The sanctuary of the Temple à redans in Ai Khanum (Hellenistic Bactria): cults and ritual practices
Abstract
98
Full text
The sanctuary of the Temple à redans in Ai Khanum (Hellenistic Bactria): cults and ritual practices Laurianne Martinez -Sève ABSTRACT One of the main sanctuaries in the city of Ai Khanum was excavated between 1968 and 1973 under the direction of Paul Bernard and its study is nearing completion. The purpose of this article is to describe how the sanctuary was organized and the different rituals celebrated within its walls. The sanctuary comprised several buildings standing around acourtyard, which formed the core of the sacred space. The lands extending on either side of the courtyard also belonged to its properties and were used for economic activities. At least two deities were honoured by rituals that generally left few traces. They received votive offerings, but also cereal -based food offerings, perhaps consumed during sacred banquets. Foundation and purification rituals can also be highlighted. KEYWORDS Ai Khanum; Hellenistic Bactria; Hellenistic Central Asia; Religious life; sanctuary; temple; ritual structures. INTRODUCTION The large sanctuary that was located at Ai Khanum on aterrace overlooking the palace area was probably one of the main places of worship in the town and is today the best known of all (Fig. 1).1 It was excavated under the direction of Bernard between 1968 and 1973, by several archaeologists including Henri -Paul Francfort who published the related finds in 1984 (Francfort 1984).2 The sanctuary comprised agroup of buildings organized around arectangular courtyard and covering asurface of about 8.000 square metres between the main street of the town and the edge of the terrace. Some well -stratified coins show that it was founded during the reign of the Seleucid king AntiochusI. The Seleucid temple, of which only the first layers of bricks were preserved, was located on the west side of the courtyard (Fig. 2). The three other sides were enclosed by walls or buildings, especially to the east, on both sides of the entrance to the sanctuary (period V). The temple and most of the other buildings were razed and rebuilt around 240 BC under Diodotus Ior his son Diodotus II (period IV, Fig. 3). The new buildings were based on the Seleucid building plan, except in the northeast corner of the courtyard where alarge portico was standing now. This is when the so -called Temple à redans was constructed on almost the same plan as the earlier temple except for the presence of two annex rooms flanking the cella. 3 The buildings (but not the temple) were entirely rebuilt once again in the first decades of the 2nd century BC, around 180–170 (period III, Fig. 4), and yet 1 This article is an updated and enriched version of Martinez -Sève 2013a. 2 Paul Bernard entrusted me the task to publish this monument several years ago, and Ihope to finish very soon now. 3 These annex rooms may already have existed in the Seleucid temple, of which only the foundations are known. STUDIA HERCYNIA XXVII/1, 98–123
99LAURIANNE MARTINEZ ‑SÈVE again under Eucratides Iaround 160–150 (period II, Fig. 5). Contrary to what was previously written, the sanctuary was not desacralized after the collapse of the Graeco -Bactrian power and the attack that struck the town around 145 BC.Ai Khanum remained inhabited by asmall population of local origin, probably for several decades, and the sanctuary was still under the control of priests who undertook maintenance operations (period I). But the cult underwent some evolutions to better fit Bactrian conceptions (Martinez -Sève 2018). Only at the end of the history of the city most of the buildings were privatized and reoccupied, except for the temple and chapel. The city included other places of worship especially asecond temple located 100 metres away outside the ramparts, astepped podium of Iranian type at the top of the Acropolis, and the Heroon of Kineas built in the lower city to house the remains of aman, who had played arole in the first times of the city (Fig. 1). Fig. 1: General plan of Ai Khanum. Drawing by J.‑C. Liger and C. Rapin.
100 STUDIA HERCYNIA XXVII/1 Fig. 2: General plan of the sanctuary– period V. Drawing by A.‑B. Pimpaud. Fig. 3: General plan of the sanctuary– period IV. Drawing by A.‑B. Pimpaud.
101LAURIANNE MARTINEZ ‑SÈVE Fig. 4: General plan of the sanctuary– period III. Drawing by A.‑B. Pimpaud. Fig. 5: General plan of the sanctuary under Eucratides I– period II. Drawing by A.‑B. Pimpaud.
102 STUDIA HERCYNIA XXVII/1 PRELIMINARY CONSIDERATIONS Alarge amount of information was collected during the excavations, but the archaeologists’ concerns and their intellectual tools were not the same as ours. They focused especially on architecture and stratigraphy, the latter being noted very carefully, which was areal innovation and progress compared to the excavation methods that were still frequently deployed at that time. They endeavoured to reconstruct the chronology of the buildings and to understand their architectural transformations from the foundation of the sanctuary under Antiochus I until its abandonment, probably during the 1st century BC.They also attempted to identify the deities to which it was dedicated, especially the god of the temple. The latter was represented in Greek style, in the form of acolossal acrolith cult statue, of which only some fragments remain, including afoot wearing asandal decorated with awinged thunderbolt. Bernard assumed that the god was figured as Zeus and assimilated to Ahura Mazda (Bernard 1974, 298), the chief of the Iranian pantheon; Frantz Grenet preferred to recognize Mithra, avery important deity in Bactria (Grenet 1991) and, recently, Francfort proposed that he was Vaxšu, the Oxus River god, but conceived in afeminine form inherited from the great Bactrian goddess from the Bronze Age (Francfort 2012). These questions were usual when the excavations took place. Since that time, issues related to the archaeology of worship and the archaeology of sanctuaries have been considerably renewed.4 Careful analysis of the remains and the use of scientific techniques allow amore precise approach to rituals through the reconstruction of religious gestures. The use of ritual objects and the symbolic dimension they conveyed have also become better understood.5 Naturally, the excavators of Ai Khanum could not take this into account. Excavations were interrupted in 1978, ayear before the invasion of Afghanistan, and never resumed. We cannot return to the field to collect additional information because of safety reasons, all the more as the site has been completely destroyed by looting. Only the documentation gathered during the excavations and the archives of the mission, which are mainly housed at the École Normale Supérieure de Paris, can be used. These archives include daily digging diaries detailing the excavation operations with lists of finds, synthesis reports where the excavators give their first interpretations, and numerous plans and photographs. Alot of information has therefore been preserved and can be exploited in accordance with the renewed issues mentioned above. When the sanctuary was discovered, no other religious buildings from the Achaemenid and Hellenistic periods were known in Bactria and Sogdiana. The temple of Diľberdzhin, dated to the end of the Graeco -Bactrian period by Irina Kruglikova (but much later by Ciro Lo Muzio: Lo Muzio 1999; cf. also Shenkar 2011, 124–125), was explored between 1970 and 1977 and first presented to the scientific community in 1974, at atime when the excavations of the Ai Khanum sanctuary had just been completed (Kruglikova 1974; 1977; 1986). Thus, when Bernard described these works in his preliminary reports and gave his first analyses, he was not aware of the existence of this other temple (Bernard 1969, 327–355; Bernard 1970a, 317–347; Bernard 1971, 414–431; Bernard 1972, 625–629; Bernard 1974, 294–298).6 Since then, new discoveries have considerably improved our knowledge of local religious life. The Oxus temple was explored between 1976 and 1991 at Takht -iSangin, the place where the Vakhsh and 4 For instance, Renfrew 1985. 5 For example, by the contributions in Kyriakidis 2007 and Raja– Rüpke 2015. 6 He mentioned the temple of Dil’berdzhin in his later articles focusing on religious Bactrian architecture (Bernard 1976; 1990; see also Bernard 1987).
103LAURIANNE MARTINEZ ‑SÈVE Panj Rivers join their waters to form the current Amu Darya, and these excavations provided awealth of new data (Litvinsky– Pichikyan 2000; 2002; Litvinsky 2001; 2010). Other discoveries are even more recent. 7 Worship platforms dating from the pre- -Achaemenid and Achaemenid period were excavated at Koktepe (30 km northwest of Samarkand; Rapin 2007, 33–38; Rapin 2017, 419–426). Three temples from the Achaemenid period are also evidenced, which contradicts the long -held assumption that local cults were mainly celebrated in the open air. One is located at Sangir tepe in the Kashkadarya Valley (60 km south of Samarkand), probably on the territory of the former city of Kish -Nautaca, and was in operation during the 4th and early 3rd centuries BC (Rapin 2017, 438–443). The second one was uncovered at Kindyktepa (near Bandikhan, Surkhandarya Valley) and dated to the 4 th century BC (Boroffka 2009, 138–141; Sverchkov– Boroffka 2009, 226–230. See also Grenet 2008, 30–31 and Rapin 2017, 443–444). The reopening of an archaeological activity at Kyzyltepa in the Surkhandarya Valley (Mirshade Oasis, Denau region) has provided new information on the settlement and the discovery of athird temple from the Achaemenid period, very similar to the one at Kindyktepa (Sagdullaev 1990; Wuetal. 2017, 298; Sverchkov– Wu 2019). The Hellenistic sanctuary of Torbulok, 40 km northwest of the present city of Kulyab (Tajikistan), is another major discovery (Lindström 2014; 2017a; 2017b; 20218). Organized on several terraces, it included open -air spaces, small buildings, and perhaps atemple. It is interpreted as asanctuary of regional importance and rural nature. Finally, the archives of the Achaemenid governor of Bactria reveal that the Achaemenid administration provided wine and flour as offerings for acult of Bel (Naveh– Shaked 2012, 35–37, 179, 184–185, letter no. C1; Tavernier 2017). The god was probably areal Bel and not aform of Ahura Mazda. These data allow us to better contextualize the monuments of Ai Khanum and to grasp the Bactrian religious landscape with greater finesse. On the premise that Zoroastrianism was the main religion in Central Asia, archaeologists have sometimes tended to interpret the religious conceptions and practices of the Central Asian people on the basis of knowledge transmitted by the clergy of the Sasanian Empire, where Zoroastrism had become astate religion conceived in arigid manner. For example, they tend to interpret temples too systematically as fire temples and cults as cults of fire.9 We are now able to better assess the specificities of Bactrian cults and to think of them within alocal context. The sanctuaries of Ai Khanum, Takht -iSangin or Torbulok were not of an equivalent nature, however, and it seems difficult to compare them with each other. Their function and status were different. The Oxus Temple was obviously visited by people from avast regional area encompassing the territories south and north of Amu Darya, coming there to celebrate along -established and renowned cult, which probably explains the development of asettlement nearby. Ai Khanum played adifferent role. From the reign of Antiochus I, it took the form of aroyal city and became the centre of the Seleucid power in Eastern Bactria, before flourishing as one of the main residences of the Graeco -Bactrian kings (Martinez -Sève 2014; 2015). Its main cults had probably been chosen at the time of its foundation, by the kings themselves or by their administration and for specific reasons. Though these cults were probably in accordance with local populations’ aspirations, they were also official and royal cults, which had an impact on the nature of the rituals performed in the sanctuary. 7 See the various contributions in the present volume. 8 Iwarmly thank Gunvor Lindström for communicating this manuscript to me before it was published. 9 According to its excavators Boris Litvinsky and Igor Pichikyan, the temple in Takht -iSangin was afire temple (even if they assumed that acult of Waters was also celebrated there): Litvinsky– Pichikyan 2000, 303–366. Bernard 1994 and Lindström 2013a, 100–101 criticized this hypothesis.
104 STUDIA HERCYNIA XXVII/1 Ai Khanum is also often conceived as acity inhabited by Greek settlers. The observations made in the sanctuary are then analysed in terms of cultural interactions. One wonders how the Greek settlers behave, whether they retained their own religious beliefs and practices or whether they adopted local ones. While qualifying the city’sinhabitants as ‘Greeks’ is already very simplistic, interpreting their behaviour in terms of cultural or ethnic categories does not help to grasp these practices in their full complexity. Although linked to acommon and general set of traditions, each cult and each sanctuary had its own specificities and must be understood according to avery local environment.10 Worshippers were not concerned about the origin of the rituals they were celebrating. They did so because such was the rule and precisely in the form that suited the deities they prayed to. Any comment on religious life in Ai Khanum therefore implies first gathering information on these rituals and on the organization of the sanctuary, which Iintend to do in this article. THE SPACES OF THE SANCTUARY Throughout its existence, the sanctuary included atemple and various other buildings set around avast courtyard that formed the heart of the sacred space proper, the place where worship was celebrated. The excavators tended to consider that the boundaries of this courtyard corresponded to those of the sanctuary. But other spaces of adifferent nature also belonged to the sanctuary. With the temple on the west side, the sanctuary comprised throughout its existence asecond religious building, of smaller size and for this reason qualified as a‘chapel’ (Martinez- -Sève 2021). First located to the south of the courtyard, it was moved to the north side under Eucratides I(period II). The sanctuary also included large rectangular halls, which were gradually replaced by increasingly Greek -style porticoes, and smaller rooms too. These buildings contained few installations allowing them to be assigned specific functions, but we will see below that traces of culinary preparation were sometimes found there. The spaces extending beyond the courtyard to the south and to the north also belonged to the sanctuary but their boundaries were not always clear. These areas were apparently devoted to economic activities for the religious personnel’ssubsistence or for the provision of food offerings to the gods. They are mainly known for the most recent periods, as the upper levels of the sanctuary were excavated on amore extensive surface than the lower ones. They were gradually occupied by private dwellings that are sometimes difficult to distinguish from the buildings belonging to the sanctuary because of their similar appearance.11 Under Eucratides (period II, Fig. 5), this housing was present throughout the south side and clearly rejected beyond alarge portico enclosing this part of the courtyard, in an area that was no longer part of the sanctuary. On the north side, however, open -air lands extended between the street and the cella of the chapel in an area that was at aslightly lower level that the sanctuary courtyard (Fig. 6). The remains of an awning -shaped construction and two fireplaces were found in this area. The cella of the chapel also communicated with an annex (Martinez -Sève 2021), to the north of which extended asmall courtyard measuring 6.30 m in anorth -south direction (its western boundary was not recognized). It included asecond awning under which storage vessels were housed, and adouble -chamber fireplace (Fig. 7). Numerous vases and several millstones were still lying on the ground, which shows that this place was devoted to food stor10 On this topic: Kaiser 2006. 11 However, these houses were perhaps inhabited by members of the sanctuary’spersonnel.
105LAURIANNE MARTINEZ ‑SÈVE Fig. 6: Plan of the northern part of the sanctuary– period II. Drawing by A.‑B. Pimpaud.
106 STUDIA HERCYNIA XXVII/1 age and culinary preparation. Access to these areas appears to have been relatively restricted, although the excavation did not extend far enough westward to know how communications were established in that direction, where was standing alarge villa, three rooms of which were excavated. Apassage probably provided access from there to the courtyard behind the annex, which in turn opened onto the open spaces. From the courtyard of the sanctuary, an access was also possible passing through asmall door that connected theses spaces with asquare building located east of the chapel (building 2.13: Fig.6). In addition, they were also accessible from the street that bordered them to the north, through arather wide gate allowing the passage of animals or bulky objects. Further north, other constructions were separated from an unexcavated public building by avast rectangular open -space (Fig. 5). They seem to have been erected as part of the same architectural program as the sanctuary from the time of Eucratides and may have belonged to its premises, unless they were part of the neighbouring public building dated to the same period. These rooms were connected to each other through many doors, which facilitated movement and circulation suggesting that they may have been warehouses.12 During earlier periods, the properties of the sanctuary also extended outside the courtyard and included light structures (low walls for instance) built near fireplaces, or slightly larger constructions.13 Four large storage jars were also discovered behind the south chapel, where 12 They may have had the same function as the building located to the south -east of the city near the bank of the Kokcha River: Lecuyot 2013, 101. Like the latter, they were subsequently reoccupied and transformed into dwellings (period I): Martinez -Sève 2013b, 141–142; Martinez -Sève 2018, 390–396. 13 For example, building 5.18, built against the wall that enclosed the courtyard to the north during the Seleucid period (period V, Fig. 2). Fig. 7: Courtyard behind the annex of the north chapel. ©DAFA.
113LAURIANNE MARTINEZ ‑SÈVE sanctuary was therefore perhaps closely associated with water. It was Vaxšu according to Francfort, but one could also think of Anāhitā (Martinez -Sève 2018, 403–404).21 FOOD OFFERINGS Many of the sanctuary’sfacilities suggest that food was stored in some of its spaces and cereal- -based food offerings were cooked, for example on the large double -chamber fireplace located in the courtyard adjacent to the annex of the north chapel (period II). Several millstones also come from this area (see below). Material was also stored inside the temple, including about fifteen large storage jars, which encouraged the excavators to assume that it was desacralized and transformed into afarm (Bernard 1969, 352–354). Four of them were grouped in the northern half of the vestibule; three others were in the north annex; two or three were in the cella, near the door to the north annex. Seven others were installed in several stages in the south annex, and half buried in an earthen fill (Fig. 12).22 At least some of these jars were used to store cereals. Earlier installations were also uncovered. Two rooms on the east side of the Seleucid sanctuary (Period V) were each equipped with half -buried jars transformed into ovens to cook food. One, 75 cm in diameter, was located in the north -western part of the large room 5.08 (north to the entrance of the sanctuary, Fig. 2), but little is known about it. The second one, 65 cm in diameter, was standing in the northeast corner of the adjoining smaller room 5.09. Alayer of pebbles with traces of fire extended all around over adistance of about 75 cm. The surrounding ground was covered with very dense ashes, of black or grey colour (5 to 10 cm thick). They had been deposited in several stages and one of the strata, 1 cm thick, contained cereal seeds. Next to the jar was an oval pit (70 × 50 cm, 60 cm deep), whose walls were covered with awhitish coating, suggesting it was used to store food. Both jars were interpreted as tandoor -jars for baking bread as found in private houses (Lecuyot 2013, 58, 116). They may have been used to cook cereal food offerings or products perhaps consumed on the occasion of sacred banquets. Alarge number of millstones were also conserved in the sanctuary and used to produce flour.23 They were of several types (Francfort 1984, 85–88, tab.47–50, pl. XXXVIII), some of them more elaborate than others, including hopper -rubber millstones whose operation was precisely described by Francfort (1984, 86–87). Hopper -rubber millstones were invented by the Greeks and constitute amarker of their presence. They made it possible to grind large quantities of grain at atime, thus for non -domestic consumption. Hence, their presence in the sanctuary is significant. The repartition tables published by Francfort list 98 more or less fragmentary millstones, including 17 hopper -rubber ones which, with the exception of two specimens, were all stored inside the temple. About twenty other millstones were found in the north chapel or in its surroundings, notably in the courtyard adjacent to its annex.24 Another twenty or so were in the rooms located to the northeast of the sanctuary’scourtyard or nearby. 21 It should be noted, however, that the archives from Persepolis document the possibility that ariver sacrifice may be made for another god, in this case Humban (Henkelman 2008, 377–380). 22 Three jars were installed at the beginning of Period II, then three others at the beginning of Period Iand the last one later, during that period. 23 Most of these objects belonged to the furniture of the sanctuary and were not introduced by the last inhabitants who reoccupied its spaces. Francfort (1984, 88) also considered the possibility that most of the millstones were part of the sanctuary’sequipment and that grain had been stored in the temple before PeriodI. 24 Five millstones were in this courtyard, which was backfilled during Period I, at the same time as the annex was being destroyed. The latter may have been replaced by aroom located at the back of
114 STUDIA HERCYNIA XXVII/1 Fig. 12: Storage jars inside the south annex room of the temple– period II. ©DAFA.
115LAURIANNE MARTINEZ ‑SÈVE These rooms were used by the individuals who reoccupied the sanctuary, some time before its abandonment, to store the many objects they recovered from its richness, especially in the temple. The equipment also included four large mortars that were probably used to grind with apestle (Francfort 1984, 90, tab.51, pl. XXXIX:10–13). The presence of millstones inside the temple and the chapel is hardly surprising since they belonged to the sacred furniture and had symbolic value. Parallels can be invoked, but for other regions and other periods. The fixed lower part of agrinding stone, from which alarge fragment remained (1 × 0.45 m), was found, for example, on the floor of the central temple in Tepe Nuš -e Jān, asettlement located in Media about 60 km south of Ecbatana and dating from the 7th century BC (Stronach– Roaf 2007, 77–78, pl. 9b, 10b). This was obviously part of the sacred equipment and allowed large quantities of flour to be produced. In this respect, the Ai Khanum sanctuary may have shared similarities with the Syro -Mesopotamian temples whose deities received daily food offerings. The tablets from the Persepolis Fortifications Archives also keep records of regular allocations provided by the Achaemenid kings for gods of their empire. These gods systematically received cereals, most often in the form of flour, but sometimes also in the form of grains that were probably ground by the priests themselves. They also received beer or wine, and more occasionally animal victims (Henkelman 2008).25 These allocations were considered as offerings, but they were consumed by the priests and assured their sustenance too (Henkelman 2008, 288–289). Aletter from the archives of the Achaemenid administration in Bactra highlighted similar practices in Bactria. This document, very similar to those from Persepolis, records alist of food deliveries, including offerings for one or more deities (Tavernier 2017, 114–115).26 Some were clearly intended for the cult of Bel referred to above (Naveh– Shaked 2012, C1, lg. 37–45, 36–37, 184–185). Other were perhaps allocated for the wind god Vāta or for the river god (Vahvī) Daityā. This last possibility suggested by Jan Tavernier is interesting, since (Vahvī) Daityā was perhaps the river bordering Ai Khanum (see above; Tavernier 2017, 101, 105–106).27 However, the document is difficult to read and these last rations were more probably attributed to individuals rather than deities (Tavernier 2017, 108–110). In any case, the offerings and the quantities allocated were comparable to what was prescribed by the administration in Persepolis (Tavernier 2017, 106–108). These food offerings were deposited inside the temples on altars or offering tables placed in front of cult statues. The small altar located in the cella of the Seleucid temple was therefore perhaps used for this purpose. These ceremonies are well known in Mesopotamia thanks to religious texts that accurately describe the daily meals prepared for the gods, consisting of beer, wine, bread, cakes, and also meat. Fumigations were carried out before and after the meal to please the gods and purify the rooms where they stood.28 These thus texts offer interthe cella, where seven millstones were also found. The chapel contained four specimens, and the others were found in the vicinity of these buildings. 25 See Henkelman 2008, 417–426 for meat offerings, 477–478 for the repartition of different categories of offerings according to the various cults, 517–520 for barley offerings, especially for the god Adad. Also Henkelman 2017, 290 for the economic role of the temples. 26 Tavernier 2017, 114–115, including neo -Assyrian, neo -Babylonian, and Elamite parallels. These were not general administrative documents as produced by the Achaemenid administration, but specific lists of religious offerings. 27 Tavernier (2017, 101, 105–106) also considers the possibility that Zairivairi, whose Fravaši is worshipped in Yasht 13.101, was mentioned here as ahero. 28 These daily meal offerings are precisely described in Linssen 2004, 129–166. We know that 648 litres of flour (mainly barley and also emmer flour) were needed every day for the offerings due to the deities of Uruk at the beginning of the 2nd century BC.
116 STUDIA HERCYNIA XXVII/1 esting parallels that should be taken into account, even though we are unable to conclude that similar rituals were celebrated in Ai Khanum. Sacred banquets could also be held in some of the rooms bordering the courtyard, or even inside the temple or chapel. During the Seleucid and early Graeco -Bactrian periods, the sanctuary was provided with large rooms where numerous people could gather for collective ceremonies, including Room 5.08, which contained afireplace -jar for cooking food. More restricted ceremonies could be organised in the cella of the Graeco -Bactrian temple, where the lateral benches, each about 4 m long, could accommodate about four people in alying position and more in asitting position. The building was also equipped with an ivory bed or armchair, the fragments of which were lying in the south side room (Bernard 1970b; Francfort 1984, 9–10, tab.1:1–9, pl. I–II).29 Astool made of stone was also kept in the temple (Francfort 1984, 90–91), and perhaps used to climb and lie on the bed or benches. At the beginning of Period III, the southern chapel underwent areconfiguration that could also suggest that banquets were there celebrated. Its facade colonnade was replaced by asolid wall, with afront door located 1 m from the northwest corner, in an off -centre position that seems alittle surprising for acult building (Fig. 4). This arrangement was perhaps adopted to make way for beds in the vestibule as in the banquet halls from the Mediterranean world. VOTIVE OFFERINGS Another ritual consisted in dedicating objects as offerings, still preserved in large numbers (Francfort 1984; Francfort 2012, 113–116). Only most of the precious ones had disappeared. They were probably recovered before the city was abandoned (Martinez -Sève 2018, 396–398, 404–405), except afew pieces such as the famous silver plate known as the ‘Cybele plate’, which had been hidden in the temple under one of the jars in the south annex room (Bernard 1970a, 339–347; Francfort 1984, 93–104). Most of the objects published by Francfort are offerings, with the exception of certain categories of material such as roofing and carpentry components. Some were part of the furniture or equipment of the sanctuary (tableware, small stone bases, millstones…), but they too were charged with asacred dimension and had probably been introduced into the sanctuary as offerings. Moreover, the difference between offerings and the sanctuary’sfurniture was probably not really formalized. This material was found in large numbers inside the temple, in agreat disorder because the interior of the building was disturbed by looters, who dug several trenches long after its abandonment (Francfort 1984, 107–111). The chapel, on the other hand, housed very few objects, probably because it had been emptied and cleaned during period I. The offerings it contained were buried in athick layer of soil that was used to backfill the entire area located to the west of the building in order to extend the courtyard northward and level it.30 Many other were also transported to the north -eastern rooms (Francfort 1984, 112–113), when the sanctuary was re -occupied by individuals engaged in recovery activities. This material consisted of precious ivory or wooden pieces of furniture, schist tableware decorated with inlays, decorated metal plates, schist pendants, jewellery, weapons, bone and terracotta figurines, four theatre masks made of clay, abone flute, and probably agold olive -leaf crown. Statues and statuettes made of clay, stone or bronze, including those of aHeracles and anaked athlete, 29 Unfortunately, it is not possible to decide whether it was aseat or abed. 30 These finds are published in Francfort 1984, without being assigned to the chapel: see the objects found in the northwest villa (rooms nos. 30 to 32 in Francfort 1984, 116), and in the area excavated between this villa and the chapel.
117LAURIANNE MARTINEZ ‑SÈVE were also standing in the temple and in the chapel. Others may have been exposed outside these buildings, for example on atwo -level mud brick podium applied against the northeast angle of the Graeco -Bactrian temple (Fig. 3). It remained in use throughout Period IV before being hidden by the rise of the ground level, and measured about 4.50 × 1.50 meter. Asimilar podium was built against the southwest corner of the north chapel (period II, Fig. 6).31 Another exposure area may also have been created during Period III on asmall terrace that was raised by about fifty centimetres along the south side of the courtyard, closed by along wall at that time (Fig. 4). This was replaced by alarge portico during period II, which in turn may have been used as an exhibition space for offerings. The deposit of offerings is one of the most widely known rituals and is very common in the Greek world. This was also the case at Takht -iSangin, where offerings were regularly removed from their original location to be gathered inside the temple and stored in bags, furniture or directly on the ground (Lindström 2013a, 103–104, 112–114; Lindström 2013b, 305; Lindström– Pilz 2013; Lindström 2016, 291–296. See also Grenet 2023; for Dil’berdzhin: Kruglikova 1986, 23–30). If we accept that the Oxus treasure was mainly composed of objects from the Achaemenid period coming from the surroundings of Takht -iSangin, these precious objects were probably also offerings that had been consecrated in aneighbouring sanctuary before the Hellenistic period. PURIFICATION RITUALS The large numbers of small limestone bases used to support incense -burners that were found during the excavations highlights the widespread practice of burning plants or aromatic in the sanctuary. Francfort listed 75 of them that he divided into eight categories according to formal criteria (Francfort 1984, 81–84, tab.41–46b, pl. XXXVI–XXXVII). The oldest known was found in athick embankment that supported the floor of the courtyard, rebuilt at the beginning of the 2 nd century BC (period III), and was therefore in use at the beginning of the Graeco -Bactrian period (Francfort 1984, tab.44:5). As we have seen, the small brick bases set around the Graeco- -Bactrian temple probably had the same function. Some of the stone bases were worn out, as if they had remained outside for along time and used in the open air too. Others were stored inside the temple, where fourteen were found at the time of excavation. Thirty -three had been moved into the rooms located in the north-eastern part of the courtyard when the sanctuary was reoccupied and probably some of them come from the temple too. On the other hand, there were none in the chapel, which is surprising. Francfort has shown that these objects were very common in Bactria both during Hellenistic and later periods, and Galina Pugačenkova also recorded them (Pugačenkova 1996). Some were found in Takht -iSangin (Litvinsky– Pichikyan 2000, 60, 75; Litvinskij– Pičikjan 2002, 21), but in smaller number than in Ai Khanum, and in Torbulok (Lindström 2017b, 170; Lindström 2021). These objects are also known outside Bactria, in Iran since the Achaemenid period, throughout the Hellenistic world, as well as in Gandhara (Boyce– Grenet 1991, 168; Lindström 2017b, 170; Lindström 2021). The perfume burners they supported were used for fumigations that could be assimilated to plant offerings. But they also served for purification rituals, as was done in Babylonian temples (Linssen 2004, 145–147). In Mesopotamia, juniper was the main ingredient for these rituals, afrequently found plant in Central Asia too. But incense or other products were probably also fumigated. In addition, these small bases may also have been dedicated in the sanctuary to commemorate the ritual act of burning fragrant plants, without necessarily having been used themselves. 31 Unknown dimensions.
118 STUDIA HERCYNIA XXVII/1 Other purification rituals were celebrated on specific occasions. The construction of the small altar in the cella of the Seleucid temple was accompanied by aset of gestures that seem to be of ritual nature. After the first brick was laid on the ground, asmall fire of twigs and straw was burned on the top of its bedding surface. The second brick was laid before the flames were extinguished and partially burned in turn. The same procedure was followed for the other bricks, and again when the altar was enlarged with new ones. Burnt remains were also observed between the bricks of one of the small bases located around the Graeco -Bactrian temple (the one at the eastern end of the north face), which could suggest that similar actions were undertaken, but the observation is uncertain. Another purification ritual may have occurred after the destruction of the cult statue of the chapel and at atime when this building seems to have become the main place of worship in the sanctuary (period I). The statue was previously resting on alarge mud -brick base built against the back wall of the cella. The bricks from the central part of the base were removed to make way for afireplace. This was not adomestic fireplace since the chapel does not seem to have been reoccupied as adwelling (Martinez -Sève 2018, 386–387, 396–399; Martinez -Sève 2021). On the other hand, this finds parallels in Sassanid texts from the 6th century AD that mention aritual consisting in the lighting of an Ādurōg or ‘Little Fire’ after the destruction and removal of astatue. This fire was supposed to purify the place where the statue had stood before, by eliminating the demon supposed to live inside (Boyce 1975a, 463; Boyce 1975b, 106–108; Boyce 1979, 110; Grenet 2008, 32). The same ritual may have been celebrated in the temple since afire was perhaps lit on the top of two adjoining bases placed in the vestibule, north of the door leading into the cella.32 Both supported clay statues, which were destroyed and whose fragments were lying nearby on the floor. They were not cult statues, but they had probably been erected there by members of the Graeco -Bactrian elites, which probably explains their destruction. Nothing can be concluded for the base of the cult statue since it was largely destroyed when the building was looted. If the removal of the statues did indeed give rise to such rituals, one may conclude that the conceptions of the populations frequenting the sanctuary after the collapse of Greek power in 145 BC underwent some changes. But it would probably be too simplistic to interpret them as manifestations of aniconism (Martinez -Sève 2018, 399–402). FOUNDATION RITUALS Some observations suggest that foundation rituals were also celebrated in the sanctuary. Asmall pit containing ashes mixed with burned residues including bones, wood, and straw was dug inside the floor of the cella of the Seleucid temple, north -east to the small altar. 33 These were probably the remains of asmall sacrifice made when the altar was erected,34 during aceremony of foundation complementing the purification ritual mentioned above. An alternative interpretation is to link the burial of this deposit with the construction of the temple itself. Another foundation deposit, consisting of anecklace of blue, white, and green glass pearls combined with turquoise pearls, was hidden under the threshold of adoor, temporarily open in the eastern enclosure wall of the sanctuary during Period I, in the south -eastern Room 32 Both bases were each provided at the top with alarge cavity, very irregular in shape, filled with soil and also with ashes, at least in their upper part. As the temple was burned down, these ashes may have formed on the occasion. But the observations made in the chapel suggests that both cavities were also deliberately made to insert afire. 33 This pit was located 3.60 m from the north wall and 2.10 m from the east one. 34 For Greek parallels: Patera 2012, 144–146.
119LAURIANNE MARTINEZ ‑SÈVE 2.06 (Fig. 6). It seems that the sanctuary official entrance was moved to this place, while the usual one was being rebuilt and condemned during the work. The new door was very carefully opened in the sanctuary wall, which was more than 2 m thick. At the threshold location, the opening went deeper on the street side than on the sanctuary side forming acavity, where the necklace was placed. The cavity was then filled with pebbles, and the whole was covered and hidden by alayer of adobe regularizing the threshold surface. There is no doubt that the necklace was intentionally deposited. Modifying the accesses to the sanctuary and thus the location of apoint of contact between the sacred and profane spaces may have been considered as apossible factor of disruption, from which it was necessary to protect oneself. Foundation deposits are awell -known component of foundation rituals for sacred buildings in Mesopotamia (Ellis 1968), but they are also documented in the Greek world. For example, adeposit of 80 bronze and silver coins, mainly of Rhodian origin, was buried in the first half of the 2nd century BC against the threshold of the cella of the Leto Temple in Xanthos (Hansen– Le Roy 1976, 321–325).35 The treasure of about ahundred coins and jewellery, kept in asmall wooden box, found in the foundations of alarge building in the Donjon area of Susa is probably of the same nature. The structures that Roland de Mecquenem excavated at this place cannot be accurately reconstructed. But the sanctuary of Nanaia was probably located there and this deposit may have been buried at the time of its foundation at the end of the 4th century or the beginning of the 3rd century BC (Martinez -Sève 1996, 175). Some coins may also have been intentionally deposited during the construction of several buildings in Ai Khanum. The most obvious case is given by three bronze coins minted by aking Diodotus and found in significant stratigraphic positions. One was inside the thick layer of clay forming the floor of the Graeco -Bactrian temple and the other two, which belonged to the same monetary series, were found in the large portico with two orthogonal wings that occupied the northeast corner of the courtyard (Fig. 3). One was inside the floor of the east wing and the other one inside the fill supporting the floor of the north wing (Bernard 1985, no. 5, 91, 97).36 Both buildings were built under Diodotus Ior his son Diodotus II, and probably on their own initiative. At that time, their coins were not the only circulating ones in the city. Alarge number of coins minted by Antiochus Icontinued to be used and it seems surprising that precisely three Diodotus coins were then lost, all three when afloor was being laid, two of which from the same series and, besides, each in awing of the same building. The assumption of intentional deposition may be proposed, but is difficult to prove, as for abronze coin minted by Antiochus I, buried 1 m deep in apebble fill at the back of the south chapel (period V) (Bernard 1985, no. 68). This fill was very thick (1.80 m) and used to reconfigure the natural relief of the terrace on which the sanctuary was built. Thanks to this coin we can date its foundation under AntiochusI. Foundation deposits may also be evidenced in the fortress of Uzundara, located near the Iron Gates in the Hissar Mountains. Bronze coins of Euthydemus I were apparently present within its walls and the hypothesis that they were deliberately put there was proposed (Rtveladzeetal. 2014, 157). While parallels can be found for the foundation rituals performed at Ai Khanum, it is worth noting that they are different from those that are known in other sanctuaries throughout the region, notably in Koktepe and Sangir -tepe. In Koktepe, small pits filled with pure sand and 35 Similar deposits from the Hellenistic period are known in Priene (inside the base of the cult statue of Athena), in Sardis (inside the base of the Artemis cult statue), and in Kastabos (under the cella in the temple for Hemithea)(Hunt 2006). 36 Afourth bronze of Diodotus (no. 101) was found during the excavation of the portico, but its stratigraphic position is less significant (Martinez -Sève 2017, 285–288).
120 STUDIA HERCYNIA XXVII/1 pebbles were dug in the remains of an older sacred platform, before the construction of the new Achaemenid one. One of these pits contained afoundation deposit composed of limestone objects, including alarge disc with ahandle dated to the Bronze Age. Other larger pits may have contained simulacra of human bodies (Rapin 2017, 424–425). Circular pits filled with pure ash, sand and pebbles were also found under the temple in Sangir -tepe, suggesting asimilar ritual (Rapin 2017, 440). The pits found under the floor of the annex room in the temple of Kindyktepa, which are interpreted as the remains of apurification ritual similar to the Zoroastrian barəšnūm ‑gāh (Sverchkov– Boroffka 2009, 228), was perhaps of the same nature. Despite the early date of the excavations, we can thus reconstitute some elements of the cults celebrated within the sanctuary of Ai Khanum. Although comparisons can be made with the evidence coming from other Bactrian and Sogdian sanctuaries, some rituals seem more specific to Ai Khanum, at least in the present state of knowledge, while others can be related to practices that were widely known in the Near East. All this contributed to create alocal religious life, with some peculiarities that were probably the result of the exceptional royal status of the city. SOURCES Lecoq, P.2016: Le livres de l’Avesta. Textes sacrés des Zoroastriens. Paris. Strabon, Géographie, Tome XII, Livre XV, L’Inde, l’Ariane et la Perse, Texte établi et traduit par P.-O. Leroy, «CUF», Paris, Les Belles Lettres, 2016. BIBLIOGRAPHY Bernard, P.1969: Quatrième campagne de fouille à Aï Khanoum (Bactriane). Comptes rendus des séances de l‘Académie des Inscriptions et Belles ‑Lettres 113, 313–355. Bernard, P.1970a: Campagne de fouilles 1969 à Aï Khanoum en Afghanistan. Comptes rendus des séances de l‘Académie des Inscriptions et Belles ‑Lettres 114, 301–349. Bernard, P.1970b: Sièges et lits en ivoire d’époque hellénistique en Asie centrale. Syria 47, 327–343. Bernard, P.1971: La campagne de fouille de 1970 à Aï Khanoum (Afghanistan). Comptes rendus des séances de l‘Académie des Inscriptions et Belles ‑Lettres 115, 385–452. Bernard, P.1972: Campagne de fouilles à Aï Khanoum (Afghanistan). Comptes rendus des séances de l‘Académie des Inscriptions et Belles ‑Lettres 116, 605–632. Bernard, P.1974: Fouilles de Aï Khanoum (Afghanistan), campagnes de 1972 et 1973. Comptes rendus des séances de l‘Académie des Inscriptions et Belles ‑Lettres 118, 280–308. Bernard, P.1976: Les traditions orientales dans l’architecture gréco -bactrienne. Journal Asiatique 264, 245–275. Bernard, P.1985: Fouilles d’Aï KhanoumIV. Les monnaies hors trésor. Questions d’histoire gréco ‑bactrienne. Mémoires de la Délégation archéologique française en Afghanistan 28. Paris. Bernard, P.1987: Publication des fouilles soviétiques du temple de Dilberdjin en Bactriane (IIe s. av. n.è.–IVe s. de n.è.). Abstracta Iranica 10, 60–62. Bernard, P.1990: L’architecture religieuse de l’Asie Centrale à l’époque hellénistique. In: Akten des XIII. In‑ ternationalen Kongresses für Klassische Archäologie. Berlin 24.–30. Juli 1988. Mainz, 51–59. Bernard, P.1994: Le temple du dieu Oxus à Takht -iSangin en Bactriane: temple du feu ou pas? Studia Iranica 23, 81–121.
121LAURIANNE MARTINEZ ‑SÈVE Boroffka, N. 2009: Siedlungsgeschichte in Nordbaktrien– Bandichan zwischen Spätbronzezeit und Frühmittelalter. In: S. Hansen– A. Wieczorek– M. Tellenbach (eds.): Alexander der Große und die Öffnung der Welt. Asiens Kulturen im Wandel. Mannheim, 135–144. Boyce, M. 1975a: On the Zoroastrian Temple Cult of Fire. Journal of the American Oriental Society 95, 454–465. Boyce, M. 1978b: Iconoclasm among the Zoroastrians. In: J. Neusner (ed.): Christianity, Judaism and Other Greco ‑Roman Cults. Studies for Morton Smith at Sixty 4. Leiden, 93–111. Boyce, M. 1979: Zoroastrians. Their Religious Beliefs and Practices. London. Boyce, M.– Grenet, F. 1991: AHistory of Zoroastrianism 3. Zoroastrianism under Macedonian and Roman Rule. Handbuch der Orientalistik. 1. Abteilung. Der Nahe und der Mittlere Osten, 8. Religion, 1/2/2/3 Leiden– Köln. Ellis, R.W. 1968: Foundation Deposits in Ancient Mesopotamia. New Haven– London. Francfort, H.-P.1984: Fouilles d’Aï KhanoumIII. Le sanctuaire du temple à niches indentées 2. Les trouvailles. Mémoires de la Délégation archéologique française en Afghanistan 27. Paris. Francfort, H.-P.2012: Ai Khanoum ‘Temple with Indented Niches’ and Takht -iSangin ‘Oxus Temple’ in Historical Cultural Perspective: Outline of aHypothesis About the Cults. Parthica 14, 109–136. Grenet, F. 1991: Mithra au temple principal d’Aï Khanoum? In: P.Bernard– F. Grenet (eds.): Histoire et cultes de l’Asie centrale préislamique: sources écrites et documents archéologiques. Actes du colloque international du CNRS, Paris 22–28 novembre 1988. Paris, 147–151. Grenet, F. 2002: Zoroastre au Badakhshān. Studia Iranica 31, 193–214. Grenet, F. 2008: Mary Boyce’sLegacy for the Archaeologists. Bulletin of the Asia Institute 22, 29–46. Grenet, F. 2015: Zarathustra’sTime and Homeland. Geographical Perspective. In: M. Stausberg– Y.S.-D. Vevaina (eds.): The Wiley Blackwell Companion to Zoroastrianism. Chichester, 21–29. Grenet, F.– Rapin, C. 1998: Alexander, Aï Khanum, Termez. Remarks on the Spring Campaign of 328. In: O. Bopearachchi– C.A. Bromberg– F. Grenet (eds.): Alexander’sLegacy in the East. Studies in Honor of P.Bernard. Bulletin of the Asia Institute 12, 1998 [2001], 79–89. Hansen, E.– Le Roy, C. 1976: Au Létôon de Xanthos. Les deux temples de Léto. Revue Archéologique 1976/2, 317–336. Henkelman, W.F.M. 2008: The Other Gods Who Are. Studies in Elamite ‑Iranian Acculturation Based on the Perse‑ polis Fortification Texts. Achaemenid History 14. Leiden. Henkelman, W.F.M. 2017: Humban & Auramazdā. Royal Gods in aPersian Landscape. In: Henkelman– Redard eds. 2017, 273–346. Henkelman, W.F.M.– Redard, C. eds. 2017: Persian Religion in the Achaemenid Period / La religion perse à l’époque achéménide. Wiesbaden. Hunt, G.R. 2006: Foundation rituals and the culture of building in Ancient Greece. Unpublished dissertation. University of North Carolina at Chapel Hill. Available on-line at https://core.ac.uk/download/pdf/210600585. pdf Kaiser, T. 2006: In Search of Oriental Cults. Methodological Problems concerning ‘the Particular’ and ‘the General’ in near Eastern Religion in the Hellenistic and Roman Periods. Historia 55, 26–47. Kruglikova 1974 = Кругликова, И.Т.: Дильберджин (раскопки 1970–1972 гг.). Москва. Kruglikova, I. 1977: Les fouilles de la mission archéologique soviéto -afghane sur le site gréco -kushan de Dilberdjin en Bactriane (Afghanistan). Comptes rendus des séances de l‘Académie des Inscriptions et Belles‑ ‑Lettres 121, 407–427. Kruglikova 1986 = Кругликова, И.Т.: Дильберджин. Храм Диоскуров. Материалы Советско ‑Афганской археологической экспедиции. Москва. Kyriakidis, E. ed. 2007: The Archaeology of Ritual. Los Angeles. Lecuyot, G. 2013: Fouilles d’Aï KhanoumIX. L’habitat. Paris.
122 STUDIA HERCYNIA XXVII/1 Lindström, G. 2013a: Votivdeponierungen im Oxos -Tempel (Baktrien)– Tradierung Griechischer Kultpraxis? In: A. Schäfer– M. Witteyer (eds.): Rituelle Deponierungen in Heiligtümern der Hellenistisch ‑Römischen Welt. Mainz, 97–114. Lindström, G. 2013b: Baktrien– Votive und Votivpraxis in den hellenistischen und kuschanzeitlichen Heiligtümern (3. Jh. v. Chr.–3. Jh. n. Chr.). In: I. Gerlach– D. Raue (eds.): Sanktuar und Ritual, Heilige Plätze im archäologischen Befund. Rahden, 299–306. Lindström, G. 2014: Torbulok, Tadschikistan. Ein neuentdecktes Heiligtum im hellenistischen Osten. Die Arbeiten der Jahre 2012 und 2013. e ‑Forschungsberichte des DAI 2014/1, 120–124. 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, 281–310. Lindström, G. 2017a: Torbulok, Tadschikistan. Die Arbeiten der Jahre 2014 bis 2016. e ‑Forschungsberichte des DAI 2017/1, 174–183. Lindström, G. 2017b: Auf der Suche nach dem Ritual. Ausgrabungen in einem hellenistischen Heiligtum in Torbulok, Tadschikistan. Das Altertum 62, 161–180. Lindström, G. 2021: The Hellenistic Sanctuary at Torbulok, Tajikistan. In: J. Lhuillier (ed.): Archaeology of Central Asia during the first millennium BC, from the beginning of the Iron Age to the Hellenistic period. Proceedings of the workshop held at 10th ICAANE held in Vienna. Wien, 89–104. Lindström, G.– Pilz, O. 2013: Votivspektren von Heiligtümern. In: I. Gerlach– D. Raue (eds.): Sanktuar und Ritual. Heilige Plätze im archäologischen Befund. Rahden, 267–274. Linssen, M.J.H. 2004: The Cults of Uruk and Babylon. The temple ritual texts as evidence for Hellenistic cult prac‑ tices. Leiden– Boston. Litvinskiy 2001 = Литвинский, Б.А.: Храм Окса вБактрии (Южный Таджикистан) II. Бактрийское вооружение вдревневосточном игреческом контексте. Москва. Litvinskiy 2010 = Литвинский, Б.А.: Храм Окса вБактрии (Южный Таджикистан) III. Искусство, художественное ремесло, музикальные инструменты. Москва. Litvinskiy– Pichikyan 2000 = Литвинский, Б.А.– Пичикян, И.Р.: Эллинистический храм Окса вБактрии (Южный Таджикистан) I. Раскопки, архитектура, религиозная жизнь. Москва. Litvinskij,B.A.– Pičikjan, I.R.2002: Taxt ‑iSangīn. Der Oxos ‑Tempel. Grabungsbefund, Stratigraphie und Architektur. Archäologie in Iran und Turan 4. Mainz. Lo Muzio, C. 1999: The Dioscuri at Dilberjin (Northern Afghanistan): Reviewing their chronology and significance. Studia Iranica 28, 41–71. Martinez -Sève, L. 1996: Une statue romaine trouvée à Suse et la chronologie du Donjon. In: H. Gasche– B. Hrouda (eds.): Collectanea Orientalia, Histoire, arts de l‘espace et industrie de la terre. Études offertes en hom‑ mage à Agnès Spycket. Neuchâtel– Paris, 171–180. Martinez -Sève, L. 2013a: Les cultes dans le sanctuaire principal d’Aï Khanoum (Afghanistan). Remarques sur la vie religieuse des Grecs de Bactriane. In: M.-C. Ferriès– M.P.Castiglioni– F. Létoublon (eds.): Forgerons, élites et voyageurs. Mélanges à la mémoire d’Isabelle Ratinaud. Grenoble, 483–504. Martinez -Sève, L. 2013b: Les maisons du quartier du temple principal d’Aï Khanoum. In: Lecuyot 2013, 137–143. Martinez -Sève, L. 2014: The Spatial Organization of Ai Khanoum, aGreek City in Afghanistan. American Journal of Archaeology 118, 267–283. Martinez -Sève, L. 2015: Ai Khanoum and Greek domination in Central Asia. Electrum 22, 17–46. Martinez -Sève, L. 2017: Les opérations d’Antiochos III en Bactriane -Sogdiane: l’apport de la documentation archéologique. In: C. Feyel– L. Graslin -Thomé (eds.): Antiochos III et l’Orient. Nancy, 271–302. Martinez -Sève, L. 2018: Ai Khanoum after 145 BC.The Post -palatial Occupation. In: A. Ivantchik– M. Minardi (eds.): Ancient Chorasmia, Central Asia and the Steppes. Cultural relations and exchanges from the Achaemenids to the Arabs. Ancient Civilizations from Scythia to Siberia 24, 354–419.