structure of scientiic revolutions. 3rd edition. Chicago: University of Chicago Press. Kuper, R., Lüning, J. and Stehli, P. 1975. Bagger und Bandkeramiker. Steinzeitforschung im rheinischen Braunkohlengebiet. Bonn: Schriten des rheinischen Museumsamtes. Kuper, R., Löhr, H., Lüning, J., Stehli, P. and Zimmermann, A. 1977. Der bandkeramische Siedlungsplatz Langweiler 9, Gemeinde Aldenhoven, Kr. Düren. Beiträge zur Besiedlung der Aldenhovener Plate II. Köln: Rheinland Verlag. Lehmann, J. 2004. Die Keramik und Befunde des bandkeramischen Siedlungsplatzes Erkelenz-Kückhoven, Kreis Heinsberg (Grabungskampagnen 1989–1994). In H. Koschik (ed.), Der bandkeramische Siedlungsplatz ErkelenzKückhoven, 1–364. Mainz: Habelt. Lehner, H. 1912. Prähistorische Ansiedlungen bei Plaidt a. d. Nete. Bonner Jahrbücher 122, 271–310. Lenneis, E. 2004. Ein bandkeramischer Großbau aus Mold bei Horn, Niederösterreich. In B. Hänsel (ed.), Zwischen Karpaten und Ägäis. Neolithikum und ältere Bronzezeit. Gedenkschrit für Viera Němejcová Pavúková, 379–93. Rahden/ Westf.: Marie Leidorf. Lienau, C. 2000. Die Siedlungen des ländlichen Raumes. 4th edition. Braunschweig: Westermann. Luley, H. 1992. Urgeschichtlicher Hausbau. Bonn: Habelt. Lüning, J. 1980. So bauten die Zimmersleute in der Steinzeit. Bild der Wissenschat 8, 44–59. Lüning, J. 1982. Siedlung und Siedlungslandschaft in bandkeramischer und rössener Zeit. Ofa 39, 9–33. Lüning, J. and Stehli, P. 1989. Die Bandkeramik in Miteleuropa: Von der Natur- zur Kulturlandschat. In J. Lüning (ed.), Siedlungen der Steinzeit. Haus, Festung und Kult, 110–21. Heidelberg: Spektrum der Wissenschat. Lüning, J. and Stehli, P. (eds). 1994. Die Bandkeramik im Merzbachtal auf der Aldenhovener Plate. Vier bandkeramische Siedlungsplätze im Merzbachtal. Beiträge zur Besiedlung der Aldenhovener Plate V. Köln: Rheinland Verlag. Masuch, A. and Ziessow, K.-H. 1983. Überlegungen zur Rekonstruktion bandkeramischer Häuser. Frühe Bauernkulturen in Niedersachsen. Archäologische Miteilungen aus Nordwestdeutschland, Beih. 1, 229–61. Meier-Arendt, W. 1979. Die Steinzeit in Köln. 2nd edition. Köln: Römisch-Germanisches Museum. Meyer-Christian, W. 1976. Die Y-Pfostenstellung in Häusern New aspects and models for Bandkeramik setlement research der Älteren Linearbandkeramik. Bonner Jahrbücher 176, 1–25. Modderman, P. J. R. 1970. Linearbandkeramik aus Elsloo und Stein. Analecta Praehistorica Leidensia 3. Modderman, P. J. R. 1977. Die neolithische Besiedlung bei Hienheim, Ldkr. Kelheim. Die Ausgrabungen am Weinberg, 1965 bis 1970. Analecta Praehistorica Leidensia 10. Modderman, P. J. R. 1985. Die Bandkeramik im Graetheidegebiet, Niederländisch-Limburg. Bericht der Römisch-Germanischen Kommission 66, 25–121. Modderman, P. J. R. 1988. The Linear Pottery Culture. Diverstity in uniformity. Berichten van de Rjksdienst voor het Oudheidkundig Bodemonderzeok 38, 63–139. Münch, U. 1999. Zur Siedlungsstruktur der Flombornzeit auf der Aldenhovener Plate. Unpublished MA dissertation, Cologne University. Paret, O. 1946. Das neue Bild der Vorgeschichte. Stutgart: Schröder. Posselt, M. and Saile, T. 2003. Durchblick in Diemarden. In J. Eckert, U. Eisenhauer and A. Zimmermann (eds), Archäologische Perspektiven; Analysen und Interpretationen im Wandel. Festschrit für Jens Lüning zum 65. Geburtstag, 309–14. Rahden/Westf.: Marie Leidorf. Purslow, D. F. 1976. Results of ield tests on the natural durability of timber (1932–1975). Aylesbury, Bucks: Building Research Establishment. Purslow, D. F. and Williams, N. A. 1978. Field trials on preserved timber out of ground contact. Aylesbury, Bucks: Building Research Establishment. Rapoport, A. 1969. House form and culture. Englewood Clifs, N.J.: Prentice-Hall. Rousseau, J. 1974. Die Kajan auf Borneo. In E. Evans-Pritchard (ed.), Bild der Völker. Malaiischer Archipel mit Philippinen, Hinterindien. Brockhaus Völkerkunde 6, 124–31. Wiesbaden: Brockhaus. Rück, O. 1999. Zwei Grundrisse aus der bandkeramischen Siedlung Witislingen „Am Wiesenberg“, Lkr. Dillingen, BayerischSchwaben. Unpublished MA dissertation, Tübingen University. Rück, O. 2001. Eine bandkeramische Siedlung bei Utzwingen, Gemeinde Maihingen, Lkr. Donau-Ries, Schwaben. Das Archäologische Jahr in Bayern 2000, 17–19. Rück, O. 2004. Zur Lage bandkeramischer Siedlungsplätze West- und Süddeutschlands – Überlegungen zum Hausbau. Archäologisches Korrespondenzblat 34, 309–19. Rück, O. 2007 . Neue Aspekte und Modelle in der Siedlungsforschung zur Bandkeramik. Die Siedlung Weisweiler 111 auf der Aldenhovener Platte, Kr. Düren. Rahden/Westf.: Marie Leidorf. Schade-Lindig, S. and Schwitalla, G. M. 2003. Die Kreispalisadenanlage des bandkeramischen Zentralortes Bad Nauheim-Nieder Mörlen. In J. Eckert, U. Eisenhauer and A. Zimmermann (eds), Archäologische Perspektiven; Analysen und Interpretationen im Wandel. Festschrit für Jens Lüning zum 65. Geburtstag, 351–58. Rahden/Westf.: Marie Leidorf. Schalich, J. 1977. Der Fundplatz. In R. Kuper, H. Löhr, J. Lüning‚ P. Stehli and A. Zimmermann (eds), Der bandkeramische Siedlungsplatz Langweiler 9, Gemeinde Aldenhoven, Kr. Düren. 183 Beiträge zur Besiedlung der Aldenhovener Plate II, 9–16. Köln: Rheinland Verlag. Schmidt, B., Gruhle, W. and Rück, O. 2004. Klimaextreme in bandkeramischer Zeit (5300 bis 5000 v. Chr.). Interpretation dendrochronologischer und archäologischer Befunde. Archäologisches Korrespondenzblat 34, 303–07. Schmidt, B., Gruhle, W., Rück, O. and Freckmann, K. 2005. Zur Dauerhaftigkeit bandkeramischer Häuser im Rheinland (5300–4950 v. Chr.) – eine Interpretation dendrochronologischer und bauhistorischer Befunde. In D. Gronenborn (ed.), Klimaveränderungen und Kulturwandel in neolithischen Gesellschaten Miteleuropas, 6700–2200 cal. BC, 151–70. Mainz: Monographien des RGZM. Seitz, H. J. 1990. Die Steinzeit in Wittislingen. Augsburg: Veröffentlichungen der Schwäbischen Forschungsgemeinschat. Sielmann, B. 1971. Der Einluss der Umwelt auf die neolithische Besiedlung Südwestdeutschlands. Acta Praehistorica et Archaeologica 2, 65–197. Sielmann, B. 1972. Die frühneolithische Besiedlung Miteleuropas. In H. Schwabedissen (ed.), Die Anfänge des Neolithikums vom Orient bis Nordeuropa Va. Fundmenta A.3., 1–65. Köln: Böhlau. Smith, G. A. and Orsler, R. J. 1996. The biological natural durability of timber in ground contact. London: Building Research Establishment. Startin, W. 1978. Linear Potery Culture houses: reconstruction and manpower. Proceedings of the Prehistoric Society 44, 143–69. Stehli, P. 1989. Zur relativen und absoluten Chronologie der Bandkeramik in Miteleuropa. In J. Rulf (ed.), Bylany Seminar 1987. Collected papers, 69–78. Prague: Archeologický ústav ČSAV. Stehli, P. 1994. Chronologie der Bandkeramik im Merzbachtal. In J. Lüning and P. Stehli (eds), Die Bandkeramik im Merzbachtal auf der Aldenhovener Platte. Beiträge zur Besiedlung der Aldenhovener Platte V, 79–191. Köln: Rheinland Verlag. Stieren, A. 1951. Bandkeramische Großbauten bei Bochum und ihre Parallelen in Miteleuropa. Berichte der RömischGermanischen Kommission 33, 61–88. Stirn, A. 1964. Kalktuffvorkommen und Kalktufftypen der Schwäbischen Alb. Forschungen zur Karst- und Höhlenkunde. Blaubeuren: Mangold. Stirn, A. and van Ham, P. 2000. The Seven Sisters of India. Tribal worlds between Tibet and Burma. Munich: Prestel. Stork, M. 1993. Tierknochenfunde aus neolithischen Gruben in der Gemeinde Ammerbuch, Kr. Tübingen. Zeitschrit für Archäologie 27, 91–104. Thiemeyer, H. 1988. Bodenerosion und holozäne Dellenentwicklung in hessischen Lößgebieten. Frankfurt/Main: Selbstverlag „Rhein-Mainische Forschung“ der Institute für Kulturgeographie und Physische Geographie der JohannWolfgang-Goethe-Universität. Wagner, E. 1995. Cannstat I. Großwildjäger im Travertingebiet. Stutgart: Theiss. Warrick, G. 1984. Reconstructing Ontario Iroquoian village organization. Otawa: Archaeological Survey of Canada. Waterbolk, H. T. 1958/59. Die bandkeramische Siedlung von 184 Oliver Rück Geelen. Palaeohistoria VI/VII, 122–61. Waterbolk, H. T. and Modderman, P. J. R. 1958/59. Die Großbauten der Bandkeramik. Palaeohistoria VI/VII, 163–71. Weiner, J. 1995. Eine zimmermannstechnische Glanzleistung: der 7000 Jahre alte Eichenholzbrunnen aus ErkelenzKückhoven. In H.G. Horn, H.G. Hellenkemper, H. Koschik and B. Trier (eds), Ein Land macht Geschichte. Archäologie in Nordrhein-Westfalen, 179–87. Mainz: von Zabern. Willms, C. 2003. Löwe, Elch und Schildkröte: Kein jungsteinzeitliches Märchen. In J. Eckert, U. Eisenhauer and A. Zimmermann (eds), Archäologische Perspektiven; Analysen und Interpretationen im Wandel. Festschrit für Jens Lüning zum 65. Geburtstag, 181–94. Rahden/Westf.: Marie Leidorf. Young, M. W. 1974. Die Bewohner der Trobriand-Inseln. In A. Forge (ed.), Bild der Völker. Australien und Melanesien, Polynesien und Mikronesien. Brockhaus Völkerkunde 1, 100–05. Wiesbaden: Brockhaus. Zeeb-Lanz, A. 2003. Das Ries und seine steinzeitlichen B e wo h n e r. B e o b a c h t u n g e n z u r n e o l i t h i s c h e n Besiedlungsentwicklung in einer Mikroregion. In J. Eckert, U. Eisenhauer and A. Zimmermann (eds), Archäologische Perspektiven; Analysen und Interpretationen im Wandel. Festschrit für Jens Lüning zum 65. Geburtstag, 293–306. Rahden/Westf.: Marie Leidorf. Ziegler, R. 1985/86. Neolithische Tierreste aus StraubingLerchenhaid, Niederbayern. Bericht der Bayerischen Bodendenkmalplege 26/27, 7–32. Zimmermann, A. 2003. Landschaftsarchäologie I: Die Bandkeramik auf der Aldenhovener Plate. Berichte der Römisch-Germanischen Kommission 83, 17–38.