Non-funerary metal deposition in Bronze Age Scotland Rachel Faulkner-Jones Graduate School of History, Classics and Archaeology, University of Edinburgh Supervisor: Dr Magdalena Midgley Acknowledgements Many people have helped with this research, both professionally and personally. I am very grateful to Trevor Cowie for his time, unparalleled access to the material held by the National Museum of Scotland, his insights and guidance during this study, without which much of the work undertaken in Chapter 3 would have been impossible. Dr Magdalena Midgley and Professor Ian Ralston, for supervision, support, suggestions and care. The curatorial and administration staff at the National Museum of Scotland for their timely responses and encouragement. Dr Sally-Anne Coupar at the Hunterian Museum, University of Glasgow, and Jean Singleton at the Scottish Borders Council Museum and Gallery Service, Tweeddale Museum & Art Gallery, for permission to use photographs of some of the artefacts held by their respective institutions. Luise Pirie at Cambridge University and all of the contributors to the Creativity in the Bronze Age conference, Magdalene College, Cambridge, April 2013, for their feedback and support when part of this paper was presented, and for invaluable insights and suggestions. John McKenna of John McKenna Sculpture Ltd, High McGowanston Studio, Turnberry, Girvan, Ayrshire, KA26 9JT, for help with the limitations and capabilities of modern and prehistoric bronze forging and casting. The staff at the Royal Commission for Ancient and Historic Monuments of Scotland, particularly their web team who keep the digital records free, accurate and up-to-date. Sarah Jones, for multiple instances of proof-reading, encouragement and support. The post-graduate community and Archaeology Society at the University of Edinburgh, for providing an outlet and a much-needed, if sporadic, social life. Alan Faulkner-Jones, for not giving up. Table of Contents List of Illust atio s………………………………………………………………………………………………… Chapte : I t odu tio to the “ ottish B o ze Age……………………. 1.1 Introduction 1.2 Scotland in the wider world 1.3 Patterns in research 1.4 Methodology: constructing the corpus 1.5 Initial theoretical considerations Chapter 2: Waterways and prominent places: the natural world and Bronze Age depositions 2.1 Landscape and archaeology: a theoretical framework 2.2 Bronze Age water deposition 2.3 Early Bronze Age deposition, 2300-1500BCE 2.4 Middle Bronze Age deposition, 1600-1000BCE 2.5 Late Bronze Age deposition, 1100-800BCE 2.6 Riverine deposition in Britain 2.7 Identifying a prominent deposition 2.8 Early Bronze Age 2.9 Middle Bronze Age 2.10 Late Bronze Age 2.11 Frequency of deposition 2.12 Long-term patterns of Bronze Age deposition: an overview 2.13 Conclusions Chapter 3: Proximities and depositions: ritual and settlement sites in the Bronze Age 3.1 The world as lived and the world as imagined: ritual in the Bronze Age 3.2 What constitutes a ritual? 3.3 Ritual and metallurgy 3.4 Bronze Age ritual 3.5 Ritual and the non-funerary depositions 3.6 Case study: ritual and deposition in the Edinburgh area during the Late Bronze Age 3.7 Ritual sites: interim conclusions 3.8 Settlement in the Bronze Age 3.9 Case study: settlement and deposition in Midlothian 3.10 Methodological considerations 3.11 Conclusions Chapter 4: Conclusions, Issues and Suggestions for Further Work 4.1 Conclusions 4.2 Issues 4.3 Future work Bibliography Appendix: Data Sets List of Illustrations 2.3.1. Early Bronze Age dirk from Loch Glashan 2.4.1. Map showing the locations of the two Middle Bronze Age estuary deposits 2.5.1. Graph demonstrating the differences between water and hoard depositions throughout the Bronze Age period . . “il e pe da t f o Mjolnir the latte half of the fi st ille iu AD depi ti g Tho s ha e 2.5.3 Legendary swords in modern mythology: Neville Longbottom and the Sword of Gryffindor from the Harry Potter series 2.6.1 Das Irrlicht, oil on canvas by Arnold Bocklin 2.7.1 Map showing the areas of Scotland over 350m above sea level 2.7.2 Drummossie Moor, to the east of Culloden 2.8.1 Crude flat axe or axe-shaped ingot from Hillhouse, Lauder 2.9.1 Rapier from the Drumcoultran hoard 2.10.1 Prominently-placed hoard depositions throughout the Bronze Age 2.10.2 Horsehope Craig, Peebleshire 2.10.3 Selection of artefacts from the Horsehope Craig hoard 2.10.4 Socketed axehead from the Horsehope Craig hoard 2.10.5 Wagon fixtures from the Horsehope Craig hoard 2.10.5 Horse or wagon fixtures from the Horsehope Craig hoard 2.11.1 Graph showing depositions made per year in Aberdeenshire throughout the Bronze Age 2.11.2 Graph showing the intervals between depositions 3.3.1. Map showing the relevant ore sources in the British Isles 3.3.2. Environmental reconstruction of the Edinburgh area during the Bronze Age 3.3.3. Table showing one poor example of ritual oversimplification 3.4.1. Dagger from a cist burial at Ashgrove, found with meadowsweet pollen 3.4.2. The composite male mummy from South Uist 3.4.3. Satellite map of Kilmartin Glen 3.4.4. Rock art at Kilmartin Glen 3.5.1. Pictish cross-slab engraving from Invergowrie, Perthshire 3.5.2. Map showing iron artefact hoards in Britain 3.6.1. . . . Map of Quee s D i e, A thu s “eat a d G os e o C es e t, Edi . . The t o s o ds f o u gh Quee s D i e, NM“ X.DQ90 and NMS X.DQ91 3.6.4 The burnt fragments of the Duddingston Loch hoard currently on display in the National Museum of Scotland, Chambers Street, Edinburgh 3.6.5 Bent tip of blade NMS X.DQ303 from the Duddingston Loch hoard . . The lo e u kle poi t o lade NM“ X.DQ 3.6.7 The second area of buckling on the inner curve of blade NMS X.DQ303 3.6.8 Buckling above the major bend on blade NMS X.DQ303 3.6.9 The complete sword NMS X.DQ303 3.6.10 NMS X.DQ305, the second-largest artefact from the Duddingston Loch hoard 3.6.11 V-shaped notching on blade NMS X.DQ305 3.6.13 U-shaped notching on blade NMS X.DQ305 3.6.14 Partial blade NMS X.DQ304 from the Duddingston Loch hoard 3.6.15 Blade NMS X.DQ304 in profile 3.6.16. The socketed spearhead tip NMS X.DQ302 from the Duddingston Loch hoard 3.6.17. The hoard from Grosvenor Crescent 3.6.18 V and U shaped notching on blade NMS X.DQ201 from Grosvenor Crescent 3.6.19 Hilt damage on blade NMS X.DQ237 from Grosvenor Crescent 3.6.20 U-shaped notch on blade NMS X.DQ237 . . The lea side of the lade NM“ X.DQ 3.6.22 The hilt of NMS X.DQ236 from Grosvenor Crescent, showing the incompletelypunched rivet hole 3.6.23 Minute edge damage on sword NMS X.DQ236 3.6.24 Minute edge damage on sword NMS X.DQ200 from Grosvenor Crescent 3.6.25 Edge damage (notching) on sword NMS X.DQ306 from Grosvenor Crescent 3.8.1 Site plan of West Acres excavation 3.8.2 Location of the Bronze Age house at Navidale in relation to the stream Alte Briste, Sutherland 3.9.1 The Duddingston Loch hoard 3.9.2 GIS-generated viewshed analysed image showing the areas visible from the settlement sites in Midlothian 3.9.3 GIS-generated viewshed analysed image showing the areas visible from the deposition sites in Midlothian 4.2.1. Orrery by April Wernham, based on the Trundholm Sun Chariot Acknowledgements Many people have helped with this research, both professionally and personally. I am very grateful to Trevor Cowie for his time, unparalleled access to the material held by the National Museum of Scotland, his insights and guidance during this study, without which much of the work undertaken in Chapter 3 would have been impossible. Dr Magdalena Midgley and Professor Ian Ralston, for supervision, support, suggestions and care. The curatorial and administration staff at the National Museum of Scotland for their timely responses and encouragement. Dr Sally-Anne Coupar at the Hunterian Museum, University of Glasgow, and Jean Singleton at the Scottish Borders Council Museum and Gallery Service, Tweeddale Museum & Art Gallery, for permission to use photographs of some of the artefacts held by their respective institutions. Luise Pirie at Cambridge University and all of the contributors to the Creativity in the Bronze Age conference, Magdalene College, Cambridge, April 2013, for their feedback and support when part of this paper was presented, and for invaluable insights and suggestions. John McKenna of John McKenna Sculpture Ltd, High McGowanston Studio, Turnberry, Girvan, Ayrshire, KA26 9JT, for help with the limitations and capabilities of modern and prehistoric bronze forging and casting. The staff at the Royal Commission for Ancient and Historic Monuments of Scotland, particularly their web team who keep the digital records free, accurate and up-to-date. Sarah Jones, for multiple instances of proof-reading, encouragement and support. The post-graduate community and Archaeology Society at the University of Edinburgh, for providing an outlet and a much-needed, if sporadic, social life. Alan Faulkner-Jones, for not giving up. Introduction to the Scottish Bronze Age 1.1 Introduction Research into the Scottish Bronze Age is currently undergoing something of a renaissance, with the development of the period-specific section of the Scottish Archaeological Research Framework (ScARF, available online at www.scottishheritagehub.com), the Scottish Wetland Archaeology Programme (Crone and Clarke 2005), many new publications, and the application of modern analytical techniques. That being said, there has not yet been a systematic appraisal of non-funerary metal depositions in the country as a whole, which sets Scotland at a slight disadvantage when compared to the rest of Britain, and much of Europe. This paper is intended to be a preliminary investigation into Scottish metal depositions comprising axes, swords and spearheads from non-funerary contexts, any observable distribution patterns, and the wider context in which Bronze Age artefact deposition can be set. It is also intended to show that depositions were not discarded, lost or hidden for purely economic reasons, through analysis of the findspots themselves and case-studies of specific sites and artefacts. It is not intended or to be a complete overview, as there is much research yet to be done on a multitude of aspects of the Scottish Bronze Age, but more of a basis from which further, more detailed work of a wider scope can be undertaken in the near future. 1.2 Scotland in the wider world Scotland has occasionally been marginalised in continental studies of the Bronze Age, to the extent where it is a el e tio ed i K istia se s The Rise of Bronze Age Society (2005), despite having a comparable land mass to Ireland, which tends to dominate British Bronze Age studies due in part to the exemplary native material record. The Scottish record is not as complete as elsewhere in Britain, and the relatively sparse population means less infrastructural development compared to, say, the Thames Valley1, leading to distinct patterns of recovery bias in the archaeological reco d elati g to pollute -pa s , o developer-led, practices over the past half-century (Everill 2007:119-120). Trade and communication within Britain and across Europe evidently flourished in the Neolithic and continued into the Bronze Age. The potential links between Ireland and Scotland become apparent in the settlement record (Cavers 2006) and are being systematically investigated ith ega ds to hal e d a ufa tu e, use a d depositio O Flahe t , Gil h ist a d Co ie 2011). As Scotland is geologically largely non-copper-bearing, extensive and robust trade links with Wales and Ireland must have been the minimum long-distance contact at the time (Rohl and Needham 1998:16-7), though a good deal of research is required on trade networks moving goods in and out of Scotland. Lead isotope analysis shows that much British bronze material was made from native metals; though Scotland was largely excluded, 1 An area developed to such an extent as to require its own archaeological team, the Thames Valley Archaeological Services, which oversees developments and planning applications (www.tvas.co.uk). there was some evidence for an ore source in lowland Scotland, at Tonderghie (a current ScARF research project), which could have been mined during the Bronze Age; in addition, small surface outcrops deemed unprofitable to mine today would almost certainly have been exploited during the Bronze Age (Rohl and Needham 1998:179, Rice 2004 passim). It is worth noting at this early stage that the dictionary definition of o e a eed so e modification to make it applicable to prehistoric research. The Oxford English Dictionary states that a sou e of etal a o l e o side ed a o e if it is p ese t i su h ua tit a d fo as to ake its e t a tio p ofita le OED : . This appli atio of a modern capitalist-i flue ed i te p etatio of p ofita le ill a ildl i p ehisto i , o apitalist so ieties, a d he ide tif i g ati e o e sou es, the o ept of p ofita le ust be disregarded, or at least handled with extreme care, and smaller ore sources need not be overlooked due to their size. A full metallic analysis from the Scottish assemblage would be phenomenally useful and should be considered a future research topic, and is listed as a ScARF recommendation for future research (website listed in bibliography). 1.3 Patterns in research The early twentieth century was the age of great syntheses, undoubtedly dominated by V. G. Childe (1935 and 1946) and Stuart Piggott (1958 and 1962), the grand masters of Scottish prehistoric research; though admirable contemporary efforts, they are now very dated and culture-history-centric (a resurgence in the culture-history paradigm can be seen in some post-p o essual app oa hes: see Hodde s o k o Çatalhöyük, for example (Hodder 2004) and the merging of interpretative and artefact studies (Mödlinger, Uckelmann and Matthews 2011:6)). The classification of hoards into foundry, smith, merchant etc. based on contents ascribes a purely economic motive for deposition and is highly utilitarian: though this model has fortunately been mostly disregarded now, there are a handful of remaining subscribers (Harding 2000). These large-scale artefact studies resulted in the publication of extensive and detailed corpora, most notably the Prähistorische Bronzefunde series in the 1980s. There was also a flourishing of metal analysis during this period, especially at Stuttgart, some of which was unfortunately lost during the reunification of East and West Germany, though enough survived to be reassessed and updated in recent studies (Kiss 2009:330). The move into more predominantly interpretative approaches occurred shortly after this, as landscape archaeology and phenomenology became popular in the 1990s. These are both interesting approaches to apply, though not specifically related to Bronze Age studies (Tilley 1994), but are not without complications in their application (Fleming 2002, Hamilton et al 2006), and have echoes in the viewshed analysis undertaken in Section 3.9. It must be noted at this point that the dualistic and segregated categories of economic/ritual and sacred/profane which have been used extensively in much of the work undertaken so far, is neither useful nor applicable in many cases (Bradley 1998:154; see also Section 1.5). Current research tends towards the localised case study approach rather than the grand overviews of Childe and Piggott, and growing interest in the domestic and settlement record as opposed to the more visually impressive hoards and burials, though these are of course still of academic value (Cowie 1988, Brück 2001 and 2008). Needham in particular has been instrumental in realigning the older chronologies with new dating methods and metallic isotope analysis (Needham 1988, Rohl and Needham 1998). Use-wear analysis of weaponry and tools affo ds a u h lea e ie of the fu tio of a tefa ts p io to thei depositio O Flahe t , Gilchrist and Cowie 2011), and detailed reappraisals of the artefacts, combined with interpretative and experimental approaches are being undertaken across Europe. The time is now ripe for a reappraisal of existing work in Scotland, and for some headway to be made into new research areas. 1.4 Methodology: constructing the corpus The research undertaken in this paper is based on existing archaeological data, collected and examined in a new way. The corpus of data (included in electronic format in the Appendix) as asse led usi g e a atio a d useu data listed i Coles o ks (Coles 1960, 1964 and 1969), two papers from the Prähistorische Bronzefunde (Schmidt and Burgess 1981, Colquhoun and Burgess 1988), the annual reports from the Treasure Trove (available online, full web address in the bibliography) and recent articles from the Tayside and Fife Archaeological Journal (TAFAJ) and the Proceedings of the Society of Antiquaries of Scotland (PSAS). Research in Scotland is therefore at a distinct advantage, as these corpora – though not perfect – are readily accessible and reasonably comprehensive, unlike (for instance) embarking upon a similar project in the Low Countries, where such databases are as yet unknown (Fontijn 2002:1). Each entry in this corpus comprises one metal artefact find or hoard, its National Grid Reference (NGR), the contents of the find, setting of the findspot and the corresponding Royal Commission on the Ancient and Historical Monuments of Scotland (RCAHMS) and/or PSAS page, if applicable. As many of the NGRs were recorded as possible, ideally up to 8 digits, giving a 10m2 area in which to work, though this obviously depended on the quality and detail of the existing records. County assignation follows the pre-1973 government districts, as most artefacts were catalogued prior to 1973 and smaller subdivisions of land make for better scales of study, especially when they can all be compared in a national overview. The Inner and Outer Hebrides and Shetland Isles are, for the most part, side-lined: there are few relevant finds, and are logistically very difficult to get to for use as case studies – where hoards or single artefacts fit the criteria listed in section 1.5, however, they were considered and included, regardless of their geographic origin. All of the data used is available and accessible by the public, and the corpus is included as an electronic appendix in this paper so that it can be consulted and scrutinised. Furthermore, this project and the data were presented at the Creativity in the Bronze Age (CinBA) conference in Magdalene College, University of Cambridge, in April 2013 for interim peer review. None of the data contravenes the Ethics Guidance established by the University of Edinburgh: no human or animal remains were discussed, and no interviews were conducted with finders or landowners.2 Methodological considerations specifically relating to the two case studies are discussed in more detail in the relevant sections. The corpus was divided into Early, Middle and Late following typo-chronologies established by Coles (1960, 1964, 1969), Burgess and Gerloff (1981), Colquhoun and Burgess (1988) and Schmidt and Burgess . Ba e s h o olog is follo ed, he e the B o ze Age is divided into Early: 2300-1500BCE, Middle: 1600-1000BCE and Late: 1100-800BCE (Barber 2002:42), corroborated by the B itish Museu s h o ologi al guideli es (http://finds.org.uk/bronzeage). The overlap allows for stylistic continuity and regional discrepancies. Anything dating absolutely or stylistically to after 800BCE is considered early Iron Age, as the transition to an iron-based economy was underway and the concurrent changes in cosmology and society were being established; conceptually ending the Bronze Age at this point is common in modern papers (see Stevens 2008:239). A handful of artefacts listed in the earlier corpora were not included because they were one of the following: fake; insufficient provenance or had no recorded NGR and therefore effectively useless to this study; grave goods; or subsequently re-dated to the Iron Age or post-Roman Atlantic Iron Age. These incidences were few and far between, but their exclusion meant that when very specific factors were investigated, small data sets would not be affected by erroneously-included sites and artefacts (see, for instance, the discussion of high and prominently-placed findspots, where the data sets in question were around 25 entries). 1.5 Initial theoretical considerations With regards to the data, the standpoint itual u til p o e othe ise Fo tij 2002:27) is disregarded; rather, artefacts included in the corpus fulfil one or more of the following criteria which makes them arguably ritual in nature: a) the artefact is part of a hoard – the purely economic theories of hoard deposition (see above in Section 1.3, and below) have been rejected as based in modern economic theory b) the artefact has been placed in the ground in a specific way (hilt up, aligned with the compass points, in a river, in a bag or bucket), or in a ritualised landscape – inside a (contemporary or pre-existing) stone circle, or where a sacred well was placed in later prehistory, for instance c) the artefact shows signs of extensive use, but is still in a useable condition 2 The full text of the University Ethics procedures can be found online at http://www.ed.ac.uk/schoolsdepartments/history-classics-archaeology/research/ethics-procedures/principles, which was last updated 14/2/2013. d) the artefact has been rendered completely unusable through fire or deliberate breaking. See Kristinsen 2002 for an overview of the differences between use-wear damage and deliberate snapping or breaking of a blade, and Illustration 3.6.9 for a superb example of deliberately broken or decommissioned hoard artefacts. The concept of removing bronze from circulation in order to preserve its value and keep prices or its prestige status artificially high is inherently post-industrial and heavily influenced by western capitalism, and would not work in a pastoral prehistoric system undergoing major economic upheaval where value systems are unknown (Stevens and Fuller 2012:708); it is a textbook example of post-Marxist theory (Darvill 1995:38). The hoards as ritual is justified as follows: the concept of peripatetic smiths and merchants is not upheld ethnographically (Harding 2000:236) and is an assumption made by archaeologists, given the relative lack of knowledge about Bronze Age social structures in Scotland (equivalent systems in Scandinavia are well documented, and may be worth investigating as a parallel or complementary model for Scotland (i.e. Kristiansen 1987)). The lack of evidence for widespread unrest in Scotland throughout the whole of the Bronze Age makes the deposition of hoards for safekeeping during revolts, invasions or tribal wars, to be retrieved later, unlikely: furthermore, this assumes that bronze was very valuable economically: why, then, were the hoards not dug up? Some (such as Duddingston Loch) were clearly not meant to be retrieved, being deposited in the centre of deep lochs. The major argument in this paper is that the specific depositional and artefact patterns, of hoards and single items, are not coincidental, accidental or economic in nature. Blunt or broken weapons could have been recycled, as evidenced by the lead isotopes from later Bronze Age items (Rohl and Needham 1998:88-9), so hoards cannot be seen as special metal refuse sites, or middens, either. For these reasons, all known hoards are considered as having some ritual aspects dictating their deposition. It is also worth considering that the distinction between ritual and profane which has been so long discussed with relation to the depositions could be argued to be entirely meaningless, as the items end up in the archaeological record regardless and drawing artificial lines between sacred and profane closes many potential avenues of research (Bell 2009, Helms 1993, Fontijn 2002 17-8). Although this is a valid consideration, it will be contended in this paper that the circumstances in which the item goes into the ground are significant, and can be identified through close examination of specific depositions and the wider context in which they are set: that is, after all, one of the major aims of this paper. This will be investigated primarily by examining the natural landscape in which the deposits were located, which will be the focus of the next chapter. Chapter 2 Waterways and prominent places: the natural world and Bronze Age depositions The significance of water in establishing depositional patterns, and in particular the importance of rivers, is a recurring theme in Bronze Age studies (Brück 2011:389, Mullin 2012, Yates and Bradley 2010, Cowie and Hall 2001, Cavers 2006:404). Furthermore, the Bronze Age can arguably be seen as an era of major landscape change (Parker Pearson 1993:60), in which the natural surroundings were reconceptualised and utilised by Bronze Age communities. The landscape would therefore be a logical place to start when discussing possible factors influencing artefact deposition. 2.1. Landscape and archaeology: a theoretical framework Landscape theory has become a popular interpretative paradigm, and several concepts must consequently be clarified before any analysis or discussion can begin. Landscape in itself is a cultural construct, born in the Italian city-states of the Renaissance and given new life by sixteenth century Dutch artists, and is (at this point) irrevocably dependent on the cultural make-up of the viewer (Cosgrove 1984:69, Andrews 1999:4,): modern, privileged, able-bodied Europeans will view the world in a very different way to someone with a physical disability, impaired sight or the cultural traditions instilled by growing up in the Bronze Age. Etymologically, la ds ape i plies a le el of sepa atio , i order that the observer might observe it, conferring a level of dislocation between the person and their surroundings that would not necessarily have been present in prehistory (Williams 1973:120). That being said, it would be a huge intellectual leap to argue that people in prehistory did not interact with their natural world in a variety of different ways, so e of hi h a ha e i ol ed og iti el sepa ati g o eself f o o e s su ou di gs. In this paper, the efo e, the o d la ds ape ill e used as a sho t-hand way of conveying the immediate surroundings, and the known natural world in which one lives, with a elatio al sig ifi a e eated th ough elatio s et ee people a d pla es Tille 1994:11). It is not meant to convey any aesthetic or implied cultural values outside of this definition. Phenomenological approaches are also treated with caution: although very useful in certain scenarios (see Cassen 2009 passim, in particular 642-657), using them with the little data available for the Scottish Bronze Age at this point may result in misleading or inaccurate conclusions: this does not, however, preclude the application of phenomenological approaches at a future date. Viewshed analysis has been conducted in the case study in Section 3.7; this again relies on being able-bodied and fully-sighted, but is arguably a more reliable analytical tool for use as part of a phenomenological approach. The significance of water, as mentioned above, has received considerable academic attention over the last fifty years. Specific to Scotland is SWAP, the Scottish Wetland Archaeology Programme, which works to study lakeside dwelling and peat marsh exploitation in the Bronze Age (Crone and Clarke 2005:5). The ritualistic importance of bodies of water is the main argument supporting its significance in prehistory, though the importance of rivers and seas in communication systems should not be overlooked. For instance, stone axe deposition in Neolithic Ireland has been linked to bodies of water (Cooney 2000:206-11). Bog bodies, water depositions and, later, written records all attest to the cosmological importance of water throughout prehistory (Glob 2004, Aitchison 1988:275, Alcock 1966). Water and metal both have liquidity as their core concept: both transform and are transformed through solid and liquid states (Stevens 2008:244). The use of rivers for depositing metal artefacts and as a method of defleshing corpses prior to inhumation or cremation highlights their status as markers of the in-between, a conduit between physical and spiritual states (Brück 2006:302). Their obvious use as geographical boundaries may have further imbued them with a ritualised form of liminality, demarcating a border between safe and unknown, life and death and the edge of the material world, as well as tribal or cultural boundaries: it is this that makes them intrinsic to prehistoric cosmologies (after Brück 2011:389). Springs and wells would also be included in this as they would have been founts of life-giving water, a tame version of the treacherous seas and rushing rivers: this may live on in the modern tradition of throwing copper coins into wishing wells. Liminality as a ritual practice is well established for the Bronze Age (Brück 2001): the next stage is to examine rivers and other bodies of water as potential foci of liminality. 2.2 Bronze Age water deposition Artefacts that were deposited in a landscape with water as a major focus are considered in this chapter. This sort of natural setting will be referred to as a waterscape, the definition for which is as follows: Waterscape is a term that refers to the wide range of contexts, often interconnected, in which water is found in the landscape including rivers, streams, lakes, ponds, pools, springs and the large variety of different types of wetland areas. Waterscapes can alter at different times of the year with flooding events or drought, and they can change over the long-term through, for example, alterations in river-channel courses. The nature of the water within different components of the waterscapes can also vary according to a number of factors: flowing water in rivers or still water in lakes and wetland contexts; whether the water is fresh or salty; whether it is clear or opaque (the colour depending on the material which the water is carrying); and the habitats to which the water is able to contribute. (Rogers 2012:329) In terms of this paper, waterscape findspots are: 1) Within a loch, river, stream, marsh or sea 2) From within 100m of a major body of water (a river course, coastal shoreline or loch edge) 3) From within 100m of at least two streams or burns (or any other smaller bodies of water) 4) From a drained marsh or bog, where the pre-draining water levels can be securely deemed to have constituted a marsh or bog. Draining for agricultural purposes does not necessarily equate to a marsh or bog, and excavation notes from such cases must be treated with care and caution Archaeological waterscape studies from the Bronze Age show that different waterscapes are used differently, depending on the people using them. For instance, in the Norfolk Fens, bronze depositions can be grouped into two distinct categories: swords from causeways, and hoards from bogs or pools set far away from the major rivers (Yates and Bradley 2010:410). Similarly, in Ireland, the change reflects a temporal transition rather than a distinction between artefact types: findspots shift from rivers to bogs as the Bronze Age progresses (Needham 1988). It is not enough, therefore, to isolate the Scottish waterscape finds and look no further: a similar distinction may have been made between various bodies of water, and between the deposition of hoards and of single finds, as the Bronze Age unfolded. 2.3 Early Bronze Age deposition, 2300-1500BCE Of the 164 artefacts securely typologically dated to the Early Bronze Age, 59 of them came from a waterscape, or 36% of the whole. Though sizeable, this does not mean water was the statistically significant factor governing deposition patterns in the Early Bronze Age. Of the waterscape finds, the majority (41/59, or 69%) are from close to burns or streams, rather than from within a body of water. Small watercourses may have changed course, or even appeared or disappeared completely in in four millennia since deposition, but in the absence of a Bronze Age water-system map, modern cartography will have to suffice. Of the remaining 18 finds, 9 are from the coast, 4 are from marshes or bogs and 4 are from within rivers or lochs. The final find is the only dirk known so far from Early Bronze Age Scotland, and was found in a bog on the shore of Loch Glashan, making it a unique findspot as well as a unique find. It could represent an early version of the Group III rapier that would become abundant in the Middle Bronze Age (Burgess and Gerloff 1981), and its position at the intersection of two different sorts of waterscape is highly indicative of the special status of the object. It may indicate a special symbolic significance accorded to the artefact upon deposition, perhaps in recognition of a new technology. Incidentally, Barber makes the excellent point that swords were by no means the pinnacle of technological evolution, and should not be treated as the glorious result of two thousand years of inferior eapo s te h olog Ba e : . Ti oth I gold holds that p og ess is a retrospective quality, that the artisan or smith creating an object does not include the o ept of p og ess i a stage of the eatio though process (Ingold pers. comm.3). 3 th From his Creativity in the Bronze Age keynote lecture from the morning of Wednesday 10 April, 2013, at Magdalene College, University of Cambridge. This author disagrees; pragmatically, the novelty of a new form of metal weapon would require the adoption of new smithing techniques and means by which the mould was created (Binggeli 2011). Cognitively, it would be something of a leap to suggest that the appearance of a new weapon or object type would be met with universal indifference, and that the craftspeople making the items were not excited or intrigued by the innovation, at least during its initial period of circulation. Nor would the potential new uses of the item be lost on its creators – less energy-intense log-hewing, for instance, or more efficient butchering methods. Hoard depositions are not particularly significant in this period. There are only 10 hoards from waterscape contexts, or around 6% of all waterscape finds, all of which contain axes. Three of the hoards are from the coast, and the remaining seven are from within 100m of waterways – there are none from within marshes, bogs, lochs or rivers. Furthermore, seven of the ten hoards are from the north-east coast of Scotland, in Banffshire, Morayshire or Aberdeenshire. This suggests that hoard deposition was an uncommon occurrence, possibly restricted to one group which lived in the north-east, and was not the right way of expressing the dominant cosmology at the time. Given the spread of bronze depositions across Scotland in the Early Bronze Age, it is unlikely that the lack of water-based hoard depositions is due to a recovery bias. Illustration 2.3.1. The Early Bronze Age dirk from the shore of Loch Glashan, currently held in the National Museum of Scotland. Note the thick brown patina, indicating a waterlogged matrix. http://nms.scran.ac.uk/database/record.php?usi=000-100-043-980-C The focus on single artefact deposition is at its greatest in the Early Bronze Age, and wanes slightly in favour of more hoard deposits in the later Bronze Age. This could be related to the huge cosmological shift that demarcates the Early Bronze Age from the Late Neolithic: the move away from multiple and secondary inhumations found in the Neolithic, to single burials and a refocusing on the individual, rather than the collective; the Early Bronze Age also marks the point at which grave goods become the norm, indicating either the personal wealth of the deceased, or some of the funerary rituals which accompanied the burial (Brück 2004:309, Thomas 2000:656). Individual funerary customs could be reflected in the non-funerary deposition record in one of two ways. Firstly, the single artefact could have belonged to one specific person, and be accorded the same rituals as the o e upo the o e s death, a fo of ultu al sig ifi a e that i ues the a tefa t ith so e of its o e s ha a te a d due respect. Secondly, the artefacts could have no specific owners – axes are a multi-purpose tool, and do not feature as heavily in heroic mythology as, say, swords, perhaps indicating that they were imbued with different, more generalised kinds of meaning. In this case, the single depositions would reflect, rather than link directly to, the single inhumations. Such a novel way of burying the dead may have taken a while to become conceptually acceptable to Bronze Age people, and by burying their tools and weapons in the same way, could have normalised a new practice. This could relate to the Loch Glashan dirk (Illustration 2.3.1), which was also a novel practice (in this case, a technological innovation, rather than social) which was potentially normalised through ritualised deposition. The Early Bronze Age is a period of experiment and innovation, which saw a new economic basis in the form of metallurgy, new funerary customs and a shift in subsistence patterns (Brück 2000, Brück 2002:55) : cosmologically, therefore, it is entirely possible that a range of ritual reasoning was applied to artefact deposition, rather than one single universal explanation. 2.4 Middle Bronze Age deposition, 1600-1000BCE There is a slight increase in Middle Bronze Age waterscape deposition compared to the Early Bronze Age – 114 artefacts are from water contexts, as opposed to the Early B o ze Age s . Fu the o e, the p opo tio of ate s ape fi dspots a kedl i eases in the Middle Bronze Age as well: there are 219 findspots overall, meaning that the 114 waterscape finds account for 52% of the total assemblage. Given that the Early and Middle Bronze Ages are of roughly the same length, this increased significance of water can be attributed to water-based cosmologies becoming the standard, building on the foundations laid in the Early Bronze Age. Continuity with the Early Bronze Age can also be seen in the contents of each deposition. It has already been shown that tentative exploration into dirk deposition was undertaken in the Early Bronze Age, and the dirk itself was accorded a special deposition spot at the interface of two bodies of water. In the Middle Bronze Age, dirks or rapiers are found singly and in hoards; the two typologies are classed together, as detailed typological research is a slippery slope that would take too long to investigate fully in this paper, and is could also indicate a false sense of order and correlation which mask subtle differences (Lene Melheim, pers. comm4.). Furthermore, much of the terminology associated with Bronze Age blade technology heavily implies specific functions which were not necessarily the intended purpose of the blade (Molloy 2011:73). All but one of the waterscape hoards contain at least one axe or palstave: there are no hoards of spearheads or dirks that do not also include an axe or palstave – for the above reasons given, axes and palstaves are 4 Following her lecture given at the Creativity in the Bronze Age conference, Magdalene College, University of th th Cambridge, April 9 -11 2013 considered as a single group. The prevalence of axes is common to all periods of the Bronze Age, and may have served to normalise a new or experimental depositional practice. There is a slight increase in hoard deposition when compared to the Early Bronze Age. 18 of the 114 waterscape finds were hoards, or 16% of the waterscape finds from this period, up from the 6% from the Early Bronze Age. 3 of the hoards were from bogs or marshes, one from each of a river, loch and the coast, and the remaining 12 are from very close to rivers or burns. One of the bog hoards, from Corsbie Moss, is the unique hoard that does not contain an axe or palstave: the assemblage comprised a sword in a scabbard with a bone or antler hilt, and a spearhead. Unfortunately, as the items were uncovered by labourers in 1854, there is not a securely recorded findspot, though they were found in a peat matrix (http://canmore.rcahms.gov.uk/en/site/57357/details/corsbie/). It is possible that the spearhead is a much earlier weapon than the sword, and that the hoard was in fact deposited much later on in the Bronze Age, which would account for it not conforming to the observable pattern of axes or palstaves accompanying all Middle Bronze Age hoards. Of further note from the Middle Bronze Age deposits are the 96 single finds. 23, or almost a quarter, of the single finds are from within rivers or lochs. Eight have coastal findspots, and two of these eight are definitely sea deposits, rather than buried on the shoreline. This is noteworthy as both of the sea deposits – estuaries in Dumbartonshire and close to Fort William – are west-facing with the far bank clearly visible; they also drain into the Atlantic Ocean and could possibly be visible from the Hebrides, though visibility would have varied according to the weather and where precisely the observer was standing. Late Bronze Age activity on the Hebrides is inferred through Iron Age house-building incorporating Bronze Age artefacts (Parker Pearson et al 2005), but is not as clearly evident in the earlier periods. It could have been that the islands were viewed as special or imbued with specific meaning – to reiterate the concept of liminality used previously, the Hebrides occupy a remote but visible position between the mainland and the Atlantic, making them the liminal space between the known and unknown, and therefore a potential influence on bronze weapon deposition. Both of these estuary deposits were single axeheads, though given the prevalence of axe deposition, this should not be viewed as particularly special in any way apart from the location of the findspots. Illustration 2.4.1. Locations (red circles) of the two Middle Bronze Age estuary deposits. Note the position of both in areas of particularly complex coastline and proximity to the western islands. 2.5 Late Bronze Age deposition, 1100-800BCE The Late Bronze Age is the period which shows the greatest change in depositional practice, as well as the most frequent depositions (see Section 2.11). Of the 231 findspots, 94 (41%) were located in waterscape contexts, which is a decrease from the Middle Bronze Age, though still more that the Early Bronze Age. This could be related to the shorter time period covered, as the Late Bronze Age only spans a maximum of three centuries. However, the slight changes in depositional practices discussed below would suggest that the Late Bronze Age patterns indicate a change in cosmological practices, rather than a modern analytical bias. Although there are slightly fewer waterscape finds with which to work, there are enough for observable patterns to emerge. Primary among them is the vast leap in significance accorded to hoards. 39 of the 94 waterscape assemblages are hoards, or fractionally over 41%: this is the largest proportion of water-based hoard deposition yet seen in the Bronze Age, or one hoard deposition in water roughly every seven years and eight months across the whole of Scotland, or around four hoard-in-water depositions per generation. It is likely, therefore, that someone living to their mid-50s would experience (or be present in a community when a deposition is made) at least once in their lifetime. The type of wetland in which the hoards are deposited does not seem to follow a fixed pattern: 35% are from bogs or marshes, 46% are within 100m of a body of water (with no significant relationship between branched streams and linear watercourses), 17% are from within a river or loch, and the remaining 2% are from the coast. 60 50 40 30 EBA MBA LBA 20 10 0 Waterscape deposition (% of % of waterscape depositions % of waterscape depositions the whole data set) that are hoards that are from within rivers etc. Illustration 2.5.1. Graph demonstrating the differences between water and hoard depositions between the Early, Middle and Late Bronze Ages. Note the huge increase in relative hoard deposition in the Late Bronze Age (middle). The sharp rise in water-based hoard deposition could indicate an acceptance of a practice which had previously only been experimental, which would suggest a nation-wide consciousness and cosmological uniformity. If this was not the case, there may have been a period of climatic deterioration which made life substantially more difficult for Bronze Age people: increased rainfall, flooding, high tides or extreme weather may have resulted in an increase in ritualised deposits in waterscapes (for a full consideration of ritual theory, see Sections 3.1-3.5 inclusive). It is of potential interest that it is during the Late Bronze Age in Scotland that the great Mediterranean and Near Eastern civilisations suffer catastrophic collapse, which could have been due (in part) to a climate change event that made drier areas arid and uninhabitable, causing massive migrations and warfare (Weiss 1982). An alternative interpretation relates to the concept of individuality expressed in the Bronze Age, discussed in Section 2.3. Brück has criticised this approach and argued for a greater sense of the olle ti e , ased o i ed e ated e ai s a d e carnation (Brück 2006:299). This author would disagree with this analysis, as the predominance of single inhumations, single cremations, single non-funerary metal deposits and the decline of monumental construction would attest to a greater focus on the individual, potentially influenced by the creation of new wealth sources through the control of the metal trade. However, Brück argues (2006, passim) that this concentration on the collective occurs during the Middle and Late Bronze Ages in Britain, and it is during this period that the sharp rise in hoard deposition occurs. It is possible that social, environmental or economic problems or declines in the Late Bronze Age forced communities into a greater level of cohesion than seen in the preceding millennia, which was then reflected in the deposition record. The single finds from the Late Bronze Age are no less interesting. For the first time in the Bronze Age waterscapes, axes are not the dominant single artefact type. Swords account for 47% of the single finds, axes for 40% and spearheads for 13%. This does, however, only equate to 22 swords in the waterscape section, or around half of the number of hoards. This point is worth labouring here, as the myths surrounding swords in rivers remain to this day: King Arthur and Excalibur is the most obvious example, though magical and/or legendary swords abound in classical and modern mythology. However, one of the most striking examples of legendary weaponry, Mjöl i , Tho s eapo apa le of le elli g mountains, is in fact an axe, despite being commonly sand erroneously depicted as a hammer. This is a rare instance of axes being accorded the same mythological status as swords, and it is interesting that it originates in Bronze/Iron Age Scandinavia, an area with strong trade and communication network links with Scotland. Illustration 2.5.2. A silver pendant, from the latter half of the first millennium AD, which depi ts Tho s lege da eapo Mjölnir. © Swedish Museum of National Antiquities, www.historiska.se Illustration 2.5.3. Legendary swords in modern mythology: Neville Longbottom and the Sword of Gryffindor from the Harry Potter series. © Warner Bros Inc. There are, in fact, only 13 Late Bronze Age swords found in rivers from Scotland, accounting for 68% of the metallic riverine finds from this period. The very justifiable concerns of Cowie and Hall, that the breadth of artefacts found in Scottish rivers range from Neolithic stonework to post-Medieval pottery suggests either a very long period wherein rivers serve as foci for ritualised deposition, or that riverine artefacts are the results of recovery bias due to soil erosion and river courses changing over time, are somewhat assuaged by this significant incidence of Late Bronze Age riverine swords (Cowie and Hall 2001). Furthermore, if soil erosion is responsible for a percentage of the riverine finds, it would follow that the artefacts would have been deposited on or very close to the river bank, or on an exposed slope in the river valley, far away from well-trodden paths, in order to become exposed and washed down into the river. This would still be classed, by this pape s ethodologi al o st ai ts, as a ate s ape fi d, a d ot ate iall alte the conclusion that bodies of water played a major part in determining where an artefact was deposited. Having belaboured this point, a sense of perspective must be borne in mind: riverine swords account for 14% of the waterscape assemblage from the Late Bronze Age, and for just over 5% of the Late Bronze Age assemblage as a whole. While significant when studied in the isolation of river finds, riverine swords are by no means the primary, or even a major, method of deposition in the Late Bronze Age. 2.6 Riverine deposition in Britain Weapon deposition in rivers is a relatively common area of research in British and European Bronze Age studies, and the findings from Scotland can therefore be placed in a slightly wider context. In the southern Netherlands, swords are conspicuously absent from burials and bog hoards, but are very common river depositions. Fontijn suggests that riverine swords in the non-ferrous Netherlands conform to a pattern of local/foreign, domestic/imported and personal/combat duality, and that by viewing the house as the centre of the Bronze Age cosmological universe, riverine swords could be interpreted as being sent back to their origins, in the outside world accessed through water-based trade routes (Fontijn 2008 passim). He also reiterates the points made earlier in this paper, that the rational/irrational dichotomy used to define ritual and economic depositions is a product of post-Enlightenment thinking, and the distinction between sacred and profane places is a modern hindrance und unhelpful when applied to prehistory (Fontijn 2008:87). Compare this to the study conducted on the River Severn and its immediate surroundings. If riverine swords were significant across northern Europe as a special kind of ritual deposition, the longest river in Britain would surely bear witness to this. However, the archaeological record shows 41 items from within a 1km radius of the Severn, and 13 items from within the river itself: 7 axes, 4 swords and 2 spearheads. Expanding the scale of the investigation yields little: 282 bronze finds (including ornaments and tools, not just weapons) are recorded from the hinterland of the Severn for the whole of the Bronze Age, which is a staggeringly small amount (Mullin 2012:49). There is a similar dearth of bronze finds from the rivers Ribble, Dee and Wye. The Severn does have daily tidal surges which refresh and renew the landscape, making it a prime candidate for a ritual deposition focus point; this may also have swept finds out into the sea. A recovery bias seems unlikely, given the scale of the work. Mullin does consider the possibility of a different river-based cosmology which would account for the lack of finds, but there is another plausible interpretation. Stable isotope analysis suggests that the salt-marsh estuary land surrounding the Severn was used as seasonal grazing land by Bronze Age communities (Britton, Mülnder and Bell 2008:2113). Although the distinction between sacred and profane areas is a product of postEnlightenment theory, it seems unlikely that a secretive performance culminating in a permanent deposition would be held in an area oft-frequented and highly accessible to the rest of the community. (This does not preclude non-permanent depositions, which would have been recovered during the Bronze Age or Early Iron Age and are therefore absent in the archaeological record). A slightly tangential comparison can be made with depositions made in the Norfolk Fens. The weapons deposited here only make sense when discussed by artefact type, when a very clear pattern emerges: single swords are found in causeways, and hoards are found in bogs and small pools far away from the causeways (Yates and Bradley 2010:410). Although this is not a riverine study, an underlying principle can be found in all of these papers. Rivers would have been major communication and trade routes (Mullin 2010:53) and are a fixed route for exploration and colonization into uncharted and potentially dangerous areas. Causeways through the bogs and marshes in the Fens would have played exactly the same role as rivers in an economic sense: domesticizing and normalising a land that is very difficult to navigate and incredibly treacherous in order to connect communities and trading networks. The patterns of sword depositions for Scotland and the Fens can therefore be see to ha e a u de l i g o elatio , a d a e e elate to Fo tij s o ept of ei g returned to the earth or land from whence they came, via the trading routes which would have brought in the ingots, if not the finished product (Fontijn 2008) The fear of marshland and the related unknown remains in (comparatively) modern mythology and folklore: the Dead Marshes with their corpse-candles and the secret safe path in Lord of the Rings (Tolkein 2005:819), the Will o the Wisp hi h guides t a ellers through dark and dangerous paths (Illustration 2.6.1), G e del s te if i g s a p lai from Beowulf (Swanton 1997:39). Without wishing to stray too far into phenomenology, marshland or boggy moors after dark still etai the a ilit to f ust ate o e s sense of direction, and it does not take a great leap in conjecture to imagine how terrifying they must have been prior to the Ordnance Survey, portable GPS devices and electric torches. Scottish marshes are overwhelmingly blanket bogs, which are impassable, inhospitable and usually located in upland areas that are difficult to access, further emphasising their liminal nature and potential sacred significance (Crone and Clarke 2005:7). Illustration 2.6.1. Das Irrlicht The Will o’ the Wisp , oil on canvas by Arnold Böcklin (1882), ©Museum Georg Schäfer, Schweinfurt Prominent places 2.7 Identifying a prominent deposition There is obviously a world of difference between a high-altitude deposition and a deposition made in a prominent area of the landscape; the two are neither mutually inclusive nor exclusive. High-altitude depositions may be difficult to see, and in a relatively flat landscape, prominent areas do not necessarily equate to high altitudes. For the purposes of this study, given the constraints on time and expenditure, high altitude finds were identified through one of two means. The primary method was to check the contour lines on the Ordnance Survey maps and/or the Edina mapping software (freely available online) for the area: a findspot located significantly above the surrounding area was classed as high-altitude. “ig ifi a tl , i this ase, ea t -1000m in hilly areas, and 200+m in flatter areas. This is by no means a perfect system, and will be refined in future work to involve more stringent boundaries and, where possible, more ground-truthing at findspot locations. However, as demonstrated by Illustration 2.7.1 (below), the majority of the sites would lie in an area 350m above sea level, and so a more stringent, albeit not perfect, s ste fo atego isi g p o i e e ust e de ised. The se o da ethod as looking at the place-names of the findspots: artefacts with secure findspots (six or more National Grid Reference digits) from places with Hill, Top or Law were cross-referenced with Ordnance Survey contour maps and/or the maps available online at www.edina.ac.uk, and a high or prominent location confirmed or rejected. Finds from moors were not automatically included: despite (usually) being high altitude, the key signifier in this section of the study is visibility and prominence, not just altitude; the high-altitude but non-prominent Culloden Moor (Illustration 2.7.2) demonstrates this well. Where a site cannot be securely confirmed or rejected based on geographical linguistics and/or contour maps, caution has been exercised and the site rejected. It would be preferable at this juncture to not muddy the water with unsound conclusions. If further work encompasses prominently-placed findspots, these sites would be revisited, preferably in person, and/or viewshed analysis would be undertaken. As it stands now, however, there is neither time nor resources to undertake such work for a comparatively small section of this study. Conclusions drawn in this chapter would be more than suited for further research, using the parameters and methods outlined here. 100km Regions over 350m Illustration 2.7.1. Map showing the regions of Scotland over 350m above sea level (hatched areas). As can be seen, given the proportion of landmass which falls into this criterion, altitude alone would be insufficient as an indicator of prominence in deposition location. Illustration 2.7.2. Drummossie Moor, to the east of Culloden. Although relatively high altitude, a fi dspot o this o a si ila oo ould ot e lassed as p o i e t , as the oo itself is flat. © Clan James, available online at http://www.clanjames.com/culloden.htm 2.8 Early Bronze Age The e a e a kedl fe e fi ds i the p o i e t pla es atego tha the e e e found in waterscapes. This may have been an issue with the selection criteria outlines in section 2.7, but casting the net wider without ground-truthing runs the risk of including findspots that are not appropriately sited. As such, the following analysis is based on the criteria outlined here. The Early Bronze Age deposits conform to a fairly strict pattern. Every single one of the 28 prominently-placed finds are axes, or axe-shaped (see below), out of a total assemblage of 159 finds. 5 of the 23 prominently-placed finds are hoards, or 22%. This is almost four times the relative percentage of hoards than from the Early Bronze Age waterscape (6%), though the actual number of findspots in question is much smaller. Caution should be exercised in drawing conclusions as to hoard deposition from this data alone, as the sample set is very small and may be statistically anomalous. However, one of the single finds does stand out as being of significant interest. The find from Hillhouse, in Lauder, is an axe-shaped bronze ingot, included in this study because (as far as the collected corpus goes), this is a unique find in Scotland, insofar as its deposition location suggests intentional and selective deposition, rather than as scrap (Illustration 2.10.2). Originally recorded as an ingot, Schmidt and Burgess reclassify it as an Early Bronze Age flat axe, though a very rough version. If the artefact is an ingot, it would be a phenomenal find for many reasons. Bronze coming into Britain is known from the Dorset shipwreck, for instance (Parker Pearson 1993:109), though their eventual fate is unclear – recycled, destroyed or deposited irretrievably in the sea. An imported ingot in the shape of its intended outcome would suggest an information transfer across the trade routes, as if certain ingots (or ore sources, or smelting artisans) made one type of artefact better than another – if there were specific metals for axes, for small tools, ornamentation etc. The oxhide is the most wellknown ingot form, but variations could well be possible, as the Bronze Age metal trade was far-reaching and highly nuanced (Muhly 1977). There is also the distinct possibility that elementality played a major role in prominently-located deposition siting in the Early Bronze Age. Two of the finds show unequivocal signs of elemental interplay. The hoard of two flat axes from Craigdhu on the south coast of Arran was place on the top of a hill facing the sea, which is only a short distance away. There is also a chambered cairn and later fort in the immediate vicinity. The sea would have been audible, if not visible, from this point, assuming the sea level has not changed drastically in the last three millennia. The find from North Plot Hill in Ratho, Midlothian, was found in a bog at the base of a mountain, raised above the surrounding area. There are also some cup and ring marked stones a few hundred metres to the south-west. The role of elementality as a transformative practice has been explored in Bronze Age Scotland, and using bogs or the sound of the sea may have contributed to the change in function of the deposited artefacts, from items inhabiting the corporeal world to artefacts buried, hidden out of sight for a (possibly) supernatural purpose (MacGregor 2008:270). Indeed, Stevens has gone one further and presented the paradigm that the elementality of a findspot represents more than one place (Stevens 2009:245), that by invoking various elements – water, fire, earth, air, a particular sound etc. – that the whole cultural biography of an artefact can be summarised and represented. This is clearly going to be almost impossible to prove, but it is interesting to note that at least 5 of the 23 prominently-placed findspots are also very close to burns, streams, a coastline or marshland. Blurring the lines between various ritual elements may indeed have been one of the minor contributing factors governing Early Bronze Age deposition. Illustration 2.8.1. The crude flat axe or axe-shaped bronze ingot from Hillhouse, Lauder: front and rear faces and in profile. © Hunterian Museum, University of Glasgow 2.9 Middle Bronze Age Again, there is a relatively small amount of prominent findspots from the Middle Bronze Age – 29 in all. Of these, 7 were hoards, or 24%. Two-thirds of the single depositions are axes, and axes are found in 6 of the 7 hoards. When the assemblage is considered as a whole, axes and/or palstaves appear in 22 of the 29 depositions, or 76%. The long-flanged palstave that became common in the Late Bronze Age has roots in these Middle Bronze Age palstaves. These figures and artefact types are close enough to the Early Bronze Age data from Section 2.8 to indicate a high level of continuity in prominentlylocated artefact deposition trends; there are, however, a handful of variations. Firstly, artefact types other than axes are found: three hoards have spearheads in them. Furthermore, the Glentrool hoard comprises a flanged axe, rapier, armlet, razor, spearhead, beads and a tanged knife, a vast array of artefacts. The Drumcoultran hoard is entirely comprised of rapiers with identical damage patterns on the butts, which were buried in an earthwork on the NW side of the slope. To return to the idea of elementality explored above, the NW slope would face the setting sun, and may represent a branch of Bronze Age cosmology concerned with celestial movements and perhaps even some form of astrology. The Drumcoultran hoard also indicates the beginnings of the distinction made between swords/rapiers and other artefact types which can clearly be seen in Late Bronze Age assemblages: akin to the dirk discussed in Section 2.3, according a new artefact type a much greater level of significance, reflected in the circumstances of deposition may have eased the artefact into wider use and acceptance in society. Illustration 2.9.1. 64cm long rapier from the MBA Drumcoultran hoard, © The Stewartry Museum, Kircudbright 2.10 Late Bronze Age Once again, the Late Bronze Age shows the most deviation from the patterns established in the Early and Middle periods. The 25 prominently located findspots suggest a similar level of importance as seen already; however, the deviations from previous patterns, as well as the much shorted time-span involved (~300, rather than ~800 years). Ten of the deposits were hoards, or 40%, or approximately double the proportion of the Early and Middle B o ze Ages hoa d depositio , a d a disti t leap up a ds f o increase seen previously. the g adual 45 40 35 30 25 20 15 10 5 0 Early Bronze Age Middle Bronze Age Late Bronze Age Illustration 2.10.1. Prominently located hoard depositions during the Bronze Age, as a percentage of all depositions from prominent places. The contents of the hoards are also worth closer examination. 6 of the 10 hoards have at least one axe in, which is broadly in line with depositions from earlier in the Bronze Age. One hoard, however, is unique in its location and contents. This is the assemblage found at Horsehope Craig near Peebles, consisting of two socketed axes, a rapier and a number of small rings and other small bronze artefacts. Piggott (1955) interpreted this as the remains of a horse-drawn cart, and dates it to the Late Bronze/Early Iron Age transition period. Cart burials are strongly associated with the Hallstatt culture of the Central European Iron Age, but usually include human remains with the cart and accoutrements (Koch 2006:888). The Horsehope Craig hoard could reflect the changing transport systems developing on the Continent and in the British Isles, at a period prior to the emergence of the horse-mounted warrior-elite that would characterise the Iron Age. Furthermore, the hoard was deposited at the pinnacle of the craig, far from any cultivatable or habitable land, possi l to e phasise its othe ess a d, as efo e, a o d a e itual s ste spe ial status before it is more thoroughly incorporated into the Bronze/Iron Age world systems. Illustration 2.10.2. Horsehope Craig, Peebleshire. The cart-accoutrement hoard was found on top of the craig. © Jim Barton via www.geograph.org Illustration 2.10.3. Part of the Horsehope Craig hoard. © Tweeddale Museum and Art Gallery, Peebles Illustration 2.10.4. One of the socketed axeheads from the Horsehope Craig hoard © Tweeddale Museum and Art Gallery, Peebles Illustration 2.10.5. Wagon fixtures from the Horsehope Craig hoard © Tweeddale Museum and Art Gallery, Peebles Illustration 2.10.5 Horse or wagon fixtures from the Horsehope Craig hoard © Tweeddale Museum and Art Gallery, Peebles 2.11. Frequency of deposition All of the work done thus far relies on the underlying principle that the deposited weapons and the act of deposition were special and not an everyday occurrence (a major factor contributing to their ritual nature, discussed in Chapter 3). The observable patterns the efo e attest to this spe ial ess , a d ot to the dail life e pe ie ed B o ze Age communities. This first principle can be examined more thoroughly in order to confirm the extraordinary nature of the depositions, rather than remaining an assumption. The chronology used in this paper has the Early Bronze Age spanning the period 2300-1500BCE. The data set typologically dated to this era comprises 164 finds for the whole of Scotland, which translates to one deposition being made roughly every five years, across the entire country. These depositions were made across 27 counties (using the pre1973 divisions: see Section 1.4), making, on average, 6 depositions per county throughout the Early Bronze Age, or one deposition per county every 133 years. This is to be used as a comparison point only, as the metal depositions are by no means evenly spread throughout Scotland. For a more accurate point of comparison, a specific county and its data should be comparatively examined throughout the Bronze Age. Aberdeenshire has been the subject of an excellent regional study of Bronze Age metallurgy (Cowie 1998), and has a welldocumented material record. Its long coastline would have been the point of ingress and egress for trade between Scotland and Scandinavia and the Low Countries, which could explain the relatively high concentration of metal deposits in the area. In the Early Bronze Age, there are 16 depositions, or one deposition made every 50 years. This seems plausible: if we are to assume that a generation lasts for ~25 years, one deposition every second generation would fit nicely with the theory of metal deposition corresponding to a p o i e t figu e s death a d i hu atio dis ussed i Section 2.3. It also fits with the idea of votive offerings in response to inclement weather conditions, discussed in the conclusion below. Moreover, this rate of deposition would mean that a person living in Bronze Age Aberdeenshire had a good chance of having a single deposition made in their lifetime, making it a highly significant and special event - even if it was not personally witnessed by every member of the community. The Middle Bronze Age (1600-1000BC) shows a similar rate of deposition. The data set lists 219 finds typologically dated from this period, or one deposition every 3 years across the whole of Scotland. This is a slight increase from the 5 years seen in the Early Bronze Age, due to the slightly shorter time period covered and the higher number of finds. The depositions were made in more places, however: there are finds from 30 counties, rather than the 27 counties that featured Early Bronze Age finds. This works out to one deposition per county every 85 years: again, slightly more frequent than the Early Bronze Age. For a more specific and accurate comparison, Aberdeenshire has 33 finds from this period, or one deposit every 18 years. This is significantly more frequent than seen in the Early Bronze Age, and can safely be described as one deposition per generation. This could be taken to indicate the increased use of bronze weaponry, more situations requiring a itual of tha ks o suppli atio , o e hiefl i hu ations or an increase in cosmological activity unrelated to any theory discussed here. Once again, the Late Bronze Age shows the greatest level of change and the highest concentration of deposition activity. In the period 1100-800BCE - the shorted subdivision of the Bronze Age - there are 231 deposition events listed in the corpus, the greatest concentration of depositions visible in the Bronze Age. Because of this, and because of the contracted time period, the rate of deposition across the whole of Scotland works out to one deposition every 16 months, a huge leap in frequency compared to the previous periods. The depositions are spread across 28 counties, slightly fewer than the Middle Bronze Age, which means an average of 8 depositions per county, or one deposition per county every 36 years. Again, this is a significant increase in the amount of metal going into the ground, and correlates with the changing patterns observed in the natural landscape discussed in Chapter 2. Returning to Aberdeenshire for the final comparison, 25 findspots are listed dating to the Late Bronze Age. Over the three centuries, this works out as one deposition every 12 years. Two depositions per generation is a huge and significant, though not quite exponential,5 increase in depositional activity, a concentration of the practices and patterns developed over the previous millennia and expressed in a very short period of time. 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 E M L Illustration 2.11.1. Depositions per year in Aberdeenshire throughout the Bronze Age. 5 t Using the formula xt=x0(1+2) , where t is the time interval, r is the rate of growth and x0 is the value x at time 0, the rate of deposition throughout the Bronze Age increases cubically, not exponentially. 140 120 100 80 In the country 60 Per county 40 In Aberdeenshire 20 0 Early Bronze Age Middle Bronze Age Late Bronze Age Illustration 2.11.2. Intervals (in years) between findspot creations (assuming a constant and steady rate of deposition) Both of the graphs above, as well as the discussion throughout, assume that there was a constant rate of deposition throughout the Bronze Age, that objects were placed in the subsurface at specific points in the landscape at relatively regular intervals, possibly correlating to the lifespan of specific individuals or objects. However, given the lack of absolute dating methods applicable to these depositions, the typological assignation of Early, Middle or Late Bronze Age is as accurate a date as is going to be available, for now. Therefore, it is possible that there were short bursts of depositional activity throughout the Bronze Age, rather than continual sporadic events. For instance, the findspots at Duddi gsto Lo h a d A thu s “eat used i the settle e t patte ase stud Section 3.9) are in very close proximity to each other; it is possible that they represent three hoard depositions made over a handful of months, weeks or even days in the Late Bronze Age, rather than one hoard deposited regularly every year and a half. This is of course speculative, but is worth bearing in mind when examining the data presented here. The discussion of ritual in Chapter 3 would suggest that the act of deposition was a sacred and special event, invested with layers of symbolism and meaning. This could be applied in one of two ways. Firstly, multiple depositions could have been made over a very short period of time to coincide with a new moon, the appearance of a comet or during a particularly bountiful harvest, for instance. Conversely, the sacred nature of the depositions could be linked to their scarcity, meaning that short bursts of multiple depositions detract from their spe ial ess a d e ui i g a pe iod of ti e to pass i et ee depositio s. O efle tio , taking into consideration the repeated patterns of depositional land use (with the continuing focus on water), the overall dearth of deposition clusters, and the long-term patterns in the items being deposited (see Section 2.12, below), the second interpretation presents a stronger and more reliable argument and the depositions should be viewed as sporadic events, not necessarily rigidly equally spaced, so that the agents making the depositions maintain a level of secrecy and otherworldliness through infrequent occurrences. 2.12 Long-term patterns of Bronze Age deposition: an overview The non-funerary metal depositions from the Scottish Bronze Age show a remarkable level of continuity throughout the Early and Middle periods, and significant changes in the Late Bronze Age across all examined aspects. The items deposited, though not the primary focus of this study, attest to this: axes are the most common object in the assemblages throughout the Early and Middle periods, but the artefact record from the Late Bronze Age is much more varied and less heavily reliant on axes, though they still feature frequently in the record. Spearhead deposition remain relatively constant throughout the whole of the Bronze Age as a minor but consistent artefact found in association with other weapons, but the dirk/rapier/sword types vary quite significantly in their rate of deposition. The single Early Bronze Age dirk has been discussed in detail in Section 2.3 because of its unique findspot context at the intersection between a loch shore and marshland. Middle Bronze Age dirks and rapiers are found more frequently, and always in association with axes and/or spearheads, which were established and common tool or weapon types and possibly offset the novel or suspiciously new forms that the rapiers were taking. None of the Middle Bronze Age hoards were solely composed of dirks and rapiers: at least one axe or spearhead is found in each hoard; single dirks or rapiers were, however, deposited individually. The form change6 from rapier to sword, whereby the blade becomes longer, wider and leafshaped, traditionally assumed to correspond with a change from thrusting rapiers to slashing sword attack techniques, occurs at the Middle/Late Bronze Age transition point. Swords become much more significant in the Late Bronze Age, when hoards of swords appear and sword deposition becomes much more common. The importance of swords has, however, possibly been overstated: axes continue to be the most frequently deposited artefact type in the deposition record. The increase in water-related depositions in the Late Bronze Age is also interesting, and suggests a concentration of cosmological activity centred on the natural landscape and metal depositions in the centuries on either side of 1000BCE. 6 I a loath to use e olutio o p og essio he dis ussi g the t pologi al diffe e es et ee apie s, dirks and swords, as one can easily be trapped by semantics into implying that the sword was the pinnacle of prehistoric metal weapon technology, which immediately plays into the male-centric warfare-focussed warrior society model. The fetishisation of the sword is unwittingly extolled by Traherne (1995) and critiqued by Barber (2002). 2.13 Conclusions It has been proven in this chapter that the natural landscape was one of the contributing factors in determining where an artefact, or group of artefacts, could be deposited. The significance of water grew steadily over the Early and Middle Bronze Ages, culminating in the more elaborate patterning seen in the Late Bronze Age. It is interesting to note that the change in Scotland from the Middle to Late Bronze Age, occurring over several generations before and after 1200BCE, coincided with the collapse of the major Mediterranean and Near Eastern civilisations. It has been suggested that the catastrophic social disintegrations seen in Egypt, Turkey and the Near East at this point may be related to a climate change event that made drier areas unmanageably arid, prompting huge migrations, resource stress and social breakdown (Weiss 1982, Tallis 1997, Nunn 2007). Environmentally-deterministic interpretations are becoming increasingly popular as climate change begins to impact modern communities (Judkins, Smith and Keys 2008), and could offer an interesting perspective on the metal deposition record in Scotland. South-west England shows evidence of becoming wetter and of peat expansion in the Late Bronze Age (c.1000BCE), requiring new food crops to be cultivated which could manage these new conditions, and possibly even relocating settlements to less waterlogged areas (Amesbury et al 2008). An estimated 30 years of drought would be sufficient to prompt mass migrations, overriding any adaptive subsistence strategies previously implemented in the affected areas (Weiss 1982). A similar stress on food sources in Scotland related to increased rainfall, lower average annual temperatures and more unpredictable weather events may have resulted in an increase of water-based depositions to try and increase harvests or stop flooding. The wettest phase in Scotland occurred between 1550-1150BCE, with peak rainfall and peat coverage at ~1400BCE (Landon and Barber 2000). This means that the transition period in between the Middle and Late Bronze Ages occurred after a prolonged period of wet weather conditions, potentially decreasing the maximum fertility of the coastal plains at Moray, Fife, Midlothian and Aberdeenshire, areas already at risk from coastal erosion which was exacerbated by changing sea levels and increased rainfall (Cowie and Shepherd 2003:161). Widespread peat coverage leads to acidification of the soil, primarily affecting arable crop production and then decreasing the availability of blanket grazing land as the soil worsens (Robinson and Dickson 1988:232). Human control and organisation of the landscape through defined areas of farmland, ritual and settlement, incurring large-scale deforestation, was characteristic of the Bronze Age (arguably more so than technological developments – Parker Pearson suggests that the Bronze Age should rather be considered as the Age of Landscape Change (Parker Pearson 1993, chapters 3 and 5 passim)). Major deforestation would have increased the chance of flooding, especially in hilly regions – as can be seen in Illustration 2.7.1, this applies to a large amount of the Scottish landmass. To have the landscape under control for so long, and for it then to turn and fail, would have been deeply traumatic for Bronze Age communities, and could be one of the reasons for the increase in intentional water-based metal depositions in the Late Bronze Age seen in Chapter 2. Taken in a slightly wider view, the changes in deposition patterning seen across the board in the Late Bronze Age could be interpreted as a period of adaption to the new weather systems, in conjunction with an increase in metal deposition designed to counteract and soften the adverse agricultural conditions, prior to the abandonment of these practices and a new way of living displayed by the crannogs, hillforts and social and cosmological changes from 800BCE onwards. The climate in Britain continued cold and wet until around 500BCE, several centuries into the Iron Age (Cunliffe 2004:27). The socio-economic impact of the Old World civilisation collapses would have had far-reaching repercussions; long-distance travel for trade and personal advancement is a hallmark of the Greek heroic tradition seen in the Odyssey and Illiad, as well as in the Celtic sagas from Ireland, France and Wales, all of which are firmly rooted in the Bronze Age, and are possibly rooted in the search for better farmland or unoccupied territory relating to climate change migrations (Kristiansen and Larson 2005:24). There are two artefacts which indicate that trade contacts between the Old World and Scotland in this period were present: a Middle Bronze Age burial at Gilchom in Angus contained Egyptian and Late Helladic beads (Piggott 1962:99). Disruption at the Mediterranean supply centre of this trade network would create a ripple effect of disruption down the chain, further restricting access to prestige imports and encouraging other supply areas to take prominence. For an elite system in Bronze Age Scotland, this could have led to the undermining of their visible consumable status symbols. Furthermore, a disruption of the supply of tin from Anatolia could have led to the increased importance of Cornish tin, reaffirming the importance of Britain in European trade systems and shifting the balance of power away from the Mediterranean and out into the geographically peripheral zones. The indirect impact of climate change could also have been cosmological significant in shaping ritual structures and deposition practices. Cloudier skies and high levels of rain may have diminished the importance of astrological components of ritual life, which is a popular interpretation for Neolithic ritual systems (Parker Pearson 1993:95); by becoming more cognitively heliocentric, the increase in sub-surface depositions becomes a logical cosmological step. To return to the idea of deposition as a way of counteracting poor weather and harvests, the possibility of volcanic activity curtailing Bronze Age developments has been raised as a potential branch of research. The investigations into the burial of Akrotiri on Santorini by a volcanic eruption show the complexity of correlating various dating techniques for a single event of this nature: dates range from 1600BCE to 1300BCE, with the earlier date being more generally accepted (Buckland, Dugmore and Edwards 1997:582). The impact of a volcanic eruption is assumed to affect tree growth and ice-core development over huge areas, causing disruption on a vast scale7. Closer to the study area, the Icelandic volcano Hekla has been put forward as a potential explanation for the changes 7 The impact of the eruption of Eyjafjallajökull in 2010 on transport systems and infrastructure was huge, even with modern technological and logistic systems in place to minimize disruption: extrapolation of the effects of a similar eruption in prehistory is not difficult. observed in the Middle and Late Bronze Ages. An eruption would lead to five years of acid rain, from 1158-1153BCE, coinciding exactly with the proposed Middle-Late transition period (Parker Pearson 1993:100). Communities living on or near acidic soils (or peat, which is also acidic) would experience major crop failures, and fluorosis would develop in cattle and domestic animals (Parker Pearson 1993:100), encouraging migration across the country to less-affected areas. However, there are several problems with this hypothesis. Settlement abandonment and migration can be seen on the chalk upland settlements in southern England during this period, which would not have been sufficiently affected by acidic rain to prompt major migrations based on crop failure (Amesbury et al 2008). More importantly, however, is the fact that although there is clear evidence for a short period of acidic rainfall during this period, there is not the same level of evidence for the cause of this being an eruption of Hekla (Parker Pearson 1993: 100). Environmental determinism is of course only one theoretical paradigm which can be applied to non-funerary weapon deposition in Bronze Age Scotland. It has been demonstrated here that the patterns observed in the natural landscape and of the objects themselves deposited can be interpreted in relation to contemporary events in Europe, as well as a phenomenon peculiar to Scotland. Although water clearly features prominently in the deposition patterns, high and prominent places do not. There are 614 findspots listed in total for Bronze Age Scotland, of which 76, or 12% of the total, are located in high or prominent places. This may, therefore, be a statistical anomaly relating to the size of the data set, rather than a conscious decision by Bronze Age people to single out high places for their depositions. The glaring exception to this would be the Horsehope Craig hoard: given the unique contents, however, it may be that much would be interpreted from its findspot, regardless of where it was in fact deposited. Prominent locations may have not been singled out for two reasons: to keep the physical ritual of deposition secret, and to ensure that the la people ould ot ha e k o he e to look afte the a tefa t as u ied. Clusters of deposition sites would suggest that the reuse of a ritual landscape imbued the land with society-wide significance, which would have been passed down from generation to generation through cultural knowledge media like stories, superstitions and folkloric beliefs. This is not the case in Scotland. The lack of deposition clusters (save for a handful of examples: see the settlement study in Section 3.9) suggests that once the artefacts were buried, they did not influence future deposition locations, making them practically unique as a ritual medium: a tangible totem that involves more than one person to create (and presumably bury, though there is no evidence at all of the ritual of deposition itself, so it could well have been an individual activity), that becomes invisible after fulfilling its ritual function (see also Section 3.4). Compare this to the other two major forms of ritual expression in the Bronze Age, burial and carving rock-art, both of which leave very visible traces of ritual activity to the surviving community. It is to the ritual aspect itself that we now turn. Chapter 3 Proximities and Depositions: Ritual and Settlement Sites in the Bronze Age 3.1 The World as Lived and the World as Imagined: Ritual in the Bronze Age The previous chapter showed that there are several patterns evident in non-funerary deposition in relation to water and the natural landscape. The ritualistic nature of the depositions themselves will now be discussed because there is little worth discussing the ritual significance of these deposits if the very nature of ritual is not explored and understood. The te itual is often used in archaeology to explain an artefact or situation in the archaeological record that does not conform to a domestic, o atio al , e pla atio , but the actual form, function, meaning or style of the ritual itself is rarely discussed: this is compounded with the long-debated etymological concerns and implications of idealism and extreme positi is i he e t i the te itual B ü k : . Without goi g i to exhaustive detail (which can e fou d full else he e: B ü k s paper from 1999 would be an excellent starting point), the major problem with using itual as a a al ti al tool i archaeology is that it is difficult to arrive at a consensus on its meaning, and equally as difficult to categorically identify in the material record. The primary aim of this chapter will therefore be to decide on the most appropriate meaning and applicatio of the te itual , and to identify where, if at all, ritual can be securely inferred from the depositional record. 3.2 What constitutes a ritual? Archaeological work concerning the theory of ritual is well-documented, to the point of exhaustion, even to the extent where it has been [semi-] jokingly lassed as the allpu pose e pla atio used he e othi g else o es to i d Bah : . It elies heavily on other disciplines and its own theoretical framework for a definition of ritual, as there are no prehistoric societies left for archaeologists to study first-hand. Anthropological studies of so- alled sto e age so ieties, such as the Sami or natives of Papua New Guinea, cannot be uncritically applied to archaeological studies. Differences in subsistence and economic practices which relate to the geography and resources (mountain societies and tundra societies operate differently in part due to their natural surroundings) mean different social models exist which are not necessarily useful to compare. Furthermore, the presentation of an indigenous culture to an outsider may not necessarily be completely accurate, and result in ethnocentric conclusions. Above all, culture is a dynamic, changing concept which would not have remained static for thousands of years. Any uncritical comparison between a living social system and a prehistoric system is therefore viewed with a high level of scepticism (Sahlins 1972 and Feil 1987). The most obvious complementary discipline would be the anthropological paradigm, where the study of ritual has led to a pletho a of pape s, sta dpoi ts a d defi itio s. O e of the est o ks o the su je t, Bell s 2009 work Ritual Theory, Ritual Practice, deals exclusively with ritual as a theoretical concept, with no real-world examples used. It is from here that the definition of ritual used in this paper will be based, albeit loosely. The majority of post-processual definitions for ritual activity focus on repetition, patterning, and formalised modes of behaviour (Burl 1987:127, Renfrew 1985, Levy 1982:21- . The ajo fla i these ite ia a e that the appea i o - itual a ti ities as well: making a cup of tea and baptising a baby both involve formalised steps which are followed in a specific order and repeated every time the process is enacted. To counter this, i atio al i te tio as added: a a tio e o es itualised he it se es o utilita ia function. However, one immediately falls into the post-Renaissance Cartesian duality trap, discussed in Section 1.3 (see also Brück 1999:317-9), wherein ritual is demoted to a sphere separate from daily life and subsistence practices. Bell proposes that rituals should be viewed as a form of labour, rather than as a symbolic or idealised expression (Bell 2009:xiv), which compartmentalises the cosmos (into sacred, profane, dead, regeneration etc.) and allows carefully circumscribed interaction between the categories (Bell 1997:44). Furthermore, ritual can theoretically be identified through its audience: Bell proposes that all rituals have an element of public performance in them (Bell 2009:15). Modern private prayers and washing (for instance, of the feet before entering a sacred place or beginning prayers), may seem to contradict this, but the size of the audience is never specified. It is entirely within the limits of this paradigm that the audience consists of one person, the agent who both enacts and defines the ritual. By considering this possibility, the concept of rituals displayed to a limited audience – as the Bronze Age depositions may have been – falls into the definition of ritual outlined here. This also encompasses the notion of labour itself being a form of symbolic behaviour, which will be discussed in Section 3.3, below. Relating to this, performance theorists view ritual as an informal, changing, processual, dramatic and indeterminate entity, rather than as a fixed and static system of symbolic codes (Bell 1997:74), which would accommodate the changing patterns of deposition observed throughout the Bronze Age. The most useful approach that Bell proposes would be that rituals occur at the space in between the critical junctures at which a pair of opposing social or cultural forces, such as order/chaos, tradition/change, individual/group, nature/culture and real/ideal meet, requiring a social or cognitive link to bridge the gap: this bridge manifests as ritual activity Bell : . To put it o e o isel , itual o u s he the o ld as li ed a d the o ld as i agi ed… tu out to e the sa e o ld Gee tz : . This o elates ith the idea of liminality explored by Brück and others (Brück 1999, 2001:389, Mullin 2012), where the space in between two concepts, such as light and dark, is celebrated and revered through ritual activity, as well as the cosmological compartment interaction discussed in the previous paragraph. In this context, the findspots discussed here can be classed as examples of ritual activity because they are located at the junction of the two opposing forces of earth and water, as well as between the sub- and supra-surfaces. This has been further interpreted as representing the space between the living and dead (Owoc 2004), though this would require a degree of overlap between two aspects of Bronze Age ritual which does not have any other obvious parallels. This concept of bridging two opposing forces through symbolic activity will therefore be the definition of ritual used in this paper. It must of course be noted that much ritual theory is inapplicable to this study, as the reasons for the objects being deposited remain unclear; notwithstanding this obstacle, ritual theory does indicate that metal object deposition played a significant role in mediating between traditions and changes while maintaining a strong sense of cultural continuity within the Bronze Age (Bell 1997:251). 3.3 Ritual and metallurgy The process of turning copper and tin ores into usable and/or decorative objects is a long and complicated one. Late Neolithic painted pottery manufacture required a level of pyrotechnological expertise that facilitated the earliest smelting of copper ores, leading to the independent development of metallurgy in many areas across Europe where painted pottery was the norm in the Neolithic (Sherratt 1993:2). The production of bronze in several places entailed the development of local stylistic traits, potentially rendering the messages imbued in locally-produced bronze products significant and understandable only in a very restricted geographical or cultural area (Sherratt 1993:18); from these centres, bronze technology spread out to outlying regions (Kristiansen and Larsson 2005:112). There is a solid argument for the ritualization of metallurgy from the earliest stages of the process, beginning with ore extraction. Little is known about copper extraction in Scotland, though there is an excellent synthesis of ore exploitation in England and Wales (Rohl and Needham 1998) which indicates the scale of trade in the southern half of the British Isles during the Bronze Age. Work is also ongoing at a small copper source in Tonderghie, in Dumfries and Galloway, which was likely exploited in the Bronze Age (the Scottish Archaeological Internet Report 43, White and Richardson 2010 [online resources section of the bibliography]). Parallels can, however, be drawn with Neolithic stone-mining, the best examples of which occur in the greenstone-rich mountains of Cumbria. Stone was extracted from areas that are difficult, if not outright dangerous, to access, ignoring larger and more obvious outcrops of flint in the mountainside: in other words, the exact opposite of the current economic practice of least effort for greatest yield (Barber 2002:103, Bradley and Edmunds 2005:71). Mo i g sto e out of the ou tai s o k fa e, the pla e spa i g the sup a- and subsurface, correlates with the definition of ritual described at the end of Section 3.2, and the deliberate choosing of stone sources with difficult access routes and in potentially dangerous areas of the mountainside could suggest that maximising output was not the primary aim of the mines. Specific stone sources could have had particular symbolic meaning for the miners, or the economic value of the quarried stone could have been related to the effort required in its extraction. Difficult access routes could have been deliberately chosen to ensure that only a very specific group of people could have undertaken the quarrying – the young, fit or those who knew the path, for instance. This could relate to the concept of secret knowledge, which is also an interpretation used for the transmission of metallurgy, discussed below. The complex interpretations of Neolithic stone extraction have parallels in the obtaining and processing of metal ores, which will now be examined. Illustration 3.3.1. Map of the major non-ferrous ore sources in the British Isles that would have constituted the nation-wide ore trade network supporting the manufacture of bronze weaponry. The sites in Scotland, most notably Tonderghie, are currently undergoing investigation, and very likely do not represent all of the small deposits exploited in the Bronze Age. After Rohl and Needham 1998:17, updated for on-going research at Tonderghie; any mistakes displayed here are mine, and are therefore not present in the original. Once the ore has been collected and the impurities removed, it is ready for casting. The addition of tin into the molten copper to make 10% tin bronze lowers the melting te pe atu e of the etal ~ ˚C, edu i g the a ou t of ha oal e ui ed to keep the fire going (1kg of copper requires 40kg of charcoal (Rohl and Needham 1998:5)). Experiments undertaken at Butser Farm have shown that bronze can be smelted in a single step (i.e. smelting the tin and copper ores together in a single crucible), but that it is very difficult to control the composition of the alloy produced; this technique may have been used, and has been documented in Iberia, but it is much more likely that the ores were smelted separated and then cast into bronze, making for a three-step smelting process.8 Adding tin also improves the flow of the molten metal and the quality of the casting (Barber 2002:122); it also makes the bronze harder and shinier, increasing its aesthetic and practical appeal (Kristiansen and Larsson 2005:108). The quantity of charcoal required for casting bronze would have meant fairly extensive forest clearance (McMeekin 2010:112): changing the landscape to create the bronze objects means that their cultural biographies are firmly rooted in one place, and the process of transformation from ore to object becomes a much wider cultural and economic phenomenon. However, Fontijn argues that because bronze can be reworked and reshaped, there remains no connection between the ore site and the artefact, that the item cannot be considered a pie e of a pla e . ‘athe , the o ze ite s should be considered parts of an extensive network of technology, trade, communication and idea exchanges, coming from across the sea, or from the south, rather than retaining a specific origin point (Fontijn, pers. comm. CinBA conference, April 2013). These two ideas a e ot utuall e lusi e: a o ze o je t o i g f o the south a i g ith it tales of forest clearance or the charcoal pits that preceded its creation, which would contribute to the cultural life story of the finished artefact. 8 Simon Timberlake, working with the Cambridge Archaeological Unit at the University of Cambridge, undertook the work at Butser Ancient Farm between 2004 and 2012, and presented his findings in poster and th th flyer format at the Creativity in the Bronze Age conference on the 10 and 11 of April 2013 at Magdalene College, Cambridge Illustration 3.3.2. Environmental reconstruction of the Edinburgh area during the Bronze Age. Note the areas of forest clearance in the valley basin. McMeekin 2010:112. Once the metal had reached the optimum temperature for casting, the smith had a very short period – 90 seconds at the absolute most, depending on the crucible material, temperature and humidity of the air and the volume of bronze being cast - to get the molten bronze from the crucible to the mould before it began to cool (John McKenna, pers. comm.). This would have been phenomenally difficult and taken great skill and years of practice; a modern skilled metalworker using silicon carbide crucibles and heat-proof gauntlets can cast up to six small moulds, around the size of a Bronze Age axehead, in one minute (Barber 2002:122, John McKenna pers. comm. 2013). Further, a skilled smith could trap a small quantity of air bubbles in the metal in order to improve the final durability (Stevens 2008:242). Metallurgy and magic often go hand-in-hand: the Greek god Hephaistos was a metal-worker imbued with great power and mysticism, and goetes, or magicians, were highly skilled metallurgists that feature prominently in Greek mythology (Blakely 2006:192-3). Similar associations appear across Africa, Scandinavia and Australia (Inomata 2001:324, Barndon 2006:99-100). The smoke, heat and smells produced by the smiths would have meant that the workshops would have been located away from domestic areas, particularly because of the risk of fire for the timber-framed structures (see Section 3.9). The practical considerations dictating the removal of workshops from homes, gardens and farmland may have led to further symbolic emphasis on the separate and special position of the smiths, feeding into the symbolic significance of metallurgy in Bronze Age society. Cumbrian Neolithic axe factories displayed this blending of respect and isolation: they were very remotely located, on the sides of high hills, which were very visible from a large area, including (presumably) domestic sites (Stevens 2008:239, Barber 2002:103). Barber makes one further point that would compound the ritual nature of metal casting. Getting the metal hot enough to cast took skill, but before the advent of thermometers (or of measurement systems at all), the smith would only know that the molten metal was ready by sight. Bronze is ready to cast at ~1200 ˚C, hi h is the poi t at hi h the hea t of the metal glows bright yellow. Barber argues that this change would best be seen at night, meaning that the smiths worked best in the dark (Barber 2002:122). The colourful, t a sfo ati e, agi al atu e of o ze etallu g le ds itself to ele e tal interpretations, as it turns dark earth into bright metal using fire, air and water (Stevens 2008:240). This could be reflected in other Bronze Age practices, such as the creation of burnt mounds (MacGregor 2008:277), although the burnt mound is the most practical and useful of basic constructions and may be an example of over-interpretation. Phenomenologically, the glowing fire with its smoke and smells, the magical transformation of molten to solid metal and the skills required to do all of this safely and competently add up to create a highly respected, highly ritualised person and process. The smiths themselves are surprisingly invisible in the Scottish archaeological record: no workshops have been found, and their social status remains open to discussion. Pe ipateti t ade s p o ide goods a d services in response to diffuse or fluctuating de a ds a d thus [ e ui e] so ial o ilit , a odel hi h ould see useful to appl to the Bronze Age trading systems (Salo 1986:7). Childe (1940) argued for peripatetic smiths, travelling from community to community to craft and trade. The hoards found across Eu ope ould the e see as s iths sto k-in-trade, buried for retrieval after whatever danger or distraction had passed. However, despite the wealth of ethnographic evidence for peripatetic behaviour, the concept of the peripatetic smith is not upheld ethnographically (Harding 2000:236), though some of the problems associated with anthropological comparisons have been discussed in section 3.1. Conversely, it would be a leap in logic to imagine that every hamlet could support its own smith, so a degree of peripatetic smithing may well have been present in the Bronze Age, even if one smith served only a small region. One of the major academic considerations of the last two decades has been whether bronze-working is an example of independent or attached specialisation. Independent specialisation caters for an unspecified demand and audience, and is governed by efficiency and security. Attached specialisation crafts specific items for powerful or influential individuals to whom the artisan is contractually obliged (Inomata : . I o ata s defi itio of atta hed spe ialisatio i fe s a high le el of so ial stratification and resource management: Helms argues that the relationship between patron and artisan would have been of equals, of two craftspeople working in different spheres, as the elite patron would have created infrastructure and a functioning society (Inomata 2001:335). Following this, it is possible that it was the elites themselves who were investing their time and power into learning metallurgical skills, to keep the circulation of bronze, and potentially even the knowledge and techniques, within the elite class. This, however, would not really mesh well with the separate and special nature of the smiths and their craft, discussed previously in this section. Even the idea of metallurgy being a secret known only to a few privileged artisans is one interpretation; without a functioning Bronze Age society to study, transmission of knowledge across and within generations remains a subject for debate. The ritual aspects of Bronze Age metallurgy are therefore shown to be complex, open to interpretation and with roots and effects in many other areas of Bronze Age life, ritual or otherwise. This is not always fully appreciated, as demonstrated here in Illustration 3.3.3. Action Ritual consequence Metallurgical processes Elemental transformation Trade and exchange Construction of a cultural biography Deposition Representation of multiple cosmological entities Illustration 3.3.3. The ritual aspects of stages of Bronze Age metallurgy, after Stevens 2008:240. This is a very simplistic set of correlations, and very likely overlooks and/or minimises some of the finer points of Bronze Age cosmology and ritual associations; it also assumes a single consequence for each action, which is an extremely naïve assumption. The discussion of ritual theory as applicable to the Bronze Age in this chapter clearly demonstrates the complexity and breadth of the issues at hand, and one should be wary of oversimplifications, such as the table reproduced here. 3.4 Bronze Age ritual There would have been several ritual systems existing as part of the rich and varied Bronze Age cosmology. Although this paper focusses primarily on non-funerary metal artefact depositions, it is worth establishing the context into which these depositions happened. The practicalities of burying the dead were augmented with elaborate ritual activities. Funeral rites during the Early Bronze Age centred on cremation and the construction of individual cairns, rather than the multiple inhumations seen in the Neolithic (Barrett 1990:181). Later, in the Middle and Late Bronze Age, cinerary urns were inserted into existing tumuli, fa u ial sites di i ished i i po ta e a d less oste tatious cremations continued, though the focus remained on the individual throughout the Bronze Age (Burgess 1980:79). Flower pollen has been found in Scottish cist burials and inhumations alike, as well as honey-based drink residue, alongside the more usual small tools and personal ornaments (Tipping 1994:138), indicating the performance aspect of funerary rituals: as the deceased was buried, public mourning rituals were enacted as tokens of respect or personal items were inhumed alongside the body. Illustration 3.4.1. Dagger from the cist burial at Ashgrove. Meadowsweet pollen was found with the remains, as well as a Beaker vessel. © Kirckaldy Museum, Fife. A replica of this dagger is on public display at the National Museum of Scotland, Edinburgh. There is also some indication that other aspects of daily life involved at least one level of ritual activity. Household refuse and personal waste deposition can be always highly controlled to minimise the risks of water contamination and the spread of disease: at sites in Peebleshire, as well as in East Chisenbury, middens were built overlaying existing structures, with tools and/or human remains at the bottom, or foundation, of the midden (McOmish 1996:73). As a parallel, human remains also appear in the foundations of a Middle Bronze Age house in Cladh Hallan, South Uist. Closer investigation showed that of the male and female buried there, both had been preserved, possibly in a bog, for several centuries prior to final deposition in the house foundations. The male is in fact composed of parts from several individuals, and the female has had a right and left molar removed and placed in her right and left hands, respectively (Parker Pearson et al 2005 passim). Preservation in a bog is an intriguing prospect, as bog burials are unheard of in Bronze Age Scotland, unlike contemporaneous Scandinavia (Glob 2004). This strange assemblage underlines the transformative nature of the Bronze Age, and the ancestor-worship that is a common theme in prehistory (Fukuyama 2011:60 for a brief overview of a vast body of work). Illustration 3.4.2. The mummified ale o posite F a ke stei od f o a foundation deposit in Cladh Hallan, South Uist. Parker Pearson et al 2005:536. Large-scale ritual expression was also a feature of the Bronze Age. Stone circles have been typologically dated to the megalithic construction period in the Neolithic, which possibly extended into the Early Bronze Age as well. Stone carvings, particularly cup and ring marks, are widely found during the Bronze Age, and one of the finest set of carvings can be found in Argyll, in Kilmartin Glen (Illustration 3.4.5). Carvings are also found in funerary structures, which likely indicates the re-use of open-air carved stones as the cosmological foundations changed over time (for a European comparison, see the work from the megaliths in Brittany by Cassen 2004). The use of stone surfaces as a medium for ritual expression has been linked to the significance of liminality in Bronze Age cosmology), as the surface lies midway between earth and sky, and are also often located near shorelines, the liminal space between earth and water. (Bradley 1997:316, Helskog 1999:57). Kilmartin Glen is the route through which people, objects and information would travel between Ireland and Scotland, making it a focus for new ideas, technologies and innovations. Illustration 3.4.3. Satellite image of Kilmartin Glen (A). Note its proximity to the sea, its high altitude, and to the northern coast of Ireland. © Google 2013 Illustration 3.4.4. Cup and ring marks at Kilmartin Glen, Argyll (top) and facing inwards at a cist burial at Balbridie (bottom). Chippendale and Nash 2002:42-44 Cup-marked stones also often appear in Iron Age souterrains in Midlothian as markers of land ownership and belonging, as their previous significance in the landscape was recognised and retained from the Neolithic through to Iron Age, though placed in different contexts (Reed 2002:326). Cup carvings themselves are indicators of shifts in ritual significance, as the major cosmological change separating the Late Neolithic and Early Bronze Age was the transition from closed, indoor ritual spaces to open, outside ritual space. The Early Iron Age (beginning c.800BC) was a period of huge social and economic transition, which involved reinventing and appropriating a number of megalithic monuments, including chambered tombs, as domestic spaces, and using Neolithic-style decorations of pottery (Hingley 1996:232-240). This theme of ritual continuity will be discussed further in Section 3.5. On a more pragmatic level, ritual significance could have been accorded to the deposited artefacts through incising or etching decorative patterns, or through tinning the artefact. Although the aesthetic, social and ritual aspects of depositing decorated artefacts has been discussed at length by Andrew Jones (Jones 2001 passim), there is no correlation between artefact deposition in Scotland and a high frequency of decoration. Decoration o l o u s o a es i the ite s i luded i this pape s o pus. Othe o je ts ould ha e been polished to a reflective surface, or have had very superficial decorations which were eroded or removed peri- or post-deposition. The Early Bronze Age showed the highest level of decorated object deposition, with 13 findspots including decorated axes – none were from prominent places, however, and only five from water locations. Furthermore, the four hoards containing decorated weapons were not solely comprised of decorated objects; the hoards comprised one decorated axe and one or two further undecorated axes. A similar pattern is observed in the Middle and Late Bronze Ages, both of which feature three single decorated axe depositions and one hoard of one decorated axe accompanied by two further undecorated axes. In both the Middle and Late periods, two of the four depositions were located in or very close to a body of water. Similar to the conclusions drawn from the prominently placed depositions (Sections 2.7-10), the frequency of decorated objects being deposited suggests that decoration was not an important factor in Bronze Age deposition – the inclusion of undecorated and decorated artefacts in hoards implies that decoration did not accord the object any special treatment, and that the purpose or significance of patterns or tinning on an axe during its above-ground life was not relevant to its post-deposition existence. 3.5 Ritual and the non-funerary depositions Examination of the corpus data with regards to ritual followed the same methodological frameworks that were utilised in Chapter 2. Findspot proximity to contemporary, i.e. Bronze Age, ritual sites was taken as indication of the presence of a ritual landscape. Incorporation of existing, i.e. Neolithic, ritual sites or items into the deposition was taken as indication of a continuation, or appropriation, of existing cosmologies which then affected the deposition. Parallels can be drawn with Scandinavia, where the reuse of Mesolithic sites in the Neolithic, and Neolithic in the Bronze Age, is documented (Tilley 1996:41). Proximity or incorporation of Bronze Age depositions into Iron Age or Pictish ritual sites would indicate a continuation of Bronze Age cosmologies or ritual significance, in one form or another, in later prehistory – akin to the reuse of carved stones in the Iron Age, discussed in Section 3.4. Using these guidelines, a critical examination of data from the corpus will therefore show how non-funerary deposition fits into the wider context of Bronze Age, and prehistoric, ritual. Given the extent of re-use of ritual landscapes in Europe throughout prehistory and the range of Bronze Age ritual features (discussed in Section 3.4), it was initially assumed (prior to examining the corpus data) that there would be a clear link between the aspects of Bronze Age ritual already discussed and non-funerary deposition. The results garnered from the data were, therefore, quite surprising. Findspots that can be considered as being in a ritual landscape that was re-used from the Neolithic or reused later on in the Iron Age are few and far between; for that reason, they are classed together when dealing with the data. For this data, a statistical significance of 0.1, or 10%, was deemed sufficient for confirming an intentional pattern. The Early Bronze Age has 14 findspots that lie on or adjacent to (within 100m) outcrops of carvings, megalithic monuments – all indicators of Bronze Age re-use of ritual landscapes -, sacred springs or wells or (in the south) Roman temples or shrines –indicators of later re-use of Bronze Age ritual landscapes. Out of 219 total findspots, itual sites equates to 6.3%. This is not statistically significant, and is most likely due to the non-ritual-intentional reuse of habitable and cultivatable land, rather than long-term cosmological patterning. Using the same criteria, the Middle Bronze Age has 17 relevant findspots, or 7.8% of the total Middle Bronze Age assemblage. Again, this is not a statistically significant amount. The Late Bronze Age, however, shows a slight increase: 29 of the 231 findspots fall into one of the categories listed above, or 12.6% of the total. This is statistically significant given the 0.1 significance used, and required further thought. Given the much shorter timespan (~300, rather than 800+ years), this could be interpreted as an increase in the long-term importance of ritual landscapes. However, upon closer examination, a more obvious interpretation appears. Of these 29 findspots, 14 are in the immediate vicinity of Early or Middle Bronze Age non-funerary metal depositions. It is therefore much more likely that the cosmological foundations governing the placement of metal deposition continued throughout the Bronze Age, leading to the re-use of special sites in the Late Bronze Age, rather than being much more long-term. The Late Bronze Age showed several major changes in deposition practices with regards to the natural landscape (see Chapter 2): a small shift in patterning relating to ritual sites could therefore be expected, but should not be taken as proof of any trans-period ritual cohesion. Of the 60 findspots from the Bronze Age that could be considered to be in a longterm ritual landscape (as described above), only 7 are found near to areas of Pictish carved stones, such as the one in Illustration 3.5.1 – two from the Early Bronze Age, three from the Middle Bronze Age and two from the Late Bronze Age. None of these finds are particularly close, as they fall on the outer limits of the 100m border set out in the criteria above. It would therefore be logical to conclude that there is no link between the Bronze Age cosmological system concerning non-funerary depositions, and Scottish Iron Age cosmological systems. Illustration 3.5.1. A Pictish cross-slab engraving from Invergowrie, Perthshire. © National Museum of Scotland Non-funerary metallic deposition also occurred in the Iron Age, in the form of hoards of iron artefacts. This is not a consideration for Scotland, however, as the recorded Scottish hoards are all of Roman origin, and the cosmological context of their deposition will therefore be Roman, rather than indigenous. Roman finds in non-Roman contexts are limited to traded objects, and are not large collections of single-material artefacts (Robertson 1970:200). This is further exemplified by the lack of hoards outside of the area of Roman contact (Illustration 3.5.2). There is, however, some suggestion that Iron Age ritual systems could have grown out of foundations laid by the non-funerary Bronze Age depositions studied in this paper. Human and artefact burials in the British Iron Age focussed heavily on deep shafts and pits, as spiritual communication occurred deep underground in the liminal spaces between surface and sub-surface (Harding 2000:315). This may have evolved from the Bronze Age practice of burying bronze items in the ground as part of their cultural life-cycle. Humans establish place from space in the landscape through oral tradition, action, pe fo a e, ate ial ultu e a d uilt fo … this a i ol e a ki g the la ds ape i olou ful ways, though it may also involve the ascription of symbolic significance to certain, often coloured, features of the landscape. This phenomenological construction should be seen to involve elements of both the surface and subsurface landscapes, and include the meaningful construction of a variety of buried soils, sediments and rocks. (Owoc 2002:129- , autho s itali s If it is a epted that the s ste of lassif i g hoa ds i to oti e, do esti , fou de s and commercial categories is rejected as being overtly economy-centric, (which is by no means a hallmark of Bronze Age studies: see Harding 2000:330-367 for a continued application of the framework, at the expense of other interpretative models), it must therefore be acknowledged that there were cosmological factors which affected deposition, other than keeping a scarce material scarce and therefore expensive and prestigious (Forbes 1950:7, Harding 2000: chapter 10). There are several theories concerning the reasons behind deposition. The deliberate abandonment of prestige items is particularly noted in non-ferrous wetland contexts, such as the Netherlands and Scotland (Fontijn 2008:87). Items displaying a long use-life that are accorded water depositions could be undergoing the final stage in their cultural life, representing the end of their own above-ground existence as well as that of their owner, or a pars pro toto (Fontijn 2008:96). It has also been suggested that the depositions represent gifts to the gods, left in sacred places, to garner favour, placate anger or atone for transgressions (Cowie 1988). Brück (2001) suggests a link et ee the s oli killi g of houses i the B o ze Age a d deli e ate s appi g of lades prior to deposition, indicating that inanimate objects were never neutral, or value-free, and that many aspects of Bronze Age life were imbued with a cultural biography. The act of deposition at a place known to be a tribal boundary could have unified the community doing the depositing, wherein the rituals and actions surrounding the deposition act as an agent for group cohesion, to protect against internal and external conflict (Mullin 2012). Parker Pearson suggests that returning the bronze items to the earth, from whence the ore came, or to the sea or river, which would have been where foreign ore, ingots and/or artefacts a e i to B itai f o , as the os ologi all logi al fi al stage of the eapo s life (Parker Pearson 1993:109). Using analogies from the Mediterranean heroic culture, prolific or famous warrior-heroes could have had their possessions stripped from them after dying, and their weapons and armour distributed as prizes for sporting events, as gifts to the a io s eti ue o sa ifi ed to the gods K istia se a d La sso : -1)9. Finally, the blurring of boundaries incurred by opening liminal spaces through deposition can be 9 This reference contains many mentions of the burial of Petroclus, in book 23 of the Iliad. Bronze Age funerary rituals are detailed in lines 20-39. Of particular note are lines 36- : Ma sheep, leati g goats, a d white-tusked pigs i h i fat e e laid out to oast o e Hephaestus fi es , espe iall i te esti g ecause Hephaistos is the Greek god of metallurgy, particularly of bronze-working. Fitzgerald, Kirk and Homer 2008:369) reflected in the blurring between ritual and domestic spheres, imbuing the landscape with meaning and anchoring a community firmly to an area: water dwellings, such as crannogs, that appear in Scotland during the Bronze/Iron Age transition period at around 800BC are held to be the result of a similar process (Cavers 2006:404). Illustration 3.5.2. Map of iron artefact hoards in Britain. Note the scarcity of Scottish finds, and that the hoards stop with the line of Roman contact. Manning 1972:225 3.6 Case Study: ritual and deposition in the Edinburgh area during the Late Bronze Age Unless otherwise indicated, the images in this section are © Rachel Faulkner-Jones, 2013, taken with kind permission of Trevor Cowie, Senior Curator in the Scottish History and Archaeology Department at the National Museum of Scotland. As such, the photographs of the bronze weapons in this section are not to be reproduced under any circumstances without the express permission of a curator or administrator from the Scottish History and Archaeology Collection, National Museum of Scotland. Ritual theory is a fascinating and complex interpretative paradigm which can be applied to Bronze Age depositions. It can be difficult, however, in aligning the theoretical approach with the data presented by the archaeological record. The longevity of ritual landscapes has been investigated in the previous section and proven inconclusive. One further method proposed here is to investigate the state of a small selection of bronze weapons using use-wear analysis. By studying the damage to the blades edges, it should become clear whether the weapons were used in combat prior to being deposited, indicating a long cultural life above-ground; whether the blades underwent any intentional damage immediately prior to being deposited, indicating a level of ritualised destruction of the object which would have transformed it into a different state, much as the Duddingston Loch hoard was transformed by fire and force prior to deposition; or whether the item was deposited in a newly-forged state, indicating its function as a largely ceremonial item both before and after deposition. Methodologically, this case study borrows heavily from the work undertaken by O Flahe t , Gil hrist and Cowie (2011), where the edges of Early Bronze Age halberds from Ireland were studied and the damage analysed in laboratory conditions. Damages were categorised into four types: v-notching, u-notching, denting and bowing (Illustration 3.6.1; O Flahe t , Gil h ist a d Co ie : . Ea h t pe of da age esulted f o a spe ifi set of circumstances. Illustration 3.6.1. BAR2255 Illus of page 42, BAR 2255. The two types of damage most significant to this study are the v- and u-notching. Notches occur when the blade suffers a high-pressure impact from another weapon edge, pushing the ate ial out i to a lip o i a ou d the ot h usually; some notching occurs where all of the material is lost, rather than malformed – see Matthews 2011:94). However, the most significant difference between the v- and u-notch types is the manner in which they are created. After careful and controlled experimentation, it has been determined that unotching occurs when two blades impact upon each other when both are in motion O Flahe t , Gil h ist a d Co ie : ; i othe o ds, u-notches are formed in (military or ritual) combat situations. V-notching occurs when a high-impact blow lands on a blade edge, but only in situations where the damaged weapon is held immobile against a rigid su fa e O Flahe t , Gil h ist a d Co ie : . This is e diffi ult to i agi e o u ring in a combat situation, and the only logical explanation would be that the blade was deliberately struck and damaged in a non-combat situation, presumably using an anvil or similar surface upon which the weapon was placed before being struck. Evidence of intentional damage prior to deposition suggests a period of ritualised destruction of the weapon and supports the hypothesis, laid out above, that the deposited items represented sacrifices or appeasements and that the act of deposition required a cognitive change in the item, from functionality above-ground to spiritual significance below-ground. The study comprises three separate deposition events, all of which date to the Late Bronze Age and occur within the modern city of Edinburgh, Midlothian. The Duddingston Loch hoard will also be used in the second case study (Section 3.9), but is of interest to this se tio e ause it has ee u e ui o all de o issio ed th ough u i g p io to ei g deposited in the middle of the loch. Immediately adjace t to the lo h is A thu s “eat, where two near-identical swords were deposited in what can be assumed to be two associated deposition events. Finally, a hoard of weaponry was discovered by labourers working on Grosvenor Crescent. Illustration 3.6.2. Map of the depositio e e ts dis ussed i this stud . A: Quee s D i e, A thu s Seat (two swords). B: Duddingston Loch (burnt hoard). C: Grosvenor Crescent (hoard). Note how all three hoards are within a 3km radius. The t o s o ds f o Quee s D i e, A thu s “eat, ill e dis ussed fi st. Both a e held in the National Museum of Scotland (NMS X.DQ90 and NMS X.DQ91), off display. They are both Ewart Park, Northern Step 2 types, and are listed in Burgess and Colquhoun as numbers 475 and 483 (Burgess and Colquhoun 1988, plates 70 and 71). Both are in phenomenal condition, and have been cleaned with extreme care which has resulted in a fine burnish without damaging the blades. Both blades show remarkably little corrosion or damage to the blade edges; indeed, the only areas of corrosion occur on the very top of the hilt, suggesting that they were placed upright in the ground, leaving the hilt exposed which resulted in the corrosion. Of greatest interest to this case are two small instances of edge damage, shown in Illustration 3.6.3. Being the only areas of damage on the blade edges, it is unlikely that the damage was caused by post-depositional corrosion. The V-shaped notch near the tip of one of the blades is characteristic of a heavy blow braced against an immovable surface, de o st ated e pe i e tall Gli h ist, O Flahe t a d Co ie, a d was confirmed through microscopic analysis, undertaken by the author at the National Museum of Scotland. A plausible scenario suggests itself, whereby the two twin swords were engaged in a single combat or performance event where the damage occurred, before being ceremonially deposited point-first in the mountain face next to the lake. Illustration 3.6.3. The t o s o ds f o Quee s D i e X.DQ top, X.DQ otto , sho i g the only two instances of edge damage on either of the blades. The v-shaped notch in the bottom is characteristic of intentional blade damage. Duddingston Loch is located immediatel adja e t to the Quee s D i e he e the two swords above were discovered. During a dredging of the loch in the nineteenth century, a hoard of metalwork was recovered from the loch bed in a leather or fabric bag which disintegrated upon recovery. The hoard consists of four complete swords which were given to Walter Scott, a spearhead and a large number of sword, spear, axe and tool fragments. Every single piece recovered from the loch displays an exceedingly high level of intentional fire damage, resulting in the weapons fracturing and deforming (see Illustration 3.6.4). Illustration 3.6.4. The burnt and fragmentary pieces of the Duddingston Loch hoard, currently on display in the National Museum of Scotland, Chambers Street, Edinburgh. Even through the glass, the poor condition of the pieces can be seen and close damage analysis would demonstrably be fruitless. As can be inferred from the above photograph, the condition of the surviving weaponry prohibits any edge-damage analysis, as – in addition to the fire damage - there has been much corrosion damage inflicted on the pieces after deposition, destroying any edge surfaces which would have retained evidence of combat or cushioned impact damage. As such, there can be no comprehensive use-wear or edge-damage analysis for the hoard as a whole. Although the presence of intentional fire damage is unequivocal (the implications of which are discussed below), there are also several instances of possible intentional damage to the weapons as well. The complete and strikingly bent sword (NMS X. DQ303) found as part of the hoard which includes the broken artefacts in Illustration 3.6.4 has lea l ee de o issio ed prior to deposition by bending the blade almost in half. The tip of the blade has also been bent to a right-angle, presumably by hammering the tip – experimental destruction of a similar weapon would shed some much-needed light on blunt-force destruction techniques (Illustration 3.6.5). Additionally, there are also two instances of stress-fractures o ips u i g a oss the lade s fa e Illustrations 3.6.6 and 3.6.7), strongly suggestive of further bending or pressure exerted on the blade in order to deform it beyond the point of repair (which was significant, given the evidence for hammering from previous repairs shown in Illustration 3.6.6). Moreover, the tear in the metal suggests that the sword suffered these two damage episodes at ambient room temperature, or thereabout; bending from fire damage would have resulted from the metal softening with the heat, and would have been unlikely to result in the gaping tears displayed on the blade. It is worth noting that there is a similar stress pattern on the most prominent bend on the blade (Illustration 3.6.8). This sword has clearly undergone massive destructive forces prior to deposition, most likely as a result of crushing between or by large stones or boulders, in a manner that does not easily correspond to an accidental or unintentional damage scenario. Illustration 3.6.5. The bent tip of the blade fragment from NMS X.DQ303 Illust atio . . . The lo e u kle poi t o the lade NM“ X.DQ . The e is also fai t e ide e of hammering or reworking the blade on the right-hand section of the image. Illustration 3.6.7. The second area of buckling on the inner curve of the bend in blade NMS X.DQ303 Illustration 3.6.8. The buckle above the major bend in blade NMS X.DQ303 Illustration 3.6.9. The complete sword NMS X.DQ303, showing the extent of the damage and the poor quality of the edge survival, which negates any attempt at an edge-damage analysis. The second sword (NMS X.DQ305) is only partial, comprising the tip and body of the blade with a sheer break before the rissaro. Again, the blade displays damage consistent with exposure to intense heat, the major difference being that the damage does not extend to one of the blade edges. As such, the two notches with lips (Illustration 3.6.11) were suitable for close examination under a microscope, and were demonstrably caused by blade-on-blade impact. This sword also displayed a bent tip, similar to the complete weapon discussed above (NMS X.DQ303). This is the only weapon in the hoard where such edge notching survives, and therefore the only instance where a deliberate blade-on-blade impact can be securely ascribed. This is arguably the most interesting piece in the whole hoard, because the edges have survived (or not been subjected to) the heat damage and/or corrosion visible on the other artefacts. As established, the two most significant types of edge damage are V- and U-shaped notching, indicative of intentional and combat damage respectively. X.DQ305 displays both of these types of damage (Illustrations 3.6.11 and 3.6.12). The most literal interpretation of this would be that the weapon was used for a period of time as a weapon in combat (whether in warfare, a ritualised combat setting similar to the fighting by re-enactment societies he e the ai is ot to kill o ai o e s opponent, or, more aggressively, in a trial-by-combat fight such as found in medieval Europe where Germanic law was enforced (Ziegler 2004) , a d as su se ue tl de o issio ed prior to its deposition. The angle of the incoming blade impact can be inferred from Illustration 3.6.11, co i g i f o uite high up a d at a ˚ a gle; su h a lo does ot easily lend itself to a combative reconstruction scenario. Furthermore, the positioning of the V-shaped notching, higher up the blade that the widest point, contrasts sharply with the positioning of the combat-related U-shaped notches, which occur much lower down the blade, closer to the tip (Illustration 3.6.12), which correlates to the type of blow and impact that would be expected from a combat situation. Illustration 3.6.10. A view of the second-largest surviving sword from the Duddingston Loch hoard (NMS X.DQ305), showing the extent of the bending damage and the nicks and lips on the edge, examined in greater detail in the following photographs. Illustration 3.6.11. The most significant area of edge damage on the blade NMS X.DQ305. V-shaped notching is characteristic of intentional damage to the blade, as it only occurs in situations where the impacted blade is not in motion. The angle of the incoming blow can be inferred from this image. Illustration 3.6.12. Notching on the other side of the blade. U-shaped notching such as this is indicative of damage sustained while both the impacted and impacting blades are in motion, such as combat (military or ritual). The third blade complete enough to warrant study (NMS X.DQ304) has been arguably the most deformed through intense heat, as shown in Illustration 3.6.15, where the blade has contracted and bent. Again, the damage to the blade edges is such that any trace analysis on the edge has been rendered impossible (Illustration 3.6.14). Illustration 3.6.14. NMS X.DQ304, showing the terrible condition of the blade edges which precludes any closer examination of pre- and peri-deposition damage. Illustration 3.6.15. NMS X.DQ304 in profile, showing the level of deformation on the blade. The final piece held by the National Museum of Scotland which was acquired from the Walter Scott collection is a fragment of a spearhead (NMS X.DQ302), where the tip has broken away from the main body of the spear. It is unclear what has caused the spear to break in such away, but it has been suggested that blunt force using a hammerstone and chisel, or stone, could have resulted in the damage shown in Illustration 3.6.16 (Cowie, pers. comm. June 2013). Illustration 3.6.16. NMS X.DQ302, the socketed spearhead tip. The notch on the central ribline above the white classification is a result of modern sampling techniques, not pre- or peri-deposition damage. The Duddingston Loch hoard is, damage-wise, very interesting. The buckling and warping of the metal suggests exposure to intense heat such as a pyre or fire. The subsequent deposition in water is redolent of a quenching or antonymic action; fire and water both being transformative processes well known to prehistoric communities, but rarely employed together on a single deposition. Indeed, the only other hoard to show a similar level of heat damage (Peelhill, Lanarkshire) also dated to the Late Bronze Age, also included a range of weaponry (28 spearheads, an axe and a sword) and was also deposited in a water context (in this case, a bog). It is possible that these two hoard depositions followed a similar line of ritual logic. Primarily, it is by no means certain that the heat which did the damage was a result of intentional burning. Given the scarcity of this type of damage – two hoards from the whole of Scotland throughout the entire Bronze Age – fire can hardly be considered as a major transformative process for bronze. It is possible that the weapons from the Duddingston Loch hoard belonged to a single person, or family, and were kept together in an armoury or great house. If this structure was then to catch fire, the weapons would be damaged but not irrevocably lost, and could then be de o issio ed th ough defo atio efo e ei g deposited i the lo h, ith its overtones of extinguishing a fire with water. This is of course supposition, but offers a viable alternative interpretation of the deposition process. The final set of weapons discussed in this case study is the hoard from Grosvenor Crescent, Edinburgh, approximately 1km from the hoard and swords found at Duddinsgton Loch and Quee s D i e that have already been discussed. The hoard was discovered by labourers during a renovation of the cellar level in Grosvenor Crescent, and consists of six swords (three complete, three fragmentary), two rings and a pin and a socketed axe. All of the artefacts are held by the National Museum of Scotland except for the socketed axe, the whereabouts of which are not known, and the swords are recorded in the Prähistorische Bronzefunde (PBF) (Colquhoun and Burgess 1988, plate 174). The hoard is shown in Illustration 3.6.17. The swords are in significantly better condition than those from Duddingston Loch; cleaning techniques in the early twentieth century have rendered two of the swords beautifully burnished and completely devoid of the original patina and any evidence for fine damage (visible on Illustration 3.6.23). The condition of all six swords is such that a close analysis of any surviving edge damage is possible, and is undertaken here in the order that they are kept when in storage, left to right on Illustration 3.6.17. Illustration 3.6.17. The hoard from Grosvenor Crescent. L-R: the three complete swords NMS X.DQ201, 237 and 236; the three fragmentary swords NMS X.DQ200, 199 and 306; two rings and a pin. Shown in the protective foam in which the hoard is kept when in storage. None of these artefacts are on display to the general public in the Museum. The socketed axe is now lost. The first sword, NMS X.DQ201 and #474 in the PBF, is a Cherwell variant of the Ewart Park type. It is complete and has a heavily weighted pommel devoid of rivet holes. The edge damage begins at the rissaro and continues down the length of the blade. However, V- and U- shaped notching only occurs on one side of the blade; the rest of the damage can be attributed to corrosion and post-recovery cleaning methods. It does however mean that smaller notches may have been present which have been masked or destroyed by corrosion damage. Illustration 3.6.18. V and U notching on the edge of blade NMS X.DQ201. Note the extent of the corrosion on the edges. The second sword, NMS X.DQ237, #547 in the PBF, is a Northern (unclassified) variant of the Ewart Park type. The damage on this weapon begins on the hilt, but this is most likely due to the close proximity of the rivet holes to the hilt edge (see Illustration 3.6.19). As the sword was deposited still hafted (one of the rivets is still in situ), an intentional blow would have to avoid any organic material wrapped around the hilt. A major indicator of intentional damage is a prominent V-shaped notch in the top of the hilt, which is unique in its placement among the swords studied in this paper (visible on Illustration 3.6.19). Illustration 3.6.19. The hilt damage on NMS X.DQ237: the rivet in situ at the top suggests that the sword was still hafted when it was deposited in the ground; the V-shaped notch immediately above it is a clear indicator of intentional damage; the lower section of the hilt displays a good deal of damage, due to either poor casting of the rivet-holes, corrosion or post-depositional damage. The edge damage on this sword is most prominent down one side only, and shows no V-notching indicative of intentional damage, though the relatively smooth lines of damage could be due to intentional blunting of the weapon (Trevor Cowie, pers. comm. 2013, also visible in Illustration 3.6.20); the corrosion, though minimal, has removed any scratching or similar which would prove or disprove this theory, however. There are two Ushaped notches, one on each edge, near the tip of the sword consistent with combat damage, meaning that the sword was definitely used as a weapon prior to it being deposited. Illustration 3.6.20. U-shaped notch near the tip of sword NMS X.DQ237. Note the long, shallow curves of damage along the blade edges, of unknown origin. Illustration 3.6.21. The lea side of the lade where there is little to no edge damage. The central notch is the second U-shaped notch, with a clear deformation of the remaining material (small lip formed). The left-hand trapezoidal notch is a result of modern sampling techniques, rather than pre- or peri-depositional damage. The third sword, NMS X.DQ236 (Ewart Park [Northern Step 2] type), has been heavily and intensively cleaned at some point after its recovery from the ground, possibly including a brass rub. The sword is a fairly standard Northern Step 2 type, with a slightly asymmetrical ricasso and two incompletely-punched rivet holes. The blade itself shows virtually no damage to the edges at all, save for a small area on the maximum blade width. This damage was minute, and required microscopic analysis. The microscopy showed a regular pattern of scratches and notches which resemble the intentional pre-deposition damage seen on the other swords investigated in this case study. However, given the minute scale of the damage (<1mm per notch), and the intense abrasion from the cleaning processes, it would be unwise to securely ascribe this damage to intentional destruction of the blade. Illustration 3.6.22. The hilt of NMS X.DQ236, showing the two incompletely-punched rivet holes. This is usually taken to indicate that the weapon was not used in combat, as it could not have been securely hafted. Illustration 3.6.23. The minute edge damage on NMS X.DQ236, which was examined using the 70x magnifier on the microscope. The scratches from the intensive cleaning can be seen running across the lade s idth at a slight a gle. The fourth sword (NMS X.DQ200, #530 in the BPF) is a Northern (unclassified) variant of the Ewart Park type, and is missing the hilt and tip. It has three instances of denting on the blade edge, where the damage slopes inwards rather than being a single uniform loss of material (notching), all of which measure <0.5mm. This damage was again investigate using microscopy, and was determined to be too small and the wrong shape to have been caused by another metal weapon. The lack of corrosion on the blade suggests that the damage was intentional, possibly caused by lithic damage. Illustration 3.6.24. The minute edge damage to blade NMS X.DQ200. Patches of bronze disease can be seen on the left. The fifth sword (NMS X.DQ199, #488 in the PBF) was very badly corroded and afflicted with bronze disease (visible in Illustration 3.6.17, the sword second from the left displaying patched of green-coloured corrosion), displayed no damage suitable for closer examination. There was, however, one rivet still in situ. The final sword (NMS X.DQ306, #513 in the PBF), a Northern Step 3 variant of the Ewart Park type, again shows extensive corrosion damage down one edge. The other edge is in significantly better condition, though there is only a single instance of edge-damage which could indicate intentional destruction of the weapon. There is a notch which has deformed outwards slightly while retaining a small amount of the material without forming a lip, making it a notch rather than a dent, though it could be classed as either v- or ushaped; the deformation makes securely classifying it somewhat problematic. It is mostly likely a U-shaped notch which has retained some of the material on the right-hand side of the curve (see Illustration 3.6.25), leading to its more unusual shape. Illustration 3.6.25. The damage to the edge of blade NMS X.DQ306 (on the left). The deformation has resulted in a notch that is not easily classified into V- and U-shaped. The angular notch on the right is a result of modern sampling techniques. Furthermore, every single one of the Grosvenor Crescent swords shows some degree of corrosive damage from the ricasso down, but very little (if any) corrosion on the hilt. Combined with the presences of some in situ rivets, it is clear that all of the swords were deposited with organic hilt-wrappings intact. This serves to emphasise their status as functioning weapons prior to deposition, further confirming the classification of U-notching as combat damage. These three deposition events were selected because they comprise the closest approximation of deposition practices in a confined geographic area in the shortest time period possible. It was expected that there would emerge some pattern or internal logic from these depositions that could be used to extrapolate the beginnings of a wider deposition pattern. What has emerged encapsulates the complexity of studying ritual in prehistory. Within 300 years and a few kilometres, we see three very different depositions being made. The Duddinsgton Loch hoard shows the greatest damage, having being subjected to intense heat and pressure prior to deposition, but the similar hoard from Grosvenor Crescent is in excellent condition. The Duddinsgton Loch hoard was deposited in water, but the two adjacent related depositio s f o Quee s D i e a e adi all diffe e t, being in pristine condition and buried with no associated other weapon types or ornamentation. The Grosvenor Crescent hoard and Duddingston Loch hoard show evidence of being used as combative weapons prior to the intentional damage and deposition, but the Quee s D i e s o ds sho a iguous da age patte s that possi l elate to a si gle ritualised combat event where both swords were used against each other before being deposited. There are, however, some commonalities which unify all of these depositions as ritualistic in nature. All of the deposits were arranged specifically in the ground: the Duddingston Loch hoard was most likely placed in a bag or sack made from organic material before being dropped i to the lake; the Quee s D i e s o ds e e – based on the corrosion patterns – deposited point-down into the ground with the hilts exposed; and the Grosvenor Crescent swords all had their organic hilts intact, and possibly even scabbards before being deposited (Anderson 1879:320). More importantly, every single deposition event contains artefacts with clear indications of intentional destruction using another weapon, characterised by the V-shaped notches. As discussed above, the fire damage present on the Duddingston Loch artefacts could have resulted from an accidental fire (rather than intentional destruction) and the weapons then further decommissioned. Regardless, the weapons underwent a stage prior to deposition that took time and effort, negating the concept of accidental loss, and served no economic purpose – if the weapon was being removed from circulation to maintain the value of bronze, why would it be notched beforehand? The weapon itself would still be functional, if a little more susceptible to fracturing and cracking, and edge-damage would be relatively straightforward to repair (as shown by sword NMS X.DQ303, Illustration 3.6.6). Therefore, even when working in as strict parameters of time and space as possible, it can be seen that there is not a single overarching ritual explanation or single interpretative paradigm for studying the nonfunerary metal depositions, though there are a handful of similarities which allow for comparative studies. 3.7 Ritual sites: interim conclusions It may be time to return to the clunky and oft-used phrase in this pape , o funerary metal depositio . I itiall , this as the u h o e ergonomic and recognisable oti e depositio , ut this ph ase as ui kl a a do ed afte esea h o e ed. It became very clear very quickly that there were many different factors at play governing the type of artefact and location of its deposition, and that these changed depending on when in the Bronze Age the item(s) went into the ground; there are also several theories as to h the a tefa ts e e e e ei g deposited, athe tha e led o eused. Voti e suggests a single reason for deposition, and that this would be relatively uniform across a long period of time and many geographical regions. The case study discussed in Section 3.4 showed that even covering a small geographical area with fixed time parameters, the ritual interpretations for weapon deposition are varied and complex, and certainly do not lend themselves to an overarching explanation. Etymologically, oti e also suggests that the artefacts were being deposited as part of a deity-worship system, as part of a vow-fulfilment scenario, which was not necessarily the case (OED ref). It is for these reasons that the more accurate, albeit wordier, phrase is used. The artefacts deposited in the Bronze Age were inhumed in a very specific, highly ritualised scenario which potentially fulfilled a range of social obligations: group cohesion, economic or spiritual debt payment, signifying the end of a cultural biography, or o pleti g the o zes jou e etu i g the etal to the su -surface. These factors have all been thoroughly explored in this chapter. However, the primary interest of this paper is findspot location and patterning, and it is clear that the aspect of ritual examined in this way, long-term use of a ritual landscape, is not a factor in Bronze Age deposition. The ritual significances of water, prominence and liminality are discussed in the relevant chapters. The long-term view was taken in this chapter in order to establish whether the artificial divisions that archaeology has imposed upon prehistory, the epoch-definitions the Three Age System and its more modern subdivisions such as the Chalcolithic, were hindering study rather than encouraging it. This was suggested by Parker Pearson, as he views the Early Bronze Age (at least) as a cosmological and cultural continuation of the Neolithic, rather than as a distinct and separate period (Parker Pearson 1993, chapters 1-3). The evidence gathered in this chapter would suggest otherwise. Bronze weapon deposition in non-funerary contexts is a distinctly Bronze Age phenomenon, and has no overlap with any other aspect of ritual life that remains visible in the archaeological record, such as rock carving or human inhumation. The pattern observed in the Late Bronze Age, of an increase in overlapping depositions, could be taken as an indication of the re-use of a ritual landscape that remained confined to the Bronze Age. It could also indicate that the cosmological reasoning behind the location of depositions did not change throughout the Bronze Age, and the overlapping depositions are instead the result of sacred places remaining sacred, rather than the conscious decision to deposit weapons close to existing depositions, which could have been marked by a tree or rock formation that has since vanished. However, I would argue that this is unlikely – Scotland is hardly short of areas featuring rivers, marshes, mountains, coastlines or liminal spaces, and the relative scarcity of deposition would have meant that the ground was not exactly saturated with metal depositions (see Section 2.11). 3.8 Settlement in the Bronze Age Great strides have been made in the last two decades in the study of settlement patterns in the Bronze Age, as excavation techniques have become more meticulous and research frameworks have been widely (though not yet uniformly) adopted (Last 2008:35). There are many well-documented excavations of settlement sites, but few are securely dated. Bronze Age houses were usually circular, clustered in small groups and often displayed a high level of sophistication: houses were constructed on artificial platforms that circumvented drainage problems before they occurred (Dunbar 2007:159). Domestic function is usually inferred by pottery (Strachan, Ralston and Finlayson 1998:71-2) as there is not always a hearth in every structure (Cook 2006:13, Jobey 1978-70:79). Pottery can be misleading, however, as several pottery deposits have been found in settlement sites that are intrusive, not diagnostic (Strachan and Dunwell 2003:154, Rees 2002:325). That being said, the roundhouse is in no way indicative of the Bronze Age: timber roundhouses with a ditch, groove and post-hole system were the architectural norm for two millennia, continuing well into the first millennium AD (Strachan and Dunwell 2003:159). Crannogs appear in the middle of the 9th century BCE at the very earliest, and cannot therefore be reasonably classified as a typical Bronze Age settlement site: they are therefore omitted from this study (Cavers 2006:398). Sites were used throughout the Bronze Age for varying purposes: the excavations at Newton Mains in the Clyde Valley showed that the initial palisade trench was overlaid by a wattle-panel roundhouse displaying multiple periods of occupation, before finally being partially overlaid by another palisade trench (Toolis 2004:476-7); excavations at Oldmeldrum in Aberdeenshire suggests a settlement site of long duration (White and Richardson 2010:24). It has also been suggested that the houses themselves were occupied for quite a short period of time, but that new dwellings were constructed nearby later, with the intervening time being used for cultivating the land that once had houses on, so that the cultural and economic links with the land were not lost (Johnston 2005:217). There is a fairly concrete case to be made for the land immediately adjacent to the houses being used for stabling or corralling cattle or pigs, based on phosphate and magnetic susceptibility testing (Terry 1995:419), small-scale wild or domestic plant cultivation, akin to a modern vegetable patch (Johnston 2005:213-4), and linked with large-scale field systems (Jobey 1978-80:94). Illustration 3.8.1. The site plan from West Acres, Newton Mearns, Clyde Valley, indicating the extent of the overlapping Bronze Age features and, consequently, the multiple phases of use. Toolis 2004:475. Ritual and settlement are deeply enmeshed in the Bronze Age: it has been suggested that the majority of permanent structures in the Early and Middle Bronze Ages were funerary or ritual constructions, while the agricultural communities lived in much more transient dwellings (Cowie and Shepherd 2003:158). Issues with the distinction between sacred and profane sites have been discussed earlier (Section 1.5), and this applies to the settlement record as well. The proximity of the houses to daily farming life has been established, but the houses themselves are ritualised. Domestic structures have been interpreted as indicators of social power in the immediate community: the house at Navidale in Sutherland features an elaborate entrance, which would have been visually striking as one approached the building (Dunbar 2007:163). A pendant found in association with the elaborate house further suggests that important figures in the community would use conspicuous wealth to reinforce and remind their neighbours of their status, whether that was through personal adornment or impressive architecture. The position of these houses in the landscape, particularly in relation to the stream of Allt Briste, suggests that the power displayed by their owners could have stemmed from control of trade routes or access across or down the stream (Dunbar 2007:159) (Illustration 3.8.2). Illustration 3.8.2. The location of the Bronze Age house at Navidale in Sutherland (red) and the stream Alte Briste (blue). Adapted from Dunbar 2007:138. Ritualised abandonment events also reflect the lack of sacred/profane segregation in the Bronze Age (Brück 2011:395). Foundation deposits in middens have already been discussed above (see Section 3.4) and this is paralleled in domestic architecture as well. Chalk phalluses and human burials are relatively common features in house foundations in Late Bronze Age England (Brück 2011:296), and the mummies in South Uist discussed above were found in a Late Bronze Age context (Illustration 3.4.2, Parker Pearson et al 2005). Although the non-funerary metal depositions that are the focus of this paper are not found in settlement contexts, the well-documented importance of liminality in the European Bronze Age means that there is solid theoretical reasoning for investigating the relationship between settlement sites and the non-funerary depositions. 3.9 Case study: Settlement and deposition in Midlothian Given the vast number of unconfirmed and confirmed settlement sites in Scotland, a comprehensive and robust study of Bronze Age settlements in relation to the non-funerary metal depositions would take more time (and more space) than allocated for this research, though would make an excellent doctoral study. As such, a case study will be made and establish potential patterns for future research to confirm, deny or merely use as a springboard for other projects. This case study will cover the Midlothian area, as there are several depositions and settlements that have been thoroughly documented and have ten-digit national grid references, facilitating viewshed analysis (see below). It will also cover the Late Bronze Age period only, as the dating of the two settlement sites is only secured to this period: earlier settlement may have occurred at both sites, but without an unequivocal date, this will remain speculative rather than secure. The two settlement sites involved are Thornybank, a rectangular structure that suggests temporary seasonal habitation (Rees 2002:318), though the deep post-holes could indicate a more permanent structure (Rees 2002:325), and Howe Mire near Musselburgh, an unenclosed grouping of six timber roundhouses (Cook 2004:133). There is a further site at Kaimes Hill in Ratho featuring, among many other things, a series of ramparts and a dense concentration of cup-and-ring carvings, but this has been tentatively dated to the Early Neolithic period (Simpson, Gregory and Murphy 2004:89). Even if the ramparts are instead from the Bronze Age, there is no evidence for an associated settlement at any point during this period, so this site is not included in this case study. Furthermore, there are several Late Bronze Age hoards in the area: Gogarburn Hospital, Water of Leith, Leith Citadel, t o at A thu s “eat, Duddi gsto Lo h (discussed in the previous case study, Section 3.6), Tynehead and Grosvenor Crescent, all listed in the database. All of these assemblages are confirmed hoards: there are no groups of artefacts suggested to be hoards, as they were all found in single matrices with clear associations. The Duddingston Loch hoard is of particular interest: deposited in the centre of the loch, which is immediately adja e t to A thu s seat, the fi dspot is e pa ti ula l lo ated i the atu al la ds ape. Moreover, every single item in the assemblage has been burnt, broken, twisted, bent, snapped or in some way damaged prior to deposition (Illustration 3.9.1). The symbolic killi g of i a i ate o je ts, e the eapo s, houses o potte , is a ell-documented phe o e o a ki g the e d of the o je t s ultu al iog aph , pa ti ula ly of domestic structures, and was firmly established by the Neolithic period (Stevanovic 1997). Illustration 3.9.1. The hoard from Duddingston Loch, displaying unequivocal intentional destruction prior to deposition in the water. © National Museums Scotland (via SCRAN). The bottom image shows the items studied in close detail in the previous case study (Section 3.6). L-R: NMS X.DQ303 (complete sword with major bending damage), NMS X.DQ304 (blade fragment, extensive heat damage), NMS X.DQ305 (large blade fragment) The first line of study was to determine whether the sites of metal deposition were visible from the settlement sites. This involves using a Geographic Information System programme, ArchGIS, and rasters available online, to perform a viewshed analysis. If they were visible, it would mean that there was a very strong likelihood of the depositions being used as a daily reminder of the ritual significance of the objects and the deposition itself, as well as the geographical and cultural distance between the members of the settlement community and the few who were involved in metallurgy. If the deposition sites were not visible from the settlements, it would underline the liminal nature of the depositions, and the gulf between daily life and the separate world in which the metal was deposited. Having isolated the sites to be used, their national grid references were located and used in the ArchGIS programme with a 50 metre raster from the Digimap archives on Edina.ac.uk. The Viewshed Tool was then run and the results are displayed below. As can be seen overleaf in Illustration 3.9.2, the deposition sites are not visible from the two settlement sites: Howe Mire is the upper white triangle, and Thornybank is just below it. The small visible area near to the cluster of three finds is the prominent and e og isa le e ti t ol a o, A thu s “eat, ut the depositio s the sel es e e lo ated well below the line of sight, at the base of the hill and in Duddingston Loch. Therefore, although A thu s Seat is visible, the deposition sites themselves are not. If the depositions were made by people living in the settlement communities here, the sacred landscapes were kept separate from the immediate surroundings of daily life. The deposition sites closest to the settle e ts a e a ou d k a a , hi h ould e a fe hou s alk fo a unencumbered, able-bodied person or group of people across a hilly but traversable landscape (some issues related to this are discussed below). This fits well with the theme of liminality which is prevalent throughout prehistoric ritual, and which features heavily in the cosmological reasoning behind water-based depositions discussed in Chapter 2. Illustration 3.9.2 (next page). Viewshed analysis of areas visible from the settlement sites (red circles) in Midlothian. 10 km The next logical step, then, would be to investigate whether the settlement were visible from the deposition sites. This would mean that during the act of depositing the hoards in the ground (or water), the agents of the deposition would be in visual contact with the rest of the community, despite their remote location. Using the same GIS methodology as previously, visibility from the deposition sites is displayed overleaf (Illustration 3.9.3) As there are many more deposition sites than settlement sites, the combined visible area is therefore much bigger than show on the previous illustration. The two settlement sites fall right on the very edge of the zone of visibility: Howe Mire (the upper triangle) is in fact in a very small area of non-visibility, just visible on the illustration, as it is located in a small hollow surrounded by higher ground. Thornybank looks to be just visible from the deposition sites located at Duddingston Loch and Arthu s “eat, though just a el . As the National Grid References are accurate up to 10 feet, this is likely the reason for the slight discrepancy: the visibility algorithm used by the viewshed tool is only as good as the data it receives, one of the issues discussed below (Section 3.10). It would therefore appear that settlement-deposition inter-visibility was not a significant factor in the depositions, and the disassociation between the sites emphasised the different cosmological worlds in which they existed. It is also interesting to note that the deposition sites themselves are not interisi le, e ept fo the luste s see at A thu s “eat/Duddi gsto Lo h a d the Wate of Leith/Leith Citadel. It seems that the concept of a ritual landscape would have been linked to the natural or cognitive cosmology of the land, rather than of depositing hoards where they could see or be seen by previous depositions. The lack of surviving visible markers for the hoards is symptomatic of this as well: unlike the Neolithic megaliths, which were very obvious and widely visible in the landscape, the depositions remaining for modern study went into the subsurface and stayed there, possibly with an organic marker like a timber monument or temporary hide-tent, but with no artificial long-lasting indicator that there was a hoard buried at that spot. It would therefore be fair to conclude that the hoard depositions represent the very last stage of a cultural biography that was visible to the living human community, and that each hoard was deposited individually and without recourse to any previous depositions. Even the two clusters that appear could have been deposited individually: three centuries is a long time, and it is possible that the hoards were completely obscured by earth or vegetation, and that the proximity of the depositions is due to long-term cosmological trends, rather than intentionally creating a cluster of hoards. Illustration 3.9.3 (next page). Areas visible from the deposition sites (red circles) in the Midlothian area. 10 km 3.10 Methodological considerations The viewshed analysis undertaken above represents one method in an arsenal of archaeological tools that contribute daily to the discipline. It is not, however, infallible. Viewshed analysis is the best tool for this particular study because modern urban development has meant that to visit the sites in question in person would yield flawed, if any, results: that is not to imply that the results of the GIS programme are beyond criticism. For instance, Illustration 3.9.3 features an odd, almost perfectly straight line of visibility in the upper left quadrant that does not correlate with the underlying contours or the distribution of the distribution points: one must assume that it is due to an issue with the algorithm, persisting despite multiple iterations of the programme. Fortunately, it does not affect the conclusions drawn, as the area of interest lies to the south-east. (See Fisher 1993 for an overview of the issues related to algorithms in viewshed analysis.) Relating to this, and briefly touched upon on in Section 2.1, is the issue that the viewshed algorithms assume that the viewer is fully sighted, standing upright and that there is no obstruction of the view from vegetation or poor weather conditions; it also assumes perfect vision for the whole length of the line of sight, rather than the ~1km range that people with good vision can see with clarity for. The viewshed analysis merely decides whether a pixel is visible from a fixed poi t o ot, leadi g to the falla of the i a ep ese tatio Fishe : . The results from the two iterations above show that inter-visibility between deposition and settlement sites is not a factor to consider, side-stepping the issue, but it is worth bearing in mind for comparative studies. 3.11 Conclusions This case study indicates that non-funerary metal depositions were located away from settled areas and with little reference to existing findspots. Given the secluded, secretive nature of metal smelting and forging discussed in Section 3.3, secluded deposition findspots mean that the cultural biography of a metal artefact in Scotland begins and ends in a heavily ritualised, isolated ceremony away from the spheres of domestic and economic daily life. This also correlates with the on-going discussion of liminality as an underlying principle of prehistoric ritual systems. Furthermore, despite being wary of the sacred/profane divide redolent of post-Renaissance Cartesian philosophy, the data does suggest that the depositions were kept separate from areas of settlement and farming, which is reflected in the lack of depositions from the River Severn discussed by Britton, Mülnder and Bell (2008) and Mullin (2012). This does not mean that domestic life was devoid of ritual significance, but it would appear that the depositions and their locations did not feature heavily in day-to-day life. This is of course based on one limited sample, from the shortest sub-division of the Bronze Age, from a small geographical area. Were a more comprehensive study to be done, more robust conclusions could be drawn concerning the relationship between the deposition findspots and their proximity to ritual and settlement sites. The major inference to be drawn from this chapter is that depositions were located according to factors not related to existing depositions or other visible aspects of Bronze Age ritual systems, but were instead a highly complex and site-specific system where ritual significance derived from the contents of the deposition, its location in the natural world and its location in relation to the hubs of Bronze Age communal life. The interplay between these factors would be symbolic of various states and events, and contributed to a very intricate cosmological system that cannot easily be condensed into a single theme or typical deposit. Chapter 4 Conclusions, Issues and Suggestions for Further Work 4.1 Conclusions The research aims outlined at the beginning of this paper have been completed: the specific depositional and artefact patterning observed in Scotland in the Bronze Age, of both hoards and single items, were not coincidental, accidental or economical in nature. As expected, the natural landscape, particularly water, played a major role in determining where weapons were deposited, which means that Scotland followed the same broad deposition patterns as the rest of Britain, Ireland and the Low Countries. It differs, however, in that there is no overarching theme or pattern to the water-based deposition, unlike depositions made in the Norfolk Fens or the Low Countries (Yates and Bradley 2010, Fontijn 2008) which follow quite prescriptive models. Furthermore, the preponderance of axes in Scottish hoards and single depositions throughout the Bronze Age attests to their continuing significance throughout the period, despite the introduction of dirks and swords into the o ze eapo epe toi e. The a es fu tio as oth tool a d eapo o responds with the dual role of deposition as an agent of social cohesion and an act of cosmological or ritual significance. The deposition sites in Scotland were kept separate from settlement and agricultural areas, possibly to reflect their status as a rare or special event, segregating the act of deposition from daily activities. This separation occurs in the Low Countries as well, reinforcing the notion of information and tradition exchange via trade routes as well as raw and worked materials, indicating a shared cultural tradition of cosmological fundamentals which are implemented region to region with minor but observable differences. It could therefore be possible to implement a rudimentary predictive model for identifying potential Bronze Age deposition sites in Scotland by maintaining meticulous watching briefs on the shores of rivers, lochs and smaller watercourses, and where possible by surveying (by fieldwalking or through aerial surveys) wider areas surrounding known Bronze Age settlements. However, given that there are many finds which were found outside of waterscapes or known settlement sites, this predictive model is by no means perfect. The lack of correlation with the contemporary ritual landscapes suggests that the non-funerary depositions were kept separate from other spheres of ritual life, such as rock art complexes and major burial sites: as this is based on the ritual sites that survive today, it is of course possible that there were organic components which determined the location of the deposition sites, such as orchards, burials or ephemeral wooden ritual structures, that cannot be recovered through survey or excavation. Furthermore, the lack of coherence with previous or future ritual landscapes indicates an insularity in the Bronze Age nonfunerary practices, that the patterns observed were unique to the Bronze Age and did not significantly influence future generations or cosmologies, and that the depositions were representative of the major social, technological and economic upheaval and change experienced by the Bronze Age communities. This duality of isolation and integration, of a shared and yet regionally distinct set of cosmological values, is a very interesting one and would be a valuable theme to pursue further. There is neither space nor time to fully investigate the ritual significance of Bronze Age deposition, as it is clearly more nuanced and complex that was expected at the beginning of this research. The case study from Section 3.5 demonstrates this: three deposition events in similar landscape settings in a small geographic area made within a maximum of 300 years, each displaying different damage patterns, pre-deposition usage and levels of intentional destruction. A wider study may result in patterns emerging, as a work using only three sites can hardly be considered comprehensive. However, the social implications of the depositions should also be considered: though briefly mentioned in this paper, that the depositions functioned as agents of social cohesion, through the transmission of secret knowledge among an elite few or through the communal act of gathering at a deposition site, should not be underestimated. In contrast, the settlement case study did not particularly exemplify the concept of visible liminality which was expected based on existing theoretical work undertaken in particular by Brück (1999): this may be a further indication of the unique regional traditions found in Scotland, or a problem of using a single case study to extrapolate a national framework (as above). The integration of Scotland into the wider European and Near Eastern world system is an important research question, and was investigated in relation to the non-funerary depositions using existing data concerning climate change and trade systems in Chapter 3. The impact of climate on human behaviour has undergone a recent resurgence, particularly in the popular press with the publication of Guns, Germs and Steel (Diamond 1998), and environmental determinism is being to become an acceptable theoretical paradigm, when tempered with other potential causative factors. Moreover, pragmatically, as sea levels change and modern populations migrate, more archaeological sites will become available to study and existing sites should be protected. The actions of migrating populations may also provide insights into the actions of prehistoric populations, who also faced similar challenges of changing weather systems and sea levels, and an ethnological study into climate change migrants could prove interesting and relevant to Bronze Age studies. The research undertaken here has therefore achieved the primary aim set out in Section 1.1: an overview of a specific time period and set of data has been completed, in preparation for a complete synthesis of the Scottish Bronze Age to be made in the next decade by interested scholars, as outlined as a major future research aim in the ScARF recommendations (online, listed in Web Resources section of bibliography). Given the recent launch of ScARF, the interest in Scotland as a distinct entity generated by the independence referendum in 2014 and the discoveries made by amateurs and academics alike, a wider study of Scotland in prehistory would be useful and timely. 4.2 Issues Research was undertaken with a surprisingly small amount of problems and setbacks. The p o le s ith the o pus o st u tio e e dis ussed i “e tio . a d aused o o e tha a fe da s o th of dela hile the data as e isited a d o e ted. A ess to museum collections, particularly Marischal Museum in Aberdeen, was logistically inconvenient, but thanks to the curators and email attachments, the images used throughout this paper were acquired with minimum inconvenience and no travelling whatsoever. Access to the collections held by the National Museum of Scotland was very straight-forward, kindly facilitated by Trevor Cowie and the curatorial staff. Accessing literature was primarily achieved through online resources, university resources, the National Library of Scotland and the Royal Commission of Ancient and Historical Mo u e ts a hi es, ith the assista e of staff a d a hi ists. T ai i g i the A hGI“ p og a e as o pleted du i g the U i e sit of Edi u gh s I o ati e Lea i g Week, as was a course in image manipulation and creation, both of which were instrumental in the finer details of pattern analysis used in Section 3.9. Should this research be taken forward into a more detailed and in-depth doctoral project, training in a relevant statistical package will be required. Peer reviews, criticisms and suggestions were achieved through presenting interim conclusions at the launch of ScARF at the Royal Society in Edinburgh in January, presentations at the University of Edinburgh and at the Creativity in the Bronze Age (CinBA) conference in Cambridge in April: these experiences were instrumental in refining, dismissing and exploring various ideas and research methods. CinBA in particular showcased the phenomenal impact that Bronze Age research has on a wide range of disciplines and creative constructs, as well as being inherently valuable as a set of archaeological investigations (Illustration 4.3.1). The single major issue encountered in this work, which was identified in Section 1.1, is that this research is specifically intended to be an introduction to some of the themes and potential objectives available to those working on the Scottish Bronze Age. This necessitated setting distinctly defined boundaries and limitations on the data studied, as there was neither time nor space to fully investigate every single avenue. Although necessary, this was on occasion deeply frustrating, as huge, potentially fascinating but tangential patterns and themes appeared that would make excellent candidates for future work. These unexplored avenues of research validate the work done here as an indicator of the vast potential inherent in Scottish Bronze Age studies, and will hopefully be investigated in coming years by other researchers. Illustration 4.3.1. An orrery based on Bronze Age cosmology, modelled on the Trundholm Sun Chariot, showing how Bronze Age creativity can inform and inspire academics and artists in the twenty-first century. © April Wernham, from page 31 of the CinBA LIVE Project catalogue (online resource). 4.3 Future work As mentioned in the previous section, this study was always intended to provide an introduction to the potential inherent in Bronze Age studies by exploring one aspect of the Scottish Bronze Age, rather than as a detailed work covering every aspect of Bronze Age life. As such, there are numerous research avenues suggested by this work that would be worth pursuing. Initially, widening the data set to include any known burial sites and artefact analysis of grave goods would reveal the link or distinction between funerary and nonfunerary patterning. This would be appropriate for a trans-Continental approach, but the dearth of Bronze Age burials in Scotland may make this a futile endeavour. The excellent record and research from Ireland could be a useful base from which to formulate initial hypothesis, but without a wide-ranging and well-preserved burial record, this research will be very constrained. Alternatively, incorporating all known settlement sites would show whether the conclusions drawn from the case study hold true, or whether the data set was too constrained to generate widely-applicable conclusions. For a more absolute data set, metal analysis could be conducted on a selection of weapons and known copper and tin sources, using a similar methodology to that employed by Rohl and Needham (1998) in order to determine the supply chains of native or continental metal. The project at Tonderghie may well result in a metallurgical study of this kind, though isotope analysis is very expensive. A doctoral proposal has been accepted for the author to further investigate the trade links between Scotland and northern Europe, in order to explore the duality of the integration/isolation which has appeared as an overarching theme (Section 5.1) in a wider and more detailed study. As part of this work, the corpora constructed as part of this study would be expanded and published online for use by other academics. Close involvement with ScARF would also be required. Furthermore, there is also the possibility of undertaking a use-wear analysis on the Scottish bronze halberds, similar to the project undertaken in Ireland by ‘o a O Flahe t i Du li , he e a epli a hal e d ould e ade a d used to split bone, in order to establish whether the halberds were routinely used in warfare and how effective they were as combat weapons. The limited number of Scottish halberds may be an issue, however, and the project would most likely form part of a wider study of warfare in the earlier Bronze Age. Duality would be an interesting theme of a textbook-style synthesis of the Bronze Age: liminality in cosmology and daily life relies heavily on the interaction between opposing forces; the definition of ritual that is most applicable to prehistoric studies is based exclusively on the bridging action required between two opposing forces; changing social and cosmological structures may well also exhibit a similar level of complementary duality. Given that the existing seminal work on Bronze Age Europe has cosmology and ritual reasoning as the overarching theme (Kristiansen and Larsson 2005), a similar work on Scotland in the Bronze Age with an abstract unifying theme could be an interesting and valuable approach. Therefore, this paper has achieved its primary purpose in demonstrating how Scotland in the Bronze Age is a dynamic and complex period in which the nuanced ritual system, incorporating the landscape, was developed. . Bibliography Aiti hso , N. . Roman wealth, native ritual: Coin hoards within and beyond Roman Britain , World Archaeology 20:2 (Special Issue: Hoards and Hoarding) pp. 270-284 Alcock, J. P. 1966. Celtic water cults in Roman Britain Archaeological Journal 122 pp.1-12 A es u , M. et al. . Bronze Age upland settlement decline in southwest England: testing the climate change hypothesis , Journal of Archaeological Science 35:1, 87-98 A de so , J. . Noti e of a e a ka le fi d of o ze s o ds a d othe o ze a ti les in Edinburgh; with notes on Bro ze s o ds fou d i “ otla d , Proceedings of the Society of Antiquaries of Scotland 13, 320-333 Andrews, M. 1999. Landscape and Western Art (Oxford: Oxford University Press) Bahn, P. 1989. Bluff Your Way into Archaeology (Horsham: Ravette Books) Barber, M. 2002. Bronze and the Bronze Age: Metalwork and Society in Britain 2500-800BC (The History Press Ltd) Ba do , ‘. . M th a d Metallu g : so e oss-cultural perspectives on the social ide tit of s iths i Andrén, A., Jennbert, K. and Raudvere, C. (eds.) Old Norse Religion in Long-Term Perspectives: Origins, Changes and Interactions (Lund: Nordic Academic Press) pp. 99-104 Ba ett, J. . The o u e talit of death: the ha a te of Ea l B o ze Age mounds in southern Britai , World Archaeology 22:2, 179-189 o tua Bell, C. 1997. Ritual: Perspectives and Dimensions (Oxford: Oxford University Press) Bell, C. 2009. Ritual Theory, Ritual Practice (Oxford: Oxford University Press) Blakely, S. 2006. Myth, Ritual and Metallurgy in Ancient Greece and Recent Africa (Cambridge: Cambridge University Press) Bradley, R. 1990. The Passage of Arms: An Archaeological Analysis of Prehistoric Hoards and Votive Deposits (Cambridge: Cambridge University Press) B adle , ‘. . Death Water: Boats and Footprints in the Rock Art of Western “ ede , Oxford Journal of Archaeology 16:3, 315-324 Bradley, R. 1998. The Significance of Monuments: on the shaping of human experiences in Neolithic and Bronze Age Europe (London: Routledge) Bradley, R. and Edmunds, M. 2005. Interpreting the Axe Trade: Production and Exchange in Neolithic Britain (Cambridge: Cambridge University Press) B itto , K., Mül de , G. a d Bell, M. . “ta le isotope e ide e fo salt-marsh grazing in the Bronze Age Severn Estuary, UK: implications for palaeodietary analysis at coastal sites , Journal of Archaeological Science 35:8, 2111-2118 Brück, J. . ‘itual a d ‘atio alit : so e p o le s of i te p etatio i Eu opea a haeolog , European Journal of Archaeology 2, pp. 313-344 Brück, J. 2001. Bod Metapho s a d Te h ologies of T a sfo atio i Brück, J. (ed.) Bronze Age landscape: tradition and transformation (Oxford: Oxbow Books) pp. 49-60 Brück, J. . Chapte Fou : What s i a settle e t? Domestic practice and residential o ilit i Ea l B o ze Age southe E gla d i Brück, J. and Goodman, M. (eds.) Making Places in the Prehistoric World: Themes in Settlement Archaeology (London: Taylor and Francis) Brück, J. . Mate ial Metaphors: the relational construction of identity in Early Bronze Age burials in Ireland a d B itai , Journal of Social Archaeology 4:3, 307-333 Brück, J. . F ag e tatio , Pe so hood a d the “o ial Co st u tio of Te h olog i Middle and Late Bronze Age B itai , Cambridge Archaeological Journal 16:3, 297-315 Joanna Brück (2006). Fragmentation, Personhood and the Social Construction of Technology in Middle and Late Bronze Age Brück, J. . Chapte - Fire, Earth, Water: An Elemental Cosmography of the Eu opea B o ze Age i I soll, T. ed. The Oxford Handbook of the Archaeology of Ritual and Religion (Oxford: Oxford University Press) Bu kla d, P., Dug o e, A. a d Ed a ds, K. . Bronze Age myths? Volcanic activity and human response in the Mediterranean and North Atlantic regions , Antiquity 71, 581-593 Burl, A. 1987. The Stonehenge People: life and death at the world’s greatest stone circle (London: J. M. Dent) Burgess, C. 1980. The Age of Stonehenge (London:Dent) Burgess, C. and Gerloff, S. 1981. 'The dirks and rapiers of Great Britain and Ireland , Prahistorische Bronzefund IV Bd. 7 (Munich: C.H. Beck'sche Verlagsbuchhandlung) Cassen, S. 2009. Autour de la Tabe. Explorations archeologiques et discours savants sur des architectures neolithiques a Locmariaquer, Morbihan. (Nantes: CNRS) Ca e s, G. . Late B o ze a d I o Age Lake “ettle e t i “ otla d a d I ela d: the de elop e t of the a og i the o th a d est , Oxford Journal of Archaeology 25:4, 389-412 Childe, V. G. 1935. The Prehistory of Scotland (London: Paul, Trench, Trubner & Co.) Childe, V. G. 1940. Prehistoric Communities of the British Isles (London:Chambers) Childe, V. G. 1946. Scotland before the Scots; being the Rhind Lectures for 1994 (London: Methuen) Chippendale, C. and Nash, G. 2002. European Landscapes of Rock Art (London: Psychology Press [Taylor and Francis Publishers]) Coles, J. . “ ottish Late B o ze Age etal o k: t polog , dist i utio a d h o olog , Proceedings of the Society of Antiquaries of Scotland 93, 16-134 Coles, J. . “ ottish Middle B o ze Age Antiquaries of Scotland 97, 82-156 etal o k , Proceedings of the Society of Coles, J. . “ ottish Ea l B o ze Age Antiquaries of Scotland 101, 1-110 etal o k , Proceedings of the Society of Col uhou , I. a d Bu gess, C. . The “ o ds of B itai , P ähisto is he B o zefu de. Abteilung IV Bd. 5 (Munich: C.H. Beck'sche Verlagsbuchhandlung) Cook, M. . Ho e Mi e: e a atio s a oss the op a k o ple at I veresk, Mussel u gh, East Lothia , Proceedings of the Society of Antiquaries of Scotland 134, 131160 Cook, M. . E a atio s of a B o ze Age ou dhouse a d asso iated palisade e losu e at Ai d Qua , Castle Ke ed , Du f ies a d Gallo a , Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society 80, 9-27 Cooney, G. 2000. Landscapes of Neolithic Ireland (London: Routledge) Cosgrove, D. 1984. Social Formation and Symbolic Landscapes (London: Croom Helm) Cowie, T. 1988. Magic metal: early metalworkers in the North-east (Aberdeen: Anthropological Museum, University of Aberdeen) Cowie, T. and Hall, M. 2001. Late B o ze Age Metal o k f o “ ottish ‘i e s: a rediscovered sword from the River Forth near Cambus, Clackmannanshire, in its wider context. Tayside and Fife Archaeological Journal 7 Co ie, T. a d “hephe d, I. . Chapte : The B o ze Age i Edwards, K. and Ralston, I. (eds.) Scotland After the Ice Age: environment, archaeology and history 8000BC-1000AD (Edinburgh: University of Edinburgh Press) C o e, A. . Chapte : Fo gi g a h o ologi al f a e o k fo “ ottish C a ogs; the adio a o a d de d o h o ologi al e ide e i Midgle , M. a d “a de s, J. eds. Lake Dwellings after Robert Munro: Proceedings from the Munro International Seminar: The Lake Dwellings of Europe 22nd and 23rd October 2010, University of Edinburgh (Leiden: Sidestone Press) C o e, A. a d Cla ke, C. . A p og a e fo etla d a haeolog i “ otla d i the st 21 centu , Proceedings of the Society of Antiquaries of Scotland 135, 5-17 Cunliffe, B. 2004. Iron Age Communities in Britain, (London: Routledge) Da ill, T. . Chapte : Value “ ste s i A haeolog , i Ca a , J., Coope , M., Fi st, A. and Wheatley, D. (eds.) Managing Archaeology (London: Routledge) Diamond, J. 1998. Guns, Germs and Steel: a short history of everybody for the last 13,000 years (London: Vintage) Drews, R. 1993. End of the Bronze Age: Changes in Warfare and the Catastrophe Ca. 1200 B.C. (Princeton, N.J.: Princeton University Press) Du a , L. . Flu tuati g settle e t patte s i B o ze Age “uthe la d: e a atio of a ou dhouse at Na idale, Hel sdale. Proceedings of the Society of Antiquaries of Scotland 137, 137-168 Edwards, K. and Ralston, I. 2003. Scotland After the Ice Age: environment, archaeology and history 8000BC-1000AD (Edinburgh: University of Edinburgh Press) E e ill, P. . Chapte : B itish Co e ial A haeolog : A ti ua ia s a d La ou e s; Developers a d Digge s i Ha ilakis, Y. a d Duke, P. eds. Archaeology and Capitalism: From Ethics to Politics (Walnut Creek, CA: Left Coast Press) Feil, D. 1987. The Evolution of Highland Papua New Guinea Societies (Cambridge: Cambridge University Press) Fisher, P. . Algo ith a d i ple e tatio u e tai t i ie shed a al sis , International Journal of Geographical Information Systems 7:4, 331-347 Fitzgerald, R. (trans.), Kirk, G. (trans.) and Homer (author). 2008. The Iliad (Oxford: Oxford World Classics) Fle i g, A. . Phe o e olog a d the Oxford Journal of Archaeology 18:2, 119-125 egaliths of Wales: a d ea i g too fa ? Fo tij , D. . “a ifi ial La ds apes: ultu al iog aphies of pe so s, o je ts a d atu al pla es i the B o ze Age of the southern Netherlands, c. 2300BC , Analectia Praehistoria Leidensia 33/4 (Leiden: University of Leiden) Fo tij , D. . Chapte B: E e thi g i its ‘ight Pla e? O “ele ti e Depositio , Landscape and the Construction of Identity in Late P ehisto i Jones, A. (ed.) 2008. Prehistoric Europe: theory and practice (Oxford: Blackwell Studies in Global Archaeology 12) pp. 86-106 Forbes, R. 1950. Metallurgy in Antiquity: A Notebook for Archaeologists and Technologists (Leiden: Brill Publishers) Fukuyama, F. 2011. The Origins of Political Order: From Prehuman Times to the French Revolution (London: Faber and Faber) Geertz, C. 1971. The Interpretation of Cultures (New York: Basic Books) Glob, P. 2004. The Bog People: Iron Age Man Preserved (New York: New York Review of Books) Harding, A. 2000. European Societies in the Bronze Age (Cambridge: Cambridge University Press) Helms, M. 1993. Craft and the Kingly Ideal: art, trade and power (Austin: University of Texas Press) Helskog, K. . The “ho e Co e tio : Cog iti e La ds ape a d Co u i atio ith ‘o k Ca i gs i No the ost Eu ope , Norwegian Archaeological Review 32:2, 73-94 Hi gle , ‘. . A esto s a d ide tit i the late p ehisto of Atla ti “ otla d: The reuse a d ei e tio of Neolithi o u e ts a d ate ial ultu e , World Archaeology 28:2, 231-243 I o ata, ‘. . The Po e a d Ideolog of A tisti C eatio : Elite C aft “pe ialists i Classi Ma a “o iet , Current Anthropology 42:3, 321-349. Jobey, G. 1978- . G ee K o e u e losed platfo settle e t a d Ha ehope ai , Pee lesshi e , Proceedings of the Society of Antiquaries of Scotland 110, 72-113 Joh sto , ‘. . A so ial a haeolog of ga de plots i World Archaeology 37:2, 211-223 o the a d este B itai , Judki s, G., “ ith, M. a d Ke s, E. . Determinism within Human-Environment Research and the Rediscovery of Environmental Causation , The Geographical Journal 174:1, 17-29 Kiss, V. The Life Cycle of Middle Bronze Age Bronze Artefacts from the Western Part of the Carpathian Basi , i Kie li , T. a d ‘o e ts, B. eds. Metals and Societies: Studies in honour of Barbara S. Ottaway (Bochum: Universität Bochum) pp. 328-336 Koch, J. 2006. Celtic Culture: A Historical Encyclopaedia, Volume 1 (Santa Barbara: ABC-CLIO) K istia se , K. . The Tale of the “ o d: “ o ds a d “ o dfighte s i B o ze Age Eu ope , Oxford Journal of Archaeology 21:4, 319-332 Kristiansen, K. and Larsson, T. 2005. The Rise of Bronze Age Society: travels, transmissions and transformations (Cambridge: Cambridge University Press) La gdo , P. a d Ba e , K. . The li ate of “ otla d o e the last ea s i fe ed from multiproxy peatland records: inter-site correlations and regional variabilit , Journal of Quaternary Science 20:6, 549-566 Last, J. . The agenda gap? Approaches to the Bronze Age in current research frameworks , Bronze Age Review 1, 34-47 Levy, J. 1982. Social and Religious Organisation in Bronze Age Denmark (Oxford: British Archaeological Reports, International Series S124) Ma G ego , G. . Elemental bodies: the nature of transformative practices during the late third and second millennium BC in Scotland , World Archaeology 40:2, 268-280 Ma i g, W. . Io o k hoa ds i I o Age a d ‘o a B itai , Britannia 3, 224-250 Maspero., G. 1903- . The Da of Ci ilisatio : Eg pt a d Chaldaea i i A.H. Sayce (ed.) and McClure, M. (trans.) History of Egypt Chaldea, Syria, Babylonia, and Assyria, (London: Society for Promoting Christian Knowledge) Matthe s, ‘. 438-46 . )e u: ha i ge s of doo i B o ze Age este Asia? , Antiquity 76, M Meeki , J. . Ea l Histo i settle e t e eath the G ass a ket i Edi Proceedings of the Society of Antiquaries of Scotland 140, 105-128 M O ish, D. . East Chise t a sitio , Antiquity 70, 68-76 Muhl , J. 73-82 u : itual a d u u gh , ish at the B itish B o ze Age-Iron Age . The Copper Ox-Hide Ingots and the Bronze Age Metals Trade , Iraq 39:1, Mulli , D. . The ‘i e Has Ne e Di ided Us: B o ze Age Metal o k Depositio i Weste B itai , Oxford Journal of Archaeology 31:1, 47-57 Needham, S. 1988. Selective deposition in the British Early Bronze Age , World Archaeology 20:2, 229-248 Nu , P. . The A.D. 1300 Event in the Pacific Basin , Geographical Review 97:1, 1-23 O o , M. . Chapte : Mu selli g the ou d: the use of soil olou as etapho i B itish B o ze Age fu e a itual i Jo es, A. ed. , Colouring the Past: The Significance of Colour in Archaeological Research (London: Berg) Owoc, M. A. 2004. A phenomenology of the buried landscape: soil as material culture in the Bronze Age of south-western Britain in Boivin, N. and Owoc, M. (eds) Soils, Stones and Symbols: Cultural Perceptions of the Mineral World (London: UCL Press) pp. 107–23. Parker Pearson, M. 1993. Bronze Age Britain (London: B. T. Batsford Ltd) Pa ke Pea so , M. et al. Antiquity 79:305, 529-546 . E ide e fo u ifi atio i B o ze Age B itai , Piggott, S. 1955. 'A late Bronze Age hoard from Peeblesshire', Proceedings of the Society of Antiquaries of Scotland 87, pp. 175-86 Piggott, S. 1958. Scotland Before History: an essay (London: T. Nelson) Piggott, S. . Chapte : T ade s a d Metal o ke s , i Piggott, “. ed. , The Prehistoric Peoples of Scotland (London: Routledge and Kegan Paul) ‘ees, A. . A fi st ille iu AD cemetery, rectangular Bronze Age structure and late prehistoric settlement at Thornybank, Midlothian , Proceedings of the Society of Antiquaries of Scotland 132, 313-355 Renfrew, C. 1985. The Archaeology of the Cult: the sanctuary at Phylakopi (London: British School of Archaeology at Athens) ‘o e tso , A. . ‘o a Fi ds from Non-‘o a “ites i “ otla d: Mo e ‘o a i Caledo ia , Britannia 1, 198-226 D ift Rohl, B. and Needham, S. 1998. The Circulation of Metal in the British Bronze Age: the Application of Lead Isotope Analysis, Occasional Paper No. 102 (London: The British Museum) ‘oge s, A. . Water and the Urban Fabric: a study of towns and waterscapes in the Roman period in Britain , The International Journal of Nautical Archaeology 41:2, 327-339 Sahlins, M. 1972. Stone Age Economics, Piscataway, New Jersey: Transaction Publishers “alo, M. 7-36 . Pe ipateti Adaptatio i Histo i al Pe spe ti e , Nomadic Peoples 21/22, “ h idt, P. a d Bu gess, C. Prähistorische Bronzefunde. Verlagsbuchhandlung) . The a es of “ otla d a d No the Abteilung IX Bd. 7 (Munich: C.H. E gla d , Beck'sche Sheridan, A. and Davis, M. I estigati g jet a d jet-like artefacts from prehistoric “ otla d: the Natio al Museu s of “ otla d p oje t , Antiquity 76, 812-825. Sherratt, A. . What ould a B o ze Age Wo ld “ ste look like? ‘elatio s et ee te pe ate Eu ope a d the Medite a ea i late p ehisto , Journal of European Archaeology 1:2, 1-57 “i pso , D., G ego , ‘. a d Mu ph , E. . E a atio s at Kai es Hill, Ratho, City of Edinburgh, 1964- , Proceedings of the Society of Antiquaries of Scotland 134, 65-118 “te a o i , M. . The Age of Cla : The “o ial D a i s of House Dest u tio , Journal of Anthropological Archaeology 16:4, 334-395 Stevens, F. 2008. Elemental interplay: the production, circulation and deposition of Bronze Age metalwork in Britain and Ireland , World Archaeology 40:2, 238-252 “te e s, C. a d Fulle , D. . Did Neolithi fa i g fail? The ase fo a B o ze Age agricultural re olutio i the B itish Isles , Antiquity 86, 707-722 “t a ha , ‘., ‘alsto , I. a d Fi la so , B. . Neolithi a d late p ehisto i st u tu es, and early medieval metal- o ki g at Blai hall Bu , A isfield, Du f iesshi e , Proceedings of the Society of Antiquaries of Scotland 128, 55-94 “t a ha , ‘. a d Du ell, A. Pete head, A e dee shi e, 137-171 . E a atio s of Neolithi a d B o ze Age sites ea , Proceedings of the Society of Antiquaries of Scotland 133, Swanton, M. 1997. Beowulf: Revised Edition (Manchester: Manchester University Press) Tallis, J. . The Pollen Record of Empetrum Nigrum in Southern Pennine Peats: Implications for Erosion and Climate Change , Journal of Ecology 85:4, 455-465 Te , J. . E a atio s at Li tshie Gutte U e losed Platfo “ettle e t, C a fo d, La a kshi e, , Proceedings of the Society of Antiquaries of Scotland 125, 369-427 Tho as, J. . Death, Ide tit a d Bod i Neolithi B itai , Journal of the Royal Anthropological Institute 6:4, 653-668 Tilley, C. 1994. A Phenomenology of Landscape: places, paths and monuments (London: Berg) Tilley, C. 1996. The Ethnography of the Neolithic: Early Prehistoric Societies in Southern Scandinavia (Cambridge: Cambridge University Press) Tippi g, ‘. . ‘itual Flo al T i utes i the “ ottish B o ze Age – Palynological E ide e , Journal of Archaeological Science 21, 133-139 Tolkein, J. 2005. The Lord of the Rings: The Two Towers (London: HarperCollins) Toolis, ‘. . B o ze Age pasto al p a ti es i the Cl de Valle : e a atio s at West A es, Ne to Mea s , Proceedings of the Society of Antiquaries of Scotland 135, 471-504 T ahe e, P. . The Wa io s Beaut : the as uli e od a d self-identity in Bronze Age Eu ope , Journal of European Archaeology 3:1, 105-144 Tringham, R. 2005. "Chapter 11: Weaving house life and death into places: a blueprint for a hypermedia narrative" in Bailiey, D., Whittle, A. and Cummings, V. (eds.) (Un)settling the Neolithic (Oxford: Oxbow Books) Weiss, B. . The decline of Late Bronze Age civilization as a possible response to climatic change , Climatic Change 4:2, 173-198 Williams, R. 1973. The Country and the City (London: Chatto and Windus) Yates, D. and Bradley, R. 2010. Still water, hidden depths: the deposition of BA metalwork in the English Fenland , Antiquity 84, 405-415 Ziegler, V. L. 2004. Trial by Fire and Battle in Medieval German Literature (London: Camden House) Online resources The CinBA LIVE catalogue, available online at http://cinba.net/exhibition/ (last accessed 15/03/2013) The Scottish Archaeological Research Framework, the recommendations for future research and the case study of the copper mine at Tonderghie, available online at: www.scottishheritagehub.com www.scottishheritagehub.com/content/46-research-recommendations http://www.scottishheritagehub.com/content/14-research-recommendations, http://www.scottishheritagehub.com/content/case-study-copper-mining-tonderghie (last accessed 15/03/13) The six Treasure Trove annual reports covering the period 2004-2012, available at http://www.treasuretrovescotland.co.uk/News_and_reports/Annual_reports/Annual_repor t_04_06.html (last accessed 5/12/2012) White, ‘. a d ‘i ha dso , P. . The E a atio of B o ze Age ‘ou dhouses at Old eld u , A e dee shi e , “ ottish A haeologi al I te et ‘epo t , a aila le o li e at www.sair.org.uk (last accessed 14/01/2013) Appendix The three data sets used in Chapter 2.