(2018) The Rationality and Mutability of Bullion in Hoard Deposits

The Origins and Identification of Bullion

The monetization of metal in the ancient Near East was the culmination of a trend towards commodification, whose origin in prehistory is murky. Most definitions of a ' commodity' have been developed from Karl Marx's distinction between use value and exchange value. The use value of a commodity is an intrinsic property of a thing desired or discovered by the society that uses it. The exchange value of a com-modity is an extrinsic property and refers to the quantitative proportion in which use values of one thing are exchanged for the use values of another (Gregory 1980: 10-11).

Archaeologists can address the use value of any object that forms part of an archaeological assemblage. Metal first became 'useful' in the Middle East when copper and malachite were initially hammered and annealed into objects of personal adornment around 9000 BC. The use value of metal increased with each subsequent technological innovation and discovery (see Strahm & Hauptmann 2009). Smelting in the sixth millennium lent a greater variety of forms and more robustness. The colour palette was expanded with the discovery of silver and gold by no later than the early fourth millennium (for silver see Helwing 2014: 415). Alloying in the late fourth millennium increased both robustness and the colour palette. By the beginning of the third millennium BC, the development of techniques such as repoussé, filigree, soldering, and granulation all created unlimited potential for decorating metal surfaces.

Assemblages from burials like the so-called Royal Tomb at Arslantepe (Frangipane 2007) show that by the early third millennium BC these technologies and materials were integral to a 'prestige goods' political-economic system in the Middle East (see Frankenstein & Rowlands 1978), hundreds of years prior to the earliest evidence for a bullion economy. Metal objects such as weapons, jewellery, and other forms of personal adornment were exclusionary objects, crafted from materials that were difficult to procure, and made with sophisticated craftsmanship that few people possessed or had access to. Metal objects were also biographical objects, and potent signifiers of an individual's social position (DeMarrais et al. 1996: 18). By the beginning of the third millennium BC, the 'use value' of metal was closely bound up with personal identity and projections of status and power.

'Exchange value' is the defining attribute of a commodity, yet elusive in prehistoric archaeological assemblages. Indeed, the earliest conclusive evidence that the quantitative proportions of one thing could be exchanged for another is textual, and does not concern metal (barley rations in Archaic Uruk IV and III texts; see below). One might draw two conclusions from this absence of evidence for exchange value for metal objects in prehistory. First, metal might have changed hands with values derived from (local) knowledge, cosmology and myths, aspects of craftsmanship, or from personal relations and biographical associations with metal objects. Thus the exchange of metal objects in prehistory may have been more closely associated with ' gift' values than ' commodity' values (see below). Second, and conversely, metal may have been commodified the moment it became useful, even if the technology did not yet exist to quantitatively measure its value. The reality probably lies somewhere between the two.

Commodification is accessible in the Archaic Uruk IV and III texts (late fourth millennium BC). Transactions of rations of barley to temple dependants show that the use values of volumes of barley (to make bread and beer) were exchanged for the use values of human labour measured in calendrical days (see Englund 2011). Metal was not recorded in accounts like barley was (although see below for the KU 3 sign in dairy accounts), but it does feature in lexical lists that in-clude terms for metal smiths and functional metal object forms like chisels, daggers, vessels, and arrow heads. The lexical lists also distinguish the relative value of different types of metal. Some of the terms for metal objects and metal smiths are qualified with the sign KU 3 , which can be used as a topographical adjective for a 'shining' , 'holy' , or 'sacred' place, or as a substantive for precious metal (KU 3a -UD = silver, KU 3a -NE a = gold; see Krispijn 2016: 9).

References to metal in the lexical lists show that metal and its relative value was of concern to temple scribes, even if those lists do not record the exchange value of the metal. Krispijn has raised the possibility that the appearance of the KU 3 sign in another Uruk IV-III text corpus, the so-called 'dairy texts' , might indicate an equivalence between a weight of KU 3 metal (1 gin 2 = 1 shekel) and a volume of dairy product (5 sila 3 = 5 litres), but concedes that the relationship between the two is speculative (Krispijn 2016: 9). And unlike the exchange value of metal recorded on documents from the mid-third millennium BC (see below) there is no supporting (archaeological) evidence from the Late Uruk period to suggest that metal was being weighed to measure value.

Sales inscribed on stone kudurrus and other documents from the Fara and pre-Sargonic periods (mid- Figure 1. Showing sites mentioned in the text. third millennium BC; see Gelb et al. 1991) provide the earliest conclusive evidence that the weight of metal had begun to be used as a measure of value and unit of account. But the same records are less clear on whether metal was actually used to make payments. In other words, it is not yet clear whether metal met the three basic functions of a currency: as a store of value; a measure of value; and a medium of exchange.

The most persuasive arguments for the use of bullion as a store and measure of value and medium of exchange in the third millennium BC combine later textual evidence from the Ur III period with archaeological data sets. It was only in the 1970s that bullion had begun to be recognized in hoard deposits in the Middle East (Dayton 1974;Powell 1978, for distribution of Early Bronze Age and Middle Bronze Age bullion deposits see Figure 1). These metal forms (see Figure 2) had previously been considered the scrap of metal smiths or jewellers (for earlier publications: see Matthews in Steele et al. 2003: 206;Peyronel 2010: 928). The initial identification of bullion was based primarily on the reading of a Sumerian term for 'ring' in Ur III treasury archives. Powell (1978) in particular drew attention to a ring/spiral metal form found in hoard deposits in Mesopotamia (see below), the so-called 'hair rings' (Lockenringe), in the archaeological literature. Lockenringe from mid-to late third millennium contexts in Syro-Mesopotamia appear to correspond with the silver, gold, and copper rings (or coils/spirals) that are referred to in the Ur III treasury archives as ḫar (see originally Powell 1978). These ḫar are mobilized in transactions, and value references to ḫar are normally recorded in five-shekel units (e.g. as ninety five-shekel silver ḫar from twenty minas; Powell 1978: 213-214). An equivalent to the ḫar is first attested in Old Akkadian as šewirum (Powell 1978: 213). The currency function of Lockenringe in the archaeological assemblages is worth considering, particularly as the coilshaped hair-like rings evoke the coil/spiral ḫar/šewirum forms recorded in Old Akkadian and Ur III texts (see also Lassen 2000: 242;Bobokhyan 2008: 63;Paoletti 2008;Peyronel 2010: 933-934).

The prevalence of pan balance weights in contemporary archaeological assemblages is a related development (see Rahmstorf 2006a;2006b; this volume; see several papers in Ascalone & Peyronel 2006). Metal objects were weighed on a scale to arrive at a commodity value by the middle of the third millennium BC. Yet all metal objects in any form were potential reserves of wealth that could have been mobilized (or liquidated) in exchange transactions (Sherratt & Sherratt 1991: 360); and the value of a metal object can be derived from more than just its weight. Value can also derive from the complexity or exquisiteness of the craftsmanship of a metal object, or the biography of that object (dependent upon who last owned the object, and from where it had come). Variables of craftsmanship and biography can potentially be disruptive to assessments of 'pure' commodity value, measured in the weight of metal. Variables of craftsmanship and biography can be reduced if metal is in the form of scrap or ingots (for discussion of ingot forms in the Early Bronze Age, see Bachhuber 2015: 160-161).

Bar and bun ingot forms have been identified in midto late third millennium contexts at Tell Brak (Matthews, in Steele et al. 2003: 203-208), Tell Khafaja (Delougaz 1952: pl. 107a, 183), Tell Taya (Reade 1968: 248), Mahmatlar (Koşay & Akok 1950), and Troy (Bobokhyan 2006: 87-90, table 5;Müller-Karpe 1994: 141-42). Large caches of Lockenringe assemblages from contemporary Tell Khafaja (as above), Tell Chuera (Moortgat 1960: 7-8, figs 11-12), Tell Asmar (Delougaz et al. 1967: 223), Tell Taya (as above), Troy (as above), and Tell Brak (Matthews 2003: 203-208, figs 6.14-19) probably also represent ingot forms (for Tell Brak see Figure 2). Scrap metal and ingot forms have been recorded together with pan balance weights in a few contexts from the third-and early second millenniums (for Larsa see Arnaud et al. 1979; for Alişar see Schmidt 1932: 36, 57-61; for Troy see Bobokhyan 2006). The regular notching on gold and electrum bar ingots from Troy have convincingly been demonstrated as a kind of mensuration based on units of 5.0 to 5.5 g (roughly ½ shekel; Bobokhyan 2006: 87-90, table 5). Similarly, detailed statistical and archaeometric analyses of scrap metal hoards have shown a correlation between the weight and shape of individual pieces and the shekel-based weight system (Ialongo et al. 2018). The unequivocal identification of bullion, in the form of scrap metal and metal ingots, and in the pan balance weights that were used to weigh these forms, suggeests that metal had become fully commodified by the mid-to late third millennium BC. In so doing metal (particularly silver, see Helwing 2014) underpinned a financial system that prevailed in the Middle East and eastern Mediterranean until the advent of coinage, two millennia later. Yet explanations for this innovation lie at the heart of a calcified debate between pro-market (formalist/metallist) versus anti-market (substantivist/chartalist) mentalities. In the following section I discuss how this ideological impasse on the origins of money in the ancient Near East reveals both paradigmatic divides between archaeological and philological approaches to the problem, and biases in interpretations of bullion hoard deposits.

Bullion and the Early Bronze Age Political Economy

The formalist/metallist position is predicated on the intrinsic value of metal, and has appealed to archaeologists for its materialist foundation (for example Sherratt 2004: 86;Renfrew 2001). The liquidity of metal, its relative scarcity, metal's luminous surfaces, and the labour intensiveness of mining and production, are all attributes of value that are appealing, in one way or another, to archaeological forms of enquiry. From this perspective, all metal objects were ' commodities' that were exchanged with a degree of rational economic calculation. The innovation of bullion in the mid-to late third millennium BC facilitated the exchange of these commodities, and has been seen as confirming Adam Smith's epochal hypothesis on the origins of money. The intrinsic value of metal was recognized and agreed upon in market-like conditions as a rational solution to the cumbersome barter systems that had supposedly existed beforehand (for prehistoric barter in particular see Kletter 2013; Renfrew 2001: 106; and for the emergence of markets in the ancient Near East, see Silver 2004).

Flaws in this formalist logic have nevertheless been exposed in anthropological and philological assessments. The most comprehensive ethnographic studies on barter, for example, are persuasive in their arguments that no pure barter economy has ever been recorded (Dalton 1982). Rather, these kinds of exchanges can only be observed on societal margins, such as between total strangers or even between enemies (Humphrey 1985). The conclusion that can be drawn from ethnography is that a pure barter economy (in prehistory) should probably be excluded as a precondition for the initial monetization of metal (see also Smith 2004: 91), whilst leaving open the possibility for barter/commercial conditions in some contact situations.

The formalist/metallist position also runs counter to much of the evidence for despotism, centralization, and indebtedness to state institutions in early Mesopotamia, precisely when and where the bullion economy is first conclusively attested. Administrative documents lend support to the substantivist/chartalist position. In early Mesopotamia the value of metal was a creation of the temple state, which required value equivalencies for materials like metal to administer payments and debts (Renger 2002;Hudson 2004;Semenova 2009). Bullion emerged from this grid of price equivalencies (Hudson 2004: 309), with metal objects being converted to a form of currency with which subjects could pay their debts. Payments into temple treasuries were sometimes couched in a language of ' giving to the gods' (Sumerian níg-ba = ' gift'; see Ur III gifts of ḫar silver rings, below), foregrounding both the ritual-cum-religious aspects of these transactions (see Wengrow 2011: 138), and the statist context within which the value of metal commodities was ultimately determined. From this perspective the value of metal was ' embedded' in the ideology of Mesopotamian elites. Metal could necessarily be invested in the cosmological and moral framework of Mesopotamian society in ways that may seem to be economically irrational.

The extent to which the value of metal commodities could have been negotiated outside of temple or palatial contexts remains hypothetical for the third millennium BC. Distinguishing ' gifts' from ' commodities' in both the documented records of exchanges in the third millennium BC and in the archaeological record is also problematic. At Ebla, for example, the Palace G archives record gifts consigned to individuals or gods (Sumerian níg-ba). Gold and silver jewellery, inlaid weaponry, gold and silver vessels, and volumes of silver were given and received as gifts in several political guises, including gifts for royal weddings, funerary gifts, gifts to commemorate the birth of a royal child and, most often, gifts delivered in diplomatic correspondence (see Viganò 1996: 57-60;Archi & Biga 2003;Peyronel 2014: 362-364). Likewise, the so-called ring monies from the Ur III treasury archives (ḫar, as above) circulated exclusively within a ceremonial framework of gift exchanges involving the king, normally with other elites but also sometimes as gifts (votive offerings) to the gods (Paoletti 2008: 151, n. 96).

Many of the metal objects listed in the gift inventories of the Palace G archives, and all the Ur III ḫar objects from the Ur III treasury texts, were nevertheless qualified and quantified by a commodity value: their weight in minas and shekels. As ' gifts' , these objects were hypothetically inalienable and possessed an indissoluble bond with their original owner. A gift necessarily carries an emotional or psychological burden, and calls for reciprocation upon receipt (Gregory 1980: 18-19, developing from Marcel Mauss). But the commodity value of these same objects suggests that they were also alienable. The emotional, psychological, or spiritual bond between the metal object and the gift giver and and/or receiver could have been broken with the liquidation of the metal object, reduced (or even smelted) to its pure mina or shekel value. The identification of commodity value in the third millennium BC raises the possibility that metal objects could have been transferred as private property, free of reciprocal obligations (Gregory 1980: 10-12, developing from Karl Marx).

Gift exchange establishes a relationship between the individuals involved in the transactions discussed above, while commodity exchange establishes a relationship between the objects. Another factor that blurs the distinction between gifts and commodities was addressed by Bourdieu (1977: 171): both are exchanged through calculated, rational self-interest. A gift transaction has the ultimate effect of projecting into the future by guaranteeing the circulation of desirable commodities. Gift exchange and commodities exchange thus meld diachronically into a seamless transaction. In the Ebla Palace G archives precious metal objects were exchanged as gifts but circulated as commodities.

The circulation of commodities raises the possibility that the value of metal objects could have been negotiated in market-like conditions. Certainly 'merchants' (Sumerian dam-gàr) did travel long distances to engage in trade, but this was done on behalf of the palaces and temples that employed and deployed them (for a recent overview of literature, see Aubet 2013: 127-128). The activities of 'merchants' described in the Ebla archives, for example, are overtly political (as above, see also Peyronel 2014: 362-364), and contrast with the private initiative that is better attested in the later Old Assyrian archives (early second millennium BC; Larsen 1976). Several merchants from Mari, for example, are revealed as high-ranking palace personnel in the Ebla archives, including one who served as a messenger (ambassador) for the king (Viganò 1996: 22, n. 62).

Neither commercial, privately sponsored activities, nor markets are obvious in the Ebla archives, but they may have been more prevalent than the record of palatial administration allows. One possible context of commercially-motivated exchange at Ebla has been suggested by Biga (2003), through a reading of the Sumerian term KI:LAM 7 as 'market' or 'fair' , recorded in contexts of monthly festivals for deities at Ebla. Biga has suggested that merchants converged at festivals and ostensibly created markets to exchange things. Thus, festivals at Ebla would provide the earliest evidence for a predominantly commercial, rather than a predominantly socio-political, context of exchange (Biga 2003: 65-68). The Ur III archives from the end of the third millennium reveal that some temple-employed merchants could operate independently in local transactions related to money lending and retail trade (see Steinkeller 2004).

Commercialization is even more hypothetical in the non-archival regions of Anatolia and the eastern Mediterranean in the third millennium BC. Yet there are reasons to suggest that the value of metal might have been negotiated in market-like conditions in the 'periphery' , perhaps more so than in Syro-Mesopotamia. In recent years, long distance trade has been the focus of considerable archaeological scrutiny in the archaeology of Early Bronze Age Anatolia and adjacent regions (see for example Şahoğlu 2005;Rahmstorf 2006a;2016;this volume;Efe 2007;Sarı 2012: 214-215;Bachhuber 2015: 150-169). Much of this literature has focused on the distribution and use of Mesopotamian weight metrology technologies (as above). The rapid uptake of this technology spread quickly across ancient western Asia and the eastern Mediterranean in the mid-to late third millennium BC, so it can only have happened in contexts of exchange (see also Rahmstorf;2006a;2006b;2016).

Two considerations support increased commercialization in this particular historical milieu. The diffusion of the bullion technology broke down ideological barriers between societies with divergent cosmologies, moralities, and regimes of value (Wengrow 2011: 141), ultimately facilitating the circulation of things like metal across an increasingly expansive network of exchanges. Bullion was a 'neutral' or purely commodified metal form (Bachhuber 2015: 161), and could be exchanged with values that were not affected by the ideologies of transactors from different cultural backgrounds.

An inverse scalar relationship may have also existed between the level of state power and the extent of commercial activity in the third millennium BC (see Smith 2004: 93). Networks of exchange extended from the powerful and centralized temple and palace states of Syro-Mesopotamia towards smaller-scale polities, for example those that inhabited the citadels of western Anatolia and the eastern Aegean (see Bachhuber 2015). Commercialization during the Early Bronze Age may have increased with distance from Syro-Mesopotamian temple/palace states. Comparable arguments have been put forward for the Assyrian trade with central Anatolia during the early second millennium BC (Larsen 1976), and for the westward expansion of maritime trade networks in the Late Bronze Age eastern Mediterranean (Sherratt & Sherratt 1991).

The same bullion technologies were used in divergent transactional contexts across the Middle East and eastern Mediterranean, on a spectrum between the less commercial and the more commercial. Likewise, the reasons to intentionally deposit bullion forms into the ground would have varied. I am wary of the archaeological tendency to frame debates between static structural oppositions (Wengrow 2011: 136), whether between market and state as above or, in the case of bullion deposition, between economically rational (mundane) and magico-religious (ritual) reasons to conceal it in the ground. By foregrounding a few contexts of bullion deposition that nevertheless resist economically rational explanations, I consider the contextual mutability of bullion forms (see Aswani and Sheppard 2003: 54) as they flowed into and out of divergent transactional contexts, including those with cosmological entities.

Interpreting Bullion Deposits

The vast majority of scrap metal and ingot finds in the Middle East and eastern Mediterranean are associated with two kinds of context: shipwrecks (which are beyond the scope of this paper); and 'hoards' (see map in Figure 1 for hoards cited in this paper). Bullion forms were deposited in hoards in the Middle East from the middle of the third millennium BC (later Early Bronze Age) to the last centuries of the first millennium (for earlier bullion hoards, see Peyronel 2010; for later bullion hoards, see Thompson 2003;Thompson & Skaggs 2013;several papers in Balmuth 2001). The highest frequency of bullion deposition is associated with two hoarding horizons: the later Early Bronze Age, which is the focus of this paper; and the Iron Age (ca. 900-600 BC). Bullion forms were often consigned to ceramic jars, or wrapped in textiles and placed beneath the floors of buildings, with a few notable exceptions that I review below.

In many respects, the study of bullion hoards has resembled the study of later coin hoards (for example in Iron Age and Roman Britain). Traditionally, coin hoards were the academic remit of numismatists who studied coins as texts and images in a monetary history. Until the 1970s there were few contextual (archaeological) approaches to coin hoards, partly because the reasons to hoard coins was seen to be self-evident. Coin hoards represented nothing more or less than the cautionary concealment of wealth (for overview see Aitchison 1988: 271-272).

In the 1980s, coin hoards entered the broader (archaeological) literature on metal hoard deposits, mostly in European and British prehistory. Distinctions that were traditionally made between 'ritual' and 'profane' hoard deposits of metal objects (for an overview, see Fontijn 2002) had begun to be made for coin hoards. Archaeologists foregrounded idiosyncratic features of some large-scale coin deposits, including their placement in or around megalithic burials that often predated the coin deposit by millennia (see Aitchison 1988: 276), clusters of coin hoards in a conspicuous and extramural landscape setting (see Bland 2013: 219-220), or the identification in hoards of coins from regions or time periods that would not have recognized them as a valid currency (Aitchison 1988: 273). Each of these examples defies economically rational explanations and has raised the possibility for the 'ritual' deposition of some coin hoards (see also Hobbs 2006).

Similarly, bullion hoards in the Middle East have been explained as 'profane' and economically rational, often as a response to some immanent attack on a city (for Early Bronze Age bullion deposits see Matthews 1994;Steele et al. 2003: 207-210; for Iron Age bullion

The mutability of bullion blurs the distinctions that are normally made between 'ritual' versus 'profane' deposits of metal objects (see Fontijn 2002: 15-18). Much of this distinction has been bound up in the formalist/metallist versus substantivist/chartalist polarization, and particular approaches to ritual that explain it as symbolic, non-functional, or impractical action, to be contrasted with rational, empirical, and practical action (see Brück 1999: 318-322). In archaeology this has led to a conception of ritual as being spatially, temporally, and conceptually removed from the mundane activities of everyday life (also Chadwick 2012).

Magico-religious practices and economic rationality need not be mutually exclusive in the interpretation of bullion hoard deposits. For example, a (third millennium) Sumerian text describes activities of silver concealment by giving voice to copper, in a dispute between the copper and silver over value:

Silver, only in the palace do you find a station, that's the place to which you are assigned. If there were no palace, you would have no station, gone would be your dwelling place . . . In the [ordinary] home you are buried away in its darkest spots, its graves, its 'places of escape' [from this world] . . . Like a god you don't put your hand to any useful work. How dare you then assail me like a wolf? Get to your dark shrines! Lie down in your graves! Thus ends copper's speech. (Translation from Kramer 1963: 265) Here silver is something concealed, in graves as funerary gifts, beneath the floors of buildings in 'places of escape [from this world]' , and in 'dark shrines' probably adorning the cult statues of gods. The most profane-sounding of these contexts, 'the places of escape' in '[ordinary] homes' , suggests acts of cautionary concealment. But such acts should not be sharply distinguished from ritual deposits in temples and graves that could also be seen as secure places to deposit wealth.

The Neo-Assyrian incantation cited at the head of this paper, titled 'To Block the Entry of the Enemy in Someone's House' , illustrates how an economically rational act of concealment could also be a magico-religious one, albeit in a different social and ideological milieu from the Sumerian example. I cite the incantation in full:

Clay pit, clay pit you are the clay pit of Anu and Enlil, the clay pit of Ea, lord of the deep, the clay pit of the great gods; you have made the lord for lordship, and you have made the king for kingship, you have made the prince for future days; your pieces of silver are given to you, you have received them; your gift you have received, and so, in the morning of Samas, I pinch off the clay NN son of NN; may it be profitable, may deposits see Gitin & Golani 2001: 30;Stern 2001: 19). Peyronnel (2010: 931) has made a rare contextual distinction between the security of public institutions, such as temples and palaces in Syro-Mesopotamia, and the insecurity of private households. For houses, where bullion hoards were often placed, metal objects were concealed under floors by households that were otherwise unable to safeguard their wealth from theft.

Yet like coin hoards, there are a few intentional bullion deposits that resist economically rational explanations. A bullion jar concealed beneath the floor of a house in the Middle Bronze Age city at Ebla resembles other bullion hoards in most respects, yet it was deposited as a funerary gift in a sub-floor burial (Baffi 1988: 4). The mortuary context of the Ebla bullion jar locates this particular assemblage in a dedicatory, non-retrievable, and 'ritual' domain. At Level 5 of Tell Brak (Akkadian period = later Early Bronze Age, see Figure 2), bar ingots together with jewellery and other metal forms were placed into a perishable container and deposited on top of the fill of a recently-abandoned temple complex. Numerous slaughtered equids and the partially-dismembered remains of three (human) individuals were also carefully placed on the same fill (Oates & Oates 2001: 41-53). Each of these deposits form part of what the excavators have described as a ' closure ritual' for this temple complex (Oates & Oates 2001: 50;Oates 2008).

The bullion deposits from Ebla and Tell Brak are not unlike the records of votive offerings of five-shekel ḫar currencies ('ring monies') in the Ur III treasury archives discussed above. Each of these contexts reveals the 'mutability' of bullion forms (see Aswani & Sheppard 2003 for ring monies in Melanesia). Objects that were produced and transferred initially as alienable commodities circulated into an altogether different sphere of exchange. Bullion metal forms were circulated through what Bloch & Parry (1989) differentiate as "long-term orders of exchange" and "shortterm orders of exchange." The former is based on regimes of value that uphold the moral and cosmological order of the society (or "long-term restorative cycles"), while the latter short-term cycle defines the regular transactions that occur in the day-to-day. The two modes of valuation do not interface or compete. When money circulates through a short-term cycle it is used as a standard of valuation to buy and sell in impersonal, self-interested, and morally equivocal transactions. Money acquired from these transactions is then invested in long-term orders of exchange and undergoes a symbolic and cosmological transformation, where it is spent (or morally purified) in economically-irrational ritual activities that nevertheless uphold the moral order of a community (see also Shennan 1993;Wengrow 2011: 141 Wiggermann 1992, Text I) In this incantation, silver is concealed in a pit, presumably beneath the floor of a house, in anticipation of a violent intrusion, but it is also a votive offering to the gods Anu and Enlil. In graves, in shrines, and in votive pits, metal objects become 'tokens of hidden power' (Graeber 1996), or objects that have been transformed into actions through communion with the cosmological realm. Pits (and graves) highlight the liminality of the ground surface, between the day-to-day routines of lived experience and those entities that inhabit the cosmological realm (Nakamura 2004: 18).

Sealing practices associated with some bullion hoard deposits also blur the boundary between the formalist/metallist (modern bureaucratic) and substantivist (pre-modern magical) rationalities of these contexts. Cylinder seal-impressed cretulae were deposited in bullion hoards at Early Bronze Age Tell Brak (Steele et al. 2003: 204) and Middle Bronze Age Larsa (together with a cylinder seal; see Arnaud et al. 1979: 17-18). The apotropaic functions of both seals and sealings are widely acknowledged (Collon 1987: 119;1997: 19-20;Charvát 1994;Simko 2015), yet the relationship between the bureaucratic function and sympathetic magic of this technology has attracted little scrutiny. When commodities were sealed by a cylinder for storage or circulation they also entered a liminal state. They became "neither used nor abandoned, neither fully given nor fully received-a moral status perhaps akin to that of things placed with the gods or with the dead" (Wengrow 2008: 15). Bullion forms, sealed by both the ' clay' matrix of the pit and the sealings they were sometimes deposited with, underwent a transformation from an object produced as a pure commodity, to a dedicatory object that was both magically effective and retrievable.

Conclusion

To what extent ethnographic analogy is appropriate for the societies of the ancient Near East and eastern Mediterranean touches upon a number of disciplinary, paradigmatic and interpretive divides. The largest divide is located between divergent academic traditions, in particular between Anglophone and Continental European approaches to the archaeological past. Ethnography, whether used as a proxy for understanding the archaeological past or as a methodology for understanding discourses around archaeology today, fits more comfortably in an Anglophone tradition that equates archaeology with anthropology. Since the 1960's archaeologists in the Anglophone tradition have explicitly or implicitly embraced the primitivist/ substantivist paradigm through their use of ethnographic analogy.

Ethnographic analogy is less appropriate in a Continental European tradition that locates archaeology in history rather than in anthropology. Consequently, discussions of bullion and metal more broadly have featured in economic histories on the origins of a metals-based financial system. A parallel historicist approach has examined the diffusion of this technology from Mesopotamia towards prehistoric Europe. The historical and historicist approaches are less wellequipped to explain the varied uses of bullion in a local, socially-mediated context, or the evolutionary origins of the technology.

A second divide can be located in the substantivist/ formalist debate. For the formalist paradigm ethnographic analogy is simply inappropriate in research on the ancient origins of a 'modern' economic activity like using metal as a measure of value, store of value, and medium of exchange. Yet, I tried to show with examples from the Mesopotamian text corpus and from archaeological contexts that metal could be invested in transactions where 'pre-modern' magico-religious values prevailed. Ethnography and ethnographically-informed archaeology can necessarily provide greater resolution on a particular transactional context: the bullion hoard deposit. Concealing bullion in the ground can represent simultaneously a rational, economic decision to safeguard wealth, and a magico-religious appeal to protect person and property.