ONCE MORE THE PAST IN THE PAST

The role of the past in various cultures has been studied intensively in anthropology, sociology, history, and psychology (eg, Silverman Reference Silverman2002; Artelius Reference Artelius2013, 23). The research topic was introduced in British archaeology in the 1980s (Evans Reference Evans1985; Bradley Reference Bradley1987). Its popularity increased rapidly. Here we introduce some approaches, concepts, and trends that are useful for the present research. Many concepts used in archaeology originate in other disciplines.

Past societies did not live in ‘empty’ or ‘untouched’ spaces, but in landscapes of ancient sites that were still meaningful (Holtorf Reference Holtorf1998, 31). Evans states that each culture creates its own context, which is reflected in the respective abandonment, reoccupation, veneration, or desecration of earlier sites (1985, 85). Some practices left traces in the archaeological record, which can be diverse. Most obvious are juxtaposed relics from differences phases. Bradley (Reference Bradley1987, 1) for example mentions Iron Age hillforts near Neolithic causewayed enclosures, Romano-Celtic temples within hillforts and churches built near Roman buildings. Williams (Reference Williams1997, 6–7) notes that Early Medieval burials are associated with various older site types.

Another type of research focuses on the reuse of ancient material culture. Examples are the copying of Neolithic pottery decorations in Iron Age Scotland (Hingley Reference Hingley1996) and the reuse of Roman objects as Early Medieval grave gifts (eg, Hunter Reference Hunter1974; White Reference White1990; Eckart & Williams Reference Eckart and Williams2003). Other studies integrate archaeological phenomena and historical, folkloric, or toponymical data. Examples are links between place names and archaeological sites (Gelling Reference Gelling1978; Meaney Reference Meaney1995), Christian legislation on ‘heathen’ burial practices (eg, Von Uslar Reference Von Uslar1972; Sippel Reference Sippel1980) and the reuse of barrows as execution or exhibition sites (Reynolds Reference Reynolds1997; Meurkens Reference Meurkens2007).

A survey of publications on the perception and reuse of cemeteries demonstrates regional differences in research history. Research in Britain has been extensive. Most influential theories and concepts originate here. Studies on reuse practices in the Anglo-Saxon period are particularly abundant (Van de Noort Reference Van de Noort1993; Thäte Reference Thäte1996; Williams Reference Williams1997; Reference Williams1998; Semple Reference Semple1998). Late prehistory and the Roman period have received less attention (but see Bradley Reference Bradley1987; Hingley Reference Hingley1996). In Scandinavia, the research topic has recently become popular (Pedersen Reference Pedersen2006; Thäte Reference Thäte2007; Wickholm Reference Wickholm2008; Artelius Reference Artelius2013). The emphasis is on the 1st millennium ad (but see Parker Pearson Reference Parker Pearson1993). Recent Danish studies demonstrate how barrows structured later (prehistoric) movement (eg, Holst & Rasmussen Reference Holst and Rasmussen2013; Løvschal Reference Løvschal2013). Studies on Early Medieval Germany are available as well (Thäte Reference Thäte1996; Härke & Williams Reference Härke and Williams1997). However, this country stands out in that both Holtorf (Reference Holtorf1998, on megaliths in Mecklenburg-Vorpommern) and Sopp (Reference Sopp1999, on different types of burial sites in northern Germany) published regional overviews of the long-term history of prehistoric burial sites (cf. Bradley Reference Bradley2002, 124–30). Studies on this spatio-temporal scale are lacking elsewhere. Most analyses focus on single sites, smaller areas or specific time frames. Some publications discuss other parts of western (France: Patton Reference Patton1996; Zadora-Rio Reference Zadora-Rio2003, 8; Low Countries: next section) and central Europe (Chapman Reference Chapman1997), but these regions have been less influential in international debates.

Several authors attempt to explain the widespread reuse of cemeteries. Some suggest the practice had various – spatially and temporally diverse – functions (Thäte Reference Thäte1996, 224; Williams Reference Williams1997, 17–22; Holtorf Reference Holtorf1998, 34). Nevertheless, most explanations share many similarities. Minor differences aside, they can be reduced to just a few interpretations that are partly interrelated and not mutually exclusive. All agree that something could be gained from the forging of links with ancient remains (Gerritsen Reference Gerritsen2003, 145). The creation of these links are generally seen as deliberate acts, and more specifically as intentional manipulations of the distant past (eg, Bradley Reference Bradley1987, 4–5; Silverman Reference Silverman2002, 5). Parker Pearson (2003) speaks of ‘the powerful dead’ to indicate that burials could be used in strategic ways by the living. Another common idea is that the ‘meaning’ of monuments was not static but continually reworked and redefined (eg, Lowenthal Reference Lowenthal1985; Bradley Reference Bradley1987; Williams Reference Williams1997, 25; 1998, 95, 100).

The strategic site location of burials near older ones is frequently interpreted as an expression of political authority, serving to legitimise, confirm, or strengthen power structures (eg, Evans Reference Evans1985, 88–9; Lowenthal Reference Lowenthal1985, 197–200, 325; Bradley Reference Bradley1987, 4–5; Lucy Reference Lucy1992). Many researchers suggest that the practice communicated the socio-cultural identity of a group or lineage (Hingley Reference Hingley1996, 241; Williams Reference Williams1997, 24–5; Parker Pearson Reference Parker Pearson1993, 94–5; Artelius Reference Artelius2013, 21, 37). Both theories are often combined, and suggest that these sites served as territorial markers (Evans Reference Evans1985, 88; Williams Reference Williams1997, 24–5; Artelius Reference Artelius2013, 35–7). The creation of funerary links does not necessarily imply direct genealogical ties with the people buried previously at these locations. Various authors agree that fictitious links were created between the present and the ‘mythical’ past (Evans Reference Evans1985, 90; Bradley Reference Bradley1987, 4; Gosden & Lock Reference Gosden and Lock1998). Specific groups may have justified their political authority by claiming to descend from mythical ancestors, heroes or gods (Hunter Reference Hunter1974; Bradley Reference Bradley1987, 4; Lucy Reference Lucy1992, 95; Spencer Reference Spencer1995).

Various archaeological studies reference the concept of ‘invented tradition’ that was introduced by the historian Hobsbawm (Evans Reference Evans1985, 89; Bradley Reference Bradley1987, 3–4; Hingley Reference Hingley1996, 240; Holtorf Reference Holtorf1998, 32). It is defined as a ‘set of practices, normally governed by overtly or tacitly accepted rules and of a ritual or symbolic nature, which seek to inculcate certain values and norms of behaviour by repetition, which automatically implies continuity with the past’ (Hobsbawm Reference Hobsbawm1983, 1). This suggested continuity is largely fictitious, and served to structure and stabilise social life in a constantly changing world (ibid., 2). Accordingly, the reuse of ancient cemeteries may be a response to social change in periods of stress (Evans Reference Evans1985, 89; Lucy Reference Lucy1992, 95; Holtorf Reference Holtorf1998, 32). This hypothesis is especially encountered in studies exploring relationships between reuse and Christianisation (Van de Noort Reference Van de Noort1993; Thäte Reference Thäte1996, 114; Härke & Williams Reference Härke and Williams1997, 25).

MONUMENT REUSE IN THE LOW COUNTRIES: PREVIOUS RESEARCH

As mentioned, research into the perception of late prehistoric monuments in the Low Countries has mainly focused on the MDS region of the southern Netherlands and north-eastern Belgium. Publications on other regions are either site-based or only briefly touch upon the subject (cf. Gerritsen Reference Gerritsen2007, 338). As an exception to this rule, Vermeulen and Bourgeois (Reference Vermeulen and Bourgeois2000) studied the relationships between late prehistoric burial sites and Roman-period habitation and burial in the region of ‘sandy Flanders’. They mainly focused on matters of continuity rather than actual reuse practices, but the sites they list form valuable input for the inventory of reused sites in Flanders assembled for this study.

The discussion in the MDS region exclusively concentrates on the perception of Late Neolithic (2850–2000 cal bc), Early (2000–1800 cal bc) and Middle Bronze Age (1800–1050 cal bc) barrows and Late Bronze Age (1050–800 cal bc) and Early Iron Age (800–500 cal bc) urnfields. Contrary to many other parts of Europe, other long-standing prehistoric monuments (eg, long barrows, megaliths) are lacking.

The first influential study was published by Roymans (Reference Roymans1995). Using archaeological, historical, and folkloric data, he argues that urnfields assumed a prominent but dynamic position in the ‘mythical geography’ of later communities. Fontijn widened the scope of the discussion by noting that the site location of various urnfields was determined by older barrows. Theunissen (Reference Theunissen1999, 102–3; Bronze and Early Iron Age), Gerritsen (Reference Gerritsen2003, 140–5; 2007; mainly Iron Age), and Hiddink (Reference Hiddink2003, 45–52; Roman period) address reuse patterns in specific times frames. All agree that the perception of ancient cemeteries was variable. Gerritsen notes that ‘ancestors do not seem to have formed a stable category in the socio-cosmic order of local communities’ (2003, 150).

The published reuse trends allow a general distinction in five phases to be drawn upFootnote ¹ :

1. 1. Late Bronze Age (1050–800 cal bc) and Early Iron Age (800–500 cal bc);

2. 2. Middle (500–250 cal bc) and Late Iron Age (250 cal BC–58/12 bc);

3. 3. Roman period (58/12 bc–ad 450);

4. 4. Early Middle Ages, divided into a Merovingian (ad 450–750) and Carolingian (ad 750–900) phase;

5. 5. High (ad 900–1300) and Late Middle Ages (ad 1300–1500).

The interpretations and concepts in the MDS studies are largely similar to those in other parts of north-western Europe. In the Late Bronze Age and Early Iron Age, reuse of older barrows was a common and diverse practice (Theunissen Reference Theunissen1999, 102–3; Gerritsen Reference Gerritsen2003, 140–5; 2007, 346). The main types of reuse are:

1. 1. interring (urned) cremations in an existing barrow;

2. 2. digging of a ring ditch around the top or base of an existing barrow;

3. 3. incorporation of a barrow in its entirety into a new mound;

4. 4. burying near or around older barrows;

5. 5. incorporating barrows in orientation lines or infrastructure. This is done by raising post alignments, establishing paths, erecting monuments in between existing ones, or orientating field boundaries towards them.

It is generally assumed that reuse for burial had a mythical rather than a genealogical background (Fontijn Reference Fontijn1996, 79–80). According to Gerritsen ancient monuments were important in the self-definition of local communities, and reuse practices were part of a legitimation strategy with respect to land claims (2003, 142–4; 2007, 341). Additionally, pre-existing sites came from an unknown past and possibly had to be ‘familiarised’ to be incorporated in the contemporary cultural landscape (Gerritsen Reference Gerritsen2007, 351–2). At some northern Dutch sites the morphology of field systems is partly determined by older barrows, but this pattern has not been documented in the MDS region (ibid., 348).

The transitional phase from Early to Middle Iron Age brought important societal changes (Roymans & Gerritsen Reference Roymans and Gerritsen2002; Gerritsen Reference Gerritsen2003). Most urnfields were abandoned. They were replaced by small, inconspicuous burial sites. Also, the interaction with older burial sites changed (Fontijn Reference Fontijn1996, 84; Gerritsen Reference Gerritsen2003, 145). Evidence for a positive appraisal is lacking. Parts of some urnfields were overbuilt (Fig. 2), or – more frequently – farmsteads were constructed close to burial monuments (Gerritsen Reference Gerritsen2003, 148, 197). Older cemeteries may not have been used to support territorial claims, legitimate authority, or construct social identities in this phase. This function may have been taken over by settlements, arable land and cult sites (ibid., 189).

Fig. 2 The Late Bronze Age urnfield of Hilvarenbeek-Laag Spul (light grey), situated in the Dutch part of the MDS region, was overbuilt by a settlement in the Middle and Late Iron Age (dark grey and black). After Gerritsen Reference Gerritsen2003 (fig. 4.30, 196)

The studies on the Roman period and Middle Ages focus on the perception of urnfields. In the Roman period, when the MDS region was incorporated in the Roman Empire, various cemeteries were linked to urnfields. Roymans concludes that these were valued positively and fully incorporated in the cultural landscape (1995, 9). Gerritsen (Reference Gerritsen2003, 197) and Hiddink (Reference Hiddink2003, 45–52) largely agree, but observe more differentiation. A few urnfields were partly disturbed by settlements. Hiddink suggests that these acts might have served to end territorial claims (2003, 52). The simultaneous occurrence of ‘disturbance’ of urnfields and their reuse as cemeteries is still poorly understood.

Spatial relations between Roman-period cemeteries and urnfields vary (Hiddink Reference Hiddink2003, 47–48). They may be situated adjacent to each other or overlap. In the latter case, burials are placed in open spaces or within older monuments. One urnfield is even fully ‘overbuilt’ by younger burials. The offered explanations for these practices are familiar: they may have served to create mythical ancestors, in order to substantiate territorial claims (Hiddink Reference Hiddink2003, 45).

The known Merovingian cemeteries are not linked to urnfields. However, some Merovingian burials occur in or near older barrows (Roymans Reference Roymans1995, 9). Also, a few Early Medieval cemeteries overlap with Roman-period burial sites (Hiddink Reference Hiddink2003, 48–9). Nevertheless, the reuse frequency of older sites is far lower than before. The MDS region was Christianised in the late Merovingian period. It is assumed that from this moment (c. ad 700) onwards, cemeteries moved towards churches (Roymans Reference Roymans1995, 9). This gradual process is still poorly documented in the Low Countries. There is no evidence for deliberate levelling of urnfields in the Carolingian period. This may indicate they were respected, or at least tolerated, in cultivated areas (ibid., 10).

Roymans describes the High and Late Middle Ages as a period of drastic reinterpretation and reorganisation of the landscape. A basic landscape structure developed into two major zones: an ‘inner’ and an ‘outer’ circle (Fig. 3; Roymans Reference Roymans1995, 18–19). The inner circle consisted of the part of the landscape shaped and cultivated by man, with church and cemetery at the centre. The outer circle is formed by the uncultivated parts of the landscape, mainly heathlands and bogs, and had negative connotations. Prehistoric cemeteries were systematically erased from the cultivated inner circle, through a ‘ritual purification’ of pagan elements. Economic motives played only a secondary role as a driving force (ibid., 10–12). Roymans’ arguments are based on historical and folkloric evidence, combined with archaeological data on some leveled urnfields.

Fig. 3 Ideological ordering of the Medieval landscape into an inner and outer circle, based on folkloric evidence from the MDS region. After Roymans Reference Roymans1995 (fig. 13, 18)

FLANDERS: LANDSCAPE AND RESEARCH HISTORY

Flanders is divided in two parts by the river Scheldt. During the Bronze Age the river flowed in a northern direction to the North Sea. Its course corresponded roughly to the modern eastern Scheldt (Vos & Van Heeringen 1997, 60–1). The largest part of the research area consists of Pleistocene coversands belonging to the Lower Rhine Plain (Ameryckx et al. Reference Ameryckx, Verheye and Vermeire1985, 243–8). The southernmost part of Flanders forms a transitional zone to heavier soils. Here, the region east of the Scheldt is covered by loess soils that continue into northern France (Sommé Reference Sommé1973, 94–5; Reference Sommé1976). The region west of the river is characterised by loamy soils that, in a southerly direction, gradually transition into loess. In the coastal parts of Flanders heavy clays occur.

The research history of late prehistoric burial sites in Flanders differs regionally. As will be explained below, we can distinguish a western (provinces of West and East Flanders) and an eastern region (provinces of Antwerp, Limburg, and Flemish Brabant; Fig. 1). Also, the level of knowledge about barrows (Late Neolithic–Middle Bronze Age) and urnfields (Late Bronze Age–Early Iron Age) differs. The site categories are not mutually exclusive. Barrows still occur in urnfield contexts, but these monuments are generally smaller and lower than before. Unless stated otherwise, the term ‘barrow’ is used here for monuments pre-dating the Late Bronze Age. Barrows occur isolated or in clusters, frequently displaying linear alignments. Based on Dutch research (Lohof Reference Lohof1991, 225; Theunissen Reference Theunissen1999, 35–6) it is assumed that only part of the population was buried in a barrow. The relative importance of cremation and inhumation probably fluctuated across space and time, but poor preservation conditions hinder detailed reconstructions (see below).

Urnfields are considered to be collective cremation cemeteries. They vary considerably in size (up to a few hundred burials) and life span (eg, Desittere Reference Desittere1968; De Mulder Reference De Mulder2011; Van Beek & Louwen Reference Van Beek and Louwen2013).

In the western part of Flanders, a few barrows were excavated in the early 19th century in the Flemish Ardennes near Ronse (Fourny Reference Fourny1985), and one after World War 2 at Ruien (De Laet & Roosens Reference De Laet and Roosens1952). Besides these visible examples, no barrows were known until the start of a long-term aerial photography project in the 1980s (Ampe et al. Reference Ampe, Bourgeois, Crombé, Fockedey, Langohr, Meganck, Semey, Van Strydonck and Verlaeckt1996; De Reu Reference De Reu2012). Over a thousand levelled barrows were discovered, that are recognised by the soilmarks of their ditched peripheral structures. We will return to the reasons behind this large-scale destruction later. Over 40 excavations in this region have yielded information on barrows (De Reu Reference De Reu2012, 68–91). This number is rapidly increasing. The number of known urnfields is much lower (De Mulder Reference De Mulder2011, 52–65). They are far less likely to be discovered by aerial photography as not all burials have (ring) ditches. Those that do are much less monumental than before. Most urnfields were discovered long before the 1980s, for example during sand extraction. Our understanding of these sites is relatively poor; hardly any have been excavated completely.

Research in the eastern part of Flanders followed a different path. In the provinces of Antwerp and the northern part of Limburg, both part of the MDS region, many more barrows survived. These monuments, often situated on (former) heathland, attracted early archaeological attention. Various barrows and urnfields have been excavated in the Campine region in particular (eg, Van Impe Reference Van Impe1976a; Theunissen Reference Theunissen1999, 42–6; De Mulder Reference De Mulder2011, 66–99).

The geographical differences in preservation and research history significantly influence the composition of our dataset. Due to the poor preservation conditions specific types of reuse cannot be demonstrated in the western region. This goes for any activity involving the mound body (eg, secondary burials). In fact, hardly any primary burials have been preserved, which very often gives rise to dating problems. Conversely, most excavation data are quite recent and frequently large-scale. Data on the eastern region also have their limitations. The quality of early barrow excavations is generally low. Research rarely involved the environment of monuments, implying that potential archaeological features in their vicinity were overlooked. As a result, detailed and reliable comparisons of reuse patterns in different parts of Flanders are hampered.

REUSE PATTERNS IN FLANDERS

So far, 62 barrow sites and 13 urnfields in Flanders have yielded information on reuse of one type or another (Figs 4 & 5; Appendices A & B). The applied selection criteria are that younger archaeological features are within a distance of 100 m from the burial monuments they ‘interact’ with, and that they are dated precisely. Various sites did not make the final inventory, as the exact date, position, or character of younger features was not fully clear. Nevertheless, almost all well-excavated late prehistoric burial sites yield evidence for later activity. As discussed above, the inventory is biased by differences in preservation and research history. Also, it is often difficult to assess how long monuments remained visible. At some sites settlement features from different phases are present. If a barrow was levelled in Carolingian times, this would imply that High Medieval settlement features at the same site are not meaningful for this study. However, the levelling of monuments does not always leave archaeological traces – let alone datable ones.

Fig. 4 Overview of all late prehistoric barrow sites (dots) and urnfields (stars) in Flanders yielding evidence for any kind of reuse between late prehistory and High Middle Ages. The site numbers refer to Appendices A and B

Fig. 5a Distribution pattern of documented reuse practices in Flanders in late prehistory and the Roman period

Fig. 5b Distribution pattern of documented reuse practices in Flanders in the Early and High Middle Ages

Ideally, younger features would reflect a recognisable attitude towards ancient relics. This is not always the case however. Placing burials in or near a barrow clearly represents a deliberate interaction. However, the presence of settlement features within a 100 m radius is less unambiguous. Is their location intentional or coincidental? We can only attempt to tackle this problem by identifying and analysing recurrent patterns and trends (Figs 5 & 6). When searching for general patterning in reuse, the sometimes complex sequences at individual sites should not be overlooked (Bradley Reference Bradley2002, 127). Therefore, we combine both. We first discuss the general trends represented in the inventory, and then give a concise summary of the finds at six informative sites that have been investigated since 2000. The latter highlight the large diversity in life histories of late prehistoric cemeteries in Flanders.

Fig. 6 Diachronic trends in reuse of barrow sites (left) and urnfields (right) in Flanders between the Late Bronze Age and High Middle Ages. Various sites yield evidence for different types of reuse, often in different phases (see also Appendices A & B). Reuse types: 1. Secondary burial in mound body or ring-ditch; 2. Redigging of existing ring-ditch; 3. Digging of new ring-ditch and/or raising of existing barrow; 4. New burials in vicinity; 5. Square monuments in vicinity; 6. Pottery deposition; 7. Gallo-Roman temple; 8. Settlement features

Middle and Late Iron Age

Both barrow cemeteries and urnfields have yielded evidence for reuse in the Middle and Late Iron Age. The picture that emerges is more diverse and complex than expected based on (previous) MDS studies.

Middle and Late Iron Age cremation burials have been found at six or seven urnfields. The sites of Destelbergen-Eenbeekeinde and Neerharen-Rekem were probably used continuously until the Middle Iron Age (De Mulder Reference De Mulder2011, 167–70, 188–9). Therefore these practices do not formally classify as reuse. However, they do indicate that the transition from Early to Middle Iron Age did not represent an abrupt break. That some urnfields remained meaningful places can also be deduced from incidental later Iron Age burials at Kontich-Duffelsesteenweg (De Mulder Reference De Mulder2011, 171–3) and (again) Destelbergen. Such finds are probably under-represented, as these burials are inconspicuous: they often lack grave gifts and ditched structures. Barrow sites were reused for burial as well. Examples are less numerous than in the previous phase. Besides the site of Ursel (Bourgeois et al. Reference Bourgeois, Semey and Vanmoerkerke1989; two cremations in a ring-ditch), no secondary burials in mound bodies are documented. At the site of Merksplas-Heizijde a complete pottery bowl and ditch were buried near a barrow in the Middle Iron Age (Gheysen & De Mulder Reference Gheysen and De Mulder2010).

A conspicuous type of reuse is formed by square ditched structures, which have been found at nine or ten barrow sites. Dating evidence mainly points to an origin between the Middle Iron Age and the 1st century ad, but a younger (Roman-period) date cannot always be discounted. According to their morphology and relation to older structures, they can be classified in three groups. The first consists of small monuments erected near barrows. Two examples are found at the site of Oedelem-Wulfsberge (Bourgeois et al. Reference Bourgeois, Cherretté and Meganck2001; Cherretté & Bourgeois Reference Cherretté and Bourgeois2002; Reference Cherretté and Bourgeois2003; see below and Fig. 7). The second group consists of the sites of Ursel-Rozestraat (Bourgeois et al. Reference Bourgeois, Semey and Vanmoerkerke1989) and Dendermonde-Hoogveld J (Vandecatsye & Laisnez Reference Vandecatsye and Laisnez2009; Vandecatsye Reference Vandecatsye2010; see below and Fig. 9). Here, Bronze Age barrows are enclosed by large square structures dated to the Middle or Late Iron Age. These are accompanied by (partly) contemporary cremation cemeteries. Thirdly, at Aalter-Woestijne an impressive palisaded ditch system incorporating two square structures was built at a barrow site in the Middle and Late Iron Age (Van de Vijver et al. Reference Van de Vijver, Keppens and Vandendriessche2013a; see below and Fig. 8). These indicate large-scale landscape reorganisation, without disturbing the ancient funerary monuments.

Fig. 7 Schematic overview of the most important features at the site of Oedelem-Wulfsberge (prov. West Flanders). After Cherretté & Bourgeois Reference Cherretté and Bourgeois2003 (fig. 1, 34)

Fig. 8 Schematic overview of the most important features at the site of Aalter-Woestijne (prov. East Flanders). High and Late Medieval settlement features are not indicated. Modified after Van de Vijver et al. Reference Van de Vijver, Keppens and Vandendriessche2013a (fig. 2, 5) & 2013b (fig. 1, 153)

All of these practices indicate that some barrows and urnfields remained important elements in the Middle and Late Iron Age cultural landscape. They were revisited, added to, or even completely restructured. However, Middle and Late Iron Age settlement features occur frequently at these sites as well. They vary from a few dispersed pits to complete farmsteads. At some sites, such as Deinze-RWZI (De Clercq & Van Strydonck Reference De Clercq and Van Strydonck2002) and Gent-Hogeweg (Dyselinck Reference Dyselinck2013), the farmhouses appear to respect the barrows. Conversely, part of the urnfield of Herk-de-Stad-Donk was overbuilt during the Middle and Late Iron Age (Van Impe Reference Van Impe1983; Gerritsen Reference Gerritsen2003, 148; De Mulder Reference De Mulder2011, 441–2). The same happened to the cemetery of Wijnegem-Blikstraat in the Late Iron Age and Roman period (De Mulder Reference De Mulder2011, 181–4).

KEY SITES IN FLANDERS

Dendermonde-Hoogveld J (prov. East Flanders)

At the site of Dendermonde-Hoogveld J (Fig. 9) an Early or Middle Bronze Age barrow was excavated (Vandecatsye & Laisnez Reference Vandecatsye and Laisnez2009; Vandecatsye Reference Vandecatsye2010). An Early Iron Age urn burial was found to the west. At some point between the 4th and 1st century bc two connected square ditches were dug, one of which encloses the barrow. A fourth use phase consists of nine Late Iron Age or (Early) Roman-period cremation burials. Three water wells and five small buildings may date to the same period. It is not clear whether they reflect a settlement, or other activities linked to the barrow and squares. An almost complete pot was deposited in a post-hole of one of the buildings. The life history of the Dendermonde site shows similarities with the well-known multi-period site of Ursel-Rozestraat (Bourgeois et al. Reference Bourgeois, Semey and Vanmoerkerke1989; Gerritsen Reference Gerritsen2003, 152–5). Here, a double-ditched Bronze Age barrow was enclosed by a square ditched monument in the final century bc or the first century ad. A second rectangular enclosure appeared slightly further north. In the same period a cremation cemetery developed. Some burials are found within the squares and in the outer ring-ditch of the barrow.

Fig. 9 Schematic overview of the most important features at the site of Dendermonde-Hoogveld J. After Vandecatsye and Laisnez Reference Vandecatsye and Laisnez2009 (appx 5, map 1)

Beerse-Krommenhof (prov. Antwerp)

At Beerse-Krommenhof (Fig. 10) a Late Neolithic and Bronze Age barrow cluster was excavated, consisting of at least ten monuments (De Smaele et al. Reference De Smaele, Delaruelle, Thijs, Verdegem, Scheltjens and Van Doninck2011; Delaruelle et al. Reference Delaruelle, Annaert, De Smaele, Thijs, Verdegem, Scheltjens and Van Doninck2012). The cemetery was reused in the Merovingian period, between c. ad 600 and 775. Most of the 36 inhumation burials are situated in barrows, or right beside them. The site was settled between the Late Merovingian period and the High Middle Ages. The oldest settlement features are contemporary with the youngest burials. The cessation of burial practices at the start of the Carolingian period reflects a change in attitude towards the late prehistoric monuments, or at least a shift in the practices that were used. However, there is no clear evidence that the barrows were levelled or disturbed at this point. The buildings rather appear to respect them.

Fig. 10 Schematic overview of the most important features at the site of Beerse-Krommenhof. After De Smaele et al. Reference De Smaele, Delaruelle, Thijs, Verdegem, Scheltjens and Van Doninck2011, Delaruelle et al. Reference Delaruelle, Annaert, De Smaele, Thijs, Verdegem, Scheltjens and Van Doninck2012 and excavation maps provided by S. Delaruelle, AdAk (Archeologische Dienst Antwerpse Kempen)

Edegem-Buizegem (prov. Antwerp)

At the site of Edegem-Buizegem (Fig. 11) a Late Neolithic barrow with palisaded ring-ditch was found (Vandevelde et al. Reference Vandevelde, Annaert, Lentacker, Ervynck and Vandenbruaene2007). Later, this barrow was fully incorporated in an impressive barrow with a ring-ditch measuring 53–54 m. This monument possibly dates to the Early Iron Age, based on striking similarities with the Dutch site of Oss-Vorstengraf. Here a Bronze Age barrow was incorporated in an Early Iron Age monument with a diameter of 53 m containing an elite burial (Fokkens et al. Reference Fokkens, Van der Vaart, Fontijn, Lemmers, Jansen, Van Wijk and Valentijn2012). In the Low Countries, this particular reuse practice has only been documented at these two sites. Unfortunately, no central burial was found at Edegem-Buizegem, probably due to poor preservation conditions. It is not exactly clear how long this monument remained visible in the landscape. Some Medieval settlement features have been found, the oldest of which probably date to the 9th century ad. A water well is situated just inside the large ring-ditch, and some features of a possible farmhouse cut it. They indicate the presence of a settlement. Probably around the 10th century a parochial church was founded just east of the excavation area. Its churchyard is roughly dated between the 10th and 12th centuries. Various burials are found immediately south of the large barrow. Some of them cut its ring-ditch. We cannot be sure, but it seems plausible that the site location of the church and churchyard were influenced by the history of this place.

Fig. 11 Schematic overview of the most important features at the site of Edegem-Buizegem. Modified after Annaert et al. 2007 (fig. 44, 55)

SUMMARY AND DISCUSSION

This paper aimed to reconstruct diachronic trends in the reuse of late prehistoric barrow sites and urnfields in Flanders. The study differs methodologically from others in North-west Europe in that it combines an evidence-based overview of regional diachronic trends (Figs 5 & 6, Appendices A & B) with a discussion of a select group of sites yielding high-quality data. It has been demonstrated that this type of research is indispensible in obtaining detailed images of the varied, complex, and ambiguous life histories of cemeteries on different spatial and temporal scales. Almost all barrow sites in Flanders have yielded information on reuse of one type or another, indicating that they were interacted with frequently – rather than being remote, forgotten, or restricted places. So far, there are no clear indications that barrows were perceived other than urnfields in later phases.

To some extent, the documented trends are comparable to those in the MDS region. This is not surprising: both regions overlap and previous studies have partly used the same data. However, there are differences as well. Some are due to significant differences in preservation and research history, but various others reflect actual differences.

The Late Bronze Age and Early Iron Age are characterised by a wide variety of reuse practices. Most have clear funerary connotations. Both sites and individual monuments are reused. Actual genealogical ties with the ‘ancestors’ buried at these cemeteries can never be proven. These ‘fictitious links’ and ‘invented traditions’ keep occurring throughout late prehistory and early history. The occurrence of (mainly Early Iron Age) settlement features at barrow sites has not been described before. Evidence for levelling is lacking, but it certainly was no taboo to settle in close proximity to ancient monuments.

The Middle and Late Iron Age patterns are more complex than expected. Some funerary monuments were overbuilt, which may indicate a negative appraisal, but reuse for burial was by no means uncommon. Therefore the hypothesis that the focal role of burial sites was taken over by settlements, cult sites, and arable land (Gerritsen Reference Gerritsen2003, 189) at the very least requires nuancing in the case of Flanders. Various sites still attracted specific, non-domestic activities. Most of these seem to have focused on the burial sites rather than on individual barrows. The same goes for the Roman and Merovingian periods as well. This might indicate that specific places had become more important than individual monuments.

The various types of square monuments appearing at older burial sites in the Middle and Late Iron Age are hardly known from the MDS region (except Ursel). At least some of them are used for burial. Evidence for this is lacking at others, possibly due to preservation conditions. However, their morphological diversity may indicate that they fulfilled different functions. They exhibit many resemblances with the variety of square ‘cult sites’ documented across north-western Europe (eg, Smith Reference Smith2001; Fontijn Reference Fontijn2002; Groenewoudt Reference Groenewoudt2011; Bradley 2011). The relationships between both categories deserve a more thorough investigation in future research. For now, these monuments clearly indicate that older cemeteries were reused, added to and even drastically modified (Aalter-Woestijne) in the Middle and Late Iron Age. Parallels for the remarkable Aalter-Woestijne site are hard to find, but at the southern Dutch site of Itteren-Emmaus two large, interconnected square monuments were erected in the Late Iron Age (Meurkens & Tol Reference Meurkens and Tol2011). These were used for burial and show resemblances with enclosed cemeteries in northern France (Meurkens & Tol Reference Meurkens and Tol2011, 200–4).

Archaeological, historical and palynological evidence demonstrates that Flanders underwent intensive socio-economic transformations in the Early and Middle Roman Period (De Clerq Reference De Clerq2009, 465–81). In order to increase agraric production, large-scale reclamations took place. Nevertheless, these drastic changes do not seem to have led to a large-scale, deliberate destruction of late prehistoric monuments – contrary to the processes that took place in the High and Late Middle Ages in the western parts of Flanders (see below). Reuse of ancient cemeteries for burial is still very common. Others are overbuilt, however, and a few sites show a combination of both practices. This ambiguous attitude towards ancient relics, manifested even on a site level, has an opportunistic feel to it. Some burial monuments were focal points for renewed funerary activity, whereas others were seemingly in the way of farmsteads, fields, and roads. Apparently there was no need to save all ‘ancestral’ monuments at these sites. The creation of links to them was optional rather than a strict rule.

The Merovingian period is the last phase to produce irrefutable evidence for a positive appraisal of late prehistoric monuments. After that, barrows and urnfields gradually lost their important position. The Christianisation process certainly played a part in that process, as postulated by Roymans (Reference Roymans1995). However, it brought about no abrupt break. Some early churches are located at pre-existing burial sites. At the site of Ouwen (prov. of Antwerp) for example, a wooden church was built on top of a Merovingian cemetery containing a number of elite burials in the late 8th or early 9th century (Mertens et al. Reference Mertens, Van Impe and Vanderpijpen1977; Annaert & Vervoort Reference Annaert and Vervoort2003; Vandevelde et al. Reference Vandevelde, Annaert, Lentacker, Ervynck and Vandenbruaene2007, 54–7). A famous foreign example is the Danish site of Jelling (eg, Randsborg Reference Randsborg2008; Holst et al. Reference Holst, Jessen, Andersen and Pedersen2012). In general, fixed churchyards only developed in most parts of North-west Europe between the 10th and 12th centuries ad (Zadora-Rio Reference Zadora-Rio2003).

In previous research the High and Late Middle Ages in the MDS region are interpreted as a period of drastic landscape reorganisation (Roymans Reference Roymans1995). This certainly holds true for the whole of Flanders. However, the character of these processes differed regionally. The population in the western part of Flanders increased significantly from the Carolingian period and especially the 11th century onwards (eg, Thoen Reference Thoen1997). An accelerated agricultural expansion took place. This was necessary to provide for rapidly growing cities like Ghent and Bruges, that were developing into major urban centres. This resulted in a structured and intensively exploited cultural landscape, in which all available land was reclaimed and parcels were split up time and again (Thoen Reference Thoen1997). Contrary to the Roman-period phase of agraric expansion, now almost all remaining late prehistoric monuments were levelled. Compared to these drastic changes, the landscape transformations in the MDS region occurred at a slower pace and at a lower scale level. These differences are key to the differential survival of late prehistoric cemeteries in both regions.

This paper focused on identifying long-term trends in a large region. Therefore it was not possible to analyse observed patterns in specific time frames within their own cultural framework. It should be kept in mind that more or less similar reuse practices appearing at different times and places may have completely different backgrounds and meanings. This hinders meaningful intra-regional comparisons – besides the already mentioned fact that regional diachronic overviews are lacking in most parts of North-west Europe. In general studies on Early Medieval reuse practices are well-represented, possibly because sites yielding evidence on that are numerous, and sometimes spectacular. Prehistoric and Roman-period practices have received far less attention. Sopp’s northern German study offers probably the best options for comparison (1999, 131, Abb. 17) for our study, but his research area is hundreds of kilometres away. Some main trends he signals are roughly similar to those in Flanders, but there are various differences as well. It is not useful to go into detail here, but just to give an example: compared to Flanders reuse in northern Germany was far more intensive in the Iron Age and Early and High Middle Ages.

The fact that more or less similar reuse practices took place over prolonged periods of time throughout North-western Europe, has led to the development of a series of largely interrelated explanations that are applied in various contexts. It is certain that many barrows and urnfields remained meaningful places in Flanders until at least the Merovingian period, that these were incorporated in the cultural landscape and that fictitious links were created with the ‘ancestors’ present there. But based on archaeological data it is impossible to decide whether these practices served to, for instance, ‘familiarise’ ancient monuments, legitimise political authority, or strengthen territorial claims. Interestingly, researchers outside the Low Countries tend to focus on reuse acts that could be interpreted as ‘positive’ appraisal forms (but see Holtorf Reference Holtorf1998, 32–3). However, various Flemish cemeteries were partly overbuilt over the course of time, or people lived between them. The same locations were used for burial later on, leading to varied, complex and ambiguous site biographies.