The interpretation of hoards and deposition has always attracted debate, the seemingly irrational abandonment of large quantities of valuable metalwork requiring an explanation. In Ireland vast amounts of gold, bronze, and other artefacts were found in bog, river, and dryland contexts, either singly or as hoards, which stands in stark contrast to the lack of artefacts found in burials and settlements. This paper explores the depositional record in Ireland from the beginning of the Early Bronze Age (EBA) around 2400 bc to the end of the Late Bronze Age (LBA) around 800 bc. Encompassing hoard, single and burial finds, a methodology is presented that allows identifying patterns, their change over time and an interpretation of the phenomenon to be developed.
THE STUDY OF DEPOSITION
Explanations for deposition can broadly be divided in two: the artefacts were profane deposits that, due to unfortunate circumstances, could not be retrieved; or they were ritual deposits that were never meant to be retrieved and thus also served a concrete purpose (eg, Bradley Reference Bradley1998). Debate has mainly focused on hoard deposits, the deliberate character of which was self-evident, as they by definition consist of two or more artefacts. Single artefacts are often seen as the result of loss, or the remnants of disturbed and unrecognised archaeological contexts. The exceptions are single finds from rivers or bogs which could, in certain regions, be biased towards certain artefact types, particularly weapons, and are more readily accepted as deliberate deposition. The acceptance of single finds as intentional deposits raises the question of how they relate to hoards.
In regions where deposition in bogs was dominant, the interpretation of hoards and single items as votive deposits was suggested early on (Müller Reference Müller1898, 422ff; Worsaae Reference Worsaae1866, 313 ff). In areas of Europe dominated by dryland hoards containing broken and incomplete objects, concealment in times of crisis was the dominant interpretation (eg, Falkenstein Reference Falkenstein1997; von Brunn Reference Brunn1980; Holste Reference Holste1937). Both viewpoints immediately raise questions about the ownership of the metalwork, the reason for its deposition, and, in the case of a profane interpretation, why the material was not retrieved. The interpretation of hoards has changed little, although attempts to move away from the original classifications of ‘founders’, ‘personal’, or ‘merchants’ hoards were accompanied by efforts to establish more objective ways of recording and analysing the material. Strict categorisations of evidence for ritual or profane deposition were devised. The first attempt (Stein Reference Stein1976) was followed by many others who introduced their own categorisations, based on assumptions about the character of sacred and profane depositions which often were completely contradictory (cf. Fontijn Reference Fontijn2002 for a systematisation of the different arguments). The main problem of these approaches was the a priori interpretation of the criteria (for critique of Stein Reference Stein1976 see von Brunn Reference Brunn1980, 93). These were either based on characteristics of exceptional hoards, anthropological parallels (Levy Reference Levy1982), or conjecture. The interpretations reached vary from a wholesale interpretation as ritual (eg Maraszek Reference Maraszek2000; Hänsel Reference Hänsel1997; Hansen Reference Hansen1994; Taylor Reference Taylor1993) to the view that both phenomena might have existed side by side (eg Huth Reference Huth1997; Eogan Reference Eogan1983; Levy Reference Levy1982; Stein Reference Stein1976, 111).
Some approaches have added a further dimension to the interpretation of ritual hoards. Authors such as Barrett (Reference Barrett1985) and Bradley (Reference Bradley1998; Reference Bradley1988; Reference Bradley1982) were concerned with the potential social function of votive deposition, such as the display of status – competitive consumption – or the destruction of valuables to control the amount of metalwork in circulation to reduce social inequalities. Bradley (Reference Bradley1982) added another perspective to these explanations by setting them into the context of competitive gift exchange and comparing it to the North American potlatch, a ceremony in which the competitive mass destruction of goods served as a sacrifice to the spirits and gods (Mauss Reference Mauss1990, 16), as well as a display and a means of acquiring social status and prestige (Gregory Reference Gregory1980). This was seen as an analogy for hoard deposition (Bradley Reference Bradley1998; Reference Bradley1982). In contrast to the prestige momentarily gained by gift giving between humans, gifts to the gods could not be returned and thus increased the prestige of the giver (Bradley Reference Bradley1982). These concepts usually go hand-in-hand with the identification of these practices as votive deposition, without necessarily addressing the nature of votive deposition itself. Similarly, studies such as those by Kristiansen (Reference Kristiansen1989; Reference Kristiansen1979) or Taylor (Reference Taylor1993) primarily strive to extrapolate information on social and economic organisation from the hoard record and are less concerned with the interpretation of its function and meaning.
All these approaches see the gain of personal prestige as motivation for deposition or destruction. The apparently reasonable, economically sound explanation of the irrational act of deposition (cf. Fontijn Reference Fontijn2002, 18–19) explains the popularity of the concept, and reference is often made to two options – votive or prestige – as though they could, even conceptually, be separated from each other (Verlaeckt Reference Verlaeckt1998; Taylor Reference Taylor1993).
The interpretation of hoards as votive deposits has gradually found wider acceptance and established itself as the dominant interpretation (eg Bradley Reference Bradley1998; Hansen Reference Hansen1994; Fontijn Reference Fontijn2002; Hänsel Reference Hänsel1997; Vandkilde Reference Vandkilde1996; Hansen Reference Hansen1991), to the point that it is now considered self-evident and a possible utilitarian function has to be proved (Hänsel Reference Hänsel1997, 15). However, it could be argued that the votive explanation cannot be proven either and remains a hypothesis. It is notable that few authors (with the exception of, for example, Hansen Reference Hansen1994; 1991; Maraszek Reference Maraszek2006; 2000) engage with the question of what the concept of votive deposition, which is directly borrowed from the Greek world, actually entails and if it is applicable to the material at hand. Instead, in many recent discussions the word ‘votive’ has simply replaced the problematic term ‘ritual’ as another placeholder for the inexplicable. The problematic concept of ritual remains. The narrower meaning of ‘votive’ actually causes more problems as it suggests a particular, religious function directed at supernatural powers for a specific purpose. This term presupposes detailed knowledge about participants, recipients, and purpose of the deposit. While using the word ‘ritual’ remains unhelpful, the way forward lies in what Brück (Reference Brück1999) called for: gaining a more detailed understanding of the archaeological record and the past practices that may have led to its formation.
Developing a contextual approach to deposition
Composition, arrangement, and find contexts of hoards are the most frequently considered aspects and remain the key data we have to work with. However, there are problems with past interpretions, as discussed above. For example, the arrangement of finds in a particular way has often been read as indicating a non-utilitarian function of the deposit (Geißlinger Reference Geißlinger1984, 327), though finds can also be arranged for purely practical reasons (Genthe 1874, 100f, quoted in Hansen Reference Hansen1994, 325) or be the result of the former presence of an organic container. The presence of food remains, ash, or animal bones in a deposit has also been used for the identification of ritual hoards (eg, Levy Reference Levy1982, 22), as they could be part of a ritual or ceremony protecting the depositional act. But any kind of deposit – be it permanent or non-permanent, retrievable or irretrievable – may have been accompanied by a ceremony that leaves no trace.
Even the most clear criterion – the divide between wet and dry contexts – can be refuted as a clear pointer for the type of act. Deposition in bogs or rivers has been seen as indicating the permanent surrender of the valuables, as their retrieval is impossible (cf. Levy Reference Levy1982, 21). However, recovery from a variety of wet contexts is theoretically possible, and a decision to deposit something in a wet milieu might have been influenced by the greater security benefit (Needham Reference Needham2001, 290–1). Bog contexts can be extremely varied (cf. Geißlinger Reference Geißlinger1984, 322–3); drier areas would allow retrieval. Concealment of valuables in bogs in times of crisis is evidenced elsewhere in historic times (Petersen 1890, 209ff in Willroth Reference Willroth1985, 16).
Neither does the reverse argument hold true. Deposition in dry places does not necessarily indicate that the objects were meant to be retrieved (cf. Kristiansen Reference Kristiansen1996, 255), no matter how accessible or prominent the spot. Physically unprotected deposits may have been protected by social or religious rules or sacred places in the landscape (Geißlinger Reference Geißlinger1984, 323). The marking of the find location and the deposit itself may also be a way of mythically increasing the value of the place and the object (Torbrügge Reference Torbrügge1985, 18).
To avoid the dependency on a fixed set of criteria, some recent studies have moved away from the individual find and its circumstances, concentrating instead on the compositional characteristics of hoards and patterns of deposition across a broader body of finds and types of deposits. Patterns in the record are seen as evidence for deliberate deposition, pointing to deposition of some material in wet contexts as the deliberate giving-up of objects, suggesting a ritual purpose (Maraszek Reference Maraszek1998, 315). However, dry-land deposition is not considered as the expression of a different depositional intention. Instead, the interpretation of wet deposits is extended to the rest of the material. This has led to a situation in which context is only of secondary importance or its relevance denied (eg, Maraszek Reference Maraszek2000, 290; Hansen Reference Hansen1994, 325; 1991, 183; Verlaeckt Reference Verlaeckt1998, 266).
The recognition of patterns in the depositional record that span different types of deposit categories suggested that single finds could also be the result of deliberate deposition. Kristiansen's observation that over the course of the Bronze Age a complementary and oscillating relationship between burials and hoard finds can be observed also underlines the notion that different parts of the depositional record may be interdependent aspects of an overall depositional system (Kristiansen 1998, see also Hansen Reference Hansen1994; Torbrügge Reference Torbrügge1970/1). The need to understand hoard deposition within the general framework of depositional practices over time has already been recognised elsewhere (eg, Bradley Reference Bradley1982) and studies integrating the different strands of depositional practice have concentrated on the patterns or structures (Hansen Reference Hansen1994; Reference Hansen1991) or selectiveness of the record (Needham Reference Needham1988). They aim to identify repeated actions in order to differentiate between traditional practices and random activities (Torbrügge Reference Torbrügge1985, 18).
Focus of this study
This study abstains from considering the ritual or profane function as a crucial question to answer. Instead, a focus on the variability of depositional practices over the course of the Bronze Age will aid further interpretation. The demonstration of non-random patterns in the record needs to be the departure point of any study that aims to identify and interpret deliberate deposition.
For this, the concept of context needs to be expanded. The treatment of artefact types across places and forms of deposits is an important aspect of depositional articulation. If an inter-relationship between hoards, single finds, and burials is assumed, the treatment of artefacts across these different categories needs to be examined.
Rather than despairing over the inconclusiveness of known criteria, an alternative data-led approach is proposed here. This paper suggests that interpretative possibilities are not so limited, once freed from the known set of options and criteria, one re-engages with the data on a broad, comprehensive basis. The identification of structured or selective deposition is not the end of the discussion. Contextual, compositional, and conditional characteristics can reveal meaningful patterns and also potentially provide information about the rationale behind the formation of the record.
This approach depends on the availability of contextual information and single finds are notoriously badly documented. This is here compensated for by creating a substantial database of the material, taking a diachronic approach from the earliest beginnings of the Irish Bronze Age to the end of the period. In addition, the patina of bronze artefacts from hoards was assessed in order to confirm the validity of information on hoard contexts or suggest contexts for finds without provenance (Becker Reference Becker2006). Comprehensiveness is crucial in order to be able to capture variations and patterns in the record; all types of artefacts that occur in these contexts, including gold and non-metal artefacts, were therefore included. This is, however, restricted by our ability to recognise and date types. So it may not be possible to date certain classes of stone artefacts, plain amber beads, or jet and shale objects, and even some metal objects that are not typologically distinctive types, when they are in a ‘single find’ context. The numerical comparison of such artefacts when found in other parts of the archaeological record is therefore not possible (cf. Needham & Burgess Reference Needham and Burgess1980, 445–7). This particularly affects non-metal artefacts for which, after the EBA, no typologically distinct forms exist. While, for example, amber beads are found in Middle (MBA) and LBA hoards, the corresponding singly deposited forms cannot be securely attributed to either period. Caution must also be exercised with some of the metal objects that are not diagnostic.
The data (Becker Reference Becker2006) was compiled from published and unpublished catalogues and sources. In total 1686 objects from metal hoards were considered, as well as a number of non-metal hoards, and 322 burials were recorded. Of these, 201 could, on the basis of typologically distinct artefacts or independent dating evidence, be assigned to either part of the EBA. For both burials (Waddell Reference Waddell1990; Kilfeather Reference Kilfeather1991) and hoards (Eogan Reference Eogan1983; O'Flaherty Reference O'Flaherty1995) as well as for most artefact types it was possible to build on existing publications (see caption for Fig. 1), whereas, for LBA spearheads, all specimens in the National Museum and the Ulster Museum were recorded as a representative sample. About 8900 single finds were recorded in the database (for a full catalogue see Becker Reference Becker2006). Single finds and hoards are rarely encountered in excavations and the recent upsurge in excavation activity has not significantly added to the corpus of EBA burials with grave goods.
Fig. 1 Dating of the main Irish Bronze Age artefact types. The phasing is based on Waddell (1998), Needham (Reference Needham2005; Reference Needham2000; Reference Needham1996; Reference Needham1990), Needham et al. (Reference Needham, Bronk Ramsey, Cartwright and Pettitt1997), Brindley (Reference Brindley2007), Eogan (Reference Eogan1964; Reference Eogan1962) and individual find associations. A full discussion of the phasing adopted here and dating of individual types can be found in Becker (Reference Becker2006)
STRUCTURED, TYPE-SPECIFIC DEPOSITION IN BRONZE AGE IRELAND
The material record of prehistoric Ireland differs significantly from that in other European regions. Most striking is the lack of artefacts from settlement sites throughout the Bronze Age and the lack of non-ceramic burial goods after the EBA. The overwhelming majority of artefacts are found in the natural landscape, in rivers and bogs and to a significant extent on dry land.
The lack of artefact association in stratified contexts creates chronological problems and the Irish Bronze Age chronology mainly relies on the phasing of the metalwork dated by comparison with typologies elsewhere, particularly in England and Scotland (eg, Eogan Reference Eogan2000; Reference Eogan1994; Reference Eogan1964; Reference Eogan1962). The chronology adopted in this study (Fig. 1) is designed to allow the comparative analysis of types and is kept broad to legitimately allow the postulation of loose contemporaneity and thus comparison of what are mostly floating, relative typo-chronological sequences, augmented only at times by radiocarbon dates. While the application of radiocarbon dates has in the case of the pottery led to a relatively stable chronological framework for the EBA (Brindley Reference Brindley2007), the metalwork phases cannot be correlated directly to this, and for the MBA and LBA no distinct typological development of pottery types exists. The relative chronology of the metalworking phases is anchored in absolute chronology by radiocarbon dated associations and artefacts (eg, Needham Reference Needham1996; 1990; Needham et al. Reference Needham, Bronk Ramsey, Cartwright and Pettitt1997; Brindley Reference Brindley2001).
Table 1 Quantity of Objects Found As Burial Goods, Hoard Components, or Singly For Each Main Group of Artefact Type in eba1/chalcolithic
The artefacts in this study reveal some striking patterns over the course of the Bronze Age. While the number of single finds increased steadily (Fig. 2a), hoard and burial events fluctuate (Fig. 2b–c). The number of burials containing grave goods increased during EBA2 (Fig. 2b), while the number of burial goods remained broadly stable (Fig. 2c), reflecting a decrease in number of finds per burial. While a rather modest range of objects, few of metal, were deposited in graves in EBA1 (Tables 1a–c), in EBA2 metal objects that might indicate status, such as bronze razors and daggers, were deposited in burials in greater numbers, as well as a small number of amber and faience objects, battle axes, a halberd, and a spearhead (Tables 2a–c). Also, for the first time, a small number of gold finds can clearly be tied to the burial record.
Fig. 2 Number of a) depositional events, ie, single finds, hoards, and burial deposits; b) hoard and burial finds containing artefacts; c) artefacts from hoards and burials in the summary stages EBA 1/Chalcolithic, EBA 2, Middle & Late Bronze Age
Table 2 Quantity of objects found as burial goods, hoard components, or singly for each main group of artefact type in the eba
The number of hoards that can be assigned to the EBA2 are small in size, a trend that continues in the MBA (Fig. 2b, Tables 9a–c) where only a small increase in hoard numbers can be observed (Fig. 2a). From this phase on, burials lacked aceramic grave goods. Gold hoards, however, were deposited in greater numbers than before (Tables 9a–c). The range of bronze artefacts deposited in this period was severely limited, whether in hoards or as individual deposits of weapons and axes (Tables 3a–c). Hoard deposition rose in the LBA to its highest level in terms of frequency of depositional events (Fig. 2a) as well as the number of objects (Tables 2b & 4). The range of objects deposited widened significantly in this period to include bronze ornaments and tools other than axes, but hoards still contributed a significantly lower amount of objects than the single find record (Table 4a–d). Throughout the Bronze Age some non-metal artefacts were deposited separately in hoards as for example Early Bronze v-perforated buttons in two bog deposits (Harbison Reference Harbison1976, 15), MBA stone moulds and LBA wooden mould patterns (Becker Reference Becker2006).
Table 3 Quantity of objects found as burial goods, hoard components, or singly for each main group of artefact type in the MBA a) bronze b) gold c) moulds and the percentage of each type in the overall assemblage
Table 4 Quantity of objects found as burial goods, hoard components, or singly for each main group of artefact type in the LBA a) bronze b) gold c) moulds and the percentage of each type in the overall assemblage
Deposition as one option
In the normal course of events, the majority of the artefacts would have been recycled and the metal would thus have gone back into circulation. Deposition has to be considered an exception to the norm (cf. Needham Reference Needham2001). And indeed, the absence of some artefact types from the spectrum of finds could be representative of selectivity in deposition, to the extent of the exclusion of certain types from deposition altogether. It is striking, for example, that after the deposition of grave goods ceased at the end of EBA2, razors all but disappear from the record, to re-appear again in small numbers in the Late Bronze Age (Fig. 1, Tables 2a, 3a, 4a). While it is possible that the production of razors was discontinued, it seems more likely that their deposition ceased. The same appears likely for some forms of early spearheads and sickles. Bronze ornaments are similarly rare. Typologically distinct bronze ornaments were not generally deposited, but have been found in the MBA hoards from Bishopsland, Co. Kildare (Eogan Reference Eogan1983, 36, no. 16), Annesborough, Co. Armagh (ibid., 27, no. 7), and Cloneenbaun, Co. Roscommon (ibid., 44, no. 26), and rare examples of moulds for casting of bronze ornaments are occasionally found on MBA settlement sites (eg, Knockadoon, Lough Gur, Co. Limerick (Ó Ríordáin Reference Ó Ríordáin1954); Corrstown, Co. Antrim (Grogan Reference Grogan2012); Ballyprior Beg, Co. Antrim (Armit Reference Armit2003)). The fact that the rare examples of bronze ornaments in hoards are often in fragmentary condition supports the notion that they were in circulation but usually not deemed suitable for deposition unless as scrap. This is the case in the Bishopsland tool hoard where both a sickle and bronze bracelet are present in fragmentary form, as well as the two torcs from the Annesborough hoard (here beside two intact plain bracelets). The general lack of Bronze ornaments in Ireland led to the discussion of these items as imports (Eogan Reference Eogan1994). In contrast it is here suggested that particular types of bronze artefacts were in use, but were recycled and only rarely entered the depositional record. This would also counterbalance the large number of gold ornaments which have dominated our understanding of levels of affluence of Irish society at the time.
Selective deposition – single finds, hoards and burial goods
The most obvious evidence for selective deposition in the Irish Bronze Age record is the clear distinction between the types of artefacts and materials found in burials, and those comprising hoards and single finds (Tables 1–4).
Some artefact types are over- or under-represented in certain contexts and if we can identify types that were deposited only or predominantly as single finds but were not included in hoards (or burials), we can assume that a selection was made. Gold and bronze artefacts (Tables 1a–b, 2a–b), as well as battle axes and v-perforated buttons (Table 1c and 2c) for example, were rarely deposited in EBA burials. Taking issues of preservation and recognition into account, the under-representation of types in the single find record compared to closed assemblages as hoards or burials is only possible in the case of recognisable, distinct forms. As the aforementioned sickles and bronze ornaments, also the absence of Late Bronze Age tools such as socketed gouges or chisels in the single-find record (Table 4a) is likely to reflect prehistoric reality, especially given the fact that neither river-dredging schemes nor a survey of Irish river and bog finds has contributed significant numbers of these types (Bourke Reference Bourke2001; Halpin Reference Halpin1984). Other types were predominately deposited as single finds and never (rapiers and dirks of the MBA) or rarely (LBA axes: 3.5%) included in hoards.
Such patterns suggest a selective process, both in the case of artefacts selected for – or excluded from – deposition in burials and those for which hoarding or single deposition was the preferred choice. This demonstrates that deposition in these latter forms is not an automatic by-product of production and circulation, but that deliberate choices appear to structure the record.
Type-specific depositional patterns
These deliberate choices can be demonstrated to structure the record further and the selectiveness can be demonstrated to be artefact type-specific, as already suggested by the broad patterns outlined above. Within the single find record it is possible to identify type-specific associations between particular artefact types and places in the landscape (Tables 5–8). After axes, weapons constitute the largest group of artefacts throughout the Bronze Age. They were usually deposited singly in wet contexts, especially in rivers. This pattern conforms to the treatment of weapons across Western Europe in general (cf. Hansen Reference Hansen1994; Bradley Reference Bradley1998; Maraszek Reference Maraszek1998). However, distinct differences are clearly present between different types of weapons and between sub-forms.
Table 5 Single finds by find context in eba1/chalcolithic
Table 6 Single finds by find context in eba2: bronze
Table 7 Single finds by find context in mba
Table 8 Single finds by find context in lba
The deposition of specialised, and particularly bladed weapons differs considerably from that of axes and spearheads; the former are found far more often in river contexts, this is particularly true for the more elaborate and little-used pieces. This tendency can be traced from the EBA to the LBA, and begins with EBA1 halberds (32%), daggers (36.4%), and stone battle axes (82.4%) coming from river contexts (Tables 5a–c) which contrasts starkly with that for metal axes at 21.2%. The trend continues in the MBA when rapiers and dirks show a high level of river deposition (66% and 90%), and an even higher rate for Lissane-type rapiers, which are particularly large, technically perfect, and probably of ceremonial function (Ramsey Reference Ramsey1989, 60). This is in stark contrast to the rate of river deposition for contemporary axes (34.7%) and spearheads (39%) which show a higher rate of bog deposition instead. In the LBA, swords (53.4%) and spearheads (57.6%) continue the pattern with both types being predominately deposited in river contexts. There is a trend for spearheads deposited in rivers to be relatively large in size (Bourke Reference Bourke2001, 113). Similarly, in the MBA some of the most elaborate and probably non-utilitarian forms of spearheads are predominantly found in wet contexts (Ramsey Reference Ramsey1989, 329). This distinction between small utilitarian spearheads and large more ostentatious examples is also reflected in the LBA hoard record, with small spearheads deposited in complex hoards and the larger examples in one-type or single-category weapon hoards. This may reflect a functional specialisation and the combination with ornaments and tools places the small pieces in a more domestic or personal context. Wear and damage suggests that the spearheads in weapon hoards were less-used prestigious ceremonial items, while the small ones were extremely worn, resharpened and often damaged, indicating frequent use. This may suggest a distinction between utilitarian and ceremonial forms as the basis of this differentiation; the two LBA spearheads decorated with gold foil found in wet contexts (Coles Reference Coles1971a) further emphasise the prestigious or ceremonial potential of larger examples. The difference between combat and hunting spearheads may also be represented in these differences (cf. Bridgford Reference Bridgford2000, 34).
The deposition of Late Bronze Age specialised ceremonial objects such as cauldrons, buckets, and horns in bog contexts (100% each), but not in rivers, mirrors that of gold artefacts. Except for one LBA bulla and the Near New Ross, Co. Waterford hoard (Eogan Reference Eogan1994, 144; Reference Eogan1983,164, no. 145), as well as the EBA Belville hoard, gold ornaments never seem to have been deposited in rivers during later prehistory (Tables 5–9; Becker Reference Becker2011). Apart from a brief phase of inclusion of sheet-gold artefacts in this phase in the EBA2 Knockane, Castlemartyr or Topped Mountain burials, gold was deposited on dry land and in bogs throughout the Bronze Age. This suggests a depositional intention distinctly different from that of bronze artefacts. Also notable is the largely separate deposition of gold and bronze in the hoard record, which only changes in the LBA, when bronze is found in association with gold with tools, bronze rings, and amber in some hoards (see below).
Table 9 HOARDS: a) EBA1/CHALCOLITHIC
A substantial proportion of axes were deposited in wet contexts; this increased over the course of the Bronze Age, while the proportion of these as river finds declined from 46% in EBA2 to only 28.4% in the LBA. Contrast this with the high proportion of river contexts for bladed weapons in this period. This in conjunction with the changes in the associative patterns in hoards (see below) suggests a change in the role of the artefacts axe over time.
Hoards
The hoard record can also be shown to be selective and structured by type-specific depositional rules. Type-specific patterns can be identified in the form of compositional groups that can be correlated with different depositional contexts. Some of these associative patterns between objects and place show a striking stability throughout the Bronze Age.
The hoard record can be broadly divided in three different groups: one-type, single category and complex hoards (Fig. 3b, Table 9a–d). While one-type hoards contain a single artefact type, single category hoards contain different artefact types from the same functional spectrum, eg exclusively weapons or ornaments. Complex hoards contain a variety of different artefact types, sometimes of different materials. These hoard types can be recognised in all stages of the Bronze Age, however, the greatest level of differentiation is visible in the LBA with its greater range of artefact types.
Fig. 3 a) sword fragment from the Park hoard, Co. Meath, with evidence for forceful bending and breaking (NMI 1974:38); b) Irish Late Bronze Age hoard types based on associative patterns
Table 9 (cont.) HOARDS: a) EBA1/CHALCOLITHIC (cont.)
Table 9 (cont.) HOARDS: b) EBA2
Table 9 (cont.) HOARDS: c) MBA
Table 9 (cont.) HOARDS: c) MBA (cont.)
Table 9 (cont.) HOARDS: d) LBA
Table 9 (cont.) HOARDS: d) LBA (cont.)
*For list of findplaces see p. 26.
Table 9 (cont.) HOARDS: d) LBA (cont.)
Table 9 (cont.) HOARDS d) LBA (cont.)
Table 9 (cont.) HOARDS d) LBA (cont.)
Table 9 (cont.) HOARDS d) LBA (cont.)
Key to findplaces 1. Ballynure, Carrickfergus, Co. Antrim; 2. Carrigogunnel, Co. Limerick; 3. Chute Hall (Clocherclemin), Co. Kerry; 4. Cork/Mallow, Co. Cork; 5. Corracanvy, near Clough Oughter Castle, Co. Cavan; 6. Drumbest, Co. Antrim; 7. Drunkendult, Co. Antrim; 8. Dungannon (Barony of), Co. Tyrone; 9. Dunmanway, Co. Cork; 10. Killarney, nr., Co. Kerry; 11. Macroom, Co. Cork; 12. Moyarta (Barony of), Co. Clare; 13. no prov., Co. Cork; 14. no prov. Co. Mayo; 15. no prov., Ireland; 16. no prov., Co. Tyrone; 17. Athlone, nr., Co. Roscommon/Westmeath; 18. Ardlow, Co. Cavan; 19. Ballycroghan, Co. Down; 20. Drumcrample, Co. Fermanagh; 21. Latteragh (Latera Upper), Co. Tipperary; 22. Relagh Bog, Co. Tyrone; 23. Carran, Upper and Lower, Co. Donegal. ; 24. Inishleague Island, Co. Fermanagh; 25. Belturbet/Stragheglin, Co. Cavan; 26. no prov., Co. Offaly; 27. Knockadoo, Co. Roscommon; 28. Youghal, Co. Cork; 29. Ballyroe, Co. Limerick; 30. Tully More (Moolagh), Co. Donegal; 31. Tempo, Co. Fermanagh; 32. Athenry, Co. Galway; 33. Silvan Park, Co. Dublin; 34. Granny, Co. Roscommon; 35. Calverstown, Co. Westmeath; 36. Ballydivlin, Co. Cork; 37. Boa Island, Lough Erne, Co. Fermanagh; 38. Forth Commons, Co. Wexford; 39. Crevilly-Valley (Connor), Co. Antrim; 40. Booltiaghadine, Co. Clare; 41. Brother's Cave, Ballygambon Lower, Co. Waterford; 42. Newgarden North, Co. Limerick; 43. Ross, Fern Field, Co. Tipperary; 44. Crossna, Co. Roscommon; 45. Killulagh, Co. Westmeath; 46. Lackagh, Co. Derry. ; 47. Frankfort, nr, Co. Offaly; 48. Derrycoagh, Lough Gara, Co. Roscommon; 49. Garryhinch, Co. Offaly; 50. Headford, nr., Co. Galway; 51. Inch Island, Co. Sligo; 52. Waterstown, Co. Westmeath; 53. Seacon More/ Seacon Hill, Co. Antrim. ; 54. Knockglass, Co. Roscommon; 55. Annagh, Co. Roscommon; 56. Portlaoise (nr. Maryborough), Co.Laois; 57. Brockagh (Braca), Co. Westmeath; 58. Ballykeaghra, Co. Galway; 59. Kilshanvy, Co. Galway; 60. Castle Coote, Co.Roscommon; 61. Athlone, nr., Co. Westmeath; 62. Killyleagh, Co. Down; 63. Cooga Lower (former Kilcommon), Co. Limerick. ; 64. Athlone, nr., Co. Roscommon ; 65. Liverourd or Oldcastle, Co. Mayo; 66. no prov., Co.Louth/Meath; 67. no prov., Co. Galway; 68. Tara, Co. Meath; 69. no prov., Co. Mayo. ; 70. Craighilly, Co. Antrim; 71. Downpatrick 3, Co. Down; 72. no prov., Co. Clare; 73. Annaghbeg/Monastereadan, Lough Gara, Co. Sligo; 74. Enniscorthy, nr., Co. Wexford; 75. Mullingar, Co. Westmeath; 76. Cogran, Co. Offaly; 77. Gorteenreagh, Co. Clare; 78. no prov., Co. Limerick; 79. Askeaton, Co. Limerick; 80. Cashel, Co. Tipperary; 81. Drissoge, Co. Meath; 82. Clonleigh, nr. Lifford, Co. Donegal; 83. Bruree, nr., Co. Limerick; 84. Kilmoyly North, Co. Kerry; 85. Ballinesker, Co. Wexford; 86. New Ross, nr., Co. Waterford; 87. Lattoon, Co. Cavan; 88. no prov., Ireland; 89. Killymoon, Co. Tyrone; 90. no prov., Co. Limerick; 91. no prov. (formerly Mull), Co. South of Ireland; 92. Arboe/Killycolpy (Lough Neagh hoard), Co. Tyrone; 93. no prov., Ireland; 94. no prov., Ireland; 95. no prov., Co. Roscommon; 96. Rathgall, Co. Wicklow; 97. Newport,,Co. Mayo; 98. Dreenan (Boa Island), Co. Fermanagh; 99. Blackhills, Co. Laois; 100. Drumany, Co. Leitrim ; 101. Moroe, Co. Limerick; 102. Coogal, Co. Sligo; 103. Boolybrien, Co. Clare; 104. Park, Co. Meath; 105. Luffertan, Co. Sligo; 106. Ballinliss, Co. Armagh; 107. Tuam, nr., Co. Galway; 108. Kilmurry, Co. Kerry; 109. Kish. Co. Wicklow; 110. Knockmaon, Co. Waterford; 111. Bootown, Co. Antrim; 112. Grange, Co. Kildare; 113. Ballygowan (Reade), Co. Kilkenny; 114. no prov., Co. Westmeath; 115. Teernagloghane, Co. Clare; 116. Lahardan, Co. Clare; 117. Doon Uppe, Co. Galway; 118. Glenstal, Co.Limerick; 119. Trillick, Co.Tyrone. ; 120. Armoy/Cromagh, Co.Antrim; 121. Derryhale, Co.Armagh; 122. Enagh East, Co.Clare; 123. Ballinderry, Co.Westmeath; 124. Ballyvadden, Co.Wexford; 125. Callanagh, Co.Cavan; 126. Killycreen West, Co.Fermanagh; 127. Killevy, Co. Armagh; 128. no prov., Co.Fermanagh; 129. Gardenhills, Co.Fermanagh; 130. Annagh/Meneysterlin, Co. Derry; 131. Money Lower. Co. Laois; 132. Ballycurrin, Co. Mayo; 133. Knockanbaun, Co.Sligo; 134. Strabane, Co. Tyrone; 135. Kinnegoe, Co. Armagh; 136. no prov., Co. Wicklow; 137. Kilbride, Co. Mayo; 138. Belfast, Co.Antrim/Down; 139. Mountrivers (Coachford), Co.Cork; 140. Killersherdiny, Co.Cavan; 141. Ballytegan, Co.Laois; 142. Oghermong, Co. Kerry; 143. Rathtinaun, Co. Sligo; 144. Tooradoo/Cnoc na bPoll, Co. Limerick; 145. Banagher/Meenwaun, Co.Offaly; 146. Scotstown, Co. Monaghan; 147. no prov., Co.Mayo; 148. no prov., Co. Sligo; 149. Dowris, Doorosheath, Whigsborough, Co. Offaly; 150. Cullen, Co. Tipperary; 151. Mooghaun North (Great Clare find), Co. Clare; 152. no prov., Ireland; 153. Tamlaght, Co. Armagh;
Table 10 LBA HOARDS
One-type hoards contain small numbers of used but complete objects (mainly one type of weaponry, axe, or ornament). These artefacts also show the greatest rate of single-find deposition and within hoards the objects are usually in good condition and display only limited evidence for breakage (Table 10a). The only exceptions to this rule are the Late Bronze Age sword hoards in the southern part of the main cluster of sword hoards: here the swords seem to have been intentionally destroyed and damaged. In the EBA, axes and halberds were deposited in one-type hoards (Table 9a, b). Only three one-type axe hoards of both MBA and LBA date (Table 9d) have been found in Ireland, whereas they are more common in Britain and continental western Europe. For example, in Britain two-thirds of all LBA hoards consist of axes only (Huth Reference Huth1997, 127). One-type sword hoards are a specific feature of British and Irish depositional practice. However, the majority of the British examples are two-piece hoards (Huth Reference Huth1997, 127), while in the Irish assemblage up to four swords can be deposited together.
LBA horns were also deposited in one-type hoards (Table 9d). Broken horns are rare, although sometimes only parts of them were deposited and the mouthpieces are usually missing, which implies dismantling before deposition. The only occasion on which a horn was deposited in a hoard with other artefacts was in the Booleybrien, Co. Clare hoard where, however, the horn was broken. The exceptional Dowris find, which also contained horns is here not considered as a hoard in the strict sense as it does not appear to represent a closed, chronologically coherent assemblage (cf. Coles Reference Coles1971b). Most cauldrons and buckets were deposited singly and complete, which is in marked contrast to the depositional mode outside Ireland and northern Britain. In southern Britain and continental Europe the majority of both Atlantic and Urnfield buckets are only found as fragments in complex hoards (Gerloff Reference Gerloff2010; Reference Gerloff2004, 126, 128). The only known associations of cauldrons and buckets in Ireland are in the unusual mass finds at Dowris and the Bog of Cullen.
One-type hoards also have a geographically distinct pattern compared to other hoards. This is clearest in the LBA, with two foci for the deposition of horn hoards in the south-west and north-east of the country (Fig. 4). Between these in the midlands, one-type sword-, spearhead-, axe-, and ornament hoards are located, with a certain degree of overlap between the sword and horn hoards in the north-east. In the MBA, one-type bronze hoards show a pattern that contrasts with that of the complex hoards and gold hoards of the period. While there is a slight shift in distribution, the general area in which these hoards are found is the Leitrim, Fermanagh, Sligo area. One-type hoards tend to be associated with wet contexts, most clearly visible in the case of weapon and horn hoards (Table 10b).
Fig. 4 Distribution of bronze one-type hoards
Single-category hoards are a feature of the Late Bronze Age, and while the lack of context information often hinders differentiation there is some suggestion that they also follow type or category specific depositional rules: LBA bronze single-category ornament hoards show a strong link with wet contexts, while those consisting only of tools were deposited in dry and specifically rocky/stony contexts (Tables 10b, 9).
Complex hoards often contain collections of material, including fragmented objects that would be suitable for recycling, unfinished objects, and raw materials. This is the case in some EBA hoards and it is striking that those that contain copper cake are from dry land (Table 9a). Within the Late Bronze Age variation within the group of complex hoards can be observed: while most tend to be deposited to a greater degree in wet contexts (Table 10b), the group of hoards that contain swords in combination with other objects stands out in having been deposited primarily in dry contexts (87.5%: Table 10b). This is also the group which shows a greater degree of damage with 90.9% of hoards containing at least one incomplete item (Table 10a). Similar patterns can be observed throughout considerable parts of the European Bronze Age and has led to suggestions of a functional divide between ritual and utilitarian hoards. While large scrap hoards as found in Britain or continental Europe are virtually unknown in Ireland, these Irish complex hoards nevertheless show breakage and the combination of objects from different functional spheres, an inherent characteristic of scrap hoards as also observed by Bradley (Reference Bradley1998).
If single finds can be considered intentional deposits that are structured by type-specific depositional rules, the fundamental question that arises is how they relate to the hoard record. Hoards cannot be regarded as a homogeneous class of finds; they appear in a variety of forms in which, for example, one-type deposits can be shown to be subject to the same depositional rules as the single finds. In the EBA and LBA, when single-weapon deposition in rivers was accompanied by one-type weapon hoards, these hoards adhere to the same depositional pattern. Also some single category deposits appear to reflect the specific meaning of the objects contained within them, while the associative patterns and condition within the range of complex hoards seems to be the structuring factor.
INTERPRETING DEPOSITION – DISCUSSION
Patterns can reflect a variety of habitual practices such as the daily movement of people, the organisation of production, or the distribution of raw materials or products, and identifying habituality or even deliberate deposition of items does not reveal anything about its purpose. Rather, it is the intentionality, the understanding that there is a purpose beyond what is achieved in the act of deposition that sets the practices observed here apart from everyday, functional processes.
The significance and meaning of type-specific deposition
The clearest indicator of intentionality is the differential treatment of artefact types through deposition in burials or in the landscape. Aceramic artefacts deposited in EBA burials in Ireland are made of non-precious, mostly non-metal materials and are defined by their everyday functionality and their personal character, such as domestic items and tools, as well as items of personal ornament made of stone or more rarely jet, amber, and other materials. Objects interred in burials are often thought to indicate social status (cf. Shennan Reference Shennan1982) and thus to directly reflect the identity of the person buried; recent debates (eg, Brück Reference Brück2004) also highlight the agency of the mourners in the selection of burial goods. Either way, the deliberate character of the selection seems likely to reference the identity of the person interred. Artefact types excluded from burials and instead deposited in the landscape would seem not relevant in the construction of the identity of an individual at this point; in the Early Bronze Age this is strikingly the case for gold ornaments, which are clearly for the decoration of body or dress. It appears likely that they relate to other social identities that may have been held by an individual in different contexts, rather than those circumscribed in the burial context. Possibly being of communal character (cf. Taylor Reference Taylor1994; Reference Taylor1970, 56), they may have also been connected with a specialised identity not permanently affixed to an individual, but representative of a particular role taken on during a particular phase of a person's life.
Stone and metal axes in Ireland, as elsewhere, were largely excluded from EBA burials, so they were not directly connected with the identity of the person at burial. There is a strong indication that stone axes were still in use in the earliest Bronze Age (Becker Reference Becker2006), though they were in deposition strictly separated from metal axes, which obviously belonged to a different depositional and transactional sphere (cf. Vandkilde Reference Vandkilde1996, 273). The same is also the case for battle-axes which, unlike metal axes, were never hoarded but deposited singly, with a clear emphasis on river contexts. Like axes and halberds, they did not belong to the range of objects that were deemed suitable for deposition in burials (cf. Simpson Reference Simpson1990), with only exceedingly rare examples of these types having been found in burials. It can thus be argued that it was the artefact type ‘axe’ that was excluded from the burial rather than the material of which it was being significant. In contrast, the large number of metal axes deposited in hoards or singly suggests that no underlying economic rationale shaped the development of this particular depositional taboo (contra Vandkilde Reference Vandkilde1998, 254–5; 1996, 26; Kristiansen Reference Kristiansen1989, 21f; O'Flaherty Reference O'Flaherty1995, 14; but see Bradley Reference Bradley1988, 250–2). Rather, axes were not related to the persona at burial. Strikingly, an item that could be considered to be of more personal character – the perforated buttons of Beaker background – were hardly ever included in burials (but at times as secondary deposits in megalithic tombs), and only one bracer was found in a burial context. Their deposition singly, or in one-type hoards, suggests that they also were connected with a particular identity that was not relevant at the point of burial.
The contrast between the treatment of gold ornaments and weapons, with deposition of gold hoards or single finds in dry or bog and the absence of gold ornaments or other gold artefacts from rivers – the primary context for the deposition of weapons – is striking. This implies that gold and bronze had a contrasting and complementary relationship, which may be reflective of different social personae.
It is tempting to suggest that if weapons can be related to the male sphere, gold ornaments might represent the female sphere in the record. This was suggested by Waddell (Reference Waddell1990, 13) based on the observation that gold hoards were predominantly deposited in dry and weapons in wet contexts. Such a gender attribution for gold ornaments is however problematic as it is based on our present day conceptualisations of the artefacts and their connection with gender. Furthermore, in EBA Britain discs were placed in Beaker burials with males and Wessex group gold plaques continued this tradition, while females were buried with amber necklaces (cf. Eogan Reference Eogan1994, 39).
The evidence for a divide between different and directly contrasting personae within the Irish Bronze Age is further underlined by the fact that LBA hoards containing gold hardly ever contain weapons. Amber appears to be conceptually closely related to gold ornaments and is rarely found in hoards containing only bronze
In contrast, two contextualised bronze sunflower and disc-headed pins were deposited in rivers, thus apparently relating them more closely to the weapons than to the gold ornaments of the period – a pattern that can also be observed in the Danish Bronze Age record (Jensen 1972, 129–31). This difference is further accentuated by the fact that pins were only rarely deposited in bronze or mixed-material hoards and were not combined with gold ornaments with no further bronze association. The clear bipolarity of the record implies that starkly contrasting roles are represented by weapons and bronze pins on the one hand and gold ornaments on the other.
Here, however it is argued that differential treatment of types of objects can be understood as a reflection of variation in their type-specific meaning that is primarily identity-based. The continuation of trends of deposition throughout the Bronze Age also indicates a great longevity of ideas about objects and the identities reflected in them. The strict depositional rules applied to battle-axes and maceheads of the EBA foreshadows that of the later weapon forms, particularly the swords. The emergence of clearer associative patterns, particularly in the LBA hoard record, appears to be due to the greater functional differentiation of object types – with the function and meaning of axes, for example, becoming narrower after the introduction of specialised tool and weapon forms. While the associative variation of the greater range of gold ornaments could at first sight reflect their move from a communal, ceremonial context in the EBA into the personal realms, the pattern of dryland or bog deposition underlines the continuity of the underlying ideas attached to objects of this material.
Places and objects were brought together in a consistent manner (cf. Bradley Reference Bradley2000, 39; Yates & Bradley Reference Yates and Bradley2010), suggesting that particular places may have held particular relevance for ideas about different social personae. The differentiation of bog and river contexts in the case of bladed weapons and gold objects suggests a complex categorisation and understanding of these places and their properties, beyond wet and dry (cf. Yates & Bradley Reference Yates and Bradley2010). Places may have been seen as the abode of supernatural spirits or entrances into the otherworld, especially potentially liminal places such as lakes, rivers or bogs. While aspects of the character of wet places are more recognisable for us today, that of most dry places has become more or less invisible: trees, springs, etc, are known from anthropological examples (Bradley Reference Bradley2000) or classical sources (Torbrügge Reference Torbrügge1970) to have been special places in some societies. Only sites of an imperishable character such as outcrops, boulders, or caves, are still visible today. A good example is the association of hoards in Ireland with outcrops, boulders, or megalithic tombs (eg Toormore: O'Brien et al. Reference O'Brien, Northover and Cameron1990), as frequently observed in the EBA record (Becker Reference Becker2006). The reasons for the relevance of places thus goes clearly beyond their simple physical properties and may have held a symbolic as well as a functional dimension (see below).
How type-specific meaning creates structures in the record – single finds and hoards
Type-specific patterns and evidence for selective deposition demonstrate that single finds can be considered to be the result of deliberate deposition. If type-specific meaning is seen as the main structuring factor in the genesis of the record, their relationship to the hoard record is a central question. Associative patterns between certain types of hoards and particular types of context have been demonstrated and the consistency of find contexts for one-type hoards and single finds consisting of the same object types pointed out. The prime examples for this are swords or other bladed weapons, which also maintain a connection with wetlands in the hoard record. However, it appears that type-specific meaning can have different degrees of variability and determining force in its depositional consequences. For example, while the rules applied to the deposition of bladed weapons are very clear and restrictive, those for personal ornaments and tools seem weaker, evident from the greater variability in their associations and depositional contexts (Table 4). These artefacts can be deposited in combination and are found in a variety of contexts, whereas complete swords for example can only be deposited with other swords and in wet contexts.
It also appears that less ‘meaningful’ objects cannot be deposited on their own in multiple item deposits. For example, the small number of tools deposited singly, together with their association with ornaments and other artefacts in the hoard record, can be seen as indicating that their relevance in the depositional system was only acquired in combination with other artefacts. This is underlined by the fact that the artefacts that are most frequently deposited singly – particularly swords – are also those that, if they occur in the hoard record, are deposited in one-type hoards. Perhaps the different depositional strategies refer to different levels of need to protect the symbolic value of an individual type of artefact: most clearly visible in the case of bladed weapons. On the other hand, variation in depositional contexts and associations of axes seems to reflect their multi-functional character, which allows single deposition but also makes them suitable for inclusion in hoards in combination with other artefacts. Rather than just their different functions, it is the connection with certain identities – which in the case of the axes may have been variable – that appears to be the determining factor.
The intentionality of the deposit seems to determine its form and mode. They reflect in their articulation – primarily visible in the selection of artefact type and place – different roles, people, genders, occasions, and events. Variation in the occasion of the deposition, the depositional purpose, and the person or subject represented in the depositional act create the patterns in the record visible to us today. It is striking that that these rules transgress morphological boundaries of hoard and single finds and determine the patterns within both categories of finds – rendering them essentially redundant if it were not for the complex hoards that deviate by their very nature from the patterns observable across the single find and one-type hoard record.
Transforming identities – the function of deposition
Why were these symbols of social identities deposited? What was the bringing together of place and object meant to achieve?
A key to the answer may lie in the differences between the types of context considered suitable for deposition. The liminal character of particular wet contexts and the clear correlation of weapons with rivers has been pointed out; their absence from burials in Ireland may suggest that they do not stand for a fixed, permanent identity connected with the individual until death. Rather, it may be shed in a rite of passage which involved the deposition of the sword in a riverine or watery place (Fontijn Reference Fontijn2002, 230). Such an explanation would conceptually come close to considerations of hoards as an alternative burial rite that has been discussed as a possible explanation for the similarity and complementary relationship between hoard deposits and burials in some areas (Eogan Reference Eogan1983; Hundt Reference Hundt1955; Aner Reference Aner1955, 41). These deposits of bladed weapons may represent the burial or shedding of a particular social, impermanent identity.
The impermanence of social identities may also be visible in the depositional treatment of ceremonial items and particularly gold neck ornaments throughout the Irish Bronze Age. EBA lunulae were deposited in the landscape and are absent from burials of high status individuals of the period. Rather than interpreting this as evidence for their communal character, they may have been connected with a specialised identity not permanently affixed to an individual, but representative of a particular role taken on during a particular phase of the person's life. In the case of the gold neck ornament, this may have been a ceremonial identity which may not even have been continuous, but episodic and perhaps taken on or shed on a regular basis throughout the year or in a community's life cycle. This finds support in the frequent evidence for the repeated folding and rolling of these objects (Cahill Reference Cahill2005) and the frequent deposition on dry ground near recognisable places in the landscape such as outcrops or megalithic tombs, which would have facilitated their retrieval (Tables 5, 9a; Becker Reference Becker2008). The concealment, retrieval, un-rolling, and re-rolling and reconcealment of the item may have played a crucial role in the adoption and shedding of a particular, ceremonial identity This re-use may have extended over considerable time, with the object passed on through generations (cf. O'Brien Reference O'Brien2012, 223) and with it the particular role or identity embodied in it. Similarly it is argued here that the deposition of LBA ceremonial items is a reflection of a similar character of use, for which permanent abandonment in a river would not be suited.
Thus it is possible to suggest that, rather than the representation, the transformation of identities is the purpose of these deposits and place and retrieval potential correlate to different objectives and contexts of these transformative processes. This also opens up a way of thinking about those deposits that have evaded the type-specific categorisations, due to their mixed nature – the complex hoards.
Transforming objects – recycling metals
In spite of there being, with the exception of the Co. Roscommon hoard, no classic founder or scrap deposits in Ireland, a small group of complex hoards contain broken objects, scrap, and raw materials. Similar patterns with limited evidence for scrap hoards and mixed hoards can also be observed in parts of Denmark (Kristiansen Reference Kristiansen1996, 258–9) and has here been related to the availability of metal, which also may be a key factor in the lack of proper large scrap hoards in Ireland.
The Irish complex hoards were predominantly deposited in dry contexts, and often in outcrops, stony, and rocky places, rock clefts, or caves – a pattern particularly notable in the EBA. In these hoards, types are combined which are otherwise stringently kept apart. Objects such as LBA swords or horns, which follow strict depositional rules in the single find record and in single-type hoards, can occur mixed with other objects deposited in dryland contexts. The absence of any association of swords with other artefacts in ‘normal’ hoards indicates that, in the case of scrap hoards, the sword has lost its original meaning and can be deposited with other types of object. One could argue that the breaking of objects removes their symbolic integrity and reduces them to their pure material value, lifting any restrictions guiding the treatment of the complete items.
While the overall composition of the scrap hoards may not be subject to strict compositional rules, the treatment of swords appears strictly regulated. In the period with the most clearly identifiable patterns, the LBA hoards containing swords are, with the exception of the Teernagloghane, Co. Clare hoard, characterised by a very specific feature: they contain a single fragment of a sword, usually in combination with complete and still usable artefacts (eg, Park, Co. Meath: Fig. 3a). It has been suggested that scrap hoards represent collections of raw material to be melted down and re-used as needed (eg, Huth Reference Huth1997). However, it has also been argued that the fact that the pieces do not match should be seen as an intentional pars pro toto deposition of the artefacts (eg, Hansen Reference Hansen1994).
The fact that only one sword fragment was ever deposited in each hoard suggests that an intentional selection was made, rather than random scrap collection. Patterns indicating that artefacts in scrap hoards were subject to specific selection and treatment have also been noted elsewhere (Turner Reference Turner1998) and, specifically, the mode of deposition of sword fragments has also been seen to display patterns elsewhere (Bradley Reference Bradley2005, 154–60; Rittershofer Reference Rittershofer1984; Verlaeckt Reference Verlaeckt1998).
The repetitive inclusion of single sword fragments in the dryland scrap hoards of the LBA can perhaps be best understood in terms of the destruction of a particular form of identity. The inclusion of raw material in some of these hoards, in combination with the sword fragments, reinforces the notion that an act of transformation was important. The fact that they were deposited in dry contexts, often in significant locations, might imply that the deposition of only part of a sword could be understood as the legitimisation of the transformation of a highly symbolic object. As Barrett and Needham (Reference Barrett and Needham1989, 138) suggested for the British Carp's Tongue hoards, the breaking of a sword may have been a requirement for their transformation into a commodity. However, the Irish evidence further indicates that not only breaking, but also a very specific mode of deposition, was required to legitimise this act and that it was not just about commoditisation, but also the symbolic properties of the object.
The significance of rocky places and outcrops may be linked to their connection with the process of raw material procurement, so the deposition of raw materials and scrapped artefacts at such places could be seen as completing the cycle. As frequently argued (eg Budd & Taylor Reference Budd and Taylor1995; Reid & MacLean Reference Reid and Maclean1995; Rowlands & Warnier Reference Rowlands and Warnier1993), the transformative process of metal production was possibly accompanied by a variety of rituals and taboos (cf. Bradley Reference Bradley2005, 150–1; Brück Reference Brück2001). As anthropological comparisons demonstrate, smelters and smiths sometimes regard themselves as facilitating a natural process of transformation (Rowlands & Warnier Reference Rowlands and Warnier1993). The fragmentation of the artefacts in scrap hoards has been seen as the ritual symbolisation of this process of transformation (Brück Reference Brück2001, 156–7): deposition might symbolically return to nature what had been taken from it. It is not the relinquishing of the artefact, but the transformation of its social dimension and the marking of liminal events that are facilitated by deposition, at the same time legitimising and facilitating the transformation from object to raw material. It has also been argued that objects were broken and kept as tokens of economic or social transactions and relationships (Chapman Reference Chapman2000, 38) which could be reconciled through deposition (ibid., 6). However, this appears unlikely for the Irish material as the fragmented artefacts are accompanied by complete artefacts and breakage is mainly restricted to swords. It seems more likely that the breaking and depositing of a single sword fragment is a reference to the act of transformation of an otherwise strongly regulated and protected object and, to legitimise its transformation, part of the artefact had to be consecrated while the missing fragments could be melted down.
The importance of copper and bronze relates not only to their functional advantages as a material but also to the physical properties that allow fragmentation and their potential for recycling and regeneration (Chapman Reference Chapman2000, 6). These properties make it likely that metalworking gained a strong metaphorical potency for conceptualisations of regeneration and death and birth (cf. Budd & Taylor Reference Budd and Taylor1995). Smelting as a transformative and regenerative process is in some cultures symbolically and metaphorically linked to human reproduction (Reid & MacLean Reference Reid and Maclean1995, 149–50). The link between fragmentation and regeneration has been made elsewhere (Brück Reference Brück2006; 2001; 1999; Turner Reference Turner1998, 134) and the deposition of fragmented metalwork as marking out liminal states (Brück Reference Brück2006; Turner Reference Turner1998, 122, 129). Turner saw scrap hoards as symbolising a transformation from life to death (ibid., 115–16), arguing that objects lose their meaning through destruction (ibid., 118).
However, the special emphasis on the treatment of swords could be considered indicative of a specialised transformative occasion, connected both to metalworking and male identity. Instead of being a direct reflection of metalworking episodes (contra Turner Reference Turner1998), which is also unlikely in view of the almost complete absence of other items associated with metalworking in the Irish hoards, they appear to represent a means of legitimising the transformation of artefacts subject to strict regulations. It may have been dangerous and difficult to destroy and combine them into new artefacts (cf. Bradley Reference Bradley2005, 163). A ritual act may thus facilitate a profane purpose – the transformation and recycling of both item and its symbolic property, in order to facilitate a utilitarian re-use of the object.
Transforming foreign identities
Gold hoards sometimes contain broken objects. The most striking examples are the MBA finds from Downpatrick, Co. Down. Two hoards containing both raw material and finished objects were deposited in a prominent location on a hilltop, each covered with a stone. The alloys of which the objects were composed are similar enough to suggest that they were produced over a short period of time, possibly in the same workshop. Four objects in hoard 1 – fragments of a torc with clear cut marks, a decorated bracelet, and two more bracelets – form a separate group with a higher copper content (Shell et al. Reference Shell, Taylor, Warner and Watling1998, 253). The two bracelets appear to have been cast from the missing half of the torc. It has been argued that the inclusion of the incomplete torc indicates an intention of future use and a founder's hoard (ibid., 256–7). The torc is so far the only of its type from Ireland and could be an import from France (ibid., cf. Armbruster Reference Armbruster2010, 135–8). A connection with metalworking is compelling in view of the compositional evidence, and the scrapping of a foreign object for the production of local types has been suggested elsewhere. In the Netherlands imported Geistingen axes were probably converted into local types and their deposition seen as facilitating or referencing this transformation process (Fontijn Reference Fontijn2002, 254–5; 2008), building on observations and suggestions made elsewhere about the treatment of foreign objects outside their original area of production and use (eg, Bradley Reference Bradley1998, 125–7; Reference Bradley1985, 697). The cargo of the Langdon Bay shipwreck, which contained objects such as Taunton-Hademarschen type axes and tanged and collared chisels, was found outside the original area of distribution of the artefacts and was probably also destined to be converted into locally acceptable forms (Muckelroy Reference Muckelroy1981, 295); similar processes may have led to the deposition of the material in the Dutch Voorhout hoard (Fontijn Reference Fontijn2008). The Downpatrick find most clearly demonstrates the actual transformation from foreign into local and rather than seeing this as a random snapshot of a particular step in the production sequence, it is suggested that the act of deposition deliberately references and legitimises the transformation of a foreign identity.
If gold artefacts stood for a very specific local identity, the need to transform imported items into appropriate forms becomes understandable. This seems to be reflected in other hoards such as the Bishopsland find or the fragment of a possibly Portuguese gold torc at Crieve-row, Co. Armagh (Stuart Reference Stuart1819, 512, fig. opposite p. 513), and could explain the general lack of imported goods alongside the clear evidence for contact with the outside world throughout the Bronze and Iron Age of Ireland.
Thus, collections of finds with a clear connection with dry contexts may represent deposits that are concerned with the transformation of items that stand for a particular identity.
CONCLUSION
The comprehensive examination of the Irish Bronze Age artefact record has revealed that single finds, burial finds, and hoard finds constitute different aspects of one depositional system in which single finds can also be demonstrated as having been deposited intentionally. The correlation of specific forms of artefact deposition with certain contexts indicates that deposition was a meaningful practice, in which distinct qualities and meaning were attached to both the artefacts as well as the depositional contexts. Type-specific depositional patterns reflect rules that were in place for different types of objects (cf. Kristiansen Reference Kristiansen1996, 256). While the retrieval potential of places seems to be of relevance in the emergence of associative patterns, the specific symbolic meaning of the different places seems to have differed as well and challenges us to look beyond the simple dry/wet dichotomy which the record at first sight appears to represent.
Different artefacts were affected to differing degrees by the recycling or deposition strategies. The lack of tool deposition, for example, indicates that a higher proportion of these were recycled than of axes or weapons. Tools and personal ornaments also show a greater variability in their associations and depositional contexts, in contrast to very clear and restrictive patterns of deposition for bladed weapons. The small number of tools deposited singly, in conjunction with their association with ornaments and other artefacts in hoards, suggests that their relevance in the depositional system was only acquired in combination with other artefacts and their own meaning was less strictly defined. This difference to other types is accentuated by the fact that the artefacts that are most frequently deposited singly are also those that, if they occur in the hoard record, are deposited in one-type hoards – such as swords or spearheads. It would therefore seem that the different depositional strategies might refer to different levels of the need to protect the symbolic value of the artefact: this is most clearly visible in the case of bladed weapons, which show the most exclusive patterns. In contrast, the great variability of depositional contexts and associations of axes seems to reflect the multi-functional character of these tools, which allows single deposition but also makes them suitable for inclusion in hoards in combination with other artefacts. With the development of specialised tool and weapon forms in the LBA the functional and symbolic narrowing of the conceptualisation of the type ‘axe’ becomes visible in the record in form of more specific contextual patterns.
The bias of weapons towards river contexts and gold and ceremonial items towards dry or bog contexts, as well as their exclusive associative patterns in the hoard record, suggest that they represented different identities, which were perceived as being of a contrasting nature. Gold and ceremonial artefacts appear to reflect communal concerns. What defines and characterises these items is that they appear to stand for identities which are not linked to the individual, as demonstrated by the EBA record and, indeed, also the lack of grave goods altogether in the later phases of the Bronze Age. While also referencing metalworking, ‘scrap’ hoards can be argued to reference and legitimise the transformation of special types of artefacts, such as highly symbolic or foreign items. The interdependency of the different classes of depositional context implies that we should, indeed, see the different modes of deposition as part of one system, with the individual modes reflecting different people, roles, genders, occasions, or events.
The contextual associations of the different artefacts in turn also suggest that places were imbued with specific meanings and differentiated not only between wet and dry but also between river and bog contexts: such an ideational landscape might encompass the notion of sacred places (cf. Knapp & Ashmore Reference Knapp and Ashmore1999, 12–13). The correlation between artefacts and locations points to significant combinations of artefacts and places with symbolic meanings, which were intentionally brought together (cf. Bradley Reference Bradley2000, 39). The equation of watery contexts with permanent surrender and dry land with potential retrieval is too simplistic to explain the complexities of the record and the correlation between find context and artefact type has a symbolic dimension that cannot be reduced to the simple physical properties of the find context. The difference in the treatment of weapons and gold artefacts is the most striking example of this. They have in common the fact that deposition in the ground or in water assigned the respective artefacts to environments that appear to have had a significance that went beyond the mere physical fact of ‘solid ground equals retrievable’ or ‘water equals irretrievable’.
However, the potential for retrieval should not be disregarded or considered unimportant. While the deposition of artefacts in rivers is a clear example of permanent disposal, the deposition of scrap material in dry contexts, often at marked locations in the landscape, represents the far end of the spectrum of possibly quite variable recovery intentions. It is therefore argued here that deposition was a transformative act – the transformation of the artefacts themselves, and of the specific identities symbolised by them, through deposition in liminal places such as bogs, rivers, rock outcrops, or burial monuments, bridging thus the divide between ritual and functional practices. In fact, the possibility of retrieval may have been important in the case of both the gold and other ceremonial items as well as the complex hoards – if, however, with different objectives.
Type-specific deposition is a reflection of the complex ideas attached to the significance of artefacts, and their symbolic meaning in the conception of personal identity and transformation processes in general. The differences in the treatment of the various types of artefact are seen to be a reflection of their inherent meaning and the social and functional context of the transformation process with which they are connected. It is only by breaking through the artificial boundary between the profane and ritual concepts that a coherent interpretation of the practice in general, including an explanation for the scrap hoards, becomes possible.
Acknowledgements
This paper is based on PhD research undertaken in University College Dublin in 2006. I would like to express my deep gratitude to my supervisors Barry Raftery and Joanna Brück, UCD. My thanks are also due to three anonymous referees for their constructive critique of an earlier draft of this paper. Many individuals and institutions were generous in providing access to material and information on unpublished finds: Muiris O'Sullivan, Gabriel Cooney (UCD), The National Museum of Ireland, Ulster Museum, Belfast, British Museum, London, Cambridge Museum, Ashmolean Museum, Oxford, Clare County Museum. Several colleagues have generously shared their knowledge with me: Ian Armit, Holger Becker (Landesmuseum Bonn), Sue Bridgford, Mary Cahill (National Museum of Ireland), George Eogan, Eoin Grogan, Hans-Eckart Joachim (Landesmuseum Bonn), Ronan O'Flaherty, Paul Mullarkey, (National Museum of Ireland), Stuart Needham and Richard Warner, Beverley Hirschel, John Ó Néill, Brendan O'Connor, Conor McDermott, Rob Sands, Stephen Harrison and Blaze O'Connor. The research was funded by a PhD scholarship by the Deutscher Akademischer Austauschdienst (German Academic Exchange service) followed by a one-year scholarship from the Irish Council for the Humanities and Social Sciences as well as a UCD Evening Degree Teaching Fellowship.
