Hoards from Hellenistic to Kushan Central Asia: Towards Some Interpretations

Hoards from Hellenistic to Kushan Central Asia: Towards Some Interpretations Lauren Morris ABSTRACT This article is concerned with the interpretation of diverse examples of hoards (intentional deposits of valuable objects) from Central Asia’sHellenistic and Kushan periods that are traditionally understood in utilitarian terms. As ameans of comparison, it first reviews useful insights from the study of hoards in Bronze Age Europe and coin hoards, where simplistic classifications and interpretations of hoards (especially as representative of strictly ritual or utilitarian behaviour) have been increasingly problematised. The case of the Begram hoard is then discussed in reference to these insights. Arguably, this find represents not the remains of apalatial treasure or merchant’scache, but perhaps rather material selected from the property of areligious institution that was deposited and abandoned as late as the early 4 th century AD, and reflecting both ritual and utilitarian considerations. The interpretations of additional cases of diverse hoards from Ai Khanoum, Taxila, the Oxus Temple, Daľverzintepe, and Mir Zakah are then reconsidered, highlighting common shortcomings in the interpretation of coin hoards in the field, the diverse ways these finds shed light on economic, social, and ritual behaviour, and the necessity for critical interrogation of frequently presumed direct links between unrecovered hoards and expected invasions. KEYWORDS Archaeological classification of hoards; intentional deposition; ritual deposits; coin hoards; Begram; Bactria; Gandhara. Deposits of valuable artefacts are widespread in the global archaeological record, and are often described as hoards in scholarship. In some fields, considerable attention has been devoted to the interpretation, classification, and explanation of hoards, i.e., what they are, and why they were deposited. Although positions tend to vacillate over time– not only in reference to the contents and contexts of hoards, but also the period and tradition within which ascholar works– an interpretative line has often been drawn between the utilitarian and the ritual: whether ahoard seems to have been deposited for adiversity of practical reasons, or if its deposition is better interpreted as an action in awider religious framework.1 Yet, problems have also been raised with the ritual -utilitarian dichotomy too, apoint Iwill explore further below. For now, Irather wish to highlight afield -specific lacuna: such interpretative issues are rarely examined in scholarship on hoards in Hellenistic and Kushan Central Asia, apoint which became clear to me through doctoral research on the Begram hoard (see Morris 2021). The present article is thus concerned with exploring how insights and approaches from the examination of hoards elsewhere may be productively applied to hoards in aCentral Asian 1 Of course, the term ‘ritual’ is slippery. All rituals do not necessarily pertain to the realm of the sacred. Moreover, there is alongstanding (and often lampooned) tradition in archaeology to use the term to describe phenomena that are either non -functional or otherwise unexplainable (see Hodder 1982, 164; Brück 1999a). Likewise, ‘ritual’ often is used as adescriptor for religion itself, rather than as an element of awider religious framework (Insoll 2004, 12). The term is retained here for its prevalence in scholarship on hoards. STUDIA HERCYNIA XXVII/1, 151–177 152 STUDIA HERCYNIA XXVII/1 context. Ibegin by clarifying my terminology, then outline some pertinent theoretical and methodological developments in the study of hoards especially in two domains: hoards of bronze artefacts in Bronze Age Europe, and of coin hoards among coinage -using populations. Ithen consider the Begram hoard as acase study, highlighting problems and patterns in the interpretation of this find in past scholarship, and consider how theoretical and methodological impulses from elsewhere combined with are -examination of the archaeological data might lead to adifferent interpretation of this material. Finally, Iconsider some further examples of different kinds of hoards in Hellenistic and Kushan Central Asia found in areas north and south of the Hindu Kush, underlining the theoretical and methodological challenges in their interpretation, and points of potential for future enquiry. WHAT IS AHOARD? There are now numerous overlapping terms in circulation in the scientific literature to describe the broad spectrum of intentionally deposited material in the global archaeological record. To cite only English -language examples, we have hoards, caches, treasures, treasure troves, structured deposits, odd deposits, unusual, symbolic, non -domestic, placed, intentional, and ceremonial deposits, and special deposits (see Richards– Thomas 1984; Brück 1999b, 152; Hill 1995, 95–96; Hamerow 2006). Ihave chosen to retain the traditional, if problematic term ‘hoard’ here for two main reasons. The first reason is that the term feels appropriate enough to describe the Begram hoard, viewed from aconventional English -language definition: ‘An accumulation or collection of anything valuable hidden away or laid by for preservation or future use; astock, store, esp. of money; atreasure’.2 The second reason is that Ido not think that imperfect terminology, in principle, inhibits the following discussion. Hansen has noted, for example, in reference to German -language scholarship of the last decades on Bronze Age hoards, that the problematic terms ‘Hort’ and ‘Depot’ themselves have not hampered efforts towards interpretation (Hansen 2002, 96). Therefore Iuse ‘hoard’ as ageneral term to describe specific archaeological finds in Central Asia which may be broadly comparable to the situation at Begram. Here, Iunderstand ahoard to be an assemblage of more than one valuable object that appears to have been intentionally concealed or deposited. Ihave chosen to focus on depositions that typically attract utilitarian rather than ritual interpretations. The latter group of finds encompasses many easier -to -interpret deposits in known contexts, such as grave goods, stūpa deposits, and groups of votive offerings or debris thereof deposited in cult buildings and sacred spaces, such as those in the Oxus Temple. Such deposits are frequently characterised by apermanent aspect, while the hoards considered here are often interpreted to have been temporarily concealed for safekeeping, and intended for later recovery. It should be noted that the archaeological residue of regular accumulation and storage of material in designated spaces should not be considered as representing hoards in aproper sense, although storage and accumulation processes are certainly of interest for thinking about hoard formation. The rest of my criteria are admittedly contrived, more in reference to how such deposits are recovered and recognised in the archaeological record, than the human activity which produced them. The ‘more than one’ criterion is artificial, in reference to the difficulty of identifying intentionality as opposed to accidental loss or discard from imperfect archaeological data. Of course, single objects may be deposited for the same reason as multiple. Likewise, 2 Hoard, n.1 in the Oxford English Dictionary Online 2018. 153LAUREN MORRIS ‘value’ is partly culturally -constructed, apoint which has been debated elsewhere in reference to hoards.3 That said, value can be conceptualised upon aspectrum, and Ido not think it is problematic, as astarting point, to consider money (especially of gold and silver, coined or bullion), jewellery, and plate as approaching the upper end of this spectrum, alongside artefacts that had considerable antiquity, rarity, or distance from their place of production at the moment of their deposition. Ideally, ahoard should have been intentionally concealed or deposited, regardless of whether the agent understood this deposition to have atemporary or permanent aspect. Hoards deposited with atemporary intentionality become part of the archaeological record only accidentally, presumably in most cases with the loss of knowledge of the hoard’slocation (e.g., the death of the owner). Yet, ahoard with temporary intentionality may become permanent if the hoard becomes worthless. In many cases, detecting the intentionality of ahoard’snon -recovery is difficult. Indeed, some coin hoards should represent accidental losses of apurse, but conventionally we do not call them anything else but hoards. VIEWS OF HOARDS ELSEWHERE: POTENTIALS AND PITFALLS As we lack arobust literature on the interpretation of hoards specifically in reference to the archaeology of Central Asia, here Iwish to highlight some important insights produced especially in two subfields of archaeology that have dealt extensively with the classification and interpretation of hoards: the archaeology of Bronze Age Europe, and the study of coin hoards. By signposting potentials as well as pitfalls already encountered in these discourses, Ihope to highlight some useful impulses that can be transferred to the interpretation of hoards in Hellenistic and Kushan Central Asia. Since the mid-19th century, scholars of Bronze Age Europe have attempted to classify and interpret the ubiquitous phenomenon of deposits of bronze artefacts. These could include avariety of whole and fragmentary implements, as well as uncast or waste metal. The debate in this field is especially instructive in its negotiation of theoretical and methodological difficulties to make sense of often inadequately -documented archaeological evidence. Hoards have traditionally been classified according to find place and contents as either ritual or utilitarian in nature, atendency which has roots already in the late 19th century (for example, see the discussion in Müller 1897, 422–443). Shortly after, further classifications of utilitarian hoards were developed. We thus see hoards classified into votive offerings, then personal valuables hidden in times of danger, traders‘ hoards for convenient and safe storage of wares, and the remnants of smithing activity (Schumacher 1903, 90, 99–100). Such attempts are not fruitless, but have depended on sometimes faulty assumptions about the contents of hoards. For example, so -called Brucherzhorte, characterised by fragmentary objects, were long regarded as utilitarian deposits related to smithing activity, but studies of recent decades point to votive interpretations (for an overview, see Hansen 2016). Another assumption current for much of the 20th century concerned chronology; hoards were regarded as assemblages of everyday articles in use at the same time, and thus helpful for questions of dating and typologies (Hansen 2016, 197). Regardless, continued observations of wide time spans between artefacts observed in single hoards, accompanied by the increasing application of archaeometric techniques to independently establish artefact and deposition ages, have helped to dismantle this notion (Hansen 2016, 196–197). 3 For critical comments in reference to hoards of the Roman period, for example, see Millett 1994, 99–103 and Johns 1995 for aresponse. 154 STUDIA HERCYNIA XXVII/1 Indeed, the arbitrary quality of many classifications is evident when one notes strong regional trends of interpretation, even when dealing with very similar archaeological evidence (Eogan 1983, 4). Yet, somewhat more successful classifications have been developed from regional perspectives with an eye to identifying patterns in the content and context of hoards. For example, Levy’sframework for distinguishing between ritual and non -ritual hoards of Bronze Age Denmark proposed polarities in location (wet places, or places with limited retrievability, restricted in practical or ritual terms, versus dry or marked places), types of objects deposited (ornaments, weapons, ceremonial objects, associations with food, etc., versus broader ranges of types, tools and raw materials), and the conditions and arrangement of the objects (complete or near -complete objects, formally arranged, versus fragmented, damaged, and broken objects) (Levy 1982). Otherwise, ideas about location and retrievability have instigated further developments in the debate through the contextual approach advocated by Bradley (see Bradley 1990; 2017). Looking to hoards in British prehistory and especially of the Bronze Age, Bradley has tracked distribution patterns of hoards over landand waterscapes with amind to paleo -environments, and indicated that the differentiation between dry and wet places of deposition is simplistic, and that the space between ritual and utilitarian can often be problematized or collapsed. Furthermore, Bradley (2017, 10) has noted that hoards tend to be unsatisfactorily explained in an anecdotal manner– recalling strongly how the Begram hoard has been treated in scholarship (see below)– and has also challenged purely practical explanations for the deposition of valuables (i.e., with the aim of safekeeping) on the grounds of the predictability of deposition locations, juxtaposed with the frequency of their survival (Bradley 2017, 28, 49). In sum, Bradley’swork highlights the potential of tracking deposition context to observe patterns in behaviour, as well as the weaknesses of inflexible classificatory schemes especially when ritual and utilitarian hoards may also respectively have elements of the utilitarian and the ritual in each. This ritual -utilitarian dichotomous model of interpretation was also criticised by Needham, among others, who instead highlighted the flow of metal through diverse exchange systems (Needham 2001). More recent studies working beyond the Bronze Age have similarly advocated for biographical approaches to the contents, arrangements, and formation of hoards, by looking at phenomena such as use wear, evidence for repairs and fragmentation (see, e.g., Garrow– Gosden 2012; Dietrich 2014; Joy 2016). This approach opens new possibilities for making claims about accumulation and deposition processes and the social meanings that may lie behind them, and does not necessarily require acomprehensive comparative dataset. In aparallel manner, the study of coin hoards has also generated asignificant volume of scholarship largely from the 20th century, developing and debating ideas about their classification and interpretation. It should be noted that sometimes what is referred to as a‘coin hoard’ in scholarship, or analysed as such, actually includes other objects. These may be items of jewellery and plate of precious metal (including bronze or pewter), but there are cases also of lead, ivory, glass vessels and clay figurines (Robertson 1974, 26). Unlike the case of Bronze Age hoards, interpretations in the domain of coin hoards have leaned almost exclusively towards the utilitarian. Aprimary reason for this is that coin hoards usually fall under the purview of numismatists, and are accordingly viewed from amonetary perspective (Aitchison 1988, 270). Classifications have also been proposed, primarily from the contents of hoards. For example, Grierson distinguished between accidental losses, emergency hoards, savings hoards, and abandoned hoards (Grierson 1975, 130–136). Here, accidental losses are understood as purses or containers with an uneven sum of money, representing currency used in daily transactions. Emergency hoards are thought to have been concealed for safekeeping in anticipation of danger (such as warfare), representing denominations in 155LAUREN MORRIS circulation at the time of the event, with according significance for dating purposes. Savings hoards are thought to include higher -value, and relatively unworn coins minted over amore extended time period. Abandoned hoards are largely classified from contextual judgments, being apparent deposits without intention of retrieval, as in funerary, foundation, or votive deposits. Asimilar scheme offered by Casey omits the ‘abandoned’ hoards, adds that purse hoards (i.e., ‘accidental losses’) should contain smaller numbers of coins of different denominations, that savings hoards may sometimes become emergency hoards, and that the inclusion of precious metals or jewellery indicates an emergency hoard (Casey 1986, 56–57). It has also been recognised that hoards may be deposited and not recovered if their contents were rendered worthless, i.e., after acoinage reform (Reece 1981, 86–87). Reece has repeatedly scrutinised the validity of such classifications outlined above, especially when the available data is only approximate. He has argued that the difference in the contents between ‘emergency’ and ‘purse’ hoards is not detectable, and also that, presuming asaving period of twenty years, this is not long enough to differentiate between coins in circulation and coins only in savings pools (Reece 1987, 61). However, he would agree that evident selection of higher -weight coins should indicate saving behaviour (Reece 1987, 63), and that the presence of die links and coins in unworn condition within ahoard are indicators of aclose presence to the beginning of monetary circulation, via mints and state servants responsible for payments (Reece 1987, 63–65). Moreover, by tracing the history of coins since leaving the mint, it is possible to make suggestions about ahoard’sowner and patterns of monetary circulation. It should be emphasised here that the practice of hoarding coins itself cannot be associated apriori with threat, although is presumed in cases of ‘emergency’ hoards.4 We have insight only into what enters into the archaeological record– presumably unintentionally in many cases– and accordingly only apartial awareness of how routine hoarding may have been. Despite the complexities in classifying coin hoards, it remains worthwhile to develop interpretations from their composition, date, frequency, and spatial distribution. It is obviously insufficient to simply plot coin hoard distribution on amap, and to draw conclusions about presumed invasions. For example, an influential study by Kent demonstrated the increase of unrecovered coin hoards during the English Civil War (1641–1649), but saw no correlation between areas of fighting and their distribution (Kent 1974).5 With reference to the potential of hoards to inform historical interpretations, Reece has pointed out that an increase of unrecovered coin hoards must first be demonstrated spatially and temporally; if coin hoards should indicate unrest, similar hoards should be absent in areas surrounding the invasion at the same date. Temporally, the absence of similar hoards from before and after the invasion should also be argued (Reece 1981, 85). Parallel to the discussion above, there are well known examples of coin hoards associated with ritual activity, although the best examples are accumulations of long -term depositional 4 The usually -cited example for this behaviour is from the diary of Samuel Pepys in 1667, presented at length in Painter– Künzl 1997. Pepys– worried by the advance of the Dutch fleet up the Thames estuary– extracted as many gold coins as he could from his wealth in London, sent them back to be buried on the family estate in Northamptonshire, and later recovered most of the hoard, after considerable difficulty locating where his wife had buried the since -rotted bags. Aitchison (1988, 272) maintains that this analogy is invalid in reference to antiquity, due to Restoration England’sspecific societal and economic context. 5 Arecent study in Besley 2015, with the aid of many new discovered hoards, however, argues for acorrelation after all. 156 STUDIA HERCYNIA XXVII/1 practice in reference to veneration of springs and other watery features.6 Coventina’sWell in Northumberland, for example, is such an accumulation of around 16,000 coins and other artefacts deposited as votive offerings between the 2nd–5th centuries AD (see generally Allason -Jones– McKay 1985). The Frome hoard from Somerset is 52,503 coins of base silver (160 kg) of the late 3rd century AD found in apot buried in afield which had hosted several springs and produced another hoard of siliquae from acentury later. As the pot could not have borne the weight of the coins, it must have been placed into the ground empty, with the coins added thereafter (Bland 2013, 224–248). Aitchison has pointed out other instances where ritual motives may be detected, namely in patterned deposits and hoards of Roman coins at funerary monuments in Ireland (outside of their circulation in amonetary economy), and also within England, likely reflecting native pre -Roman ritual practice (Aitchison 1988, 275–277). Significantly, Aitchison also observes that the contents of most coin hoards exhibit degrees of selection– e.g. the inclusion of antique and exotic coins, or those with less intrinsic value– which do not always follow monetary criteria (Aitchison 1988, 272–273). ACASE IN POINT: THE BEGRAM HOARD To see how the above insights and examples can provide productive impulses for the interpretation of hoards in Central Asia, we can first consider the case of the Begram hoard. This necessitates afew words on the context of its discovery. During Joseph Hackin’sdirectorship of the DAFA, excavations undertaken at the urban site of Begram’s‘new royal city’ (i.e., the southern rectangular tepe, 150 × 450 m) focused extensively on the excavation area Site II (1937–1940). This area was dominated by alarge edifice (Fig. 1), at the centre of which were two rooms (rooms 10 and 13), the original entrances of which had been walled up. In these rooms, hundreds of objects were recovered in 1937 and 1939, many having been produced in diverse workshops across Afro -Eurasia. The finds from rooms 10 and 13 were published in two volumes of the Mémoires de la DAFA report series, the latter volume after the deaths of the excavators Joseph Hackin, Ria Hackin, and Jean Carl in 1941 (MDAFA IX and XI, respectively Hackin– Hackin 1939; Hackinetal. 1954). Already since the late 1930s, these finds from Begram have often been described in the literature as atreasure (trésor) or hoard (e.g., Auboyer 1938, 217), if sometimes hedged in more recent works with the qualifier ‘so -called’ (e.g., Mehendale 1996, 51). The Begram hoard has attracted much scholarly interest from the moment of its discovery, particularly for its remarkable contents produced within the orbit of the Roman Mediterranean, the Indian Subcontinent, and Han China. But evaluating the contents and context of relevant material excavated at Site II from the published reports has never been straightforward, apoint which has clearly impeded scholarly appraisals of the nature of this find.7 Indeed, the Site II structure had also been explored during Meunié’swork 6 On coin deposition in reference to spring veneration within and beyond the Roman Empire, see Sauer 2011. 7 For along time there was neither asynthesis of the excavations at Begram nor an easily -consultable account of the finds from SiteII. The finds from the Hackin excavations are dispersed in MDAFA IX and MDAFA XI, alongside finds from other areas, where they are recorded with adescription, dimensions, depth below surface, date of registration, allocation to Kabul or Paris, and sometimes an indication of association or other contextual detail. Furthermore, asingle catalogue entry does not correspond to asingle object in very many cases. In MDAFA IX, the finds are listed according to order of registration, and in MDAFA XI according to material class. The matter is further complicated as many documentation errors have crept into MDAFA XI prior to publication. For the scope 157LAUREN MORRIS in 1938 (Hackinetal. 1959, 103–105), and Ghirshman’sexcavations in 1941–1942 (Ghirshman 1946, 28, 67–69). In reference to unpublished archival documentation, Ihave now recently re -organised and analysed this material at length elsewhere (see Morris 2021). Therefore, Ilimit my following remarks to abasic account of the material, the main problems emerging from past scholarship on the nature of the hoard itself, and some findings of my longer term research on this point. Of course, the Begram hoard is foremost synonymous with the famous finds from rooms 10 and 13. More specifically, in room 10 (6 × 8.4 m) these were dispersed between horizons 2.10–2.70 m below afixed surface point (see Morris 2021, 167–172), and are largely represented by avast quantity of glass vessels, two bronze aquariums, leaded brass basins, anthropomorphic bronze balsamaria, alabaster vessels, worked ostrich eggs, and parts of ivory and bone furniture such as footstools, furniture legs, isolated plaques, and perhaps elements of an ornamental railing. Seven copper coins were also found here in association with the hoard, all at ahorizon of 2.60 m (Morris 2021, 172–174, 378–379). The finds of room 13 (6 × 9 m) exhibit even more diversity, and were dispersed between horizons 2.30–2.80 m below the surface. Fewer glass vessels are present, with more copper alloy vessels, many isolated cast bronze decorative elements, abalsamarium, figurines and utensils, over fifty plaster casts, then lacquerwares, ivory furniture elements (primarily backrests and afurniture leg), two examples of glazed pottery jugs, vessels made of porphyry and of rock crystal, some unworked lapis, amarine shell, and some fragmentary iron artefacts, including two arrowheads, as well as six copper coins. Plans of rooms 10 and 13 and the distribution of the finds within were published only in 1953 and 1954 by Hamelin (Hamelin 1953, pl. II; Hamelin 1954, pl. XLI), who worked on the excavations briefly in 1939 (Hamelin 1954, 154). For various reasons, these plans can only be regarded as broadly accurate (Morris 2021, 53, 67–69). Curiously, however, asmall amount of objects comparable to those in rooms 10 and 13 were also found in other rooms of the Site II structure, especially in the nearby room T, which had apparently been left unsealed. Although Ghirshman’sfinds from this same room have now long been clearly reported (Ghirshman 1946, 67–79; Morris 2021, 175), these and further relevant finds from the Hackin excavations have largely remained overlooked in scholarship on the hoard. This has presumably not been helped by an error in the preparation of MDAFA XI that caused the findspots of all catalogued objects from 1940 to be shifted to the previous catalogue entry in the printed report (Morris 2021, 72–73). Having reviewed the data, it is now clearer that the relevant deposits in room T, beginning from the horizon at 2.50 m below the surface, included (among other objects) some bronze figurines, detached metalwork elements, an ornately decorated box, and two glass vessels– one remarkably appearing to have been attached to two gold spouts in the form of elephant masks (see Morris 2021, 174–176). Furthermore, abronze decorative element and the remains of aquadrangular device, being perhaps aminiature shrine, were found at comparable horizons in the blocked off central corridor (i.e., directly to the west of rooms 10 and 13), as well as some coins, apottery lamp, and aterracotta spindle whorl (Morris 2021, 176). The few other comparable finds from elsewhere in the building, unfortunately difficult to more precisely localise, include abronze decorative corner element of the same type found in rooms T and 13, as well as abronze spoon and two gold pendants (Morris 2021, 176–177). These additional finds make clear that we must think of the rich and diverse material deposited at Site II in amore expansive way– certainly beyond the confines of its two famous of these legacy data, their re -organisation, and anew catalogue of the hoard objects, see Morris 2021, 50–56, 61–77, 148–177, 185–393. 158 STUDIA HERCYNIA XXVII/1 walled -up rooms. Accordingly, Iunderstand the Begram hoard broadly as aset of assemblages of intentionally deposited objects– typically distinctive and valuable– that had been sealed in rooms 10 and 13, as well as placed in room T and the central corridor. However, my position is partly apractical and artificial one, and Iacknowledge that boundaries between ‘hoard’ and ‘not -hoard’ in this space are frequently ambiguous (Morris 2021, 166–177). Iwill outline the implications of this alittle further below. By this juncture, Ihope to have already conveyed asense of the difficulties of the surviving archaeological data from Begram, as well as the interest and importance of the finds that they document. This precise imbalance of affairs seems to have shaped much of past scholarship on the hoard; indeed, although only some of the finds from Site II might cohere well with apopular or archaeological conception of ahoard, Iwould contend that decades of scholarship has effectively made them into an ideal hoard or treasure. To clarify what Imean by this, the key problem is that most scholars writing about these objects, beginning already with the Hackins and their collaborators, typically did not engage deeply with their archaeological find Fig. 1: Plan of Site II at Begram’sNew Royal City after excavations in 1937–1942, composed by Marc Le Berre in 1947, Délegation archéologique française en Afghanistan (Hackinetal. 1954, 9). 159LAUREN MORRIS context. Certainly, the data produced by the Hackin excavations are difficult and imperfect, if also internally consistent (Morris 2021, 46–77), but this tendency of scholarship has persisted even despite Ghirshman’slater excavations at Begram that developed aperiodisation for occupation at the site, Begram I–III (Ghirshman 1946). Indeed, Ghirshman proposed that the relevant phase Begram II ended with aputative invasion of Shapur Ibetween 241–250 AD (Ghirshman 1946, 100), but several rejected this date because it seemed too late in comparison to the hoard objects (see Morris 2021, 16). Most work on the Begram hoard has instead been concerned with the dates and places of manufacture of its constituent objects largely in reference to comparable material found elsewhere, as well the broader historical phenomena their existence might illuminate; these variously include the development of diverse kinds of arts (e.g. Indian ivory carving, Roman luxury glass production), wider patterns of Indian Ocean or ‘Silk Road’ trade, the role of the Kushans as middlemen of such transit trade, the kinds of sources that could have contributed to Gandharan art, or the Kushans‘ cosmopolitan or eclectic taste (see Morris 2021, 13–20). Essentially, the Begram hoard has become progressively conceptually decontextualised in scholarship and transformed into an ideal collection of wondrous and mysterious objects, as well as ablank slate for diverse interpretations that are rarely extensively rationalised by their advocates. The long scholarly impasse on the date of the hoard is aparticularly remarkable symptom of these patterns. Sometimes reference has been made in this debate to coin finds and Ghirshman’speriodisation, but arguments have always largely revolved around comparisons for the hoard objects themselves– aperennially difficult point, considering that many of the relevant objects are extremely unusual or unique in the global archaeological record (see discussions in Morris 2017, 79–81; Morris 2021, 14–15). In short, there have long been two competing positions: alate date, and an early date. Proponents of the late date include those who have argued that at least some of the hoard objects might have been produced in the later 2nd or even 3rd or 4th centuries AD (e.g., Coarelli 1962, 319; Menninger 1996, 91), while proponents of the early date consider the hoard objects as more or less coeval, having been produced in the 1st or the 1st–early 2nd centuries AD (e.g., agreed upon by the contributions in Hackinetal. 1954, 14, 54, 108, 155; more recently in Whitehouse 1989, 99; Mehendale 1997, § 5.5; Whitehouse 2001, 444). The early date has been most popular in scholarship of the last decades, and Whitehouse and Mehendale also extended their arguments to argue that the hoard should have been concealed within ageneration of 100 AD (Whitehouse 1989, 99; Mehendale 1997, § 6.4). These kinds of arguments, of course, comparatively recall the problematic assumption in some 20th -century scholarship that articles deposited in hoards of Bronze Age Europe should be coeval. Moreover, Ihave demonstrated elsewhere that three post -Vasudeva Oesho with bull coins found in room 10 in association with the hoard objects provide aterminus post quem for their deposition at least around c. 260 AD (Morris 2017, 81, 89, 95–97). Indeed, already from the perspective of the relevant material deposited in the Site II structure alone, Ialso support the late date: the majority of these objects might have been manufactured in the 1st–2nd centuries AD, but some might have been produced already in the second half of the 1st century BC, and others at least in the second half of the 3rd century AD (Morris 2021, 387). In reference to the hoard’sfind context, the question of the function and development of the Site II structure has received limited attention in past scholarship. This building has traditionally been taken to represent apalace, although some doubt has also been cast on this identification. Certainly, its large size, original rectilinear structure, and powerful walls of pakhsa on amixed stone socle, as well as traces of decorative wall painting in room 13 (Hamelin 166 STUDIA HERCYNIA XXVII/1 (often inlaid or accompanied by precious stones) and silver plate, although they sometimes include coins, and copper ornaments or plate, or other objects. 12 Yet, they were not necessarily all buried in anticipation of invasion. Sirkap was evidently awealthy city, and concealment of wealth within household confines may have been the most viable option for many to keep their belongings safe (see the discussion of treasure -troves in the Manusmṛti above). Because of their find contexts, these hoards appear to represent individual or familial wealth, rather than institutional. Yet, further information can be picked out from the contents of these hoards. For example, jewellery frequently does not appear in complete sets, and in one case, apair of silver bangles were found damaged and twisted (Block D’ House 3 Deposit D; Marshall 1951, 188). This indicates the hoarded jewellery was not in regular use, but had been accumulated elsewhere prior to deposition. Furthermore, three hoards found within another house (Block D’ House 4; Marshall 1951, 188–189) reveal abroader conception of valuables, including items related to acraft and perhaps also objects of private use; one contains primarily copper dies and ornaments apparently for jewellery production (Deposit G), the second is asum of 120 copper coins (Deposit F), and athird contains adiverse assortment of gold jewellery, silver and copper plate, copper coins, stone weights, and arange of elements, pieces and fragments of glass, agate, rock crystal, lapis, shell, and jade (Deposit E). Additionally, the frequent presence of silver plate– apparently antique in some cases– in these hoards is remarkable viewed from Sirkap II’seconomic context; from the end of Azes II’sreign, silver coinage circulating around Gandhara was progressively debased (first observed in MacDowall 1977). The hoards in this stratum thus appear to indicate that plate and jewellery of purer silver was being deliberately stored simultaneously to this debasement, perhaps being understood as amore stable form of wealth. Finally, even these apparently utilitarian hoards may have been formed in reference to ritual considerations. Awell -known bronze statuette of Harpocrates was found buried two feet above ahoard of silver plate and gold jewellery in an earthenware pot (House 1E; Marshall 1951, 159–160), which does not appear accidental. Harpocrates held multifaceted appeal by the time he entered into common personal worship in the Graeco -Roman world, and other statuettes of this deity found in Central Asia suggest at least abroader phenomenon of his private worship. Although the aspect Harpocrates takes in this context is not obvious, perhaps it is aprotective one. If we look again to Begram, we see another example of acoin hoard which appears (at first glance) to relate to the abandonment of the city at BegramIII. It was found clearly associated with the third occupation level recorded by Meunié at the city entrance, and contained 65 post -Vasudeva copper coins (Oesho with bull, seated Ardoxsho types), matching the site finds recorded for this occupation layer by Ghirshman (Hackinetal. 1959, 112, fig. M1, no. 78; Ghirshman 1946, 43, 85–86). However, its location tells us adifferent story: the hoard was found directly in the centre of the city’sentrance. Interpretations as an ‘emergency’ hoard buried for safekeeping, or an accidental loss both seem to fail from apractical perspective. Rather, perhaps this hoard could be interpreted as aritual foundation deposit after renovations at the city gate. 12 See, for example, Marshall 1951, 147, 159–160, 155–157, 180, 180–181, 186, 187, 188–189 for hoards in House 3b (sq. 31.47’), House 1E (sq. 77.66’), House 2D (sq. 58.47’); Block G’ (sq. 108.87’); Block G’ (101.86’); Block D’ House 2 Deposit A(sq. 64.92’); Block D’ House 3 Deposit C (sq. 62.113’); Block D’ House 3 Deposit D (sq. 62.113’); Block D’ House 4 Deposit E (sq. 59.114’); Block D’ House 4 Deposit F; Block D’ House 4 Deposit G (sq. 58.116’). 167LAUREN MORRIS Turning back to Ai Khanoum, another coin hoard that has been linked with invasion can reveal more nuanced information about economic and political relationships. This hoard of 663 Indian debased silver punch -marked coins and 6 silver drachms of the Indo -Greek king Agathocles was found in an oversized ceramic flask buried insufficiently in ahole in the floor of room 20 of the palace, overlaid with two stone mortars. 13 Noting heavy re -use by later occupants in room 20, Audouin and Bernard suggested that these occupants had found separate hoards elsewhere, and had concealed the present assemblage because of ‘une catastrophe soudaine, sans doute une ruée de nomades’ (Audouin– Bernard 1973, 241). As both coinages derive from adifferent currency system– and thus cannot reasonably be thought to mutely reflect widespread circulation patterns at Ai Khanoum– Audouin and Bernard instead regarded the Agathocles coins as part of asingle transfer of cash collected in the Punjab, indicating the beginning of active economic exchange between Bactria and south of the Hindu Kush (Audouin– Bernard 1974, 39). Within the punch -marked coins, they remarked upon the predominance of series Ia and Ib (76% of the hoard, now referred to as type GH575; Errington 2003, 80) as indicating aprovenance from Taxila, a‘thésaurisation rapide’, and commercial relations (Audouin– Bernard 1973, 287–289). However, the die links and wear of the coins in this hoard may reveal amore precise story. The six coins of Agathocles are virtually pristine, of which four (nos. 3–6) share the same obverse and reverse dies, and two share the same obverse die (nos. 1–2). While such observations about die links do not work for punch -marked coins– they are already classified into tight groups according to combinations of punches– Audouin and Bernard commented on the remarkable lack of wear on the GH575 coins, despite the initial blurry appearance of obverse signs, apparently caused by striking on softened or anon -homogenous metal (Audouin– Bernard 1973, 277). With reference to the discussion above, such die links and lack of wear indicate aclose link to the beginning of monetary circulation, via mints and state servants responsible for payments. Here, Imean alink in a‘network’ sense, rather than physical proximity. Accordingly, Iwould be inclined to interpret the constituent parts of this hoard as derived from direct and formal state payments (such as taxes or tribute) conveyed to Bactria, rather than reflective of broader economic or commercial relations. This aligns well with interpretations of the records on vases in the treasury of round sums of 10,000 kasapana as tribute or taxes extracted from domains conquered by EucratidesI.14 Close attention to the contents of coin hoards elsewhere can also reveal ritual associations. Remaining on the topic of punch -marked coins, the second hoard reported at Bhir Mound (Taxila)– virtually achance find near the dig house– is agood example. This hoard included 166 debased silver punch -marked coins, an unworn gold stater of Diodotus (Marshall 1951, 843), and ‘several pieces of gold and silver jewellery, besides some miscellaneous objects and anumber of small pearls, amethysts, garnets and pieces of coral’ (Marshall 1951, 110), to which Marshall adds further details later. Of the punch -marked coins, 90% are the familiar type GH 575 (Errington 2003, 80). Errington suggests this may indicate ashort time span for formation (Errington 2003, 84), but as above, we may consider aclose link with the beginning of monetary circulation. Indeed, just as for the punch -marked coins from the hoard at 13 Thirteen additional punch -marked coins were found in the earth covering this deposit and are considered together with the hoard in Audouin– Bernard 1973, 240–241. 14 For the relevant inscriptions see Rapin– Grenet 1983, 326–336, nos. 4c, 5, 6, 8d, 9–11, 12a. For their interpretation as tribute see Rapin 1992, 283 (although Rapin also suggests here that they may be booty, which is not impossible), and Mairs 2014, 50 in considering these also as sums extracted via taxes or tribute. 168 STUDIA HERCYNIA XXVII/1 Ai Khanoum, Walsh remarked that the coins in this hoard were not worn, despite the blurred appearance of the obverse strikes (in Marshall 1951, 851). Yet, the inclusion of the Diodotus stater– outside of its area of circulation and being older than the punch -marked coins– also demonstrates selectivity, which (in consideration of the other hoard objects) is not guided by solely monetary criteria. In addition to the above, there was agold and lead amulet, two miniature silver ‘reliquaries’, gold beads, pendants, disks and part of abracelet, two copper bracelets (one gilded), stone seals, two shell pendants, one in the form of atriratna, adrop of green glass, and asmall earthenware vessel (Marshall 1951, 111). This curious range of material recalls the saptaratna (the seven precious substances), which was probably understood in early northwest South Asia as gold, silver, blue precious stones, transparent precious stones, red jewels, peals, and coral. Together, they are conceptually associated in Buddhist texts with wealth, happiness, and paradises, and were considered appropriate donations to accompany stūpa relic deposits (Liu 2009, 182–183).15 Similarly, triratna beads deposited at Dharmarajika (Taxila) stūpa B6 indicate these were also ‘considered appropriate for making offerings in the relic cult’, presumably having religious and amuletic connotations (Rienjang 2017, 285). It is thus possible that this hoard is composed of intended donations for such astūpa deposit and perhaps also donations to the Buddhist community, which– in the case of the punch -marked coins– may have been given by an agent close to the beginning of monetary circulation. Yet, although Cunningham pointed to at least three stūpas of unspecified periods in parts of Bhir Mound (Dar 1993, 107, n. 55), we have no more contextual information to further explain the deposition of this hoard. It is worth mentioning here that, besides the cases of stūpa or foundation deposits, hoards of coinage found in Buddhist monastic contexts do not automatically have to be associated with donations or ritual. Despite the rule against the handling of gold and silver and engagement in ‘buying and selling’ by monks and nuns in vinaya texts, there were very numerous exceptions; in particular, abroad range of economic activities for monks are sanctioned in the Mūlasarvāstivāda ‑vinaya (referring to atime frame between the later Kushan period to the 6th century AD in North India) (Schopen 2000, 103–105). Thus we should not be surprised to find coined money circulating and hoarded occasionally in monastic contexts in the forms of private wealth. In considering coin hoards and ritual associations, the four found in the Oxus Temple (Zeymaľ 1997) provide important evidence that coin hoards can also act in avotive manner. Yet, some appear to have simultaneously been formed with more practical motives. Through close re -examination of excavation documentation, Lindström has recently observed that two containing Kushan coins (Zeymaľ’shoard 3 and hoard 2 respectively) were deposited directly above ‘Bothroi’ 4 and 3, which appear to have been previously emptied of valuable and recyclable offerings during Kushan -period works (probably in the 2nd century AD). Lindström (2016, 302–303) suggests that the coin hoards were accordingly deposited as compensatory offerings (all sanctioned by the cult personnel), presumably of lesser value. Additionally, the curious almost five -century span of coinage in hoard 2– acopper coin of Antiochus I(1), Euthydemus I(1), silver imitations of Phraates IV (69), silver and copper Heraios coins (1, and 3), and copper coins of Vima Kadphises (2)– indicate, at the very least, along accumulation elsewhere prior to deposition. 15 For acomparative excavated example see the deposit of gold sheets and disks, silver sheets and disks, apiece of quartz, crystal beads, agreen beryl bead, green glass beads, apiece of red garnet, pearls, and apiece of turquoise accompanying bones (the relics) at Kalawan, Taxila in Marshall 1951, 327. 169LAUREN MORRIS Afurther point worth raising is that even when objects in hoards are inscribed, this information is only partially helpful. To cite acase that looks fairly straightforward, of the nine silver bowls (some also gilded) published by Falk, reportedly found in ahoard at the Mohmand Agency (Pakistan), three were inscribed. The first two clearly refer to votive offerings; the first is inscribed with Greek and Gāndhārī texts recording its dedication by Kalliphon to *Chaos/Boa (No. 2), the second with aGāndhārī inscription indicating the object’sdedication by Samangaka (samagakeṇa… karavite) who had become an adorant (No. 8). The third, in Greek is translated as ‘Through [Διὰ] Phantoklēs, the Meridarchos. Dr(ahmas) 40’ (No. 4) and interpreted by Falk as also accompanying adedication (Falk 2009, 26–27, 29–31, 34–35). Falk deduces that the hoard indicates the existence of ashrine and/or temenos where such dedications could occur (Falk 2009, 40). Ido not dispute this, but would only add that we do not know whether the hoard itself was found in such aspace, and the third text, moreover, is more ambiguous in content and may refer to another process. Similarly, it often remains unclear whether acommon formula encountered in inscriptions on silver plate in other hoards– being apersonal name in the genitive case, sometimes accompanied with titles or relations, and sometimes the weight of the object– should be interpreted as representing the owner or donor of said object. The hoard of gold jewellery and silver plate from House 2D at Sirkap resembles others discussed above from stratum II in content, but six of its vessels are inscribed with various combinations of this formula, with five different names of owners/donors represented, including the kṣatrapa Jihoṇika (Marshall 1951, 155–157).16 The number of different names represented, and the proximity of the house and hoard to the so -called ‘apsidal temple’– which Marshall interpreted as aBuddhist gṛha- ‑stūpa, but is more plausibly atemple dedicated to aBrahmanic/Shaiva cult (Colliva 2007, 24–25)– led Marshall to consider, but not commit to the possibility that these were temple donations, and that the house was occupied by someone connected to the temple, or temple management (Marshall 1951, 155–156). Yet, the presence of asilver plate inscribed with the name and title of Aśpavarma the strategos in another household hoard (Block D’ House 3 Deposit D) lends support to the private ownership interpretation (Marshall 1951, 188). Acomparison with the unprovenanced silver hoard of 45 pieces of silver plate, primarily composed of drinking vessels (some fragmentary) reported from Buner in Pakistan is warranted here, as it exhibits objects inscribed with the names of six different owners/donors. Baratte already took note of the chronological delays involved in the assemblage of this hoard from an art -historical perspective, and wondered how its formation took place as the names appear to be totally unconnected (Baratte 2002, 46–50). Falk variously referred to the names as those of owners and donors, although he indicates the former in his translations with ‘[Property]’ (Baratte 2002, 58–60). Although it is possible that this hoard derives from adeposit of donations, the familiar material, the traces of damage and repairs, and the epigraphic formulae, invite comparison with the Sirkap II hoards and interpretation as objects of individual wealth. This hoard thus may also relate to the economic and historic circumstances connected to the accumulation, deposition, and non -retrieval of the Sirkap II hoards. Another difficult case is the hoard of 115 gold objects in DT-5, the ‘residence of awealthy city -dweller’ (‘Жилой дом богатого горожанина’) at Daľverzintepe. The hoard had been buried in an earthenware pot beneath the phase 2 floor in room 13 (6.8 × 2 m, separated from room 9 during renovations), and was never recovered after the house succumbed to afire (Pugachenkova– Rtveladze 1978, 35). The hoard included avariety of intact, bent, and fragmentary 16 Here, Ifollow the reading of the ka 191 in the Jihoṇika inscription as aweight– 191 ka(hapaṇa/ha‑ vaṇa) (Skt. ārṣāpaṇa)– rather than adate as in Cribb 1999, 196–197. 170 STUDIA HERCYNIA XXVII/1 pieces of jewellery, as well as blanks in the form of threads, wire, plate, and tubes, disc -shaped ingots, and rectangular ingots, 10 of these featured Gāndhārī inscriptions (Nos. 1, 2, 4–11). Pugachenkova suggested that some objects may come from ajewellery workshop, and that– as some jewellery had been broken up and flattened in an apparent attempt to squeeze more into its container– the hoard was connected with extraordinary events. Further, she noted similarities between some of the jewellery and that found at Sirkap, suggesting the house’sowner had belonged to the Kushan military nobility, and had brought back this jewellery and the ingots while participating in the campaigns into Northwest India (Pugachenkova 1976, 65–66). Unfortunately, the interpretation of the ingot inscriptions is not straightforward, not least because the published plates are not entirely clear. At least, names and weights are evident: ‘was given by Mitra’ (mitreṇa dito) appears on four (nos. 1, 2, 4, 5) and Vorobyova‐Desyatovskaya indicates this may denote atreasury official (Vorobyova -Desyatovskaya 1976, 76). Yet, her reading of the names of recipients in parts of nos. 2, 4, 6, 7, 9, and 10 (including for nos. 6–10 ś[r]amanai ‘Buddhist monk’) should be reviewed. 17 Based on the latter, Vorobyova‐Desyatovskaya had suggested that the ingots and jewellery may derive from gifts to the Buddhist community, intended to have been used for casting statuettes of the Buddha and ornaments for temple sculpture (Vorobyova -Desyatovskaya 1976, 78). Indeed, while Mitra can, sensu lato, be described as adonor here (as he has given something), aBuddhist or religious association is not necessary, and the question to whom and for what purpose these ingots was given should remain an open question until anew edition can be produced. Furthermore, there is no need to presume the deposition of this hoard was provoked by extraordinary circumstances alone; room 13, judging from its dimensions, appears well -suited to astorage function. Moving beyond inscriptions, the interrogation of hoards deposited in difficult -to -access and/or watery places, which globally tend to have ritual associations (if not exclusively), may prove fruitful for the future. For example, two relatively recently -reported coin hoards, the Khauzikhelai hoard and the Vaiśali hoard, were apparently respectively recovered in abroken vase in the bed of the Swat River and in an earthen pot in ‘the dried river bed’ by villagers collecting clay (respectively Bopearachchi 1995, 624 and Bopearachchi– Grigo 2001, 22). Yet, to determine whether they were originally deposited in rivers would require more precise verification of these reports, and indications that both rivers ran along the same course in antiquity. The potential is more salient in reference to Mir Zakah Iand II, the two deposits associated with aspring near the village of Mir Zakah, lying along the route connecting the Gardez region to the Kurram valley (Curiel– Schlumberger 1953, 67). The finds from Mir Zakah and the religious practices they speak to are discussed by Grenet (2023) in detail in this volume; here, Irestrict my remarks to the interpretation of these deposits as hoards. Indeed, it is the scale of these deposits which reiterate the high stakes of avoiding close interrogation of hoard formation processes in Central Asia. With an estimated 555,000 silver, copper, and gold coins (about four tons in weight), and 350 kg of gold and silver objects including statuettes, plate, jewellery, and votive plaques appear, they may constitute the most valuable hoard of antiquity (Bopearachchi– Flandrin 2005, 155). The chronological span of their contents is wide, including objects from the Achaemenid to the Kushan period. Yet, the picture is only partially documented. Limited excavations could at least be undertaken at Mir Zakah I(1947–1948; Curiel– Schlumberger 1953, 68–69, 93–99). Mir Zakah II (1992–1993) came to wider attention 17 Falk in Baratte 2002, 54, n. 3 demonstrates that Vorobyova‐Desyatovskaya has separated aweight measurement into two words in the cases of inscription Nos. 2 and 4, and that ś[r]amanai should be put to doubt. 171LAUREN MORRIS through acquisitions made by the Miho Museum, then represented as coming from Bactria, and accordingly considered by some as asecond part of the Oxus Treasure (as in Pichikyan 1998). Bopearachchi’sresearch has been instrumental in recovering any conception of the second deposit, tracing its path from the villagers’ excavations to the global antiquities market (see, for example Bopearachchi 1995, 612–616 and Bopearachchi– Flandrin 2005). Here Iwould like to offer only afew remarks about Mir Zakah’sclassification and interpretation. Curiel and Schlumberger first considered that Mir Zakah I’sburial and abandonment could be related to the Sasanian invasion, from the presence of coins of Vasudeva I(Curiel– Schlumberger 1953, 90). After the 1948 excavations, they suggested that the remains of rough stone walls and wooden beams had represented two sacred basins, into which small objects and coins were thrown as offerings, due to predominance of small denominations, votive plaques, and the chronological span of the coins, many of which were unworn (Curiel– Schlumberger 1953, 99). Bopearachchi cast doubt on this conclusion after Mir Zakah II’semergence, citing the high value of the objects, the absence of structures in the vicinity, and the unlikelihood that religious officials would have never cleared the basins (Bopearachchi 1995, 615). He instead suggested that the deposits are composites of collections from different origins and periods– for example, the Zoroastrian votive plaques from alooted Achaemenid temple or treasury in the Oxus valley, the Graeco -Bactrian, Indian, and Indo -Greek objects and coins from sacked treasuries or temples– that were deposited at the end of the first quarter of the 3 rd century AD (Bopearachchi– Flandrin 2005, 163–166, 245; see also the discussion and similar conclusions in Grenet 2023). Indeed, considering the interpretation of the Begram hoard that Ihave outlined above, the possibility arises that the Mir Zakah deposits– and even the earlier Oxus Treasure, with its findspot only very generally localised (most plausibly) at Takht -iKuwad/Qobad18– could be parallel manifestations of the kind of materials plausibly extracted from the property of the putative temple at Begram’sSite II before its sealing and abandonment by absconding elites: items of high -value, portable, and convertible wealth (see Morris 2021, 410–412). That being said, it simultaneously appears that the Mir Zakah deposits were also ritually formed, at least in part. The spring at Mir Zakah itself may have been aspecial or ritually- -loaded place. It is also possible that the non -recovery of the deposits was not afluke of fate; the vast volume of material involved makes it difficult to imagine that its deposition was covert or fast, and occurred unnoticed by others. This is essentially speculation, but the excavation results from 1948 certainly deserve re -examination. Curiel and Schlumberger noted that the ‘basins’ of Mir Zakah Iwere not cemented or paved, but formed instead anatural pond or collection from aspring (Curiel– Schlumberger 1953, 99). It is possible to be more specific here. Evidently, there was anatural spring at the site– the preservation of the wooden beams indicates atotally waterlogged environment– and the excavated structures appear to represent an artificial catchment installation, even if its precise function is not clear. The covered wooden gutter running along the wall GH is possibly an overflow drain. However, while almost all finds were reported in the thin layer (6–10 cm) of greenish sand excavated at Site II and noted at Site I, this cannot be the sole origin of all finds reported from Mir Zakah. The ‘terrain humide et gravillonneux’ immediately under this deposit, regarded as sterile ground by the excavators (before the trench refilled with water) (Curiel– Schlumberger 1953, 94, 99), also seems like an unsuitable foundation for acatchment installation, and may be further sedimentary infill. The finds besides coins, jewellery, and plate also deserve more attention. How should the ironwork, wood fragments, and pottery sherds picked up by Kohzad (Curi18 See also the discussion of this find in Grenet 2023; Wu 2023. 172 STUDIA HERCYNIA XXVII/1 el– Schlumberger 1953, 69) be interpreted? Likewise, the raised portion of Site IV abutting onto wall GH, interpreted as infill, contained ‘quelques monnaies, des tessons de poterie et des ossements d’animaux’ (Curiel– Schlumberger 1953, 94). What coins were these, and were they analogous to those found the deposit? Finally, although the Mir Zakah deposits are exceptional, like the Begram hoard they should not be treated as isolated phenomena. Indeed, as Ihave noted above, they may even represent parallel manifestations of similar processes, although the deposition of the Begram hoard does not seem to be directly linked to the Sasanian invasion and the retreat of Kushan power to India. For now, Ihighlight this example as only another reminder that proposed ties between archaeological abandonment events and invasion scenarios are frequently weak, and that this genre of hoard explanation remains practically anecdotal unless sustained from awider archaeological perspective. CONCLUSION Many hoards in Hellenistic and Kushan Central Asia– especially those usually interpreted in utilitarian terms– have only been subject to cursory theoretical and methodological examination in scholarship. The aim of this article has thus been to explore how insights and approaches from the examination of hoards in other domains of archaeology can be productively applied to aCentral Asian context. After highlighting especially pertinent developments in the study of hoards in Bronze Age Europe as well as coin hoards among coinage -using populations, Iexplored how these might provide some impulses for adifferent interpretation of the Begram hoard alongside are -examination of the archaeological data. Thus, instead of an inadvertently abandoned palatial treasure or atrader’sstock, Itentatively hypothesise that the Begram hoard may rather represent aselection of objects from the property of areligious institution localised in the Site II structure, with objects representing portable, high -value, and convertible wealth having been removed from this assemblage. The intentional abandonment of valuable objects in the Site II structure, and the arrangement and sealing of many of these objects in rooms 10 and 13 specifically, appears to have coincided with the wider abandonment of this part of the city by its elites. As Ipropose, this occurred perhaps in the early 4th century AD, and was probably predicated by aprocess of urban decline over the previous century. Among other points, the case of the Begram hoard highlights difficulties with classifying hoards strictly along aritual -utilitarian dichotomy, probably reflects both ritual and utilitarian considerations in its formation, and contained objects circulating between ritual and non -ritual domains over time. To provide further context for these interpretations, Ihave also considered adiversity of Central Asian hoards from the Hellenistic and Kushan sites of Ai Khanoum, Sirkap and Bhir Mound at Taxila, the Oxus Temple, Daľverzintepe, and Mir Zakah, as well as hoards with little contextual information from the Mohmand Agency and Buner. Throughout, Ihave argued that hoard formation processes should take amore central role in numismatic discourse, and moreover that hoards in Central Asia are aremarkably diverse archaeological phenomenon, and often tend to resist easy classification into strict utilitarian or ritual boxes; indeed, they also shed light on economic, social, and ritual behaviour in unexpected ways. Coin hoards do not mutely reflect neutral economic information, seemingly utilitarian hoards may be ritually formed, and inscriptions may be only partially helpful in matters of interpretation. So far, Ihave noted some issues of interest for future research, including amore nuanced approach to coin hoards and the investigation of deposits in watery features. One key future challenge that Ihave only signalled briefly, however, will be to gain insight on broader patterns of hoard- 173LAUREN MORRIS ing in specific times and places. This task would likely benefit from the application of GIS and statistical methods tailored to the imperfect data at our disposal. Then it may be possible, for instance, to determine if hoarding practices are responsive to landand waterscapes, if certain varieties of hoards are prominent in certain times and places, if broader patterns in coin hoard formation can be detected with potential interest for economic history, and finally which groups of unrecovered hoards might be explained by invasions after all. ACKNOWLEDGEMENT My thanks are due to Henry Albery, Olivier Bordeaux, and Mamta Dwivedi for their remarks on parts of this article, to Stefan Baums for helpful discussions of the Daľverzintepe inscriptions, and especially to Lara Fabian for her comments on earlier drafts. BIBLIOGRAPHY Aitchison, N, 1988: Roman Wealth, Native Ritual. Coin Hoards within and beyond Roman Britain. World Archaeology 20/2, 270–84. Allason -Jones, L.– McKay, B. 1985: Coventina’sWell. AShrine on Hadrian’sWall. Oxford. Ambersetal. 2014 = Ambers, J.– Cartwright, C.– Higgitt, C.– Hook, D.– Passmore, E.– Simpson, S.J.– Verri, G.– Ward, C.– Wills, B.: Looted, Recovered, Returned. Antiquities from Afghanistan. Oxford. Audouin, R.– Bernard, P.1973: Trésor de monnaies indiennes et indo -grecques d’Aï -KhanoumI. Les monnaies indiennes. Revue Numismatique 15, 238–289. Audouin, R.– Bernard, P.1974: Trésor de monnaies indiennes et indo -grecques d’Aï Khanoum (Afghanistan) II. Les monnaies indo -grecques. Revue Numismatique 16, 6–41. Auboyer, J. 1938: French Excavations in Indo -China and Afghanistan (1935–1937). Harvard Journal of Asiatic Studies 3/3-4, 213–222. Baratte, F. 2002: East and West. ACentral Asian Silver Hoard from the Parthian Era. London. Besley, E. 2015: Mapping Conflict. Coin Hoards of the English Civil War. In: J. Naylor– R. Bland (eds.): Hoarding and the Deposition of Metalwork from the Bronze Age to the 20 th Century. ABritish Perspective. British Archaeological Reports, British Series 615. Oxford, 181–200. Bland, R.2013: Presidential Address 2012. Hoarding in Britain: An Overview. British Numismatic Journal 83, 214–238. Bopearachchi, O. 1995: Découvertes récentes de trésors indo -grecs: nouvelles données historiques. Comptes rendus des séances de l’Académie des Inscriptions et Belles ‑Lettres 139, 611–630. Bopearachchi, O. 1999: La circulation et la production monétaires en Asie Centrale et dans l’Inde du Nord- -Ouest (avant et après la conquête d’Alexandre). Indologica Taurinensia 25, 15–121. Bopearachchi, O.– Flandrin, P.2005: Le portrait d’Alexandre le Grand. Histoire d’une découverte pour l’hu‑ manité. Paris. Bopearachchi, O.– Grigo, K. 2001: Thundering Zeus Revisited. Oriental Numismatic Society Newsletter 169, 22–24. Bordeaux, O. 2018: Les Grecs en Inde. Politiques et pratiques monétaires (IIIe s. a.C.– Iers.p.C.). Bordeaux. Bradley, R.1990: The Passage of Arms. An Archaeological Analysis of Prehistoric Hoards and Votive Deposits. Cambridge. Bradley, R.2017: AGeography of Offerings. Deposits of Valuables in the Landscapes of Ancient Europe. Oxford. Brück, J. 1999a: Ritual and Rationality. Some Problems of Interpretation in European Archaeology. European Journal of Archaeology 2/3, 313–344. 174 STUDIA HERCYNIA XXVII/1 Brück, J. 1999b: Houses, Lifecycles and Deposition on Middle Bronze Age Settlements in Southern England. Proceedings of the Prehistoric Society 65, 145–166. Cambon, P.2006: Afghanistan, les trésors retrouvés. Collections du musée national de Kaboul. Paris. Casey, P.1986: Understanding Ancient Coins. An Introduction for Archaeologists and Historians. Norman. Casey, P.– Reece, R. eds. 1974: Coins and the Archaeologist. British Archaeological Reports, British Series 4. Oxford. Coarelli, F. 1962: The Painted Cups of Begram and the Ambrosian Iliad. East and West 13/4, 317–335. Colliva, L. 2007: The Apsidal Temple of Taxila. Traditional Hypothesis and Possible New Interpretations. In: A. Hardy (ed.): The Temple in South Asia. Proceedings of the 18th Conference of the European Association of South Asian Archaeologists, London 2005, II. London, 21–27. Cribb, J. 1999: The Early Kushan Kings. New Evidence for Chronology. Evidence from the Rabatak Inscription of KanishkaI. In: M. Alram– D. Klimburg -Salter (eds.): Coins, Art, and Chronology. Essays on the Pre ‑Islamic History of the Indo ‑Iranian Borderlands. Beiträge zur Kulturund Geistesgeschichte Asiens 31. Wien, 177–205. Cribb, J. 2021: The Monetary History of Begram. In: E. Errington: Charles Masson. Collections from Begram and Kabul Bazaar, Afghanistan, 1833–1838. London, 81–126. Curiel, R.– Schlumberger, D. 1953: Trésors monétaires d’Afghanistan. Mémoires de la Délégation archéologique française en Afghanistan 14. Paris. Dar, S. 1993: Dating the Monuments of Taxila. Studies in the History of Art 31, 103–122. Dietrich, O. 2014: Learning from ‘Scrap’ about Late Bronze Age Hoarding Practices. ABiographical Approach to Individual Acts of Dedication in Large Metal Hoards of the Carpathian Basin. European Journal of Archaeology 17/3, 468–486. Eogan, G. 1983: The Hoards of the Later Irish Bronze Age. Dublin. Erdosy, G. 1990: Taxila. Political History and Urban Structure. In: M. Taddei (ed.): South Asian Archaeology 1987. Proceedings of the Ninth International Conference of the Association of South Asian Archaeologists in Western Europe, held in the Fondazione Giorgio Cini, Island of San Giorgio Maggiore, Venezia. Rome, 657–674. Errington, E. 2003: ASurvey of Late Hoards of Indian Punch -Marked Coins. The Numismatic Chronicle 163, 69–121. Falk, H. 2009: Greek Style Dedications to an Indian God in Gandhara. Indo ‑Asiatische Zeitschrift 13, 25–42. Filigenzi, A. 2019: Non -Buddhist Customs of Buddhist People. Visual and Archaeological Evidence from North -west Pakistan. In: B. Kellner (ed.): Buddhism and the Dynamics of Transculturality– New Approaches. Berlin, 53–84. Foucher, A. 1925: Notes sur l’itinéraire de Hiuan -tsang en Afghanistan. In: Études asiatiques publiées à l’occasion du 25 e anniversaire de l’École française d’Extrême ‑Orient. Publications de l’École française d’Extrême -Orient 19. Paris, 257–284. Garrow, D.– Gosden, C. 2012: Technologies of Enchantment? Exploring Celtic Art: 400 BC to AD 100. Oxford. Ghirshman, R.1946: Begram. Recherches archéologiques et historiques sur les Kouchans. Mémoires de la Délégation archéologique française en Afghanistan 12. Paris. Glenn, S. 2015: Royal Coinage in Hellenistic Bactria. ADie Study of Coins from Euthydemus Ito AntimachusI. PhD Dissertation, University of Oxford. Oxford. Gorin 2011 = Горин, А.Н.: Клад медных монет позднекушанского времени из Южного Узбекистана. Scripta antiqua. Ancient History, Philology, Arts and Material Culture 1, 369–402. Grenet, F. 2023: The Votive Plaques in the Oxus and Mir Zakah Treasures: What They Tell About the Cults. Studia Hercynia 27/1, 54–64. Grierson, P.1975: Numismatics. Oxford. Hackin, J.– Hackin, R.1939: Recherches Archéologiques à Begram. Chantier no. 2 (1937). Mémoires de la Délégation archéologique française en Afghanistan 11. Paris. 175LAUREN MORRIS Hackinetal. 1954 = Hackin, J.– Hackin, R.– Carl, J.– Hamelin, P.: Nouvelles recherches archéologiques à Begram (ancienne Kâpici) (1939–1940). Mémoires de la Délégation archéologique française en Afghanistan 11. Paris. Hackinetal. 1959 = Hackin, J.– Carl, J.– Meunié, J.: Diverses recherches archéologiques en Afghanistan (1933–1940). Mémoires de la Délégation archéologique française en Afghanistan 8. Paris. Hamelin, P.1953: Matériaux pour servir à l’étude des verreries de Bégram. Cahiers de Byrsa 3, 121–156. Hamelin, P.1954: Matériaux pour servir à l’étude des verreries de Begram (suite). Cahiers de Byrsa 4, 153–183. Hamerow, H. 2006: ‘Special Deposits’ in Anglo -Saxon Settlements. Medieval Archaeology 50, 1–30. Hansen, S. 2002: Über bronzezeitliche Depots, Horte und Einzelfunde. Brauchen wir neue Begriffe? Archäolo‑ gische Informationen 25/1-2, 91–97. Hansen, S. 2016: AShort History of Fragments in Hoards of the Bronze Age. In: H. Baitinger (ed.): Materielle Kultur und Identität im Spannungsfeld zwischen mediterraner Welt und Mitteleuropa. Akten der internationalen Tagung am Römisch -Germanischen Zentralmuseum Mainz, 22.–24. Oktober 2014. RGZM Tagungen 27. Mainz, 185–208. Hill, J. 1995: Ritual and Rubbish in the Iron Age of Wessex. Oxford. Hodder, I. 1982: Symbols in Action. Ethnoarchaeological Studies of Material Culture. Cambridge. Holt, F. 1981: The Euthydemid Coinage of Bactria. Further Hoard Evidence from Aï Khanoum. Revue Numis‑ matique 23, 7–44. Holt, F. 2005: Into the Land of Bones. Alexander the Great in Afghanistan. Hellenistic culture and society 47. Berkeley. Insoll, T. 2004: Archaeology, Ritual, Religion. London. Johns, C. 1995: Treasure. Romano -British Precious Metal Hoards. Some Comments on Martin Millett’sPaper. In: S. Cottam– D. Dungworth– S. Scott– J. Taylor (eds.): Theoretical Roman Archaeology Conference 94. Proceedings of the Fourth Annual Theoretical Roman Archaeology Conference, Durham 1994. Oxford, 107–117. Joy, J. 2016: Hoards as Collections. Re -Examining Snettisham Iron Age Hoards from the Perspective of Collecting Practice. World Archaeology 48/2, 239–253. Kent, J. 1974: Interpreting Coin Finds. In: Casey– Reece eds. 1974, 184–200. Lecuyot, G. 2013: Fouilles d’Aï KhanoumIX. L’habitat. Mémoires de la Délégation archéologique française en Afghanistan 34. Paris. Levy, J. 1982: Social and Religious Organisation in Bronze Age Denmark. British Archaeological Reports, International Series 124. Oxford. Lindström, G. 2016: Der Oxos -Tempel. Räumliche Aspekte und Kultkontinuität im Spiegel der Votivpraxis. In: S. Hansen– D. Neumann– T. Vachta (eds.): Raum, Gabe und Erinnerung. Weihgaben und Heiligtümer in prähistorischen und antiken Gesellschaften. Berlin Studies of the Ancient World 38. Berlin, 281–310. Liu, X. 2009: Buddhist Ideology and the Commercial Ethos in Kuṣāṇa India. In: J. Hawkes– A. Shimada (eds.): Buddhist Stupas in South Asia. Recent Archaeological, Art ‑Historical, and Historical Perspectives. SOAS Studies on South Asia. New Delhi, 117–191. MacDowall, D. 1977: The Context of Rajuvula the Satrap. Acta Antiqua Academiae Scientiarum Hungaricae 25, 187–195. MacDowall, D.– Taddei, M. 1978: The Pre -Muslim Period. In: F. Allchin– N. Hammond (eds.): The Archaeo‑ logy of Afghanistan. London, 233–299. Mairs, R.2014: The Hellenistic Far East. Archaeology, Language, and Identity in Greek Central Asia. Berkeley. Marshall, J. 1951: Taxila. An Illustrated Account of Archaeological Excavations. Cambridge. Mehendale, S. 1996: Begram. Along Ancient Central Asian and Indian Trade Routes. Cahiers d’Asie centrale 1-2, 47–64. Mehendale, S. 1997: Begram. New Perspectives on the Ivory and Bone Carvings.,Ph.D. Dissertation. PhD Dissertation, University of California, Berkeley. Berkeley. Menninger, M. 1996: Untersuchungen zu den Gläsern und Gipsabgüssen aus dem Fund von Begram (Afghanistan). Würzburger Forschungen zur Altertumskunde 1. Würzburg.