1The Araxes basin is rich in rock-salt deposits (fig. 1), a fact that has certainly been an economic asset for the South Caucasus at different times in history. However, all these deposits have not been exploited in the same periods and in the same way. The contrasts are sometimes striking: from the tunnel-ridden salt plateau of Tuzluca (Tuzdağı) near Iğdır in Turkey (fig. 2), to the dune-like slopes of Duzdağı in Nakhchivan (fig. 3), there seems to be a world of difference. Yet, these two salt mines are only 220 km apart.

Fig. 1 – Main rock-salt deposits in the Caucasus, western Iran and eastern Anatolia (C. Marro, map background: M. Sauvage, CNRS).

Fig. 2 – The mine of Tuzluca (Tuzdağı, Turkey) seen from the south-east (C. Marro).

Fig. 3 – The mine of Duzdağı seen from the south, 08/2013 (MBA, C. Marro).

2The differences also concern the remains left by mining activities over the salt deposits: very few artefacts could be collected from Tuzluca during our brief survey in 2003 together with the University of Van (Marro, Özfırat 2005; Marro, Michel 2013, p. 361), while hundreds of macro-lithic tools and thousands of ceramic potsherds were registered during the survey we carried out at Duzdağı together with the Azerbaijani National Academy of Sciences between 2008 and 2011 (Marro, Bakhshaliyev, Sanz 2010, maps 10.2, 11.1; Hamon 2016, tab. 2). The salt mine of Duzdağı is certainly conspicuous for its wealth in archaeological artefacts, a fact that probably has chronological but also technological explanations.

3Interestingly enough, most of the pottery retrieved from Duzdağı, whether it be during the survey or the excavations, dates back to the Kura-Araxes period. This certainly testifies to the importance of salt extraction in the Kura-Araxes economy; however, it may also reflect the specificities of salt extraction during the 4th‑3rd millennia BCE. But in what way? Are these containers directly linked to salt processing or were they used for other purposes? Ceramic vessels clearly played a major role in salt mining during the Kura-Araxes period at Duzdağı, but very few clues are available as to their actual function.

4This paper will focus on the Late Prehistoric ceramic assemblages from Duzdağı, with a special emphasis on the Kura-Araxes corpus. Its main purpose will be to assess the significance of Kura-Araxes pottery in salt-mining activities through the comparative analysis of the ceramic containers used on the salt dome before the Sassanid period. In a second step, we will see how far the function of ceramics retrieved from different mines (salt or metalliferous) reflects the constraints, work organization and technological system of Kura-Araxian mining activities.

The Kura-Araxes repertoire from Duzdağı

The general corpus: main features of the survey assemblage

5Kura-Araxes ceramics have been collected from all the transects (here called “windows”; Gonon et al., this volume, fig. 3) surveyed between 2008 and 2011, except for Window 3 (W3), which is not located on the salt dome stricto sensu, but some 3 km to the east, near the spring of Çehri Xor. Our purpose in exploring W3 was to look for miners’ camp sites, which we did not find. Instead, we found a collection of potsherds around the spring that were dated between the Early Iron Age and the medieval period. The absence of early pottery at Çehri Xor has two possible explanations: either the spring did not exist before the 2nd millennium BC, or this area had not been invested by Chalcolithic or Bronze Age communities.

6The largest concentration of Kura-Araxes pottery was found in Window 2 and Window 6, and to a lesser extent, in Window 1 and Window 5. The assemblage collected at Duzdağı in these areas displays interesting specificities when compared with the Kura-Araxes ceramics usually retrieved from settlements or funerary monuments: the corpus from Duzdağı only contains jars and pots, with an overwhelming predominance of jars (95%). Bowls, lids, pot stands and andirons are absent. Surfaces (exterior and interior) are predominantly buff, beige or cream-coloured; the vessels are sometimes slipped, but rarely burnished. Only 5% of the pottery matches with the black-burnished brand that usually typifies the Kura-Araxes pottery in the Caucasus, western Iran and eastern Anatolia. The black-burnished surface treatment at Duzdağı is only attested on pots, never on jars. These jars have fairly standardized measurements, with a narrow neck usually between 12 and 16 cm in diameter. The so-called Nakhchivan lugs are very common, both on jars and pots.

The corpus from M1 and M11

7Several excavation trenches were opened in Window 2, with the hope of finding salt-extraction facilities in connection with the concentration of Kura-Araxes potsherds collected over the south-western slopes of the dome. As a result, two mining cells were brought to light in M1, in front of which a clayish surface was probably used as a pithead for tool-making or tool-repairing (Gonon et al., this volume; Hamon et al., this volume). Another pithead, which yielded one complete Kura-Araxes jar (fig. 4‑5), stone tools and hearth remains, was found in M11 (first labelled S1 in 2011 when the sounding was opened) immediately to the east of W2 (Gonon et al., this volume, fig. 11), but the corresponding mining cells had not been reached by the end of the excavations in 2016.

Fig. 4 – Kura-Araxes jar from Duzdağı, M11-Sd. 2002 (MBA, R. Fleury).

Fig. 5 – Kura-Araxes jar from Duzdağı, M11-Sd. 2002 (MBA, K. Alhamid).

8The pottery corpus of M1 and M11, however, compares well with the assemblages collected in W1, W2, W5 and W6: the bulk of the pottery is composed of Kura-Araxes ceramics (70%), with a minority of Iron Age potsherds (30%). The Kura-Araxes assemblage is mainly composed of jars and pots (fig. 6, 7.1-2, 8), as is the case over the rest of the dome; the latter tend to be black-burnished (fig. 9), while the jars are buff, beige or cream-coloured. All were crafted with coils. The complete vessel found in M11 (S1) suggests that the jars were probably three-lugged, a system that facilitates transport with a rope, as shown by contemporary examples. In the case of the jar from M11, however, the fragments of a shell glued with bitumen were found inside one of the three Nakhchivan lugs: the utility of this shell for transport purposes has so far not be elucidated (fig. 10), while its location inside the lug seems to preclude its use as an element of decoration.

Fig. 6 – Kura-Araxes jars from Duzdağı (MBA, K. Alhamid).

Fig. 7 – Kura-Araxes jars from Duzdağı (MBA, K. Alhamid).

Fig. 8 – Kura-Araxes pots from Duzdağı (MBA, K. Alhamid).

Fig. 9 – Kura-Araxes pot from Duzdağı, M1‑011 (MBA, C. Marro).

Fig. 10 – Kura-Araxes jar from Duzdağı, M11‑Sd. 2002, shell glued with bitumen (MBA, C. Marro).

9The dimensions of the jars from M1 and M11 are fairly standardized: the neck-rim measures between 13 and 16 cm in diameter, while the bases, always flat, usually measure around 8‑10 cm (fig. 5‑7). The volume of the M11 jar is about 18 litres, for a total weight of 7‑8 kg (if we include the missing sherds). Not much may be said about the use of this jar from its stratigraphic context, since it was found smashed over the ground, possibly in situ, in the pithead area. However, from its shape (narrow neck and three lugs), it is probable that this jar was used for the transport and/or storage of liquids.

10An almost complete black-burnished pot (fig. 8.2, 9) was found in a filling layer (M1-011) just above one of the cells (F3) that was first created by erosion and then hollowed out by mining in SL2 (Gonon et al., this volume, fig. 16). No information can be obtained from its stratigraphic position, since it was not found in situ, but its shape, size and surface treatment suggest that this pot was used for purposes only indirectly linked to salt mining, possibly for serving food to the miners: it contains from 1.1 litres when filled up to the neck, to a maximum of 1.5 litres when filled up to the rim.

11Like the jars, the pots attested at Duzdağı thus seem related to logistics (consumption of food and liquids) rather than to the actual mining or processing of salt. Although ceramic containers are usually used in the production of salt for boiling the brine tapped from salt springs or other sources of liquid, as in the case of briquetage, terra-cotta pots can also be useful in the chaîne opératoire involved in the exploitation of rock-salt deposits: at Bronze Age Baile Figa, for instance, a major salt outcrop located in Romania, brine evaporation through the use of pierced wooden troughs was preferred to actual mining for salt.1 This technique may possibly have required the use of containers, and hence possibly terra-cotta containers, to collect the salt once the brine had evaporated. But this is not the technique used at Duzdağı, where most containers have a narrow opening that could not be of much use for collecting evaporation salt. These containers do not seem to fit anywhere in the technical sequence of the chaîne opératoire set up for salt production at Duzdağı. For this reason it might be useful to turn to earlier and later ceramic assemblages to see how far they compare with the Kura-Araxes repertoire.

Comparisons of the Kura-Araxes ceramic assemblage with earlier and later repertoires

Late Chalcolithic period

12Very few diagnostic potsherds dating to the Late Chalcolithic were found during either the survey or the excavations: most pieces are body sherds that offer no clue whatsoever to the vessels’ shapes. These potsherds are nonetheless interesting because they mostly belong to the Ovçular Tepesi culture, displaying Chaff-Faced, sometimes even comb-scraped surfaces (Gülçur, Marro 2012). Two body sherds recall the patchy smoke-blackened pottery that is also attested at Ovçular during the Late Chalcolithic (fig. 11.1‑4); at Ovçular, this smoke-blackened effect sometimes appears as a clear case of decoration. An interesting potsherd decorated with an annular band with oblique impressions (fig. 11.2) could be a little earlier (4600‑4500 BC?) than the Ovçular pottery, since it recalls the coarse craftsmanship typical of the Chalcolithic pottery of Uçan Ağıl (Gailhard et al., this volume). It must be stressed here that these sherds are the earliest ceramic examples so far collected from Duzdağı; not a single potsherd dating to earlier periods could be found over the slopes of the dome, in spite of the presence of a large Neolithic settlement (Kültepe I), 7 km to the north-east of the salt deposits.

Fig. 11 – Late Chalcolithic pottery from Duzdağı (MBA, K. Alhamid).

13As far as rim potsherds are concerned, the available examples (fig. 11.5‑6) all belong to large bowls, which suggest a use related to food consumption rather than salt processing or salt extraction. At all events, the use of ceramic vessels at Duzdağı during the Late Chalcolithic seems to have been markedly different from that of the Kura-Araxes period.

Late Bronze/Early Iron Age

14The Late Bronze/Early Iron Age pottery only represents about 15% of the total ceramic bulk retrieved from the excavations in M1. No complete vessels are attested in M1, while the potsherds are usually fairly small. A tumbler in perfect condition was nonetheless found in the vicinity during the survey in 2009 (fig. 12.5). It is clear from this tumbler that at least some of the ceramic vessels used on the mine during that period were meant for drinking purposes. The rest of the assemblage however, whether it be from M1 or the rest of the dome, comes from medium-sized, short-collared pots, often displaying some basic incised decoration such as annular grooves or oblique stitches (fig. 12). One piece of a narrow-necked jar collected from W5 (C. 1196) shows that ceramic water pitchers were also used at Duzdağı during the Late Bronze/Early Iron Age, but this is the only example to be found (fig. 12.1). As in the Kura-Araxes period, the LBA/EIA assemblage only represents a fraction of the general repertoire usually found on settlements or in graves, but judging from the available shapes, its function seems on the whole quite different: the short-collared pots are adapted for serving or containing food rather than for transporting or storing liquids. Moreover, the LBA/EIA corpus of Duzdağı, most of which is grey-coloured, is similar to the ceramic assemblages attested on settlements or in graves during that period, unlike the Kura-Araxes assemblage. The vessels were crafted by hand, but the use of a tournette is at times visible, especially for producing the parallel annular grooves.

Fig. 12 – Late Bronze/Early Iron Age pottery from Duzdağı (MBA, K. Alhamid).

Late Iron Age/Early Antiquity

15The Late Iron Age/Early Antiquity ceramic repertoire from Duzdağı is more difficult to analyse since very few diagnostic sherds were collected, apart from one potsherd and one miniature pot in M9, and a series of potsherds in M1, which seemingly belong to some family of large, belted, barrel-shaped vessels (fig. 13.3‑4).

Fig. 13 – Late Iron Age, Early Parthian pottery from Duzdağı (MBA, K. Alhamid).

16In M1, this pottery amounts to about 15% of the total bulk, just like the Late Bronze/Early Iron Age ceramics. From the available shapes and ¹⁴C readings, the assemblage from M1 seems slightly later (ca 200 BC-30 AD) than the Late Iron Age corpus from M9 (ca 400‑37 BC), which, admittedly, is scanty. All the pottery that was collected in M1 and M9 is black and lightly burnished on the exterior; the paste and temper (medium grits) that were used to craft the barrel-shaped jars are fairly standardised. These vessels were probably made by hand with the use of a slow rotating device. The piece of a flat bottom (21.7 cm in diameter) with a hole pierced slightly off-centre before firing was found in M1‑022; the inside part of this fragment was partly stained with bitumen (fig. 13.6). Judging by the temper, this potsherd may be the base of a barrel-shaped jar. The hole near the centre suggests that this jar may have contained some liquid, but this liquid cannot have been bitumen, since only part of the hole is stained. Whatever the function of this vessel, it appears from the size and the general composition of the Late Iron Age/Early Antiquity repertoire that the use of ceramics at Duzdağı during that period was on the whole fairly limited; even more so, it seems, than during the LBA/EIA period.

Interpretation

17It is clear from the available data that the use and thus the importance of ceramics on the salt mine of Duzdağı changed greatly over time: if we exclude the medieval pottery, which is not considered in the present study, the largest corpus dates back to the Kura-Araxes period, followed by the LBA/EIA and the LIA/Early Antiquity assemblages. The very small number of 5th millennium BC potsherds may easily be explained by the limited investment Late Chalcolithic communities had in salt extraction: salt was probably collected for daily needs from the surface, possibly even without the use of extracting tools. This kind of opportunistic exploitation system may reflect the casual use Late Chalcolithic communities made of salt – a use, however, that seems significantly greater than during the Neolithic, since not a single Neolithic potsherd has so far been found over the slopes of the dome.

18The relatively small amount of 2nd‑1st millennium sherds (Late Bronze and Iron Ages), on the other hand, may be explained quite differently: the vessel assemblage of the LBA/EIA period, in particular, seems mostly focused on containing/serving solid matter (food?); apart from two examples that clearly pertain to the individual consumption of liquids, probably water (fig. 12.1‑5). Since the cutting and processing of salt slabs evidenced in M1 for the LB/EI Ages implies a massive work investment (Gonon et al., this volume), the relative scarcity of ceramic potsherds in that area cannot be interpreted as indicative of small-scale mining activities, as is the case during the Late Chalcolithic period. It is more likely that the physiological needs of the miners, in particular thirst, were met by keeping water in alternative types of containers, possibly skin gourds. In other words, the quantity of LBA/EIA ceramics in M1, but also over the rest of the dome, should be interpreted as the evidence of a specific technological system, where clay containers only played a minor part in the logistics set up for providing food and water.

19A similar analysis may be carried out for the Late Iron Age/Early Antiquity: even if the barrel-shaped vessels may have been used for storing water, it is probable that most of the miners’ water needs in that period were provided for by other, non-ceramic, containers.

20Considering the Late Prehistoric data from Duzdağı over the longue durée, the Kura-Araxes ceramic assemblage is certainly characterized by functional specificities that mark it out from earlier and later assemblages. The fact that none of the jars from Duzdağı are black-burnished, unlike the pots, is probably indicative of this function, which judging by the jars’ shapes, was not directly related to salt mining: the narrow necks, the size of the jars, the absence of organic residues and the ubiquity of these jars all over the salt dome, all suggest that these vessels were used for procuring water for the miners: the arid conditions that prevail over the salt dome of Duzdağı, together with the absence of fresh water springs in the vicinity2, certainly make the procurement of water the main necessity for anyone seeking to work at Duzdağı.

21In confirmation of the identification of these jars as water containers, it is interesting to note that ceramic vessels that are very similar in shape, size and weight are used nowadays in contemporary Ethiopia by villagers who carry them on their backs over several kilometres to bring water home (fig. 14‑15).

Fig. 14 – Water jar from Dembetcha, Ethiopia, 2012 (C. Marro).

Fig. 15 – Queuing for water at Dembetcha, Ethiopia, 2012 (C. Marro).

22Thus, contrasting with the practices of the Late Bronze or the Iron Ages, where water was probably carried in gourds made of organic material that have left no traces, ceramic water jars were used during the Kura-Araxes period to provide for the water needs of the miners. It should be noted that these jars, whose shapes and dimensions are fairly standardized, display a wide range of grit combinations, probably temper, which are visible in the clay to the naked eye: the temper range exemplified in the Kura-Araxes pottery from Duzdağı is much wider than the variety of tempers attested in the clays from Ovçular Tepesi, for instance. This macroscopic observation awaits further investigations through petrographic analyses, but it certainly has implications in terms of mining organization, which suggests that the Kura-Araxes miners from Duzdağı came from different villages, possibly in separate groups. It should be noted that a similar hypothesis has been put forward independently by R. Badalyan in his analysis of the ceramic repertoire of Shengavit, a major Kura-Araxes settlement located next to the salt dome of Yerevan in Armenia: Badalyan analyses the unusual coexistence of different ceramic “regional sub-styles” at Shengavit as the indicator of its economic importance, which he links to salt exploitation. As an economic hub, he argues, Shengavit may have attracted villagers from various settlements in connection with the exploitation of salt (Badalyan, this volume). In support of this hypothesis, is the fact that water jars similar to the Duzdağı examples have indeed been found on the site of Shengavit (Badalyan et al. 2015, tab. 26, p. 287) – even if none has apparently been found on the salt dome itself.3 The jars from Shengavit are buff-coloured and lightly burnished, just like some of the Duzdağı examples.

Comparison of the Kura-Araxes pottery from Duzdağı with the ceramic vessels of metalliferous mines

23The specificity of the Kura-Araxes ceramic repertoire from Duzdağı is further confirmed by the data gathered from metalliferous mines, in particular from the gold mine of Sakdrisi in Georgia.

24Sakdrisi is located in the Kura basin, in a well-watered, wooded environment (fig. 16), not far from the town of Bolnisi. This archaeological site was investigated by a German-Georgian team between 2004 and 2013, before being destroyed by a private modern mining company in 2014. The gold deposits of Sakdrisi were first exploited on a systematic basis during the Kura-Araxes period, between ca 3200 and 2500 BCE (Stöllner et al., this volume; Gambashidze, Stöllner 2016, p. 36).

Fig. 16 – The gold-bearing hill of Sakdrisi, August 2011 (C. Marro).

25At Sakdrisi, the gold particles are found in contact with hematite and quartz in sub-vertical, compact veins that run through a substratum basically composed of rhyolite. Traces left over the mine’s walls, together with a series of experiments carried out by S. Timberlake, have shown that the gold from Sakdrisi was retrieved through fire-setting; a technique that requires a considerable amount of work and expertise since it implies crushing the rocks into powder before panning in order to retrieve gold dust (Gambashidze, Stöllner 2016, pp. 53‑55).

26The Kura-Araxes ceramic assemblage collected from Sakdrisi is fairly limited in size and poorly preserved.4 According to N. Otchvani, this assemblage presents overall similarities with the pottery repertoire from the settlement of Dzedzvebi, while the clay composition of both testifies to a common origin (Otkhvani et al. 2012, tab. 1). The two sites were certainly frequented by the same groups of people: Dzedzvebi appears as a miners’ settlement dedicated to the exploitation of the rich metal ore deposits of the Bolnisi area, in particular gold (Stöllner et al., this volume).

27However, there are clear differences in size and state of preservation between the ceramic corpus of Dzedzvebi and that of Sakdrisi: the Kura-Araxes assemblage from Sakdrisi is limited to a few hundred potsherds, which are fairly small and display worn-out surfaces, while the pottery retrieved from Dzedzvebi shows the full array of a settlement’s repertoire, with a variety of shapes such as jars, pots, bowls or lids and a number of complete or almost complete pots (Otkhvani et al. 2012; Otchvani et al., this volume). This contrast could merely be attributed to the difference in nature between the two sites, since we are comparing a settlement with an industrial site, a mine, where people did not actually live. But the differences between the two assemblages may also relate to the extremely different functions of the ceramics used on each site.

28The specificity of the Sakdrisi repertoire for instance is even more striking when it is set against the Kura-Araxes corpus from Duzdağı: in both cases, the ceramics come from mining works, but they have very little in common in terms of preservation, typology and quantity. It is evident that the Kura-Araxes potsherds from Sakdrisi come from a limited number of ceramic vessels, illustrated by narrow and wide-collared jars, a few pots and some decorated pottery (incised or relief decoration). Many of these sherds display eroded surfaces (Otchvani et al., this volume, fig. 6). The Kura-Araxes pottery from Duzdağı on the other hand is both much more abundant and better preserved; but it displays a much narrower range of vessel types, where decoration is virtually absent. This of course could be linked to chronological, regional or socio-symbolical parameters, since incised or relief decoration is rare in Nakhchivan and belongs rather to the end of the Kura-Araxes period, towards 2600‑2400 BCE.

29However, on comparing the two assemblages, the general impression is that the ceramic repertoire from Duzdağı is certainly narrow, but reflects specific aims and standards: beige or buff-coloured, three-lugged jars are meant for water, while plain5, black-burnished pots are reserved for solid matter (food?). The ceramic repertoire from Sakdrisi on the other hand seems to be made from bits and pieces that have been chosen more or less at random for one specific need (Gambashidze, Stöllner 2016, p. 35; Otchvani et al., this volume). According to T. Stöllner, this need could be the use of ceramic potsherds as tools for scraping the walls of the rock face, in order to expose the richest gold-bearing veins in the mine (Stöllner, pers. comm.); some of these sherds may also have been used for collecting the ore grinding remains from leathery surfaces (Otchvani et al., this volume). Almost any potsherd could fulfil this task, provided that it has the right shape and texture; this would explain why the potsherds from Sakdrisi are so varied but also so worn down: these sherds should be considered as tools of ré-emploi. This means that the ceramics brought over to Sakdrisi probably arrived already in a broken state, even if a few complete vessels may also have been used on the spot for containing water or food (Otchvani et al., this volume).

30It is clear from the above that the functions of Kura-Araxes ceramics from Duzdağı and Sakdrisi are quite different, although the two sites are partly contemporary. This is explained of course by the fact that we are faced with a salt mine in one case and a gold mine in the other: a potsherd would not be of much use for exposing the salt deposits of Duzdağı. But another important difference between Duzdağı and Sakdrisi relates to climatic conditions, which render work at Duzdağı quite impossible without regular and abundant water supplies: Duzdağı is located in a dry, barren zone where water resources are either salty or non-existent; there is no vegetation apart from rare salt-compatible species such as Tamarix. Sakdrisi on the other hand is surrounded by deciduous and pine forests, with the Mashavera river flowing at the bottom of the gold-ore bearing hills. This suggests that the organization of early mining depends at least as much on environmental conditions as on the type of resources that are actually mined.

31Another case in point is illustrated by the copper-ore deposits, which are attested in many places throughout the Caucasus. In the Araxes basin, copper beds are mainly located in eastern Nakhchivan, south-eastern Armenia and north-western Iran (Laffitte 1982, maps 6, 9). In Nakhchivan, these copper beds have been investigated between 2007 and 2016 through the systematic sampling of the ores with a view to tracing back the origin of the metal used by the ancient coppersmiths in Nakhchivan and beyond.

32This study, which is still in progress, has shown that extractive metallurgy was known to Nakhchivani coppersmiths as early as the second half of the 5th millennium BC and that the copper used by Late Chalcolithic smiths was local (Gailhard et al. 2017). In spite of the evidence, which demonstrates that at least part of the ore used at Ovçular Tepesi actually came from the copper beds of Misdağı, not a single Late Chalcolithic potsherd could be found in the vicinity of these deposits. This is also true for the Early Bronze Age: no pottery belonging to that period could be found either, but this situation is less striking since the origin of the copper used by Kura-Araxes metallurgists in Nakhchivan has not been identified yet (Gailhard et al., this volume).

33Investigations carried out elsewhere in the Caucasus or in eastern Anatolia tend to have similar results: ceramic potsherds are very seldom attested in the vicinity of copper mines, whatever the period considered (Palmieri, Sertok, Chernykh 1993; Courcier, pers. comm.). In point of fact, no archaeological artefacts have so far been found next to the copper deposits sampled in Nakhchivan; most of the copper exploitation in Late prehistory seems to have been carried out by hand-collecting the ore-bearing rocks from the surface. It is certainly hazardous to draw conclusions from the absence of evidence, but considering the fact that ceramic potsherds are attested in greater or lesser quantity both at Duzdağı and Sakdrisi, in particular for the Kura-Araxes period, it seems possible that the total absence of pottery near Nakhchivani copper beds actually reflects their specificity: most of them are easily accessible, in particular to shepherds, since most of these beds are located near major pasturelands (fig. 17). Moreover, the copper-bearing rocks are visible from the surface: they do not have to be mined. These rocks may have been collected with a minimum of logistics, and thus a minimum of tools, ceramics included.

Fig. 17 – The copper deposits of Göygöl, 2016 (MBA, C. Marro).

Conclusions

34The interpretation of the presence of pottery on a mine, and in this particular case, on early Caucasian mines, is certainly not straightforward. The greater or lesser quantity of ceramic potsherds does not necessarily reflect the intensity of mining activities: much depends on the actual function of ceramics on a particular mine, and this function may be related to several factors, which include technology, accessibility, and environmental conditions.

35As concerns Duzdağı, the ubiquity and quantity of Kura-Araxes ceramics over most of the salt dome is certainly indicative of the importance of salt in the Kura-Araxes economy, but the available evidence, in particular the information retrieved from the technological analysis of the ceramics, is complex: it suggests a series of relatively small-scale mining episodes rather than a large-scale mining system organized from the top by some regional authority. This hypothesis of multiple, small mining groups dispatched over the slopes of the dome – not necessarily in the same period or at the same time – matches the rare information available on the Kura-Araxes political and economic system, which points to segmented village societies rather than a proto-state regional formation. A regional authority in the frame of an integrated political economic system probably did not develop in the Caucasus until the second half of the 2nd millennium and possibly much later.

36At the other end of the 3rd millennium, the case of the Middle Bronze Age is more difficult to assess. Middle Bronze Age pottery is virtually absent from the archaeological record of Duzdağı, which is surprising since Middle Bronze Age cultures (ca 2300‑1700 BCE) developed in the wake of the Kura-Araxes phenomenon. Moreover, research carried out in eastern Anatolia and Armenia has shown that the Middle Bronze Age correlates with the progressive disappearance of permanent settlements from the highlands (Marro 2011). The collapse of sedentary life in the Caucasus and eastern Anatolia is followed by an increase in nomadic pastoralism, which left numerous funerary monuments and cemeteries in the uplands, but no settlements in the valleys (Özfirat 2001) – with the notable exception of Nakhchivan, where several such settlements are nonetheless attested (Belli, Bakhshaliyev 2001). Thus, one would have expected this increase in pastoral activities all over the highlands to be reflected in the salt exploitation of Duzdağı, since salt is a major staple in animal herding. But this is not the case; in point of fact, we are even faced with the opposite situation: as already said, the number of ceramic potsherds at Duzdağı falls dramatically in the Middle Bronze Age to a handful of pieces; they are not even concentrated in a particular area.

37No evidence has so far been found that could explain the near-absence of Middle Bronze Age pottery at Duzdağı; we may only surmise that the elegant black-on-orange painted ware typical of Nakhchivan and eastern Anatolia was kept for grand occasions such as funerals or banquets, not for working on a salt mine. As in later periods, Middle Bronze Age miners may have used water gourds made of perishable material to work at Duzdağı – but of course we have no proof to support this hypothesis.

38A similar case is that of the Neolithic, except that not a single sherd resembling any of the Neolithic pottery assemblages known in the area – Shomu-Shulaveris or Kültepe I – has so far been found at Duzdağı: does this mean that salt was not used during the Neolithic period in the Araxes basin? This seems unlikely since herding, in particular caprine herding, already had a major role in the local economy, whether in Armenia at Aratashen-Aknashen or in Nakhchivan at Kültepe I.6 At any rate, the absence of Neolithic pottery at Duzdağı is intriguing, since the settlement of Kültepe I is located only 7 km from Duzdağı as the crow flies: the absence of Neolithic traces of exploitation at Duzdağı so far remains unexplained, unless we again postulate the use of containers made of organic material.

39To come back to the function of Kura-Araxes pottery in Caucasian early mining, the comparison between the assemblages from the salt mine of Duzdağı and the gold mine of Sakdrisi has shown that pottery may be used in very different ways by communities that presumably shared the same technological universe, but exploited different mineral resources in widely distinct environmental settings. Not surprisingly, the specific functions ceramics had at Duzdağı and Sakdrisi are reflected in their stone tool assemblages, which are quite distinct (Hamon et al., this volume; Tamazashvili, this volume). It is clear that working at Duzdağı and Sakdrisi required both different logistics and mining strategies. But unlike ceramics, stone tools may directly reflect the qualities of the raw material that is being mined: hard or soft, bed-layered or mixed with other minerals. The tools used for mining will be crafted accordingly and may reflect the specificities of the raw material and the technology involved in mining. The information provided by pottery on the other hand is often indirect: it may give clues about the technological system set up for exploiting a particular material, as is the case in Sakdrisi for instance, but the information usually retrieved from its study rather relates to logistics, work organization and possibly to the mine’s environment.

40In the case of Duzdağı and Sakdrisi, it is clear that both the environmental setting and the specificities of the raw material being mined had a major importance in the way ceramics were used on a mine by Kura-Araxes communities.