Navigation – Plan du site

AccueilNuméros45-1ArticlesNew light on the Late Prehistory ...

Articles

New light on the Late Prehistory of the South Caucasus: Data from the recent excavation campaigns at Kültepe I in Nakhchivan, Azerbaijan (2012-2018)

Catherine Marro, Veli Bakhshaliyev, Rémi Berthon et Judith Thomalsky
p. 81-113

Résumés

Cet article analyse les résultats d’un cycle de fouilles menées sur le fameux site de Kültepe I au Nakhchivan par une équipe franco-azerbaïdjanaise (2012-2018). Après avoir comparé la stratigraphie mise sur pied par les archéologues soviétiques dans les années 1950 avec les niveaux d’occupation nouvellement exhumés, les auteurs débattent des processus de néolithisation ayant conduit au développement d’une économie de production au Caucase. La comparaison des données de Kültepe I avec celles des autres sites néolithiques fouillés dans les bassins de l’Araxe et de la Kura montre que l’adoption du mode de vie néolithique au Sud Caucase résulte probablement de plusieurs événements, dont la migration de peuples en provenance d’Iran et de Mésopotamie. L’analyse conjointe des données obtenues par la fouille des niveaux Kuro-Araxe et des données collectées par les Soviétiques montre par ailleurs que le groupe humain qui s’est installé à Kültepe I dans le dernier quart du IVe millénaire est culturellement distinct de ceux vivant à proximité de ce site, en particulier distinct du groupe établi à Kültepe II, où des niveaux d’occupation Kuro-Araxe datant de la fin du IVe millénaire sont aussi attestés.

Haut de page

Texte intégral

1Manuscript received the 8th December 2018, accepted the 17th April 2019

2The site of Kültepe-Nakhchivan, also called Kültepe I, has been considered as a key-site in Caucasian archaeology ever since it was excavated by O. Abibullayev during the Soviet era when Nakhchivan was part of the USSR. The fame of the site is partly due to its size since this artificial mound (or “tepe”) used to tower at least 12 m above the modern surrounding village, on the left bank of the Naxçıvançay river, some 8 km to the north of Nakhchivan city (Abibullayev 1959: 11).

3But the importance of Kültepe I is also due to its many occupation levels, which start in the Neolithic and encompass several millennia of Caucasian history until the Iron Age at least. This is a rare case north of the Araxes river, where most settlements hardly exceed 5 m in height and usually comprise occupation layers belonging to one or two chronological periods at the most. Lastly, Kültepe I stands out because some Halaf pottery was brought to light in its earlier levels (Abibullayev 1982: table XXII:1-2), thus establishing a link with the Neolithic cultures of Syro-Mesopotamia and the prestigious narrative attached to them.

4As a result, Kültepe I has been widely cited in all the scientific literature dealing with Caucasian Late Prehistory (Burney and Lang 1971, Kushnareva 1997, Sagona 2018), notwithstanding the paucity of data made available by the Soviet team, whose excavations lasted over 13 seasons (between 1951 and 1964), but resulted in only two general reports that partly contradict each other on chronological matters (Abibullayev 1959, 1982). This situation has led to some speculation and misinterpretation, all the more so as Kültepe I has often been mistaken for Kültepe II by western scientists: unlike Kültepe I, occupation at Kültepe II is restricted to the Bronze and the Iron Ages. The site, which was excavated between 1962 and 1989 by V. Aliyev (Ristvet et al. 2011: 10), is located at the confluence of the Naxçıvançay and the Çehriçay rivers, some 5 km to the north of Kültepe I. The fact that many settlements in the area are called by the name of “Kültepe” is due to the presence of a thick layer of ash (“kül” means “ash” in Turkish) in the Kura-Araxes occupation levels, which renders the use of additional labels necessary to differentiate between the settlements named “Kültepe” in Azerbaijan and in Iran. These ashy levels are not visible at Kültepe I any longer because most of its Kura-Araxes layers were dug out by the Soviet excavations, as we realised on working on the site.

5After excavating at Ovçular Tepesi, a Late Chalcolithic and Early Bronze Age settlement located on the left bank of the Arpaçay river in Western Nakhchivan (Marro et al. 2009, 2011, 2014), it seemed relevant to turn to earlier periods in order to analyse our data from a wider slant, in particular as regards the complex relationships between the South Caucasus and the Near and Middle East (Cucchi et al. 2013). Since Kültepe I was the only settlement where early levels were known to exist in Nakhchivan, we decided to launch a new excavation programme focused on this particular site. Another reason for this choice was the presence of stone mining tools very similar to those found on the salt mine of Duzdağı, which we had started to investigate in 2008 (Hamon 2016). A selection of these tools had been published by Abibullayev (1982: 6-8, table XIV) and interpreted as copper mining tools; thus, it seemed important to investigate what might have been a miners’ site, located both near the salt mine of Duzdağı, but also very near the copper deposits of Misdağı.

6Kültepe I is the largest Late Prehistoric site in Nakhchivan. It is located in a strategic position along a major north-south axis, which links the Iranian plateau via the Culfa (Djolfa) pass to the Lesser Caucasus, in particular to the rich pasturelands of the Upper Karabagh and the obsidian beds of the Zangezur area (fig. 1). During the Middle Ages, a caravan route followed the Naxçıvançay valley up to the pass of Batabat, where a caravansaray, now barely visible, was located in the vicinity of a worn-out cuneiform inscription, probably dating back to the Iron Age. At any rate, it is clear that the Naxçıvançay valley must have been well-trodden for a very long time.

  • 1 A few caravansarays located in the villages bordering the foothills, as at Qarabağlar, testify to t (...)

7It should be noted that Kültepe I is also located along an important East-West axis that links the Iranian plateau to Anatolia: indeed, until the beginning of the 20th century, east-west connections through the Araxes valley ran along the southern piedmonts of the Lesser Caucasus, where a number of water springs are available, and not along the Araxes river itself, which is treeless and bordered by marshes in several places.1 This piedmont route is still used nowadays by shepherds and their flocks descending from the high pasturelands on their way back to their winter quarters, some of which are located near Şarur in Western Nakhchivan: after descending from the Sirab mountains, they turn north to reach the edge of the piedmonts through the Naxçıvançay valley, instead of heading west towards the Araxes river.

  • 2 Bakhshaliyev V. (forthcoming), Archaeological excavations at Nakhchivan Tepe. In: Marro C. and Stöl (...)

8It is thus no wonder that Nakhchivan’s largest site should be located at the crossroads of two major northbound and westbound axes. Many other sites are in fact attested in this area, sometimes buried under a thick layer of alluvium, such as Uzunoba or Yeni Yol (fig. 1): these two Chalcolithic settlements lie respectively opposite Kültepe I on the right bank of the Naxçıvançay (Uzunoba) and at the confluence of the Naxçıvançay and the Şorsu rivers (Yeni Yol), some 2 km to the north of Kültepe I. A third example is Nakhchivan Tepe, also a Chalcolithic settlement located by the Naxçıvançay, but closer to the Araxes river. In this particular case, the site is not covered by alluvium but has been severely eroded by the river. This settlement is particularly interesting since it has yielded a wealth of Dalma ware belonging to the impressed-decoration period, so far a unique example in the area.2

Fig. 1 – Regional map with sites mentioned in the text

Fig. 1 – Regional map with sites mentioned in the text

Map O. Barge and C. Marro

9Apart from Kültepe I, it is worth noting that another large site, Kültepe-Hadishahr, stands on the Iranian side of the Araxes valley, in the Culfa pass itself. Kültepe-Hadishahr is a multi-layered mound whose earliest levels probably date back to the Early Chalcolithic, as suggested by the presence of painted Dalma-ware (Abedi et al. 2014: 37, fig. 6 and 9-11). This settlement controlled the Culfa pass so that its strategic function was probably as important as that of Kültepe I: no wonder then that Kültepe-Hadishahr is just as large, if not larger, than Kültepe I, since it reaches a height of 19 m for an area of 6 ha (Abedi et al. 2014: 33).

10It is in fact difficult to estimate the original extent of Kültepe I since most of the site has been destroyed by earth-pitting and covered by modern houses. But the major difficulty in reconstructing its size is linked to the massive alluvial and colluvial processes that have buried the settlement’s earlier levels: from the description of the site by Abibullayev, we understand that the mound’s total thickness must have reached ca. 22 m, of which only 12 to 13 m were visible above ground at the time of the excavations. In 2015, we opened a sounding in the southern part of the site and proceeded to some coring near the western road to estimate the site’s overall dimensions (fig. 2): these tests suggest that the archaeological layers have been disturbed by levelling and terracing activities in many places, as shown by a few erratic radiocarbon readings (table 1, especially in Chantier B and D): unless a series of excavation squares is opened along its edges, the site’s exact extent will remain difficult to assess. Abibullyaev estimated the site’s size at the time of his excavations to be at least 1.5 ha (Abibullayev 1982: 17), which he thought was a low estimation—and he was probably right.

Table 1 – Radiocarbon dates, presented in stratigraphic order for each excavation area

Table 1 – Radiocarbon dates, presented in stratigraphic order for each excavation area

The dates were calibrated with the OxCal programme (Bronk Ramsey 2009), using the calibration curve IntCal13 (Reimer et al. 2013)

R. Berthon

11As we soon discovered, most of the occupation levels still visible at Kültepe I belong to the Neolithic period: later archaeological layers were gradually destroyed through time either by the villagers, who pitted the mound’s soil from the sides, by an earthquake in 1931 or by the Soviet excavations, which levelled the mound almost flat, and even left a vast depression in the northern zone of the settlement. For this reason, this paper will mostly focus on the Neolithic period.

The excavations: aims and strategies

  • 3 This work was part of a research programme funded by The Shelby White and Leon Levy Program for Arc (...)

12On resuming the excavations of Kültepe I, we had two goals in mind: the first was to identify the former Soviet excavation area, both horizontally and vertically, with a view to matching our data with those published by Abibullayev in 1959 and 1982.3 Since the Soviet system of registering archaeological artefacts only relied on measuring the depth of each object or stratum from the summit of the mound as it stood in 1951, much of our initial work was devoted to identifying the absolute elevation of the main Soviet chronological phases, especially as the 1951 summit of the mound has been entirely dug out. For this reason, we reopened the main excavation square of the Soviet era (“Area 5”) down to the virgin soil and scraped the dirt off the stratigraphic profiles to the south and the west of the former trench (fig. 3).

  • 4 This work was supported by the French Ministry of Foreign Affairs (“Mission archéologique du bassin (...)

13Our second goal was to work on the stratigraphy of this exceptional site and compare it with that of Ovçular Tepesi in order to set up a diachronic regional cultural sequence that could be used as a basis for future work in the Araxes basin at large.4 Another aim was to investigate the storage facilities of the earlier occupation levels, with the hope of finding the remains of house mice (Mus musculus) to compare with the bones of the many rodents collected at Chalcolithic Ovçular: drawing on the evidence obtained from the latter, we supposed that the Kültepe I mice could be of some help for reconstructing the processes that led to the adoption of a productive economy in the Caucasus (Cucchi et al. 2013). At the same time, we hoped to uncover Kura-Araxes levels on a sufficiently large scale for investigating the socio-economic context associated with the mining stone-tools brought to light by the Soviet excavations.

14So, eight squares (Chantiers A-G and J) of different sizes were opened altogether (fig. 2). Chantiers A and C were placed where the high mound had once stood; they more or less correspond to the Soviet Area 5. In this zone, we first emptied the former excavation pit, which had been filled with soil and rubbish over the years, and then opened a trench down to the virgin soil (Chantier C), where we found the bottom of Area 5, as well as the settlement’s first level, which marked a striking contrast with the yellowish silt of the virgin soil. The south section of Chantier C, where the accumulation of archaeological layers could be exposed over a depth of 5 m, was sampled in order to obtain organic material for 14C dating and to match our sequence with the Soviet data.

Fig. 2 – 2012-2018 excavation areas at Kültepe I

Fig. 2 – 2012-2018 excavation areas at Kültepe I

CAD P. Lebouteiller and G. Gadebois

15In Chantier A, we opened as large a square as possible in order to identify what remained of the former high mound. The shape of this square was very much constrained by the pits opened by the villagers and previous archaeologists, but it was clear from the beginning that the archaeological remains that were still visible belonged to the very top of Soviet Stratum 1. Since parts of the mound seemed to be better preserved to the south of Chantier A, we opened a sounding under what had been a parking place in the hope of finding Kura-Araxes levels in that area (Chantier G).

16Chantiers B, D-F and J were placed in front of the new mosque, in a zone that is now quite flat. This area was a garden at the time of the Soviet excavations and was not investigated (Abibullayev 1959: fig. 3). However, Chantier B proved to contain very disturbed levels, probably resulting from terracing activities linked to the building of the mosque at the beginning of the 21st century. Chantiers D-F and J on the other hand yielded in situ archaeological remains over a fairly wide area. In Chantier E, we opened a deep sounding in order to compare the stratigraphy on the two sides of the former high mound. Lastly, we opened a narrow trench in what was left of the high mound to the south of the site, in order to investigate the Kura-Araxes occupation phase (Chantier G). In total, an area of about 350 m2 had been excavated by 2018, but this surface varies greatly depending on the period considered.

Fig. 3 – Location of former Soviet “Area 5”

Fig. 3 – Location of former Soviet “Area 5”

Photo C. Marro, MBA

The Neolithic occupation levels (ca. 6200-5000 cal. BC)

17As noted above, it became clear on resuming the excavations at Kültepe I that most of the archaeological layers still visible under the surface of the site belonged to Soviet Stratum 1. We observed that the massive alluvial and/or colluvial processes that led to a thick accumulation of sediments at the bottom of the site had stopped at the top of the Neolithic mound: later archaeological levels have not been surrounded by geological deposits. For this reason, only the Neolithic levels have been shielded, so to speak, from earth-pitting and opportunistic plundering over the years.

Chronological issues

  • 5 Or “Eneolithic” in the Russian terminology.
  • 6 Abibullayev’s second report was published in 1982 but since he died in 1970, this report was presum (...)

18The dating of the earliest occupation levels of Kültepe I has led to some controversy, partly because no adequate 14C readings were obtained by the Soviet excavations, and partly because the ground on which the Caucasian Neolithic had been defined was rather shaky, being mostly based on stylistic analogies with Near Eastern cultures. O. Abibullayev himself had first attributed what he called “Stratum 1” to the Neolithic (Abibullayev 1959: 14) and divided this remarkable stratigraphic sequence (between 9.2 and 9.4 m thick) into two main stages, labelled “Stratum 1a” and “Stratum 1b”, the former corresponding to layers located between -21.1 and -19 m and the latter to those located between -19 m and -12.8 m in depth. For reasons that are not totally clear, possibly because of a 14C dating obtained from a piece of charcoal sampled at -18.2 m, which pointed to the beginning of the 4th millennium BC (3807±90 BC; Abibullayev 1982: 191), Abibullayev changed his mind and attributed the whole Stratum 1 to the Chalcolithic5 in his second report. Abibullayev tried to justify this new chronological attribution by drawing parallels between the rare painted ceramics of Kültepe I and the Halaf pottery from Tilki Tepe in Eastern Anatolia. This argument is rather odd, since even in the late sixties6 the Halaf period was roughly (and mistakenly) dated to the end of the 6th millennium BC, not to the beginning of the 4th millennium BC (Garelli 1969: 62-63).

19The dating of Kültepe I, however, was later challenged by Kushnareva, who pointed out that part of the archaeological assemblage, namely biconical spindle whorls and a grooved-sickle made of antler, showed close parallels with similar artefacts from Hajji Firuz, a Neolithic site located south of Lake Urmiah in Iran (Kushnareva 1997: 30): for this reason, Kushnareva suggested that Stratum 1 of Kültepe I should rather be dated to the Late Neolithic period, which she placed in the 6th millennium BC (Kushnareva 1997: 22).

20More recently, Seyidov challenged both interpretations, as he attributed the earlier layers of Kültepe I (the former Stratum 1a) to the Neolithic and the later levels (Stratum 1b) to the Early-Middle Chalcolithic, which he placed within the 5900-4300 BC time span, without actually explaining why (Seyidov 2003: 22-28).

21The new chronological time span derived from the French-Azerbaijani excavations draws on a series of 32 radiocarbon dates, 28 of which were sampled in levels corresponding to Soviet Stratum 1 (table 1). The radiocarbon dates obtained from these levels are on the whole coherent and cover the 6200-5000 BC time span. The general coherence of the chronological sequence is marred, however, by two readings from Chantier B (LTL14889A and LTL14890A), where a piece of charcoal sampled at 947.7 m (LTL14890A) was dated much earlier than a sample taken from a lower locus (LTL14889A). Presumably, this may be explained by the terracing activities carried out during the construction of the modern mosque, as shown by the disturbed layers visible in the northern profile of the excavation square. This is also the case for LTL14936A, a piece of charcoal sampled from Chantier D in the north-west corner of the excavation area, close to Chantier B: it produced a date that seems much too early. As for sample LTL16899A from Chantier A, which stands completely out of range (5581±45 BP, 4494-4343 cal. BC 2σ), there is no explanation for its odd dating other than possible preservation problems of the sample itself.

22Three readings obtained from Chantier E are of special interest (LTL16900A, LTL18618A, LTL18619A), since they suggest a very early date for the foundation of Kültepe I. They come from the earliest occupation level of the deep sounding (E-362), and point to the last quarter of the 7th millennium BC (in median dates: 6228, 6138, and 6107 cal. BC). This is about 200 years earlier than the dates obtained from the earliest level of Chantier C, where two charred seeds were sampled (LTL15113A and LTL15114A). The difference in altitude between the two soundings (943.2 m versus 944.48 m) does not necessarily explain the 200-year difference between the two groups of readings, since the Neolithic settlement follows a gentle slope that is clearly visible in the north and south sections of Chantier E. Even if it is located some 1.3 m lower than the base of Chantier C, the earliest layer of Chantier E could belong to the same occupation level as the bottom end of Chantier C. However, this is not what is suggested by the radiocarbon dates obtained from Chantier E: rather, it seems that the occupants of Kültepe I first settled to the west of Soviet Area 5.

Stratigraphy and architecture

23A major point of interest in the stratigraphic sequence evidenced in Chantiers C and E is the marked change visible ca. 945 m between a series of dark, organic layers extending from the first settlement level (ca. 943 m in Chantier E) up to the height of 945 m, and the lighter, silty layers found above that height. This change has been dated in both cases ca. 5900-5800 BC. Since this change corresponds to apparent specificities in the architecture, we have tentatively grouped the dark organic layers under the label “Level 1”, while we will refer to the lighter, silty layers as “Level 2”.

24The question arose, of course, as to whether the change observed in the deep soundings corresponds to the distinction made by Abibullayev between his Stratum 1a and Stratum 1b ca. -19 m in depth. The thickness of Stratum 1a (22-19 m = 3 m) would place its upper surface at ca. 947.5 m in Chantier C (former Soviet Area 5) according to the following calculation: 944.5 m (ground level of Chantier C) + 3 m. But this would place the divide between Stratum 1a and Stratum 1b about 2,50 m higher than the stratigraphic change observed in Chantiers C and E, so that the putative correspondence of Level 1 with Stratum 1a and Level 2 with Stratum 1b does not seem to work out. Moreover, no clear change was visible in the stratigraphic sequence of Chantier C around 947.5 m asl. Abibullayev himself abandoned the distinction between Stratum 1a and Stratum 1b in his 1982 report anyway, arguing that his previous analysis was based on a 30 m2 excavation square and that there was in fact no real ground for such a distinction once the excavation area had been enlarged (Abibullayev 1982: 24).

25Whatever its relationship with Soviet Strata 1a and 1b, the change in sediment colour and texture noticed ca. 945 m in Chantiers C and E is certainly significant since it parallels certain changes observed in the archaeological assemblages between Levels 1 and 2.

  • 7 Uzdurum M. (2013), Aşıklı Höyük yerleşmesinde ateş yerleri ve kullanımı [Fire places and their func (...)

26This is the case with the architecture for instance, although the available evidence should be handled with caution since the earliest occupation levels of Kültepe I were investigated by the French-Azerbaijani team over a very small area: 12.50 m2. This figure, however, is compensated for by the height of the stratigraphic sequence in Chantier E, visible over 6 m in the eastern section of the square and 5.75 m in the north and south sections (fig. 4-6). The most striking element visible in the sections is the absence of mud-built architecture in Level 1, whereas Level 2 is marked by the regular occurrence of mud houses. One such building was partially brought to light in the deep sounding itself (Chantier E) at 945.3 m: it is a circular house (E-242) measuring some 4 m in diameter, which contained a half-buried storage jar and a circular hearth built with pebbles (fig. 4 and 7). Two building levels are visible: the walls of the first house (ca. 945.3 m) seem to have been built with mud, while the second building (ca. 946.3 m) is made of large mud lumps called “möhre” in Azeri Turkish. As for Level 1, a series of postholes bordering a dark patch found at the bottom of the sounding, possibly a fragmentary floor, suggests the existence of simple huts (fig. 7). Moreover, the presence of terrace walls carved into the yellow alluvial soil (E-294), in Chantier C and Chantier E (fig. 8), suggests the existence of semi-buried houses in the earliest occupation phase of the site. Judging by the presence of at least two levels of post-holes planted in the alluvium, visible in the north section of Sounding E (fig. 4: E-358-360 and fig. 9), there should be at least two occupation levels of semi-subterranean architecture in this phase. Level 1 is otherwise characterised by a series of deep, concave hearths made of large river-stones and coated with mud. Certain hearths may at first be mistaken for large postholes, since they are built from pits dug into the soil, and then lined with a thick layer of grey clayish mud and a few stones (fig. 10). But the occasional presence of charcoal along the clay shows that their use involved fire; these hearths actually correspond to features referred to as “fire-pits” in the literature dealing with the Pre-Pottery Neolithic in the Near-East and Anatolia.7 Hearths lined with stones are also attested in Level 2, but they are built over a flat surface, not in a hollow or in a pit.

Fig. 4 – Chantier E, deep sounding (north section)

Fig. 4 – Chantier E, deep sounding (north section)

CAD G. Gadebois

Fig. 5 – Chantier E, deep sounding (South section)

Fig. 5 – Chantier E, deep sounding (South section)

CAD G. Gadebois

Fig. 6 – Chantier E, deep sounding (East section)

Fig. 6 – Chantier E, deep sounding (East section)

CAD G. Gadebois

Fig. 7 – Chantier E, with Sub-level 2B (main square), Sub-level 2A and Sub-level 1B (deep sounding)

Fig. 7 – Chantier E, with Sub-level 2B (main square), Sub-level 2A and Sub-level 1B (deep sounding)

CAD G. Gadebois

Fig. 8 – Chantier E, deep sounding (south-western angle), Neolithic Level 1

Fig. 8 – Chantier E, deep sounding (south-western angle), Neolithic Level 1

Washed-down wall of a semi-subterranean dwelling carved into the yellowish silt (E-294)

Photo C. Marro, MBA

Fig. 9 – Chantier E, deep sounding, Neolithic Level 1

Fig. 9 – Chantier E, deep sounding, Neolithic Level 1

Post-holes (E-358 and E-360) are visible in the north section

Photo C. Marro, MBA

Fig. 10 – Chantier E, deep sounding, Neolithic Level 1

Fig. 10 – Chantier E, deep sounding, Neolithic Level 1

Fire-pit (E-270)

Photo C. Marro, MBA

  • 8 As suggested by the size of the hearth (1 m × 0,8 m preserved over a height of 0.8 m) and the prese (...)

27Again, the small size of the sounding opened in Chantier E precludes any definite interpretation of the architecture of Level 1. It should be stressed, however, that by and large the evidence brought to light by the 2012-2018 excavations matches the description of the architecture of Stratum 1 given by Abibullayev in his 1982 report: apart from two mud walls belonging to a fragmentary rectangular building and containing a large hearth (a kiln?8) found at -21.4 m, the first occupation layers of Stratum 1 are characterised by hearths and pits, with no traces of mud-built houses. All the circular mud-buildings described by Abibullayev only appear from the depth of -17,2 m (ca. 949.2 m asl) upwards (Abibullayev 1982: 87-94). In the south section of Chantier C however, which includes the area excavated by the Soviet team, mud-buildings were visible from 945.7 m asl upwards, but they were apparently not identified as such by Abibullayev.

  • 9 See table 1: LTL16015A, 6473±45 BP, 5520-5330 cal. BC 2σ.

28If we consider the whole area uncovered by the new excavations, which yielded architectural units in Chantiers A, D to G and J, the main features brought to light at Kültepe I are circular buildings made of mud or “möhre. It is not easy to distinguish between these two construction techniques since there is almost no difference in colour or texture between the soil of the mound and the mud used as building material, which suggests that a limited quantity of organic material was added to the soil when preparing the mud. The mud modules, which are too large and irregular to be called bricks, are most of the time only visible in the profile sections. A large circular building (D-025; diam. 7 m) was partly brought to light in Chantier D, in the upper layers of the Neolithic occupation (ca. 5500-5250 BC), but it was void of domestic features except for a circular hearth. Following the tradition visible in the earliest layers of the mound, occasional hearths coated with mud and paved with pebbles are found in outdoor areas: the stones used for these hearths are usually smaller (10-12 cm) than in the earlier examples, but the hearths themselves display similar overall dimensions (ca. 60 cm in diameter). A striking exception is the oval hearth found in Chantier A (A-167), which is made of large river-stones (20-30 cm) for an overall size of 1,60 × 1,80m (fig. 11). This hearth belongs to Level 2: it is dated to the second half of the 6th millennium9 and was found in an outdoor context disturbed by modern pits and earlier excavations. Its size and other peculiarities mark it as a remarkable feature: it contained a thick ash layer (8 cm), while the soil under the ash was burnt red to a depth of 12 cm. No clue was obtained from the ash layer, which was sampled and wet-sieved, that could explain the function of this unusual hearth, but the heat maintained in this place must have been particularly intense.

Fig. 11 – Chantier A, large hearth (A-167) lined with river-stones, Neolithic Level 2

Fig. 11 – Chantier A, large hearth (A-167) lined with river-stones, Neolithic Level 2

Photo C. Marro, MBA

29Surprisingly enough, virtually no features relating to food storage or waste disposal have been found at Kültepe I. A couple of half-buried vessels, like the one found in building E-242 (fig. 12 and fig. 13.7), may have been used for storage, but its dimensions remain fairly modest (E-312, ca. 52 litres) and it appears as an isolated vessel: the absence of an organised storage system actually seems to be the norm in the Caucasian Neolithic, with the major exception of Gadachrili Gora in the Kura valley, where multiple circular mud storage-basins are attested both in outdoor areas and inside the buildings (Hamon et al. 2016: 163, fig. 12 and 19). Isolated mud basins also occur at Hacı Elamxanlı, or in certain compounds at Göy Tepe, which may also have served as storage containers (Guliyev et Nishiaki 2014: fig. 3; Nishiaki et al. 2015: fig. 9).

Fig. 12 – Storage jar (E-312) from House E-242, Neolithic Level 2

Fig. 12 – Storage jar (E-312) from House E-242, Neolithic Level 2

Photo N. Gailhard, MBA

Fig. 13 – Plain ceramics from the Neolithic period

Fig. 13 – Plain ceramics from the Neolithic period

(a) Vessel colour; (b) Paste colour and temper; (c) Surface treatment; (d) Decoration if any; (e) Specific observations if any (f) Periodisation/stratigraphic attribution. E: exterior surface; I: inside surface
1. Locus KT 15-F-085: (a) E: mottled beige and buff; I: beige (b) buff, medium straw-tempered; (c) E: plain, I: slightly smoothed; both sides chaff-faced; (d) none; (e) heterogenous paste from poor clay-kneading; (f) Neolithic, Level 2.
2. Locus KT 15-F-031: (a) E: buff; I: buff; (b) black core; medium straw-tempered with a few mineral inclusions (e.g., quartz); (c) E: buff, evenly burnished; I: unevenly burnished (upper body) or smoothed (lower body); both sides chaff-faced; (d) none; (e) none; (f) Neolithic, Level 2.
3. Locus KT 14-A-040: (a) E+I: light buff; (b) buff with grey core, medium straw-tempered with mineral inclusions (c) E+I unevenly burnished; matt burnishing; both sides chaff-faced; (d) none; (e) heterogenous paste from poor clay-kneading; (f) Neolithic, Level 2.
4. Locus KT 14-A-090: (a) E: cream, I: buff; (b) darkish core, medium straw-tempered; (c) E: cream wash, evenly burnished, matt burnishing; I: unevenly burnished; both sides chaff-faced; (d) none; (e) none; (f) Neolithic, Level 2.
5. Locus KT 14-A-077: (a) E: mottled cream and buff; I: light buff; (b) brown, medium straw-tempered with a few mineral inclusions; (c) E: cream slipped; E+I: light burnishing? Matt surface. Both sides chaff-faced; (d) none; (e) none; (f) Neolithic, Level 2.
6. Locus KT 15-A-161: (a) E: mottled cream and buff; I: beige (neck) and drab (body); both sides chaff-faced; (b) dark core; medium mixed-tempered ware with a few coarse grits; (c) E+I (neck): unevenly cream- slipped, unevenly burnished, matt surface. I (body): drab, smoothed. (d) none; (e) Vessel shaped with a bat; (f) Neolithic, Level 2.
7. Locus KT 17-E-312: (a) E+I: buff; (b) buff, medium straw-tempered; (c) E: unevenly burnished, chaff-faced; I: (worn out surface); (d) none; (e) use of coil technique visible through a shallow groove; (f) Neolithic, Level 2.

CAD F. Quliyeva and T. Tiriş

30As far as waste disposal is concerned, refuse-pits are also conspicuous by their absence at Kültepe I. The only pits that have been found were filled with large grinding stones, some of which had visibly been used to grind red ochre. The absence of refuse-pits may partly explain why the remains of so few rodents have so far been collected across the settlement.

The pottery

  • 10 Chips have not been included in this count. A potsherd is considered to be a chip when at least one (...)
  • 11 This refers to the filling layers located between 943.2 m and 944.8 m asl., in the northern half of (...)
  • 12 See table 1; please note that LTL16905A and LTL18620A have been excluded from the dating of Level 1 (...)

31The pottery from the Neolithic layers of Kültepe I offers a remarkable continuity in terms of shapes and crafting techniques throughout the period, which covers at least a thousand-year timespan. It should be stressed that ceramic vessels are attested in large quantities from the very beginning of the stratigraphic sequence: almost 2000 potsherds10 were retrieved from 5.5 m3 of filling debris in the earliest occupation levels of Chantier E,11 which have been dated by three 14C readings to the 6200-6000 BC time span.12 Moreover, the density of potsherds remains more or less stable throughout the sequence. Most of the repertoire displays basic shapes, usually simple bowls, hole-mouthed pots or cylindrical vessels of various heights (fig. 13). In Level 1, a number of pots seem to have had a carinated body, but no complete vessels have been found so far (fig. 14.5-6). With rare exceptions, this pottery is vegetal-tempered; the exceptions include mixed-tempered ware (grit and straw), but a few cases of crushed-bone temper are also attested. Different kinds of wild Poaceae or a mixture of Poaceae and straw are often used as vegetal temper in the earliest levels of the Neolithic sequence, whereas straw is prevalent in the later levels: this evolution is perceptible from Locus E-250 onwards, which marks the divide between Level 1 and Level 2 (table 2).

Fig. 14 – Ceramics from various periods (1-2) Kura-Araxes; (3-4) Late Chalcolithic; (5-6) Neolithic Level 1

Fig. 14 – Ceramics from various periods (1-2) Kura-Araxes; (3-4) Late Chalcolithic; (5-6) Neolithic Level 1

1. Locus KT 15-G-030: (a) E: black (upper wall) and grey (lower wall); I: grey; (b) dark core, fine mixed-tempered ware; (c) E: lightly burnished, I:  smoothed; (d) capsule-shaped relief decoration; (e) none; (f) Early Bronze Age.
2. Locus KT 15-G-008: (a) E+I: cream; (b) buff, fine vegetal-tempered ware; (c) E+I: cream-slipped; (d) anthropomorphic face in relief over upper end; (e) none; (f) Early Bronze Age.
3. Locus KT 16-G-107: (a) E+I: beige; (b) dark core, fine mixed-tempered/ mineral temper: fine black grits; (c) E+I: slipped and smoothed; (d) finger traces of brown paint (e) none; (f) Late Chalcolithic.
4. Locus KT 15-G-026+G033: (a) E+I: beige; (b) beige core, medium straw-tempered ware, (c) E+I: scraped and smoothed; (d) none; (e) none; (f) Late Chalcolithic.
5. Locus KT 18-E-358: (a) E+I: cream; (b) light buff, medium vegetal-tempered ware with a few mica inclusions; (c) E: cream-slipped, burnished, I: cream-slipped, smoothed; (d) E: red-brown painted geometric motif; (e) use of watery paint; (f) Neolithic, Level 1.
6. Locus KT 18-E-352: (a) E: cream, I: buff, plain; (b) buff core, medium vegetal-tempered ware with a few mica inclusions; (c) E: cream-slipped, burnished; I: buff, plain; (d) red-painted geometric motif, possibly finger-painted (uneven coat of paint); (f) Neolithic, Level 1.

CAD F. Quliyeva and T. Tiriş

Table 2 – Evolution of the ceramic assemblage during the Neolithic at Kültepe I (ca. 6200-5000 BC)

Table 2 – Evolution of the ceramic assemblage during the Neolithic at Kültepe I (ca. 6200-5000 BC)

C. Marro

32The Neolithic pottery of Kültepe I usually displays a low level of craftsmanship, with irregular surfaces that show traces of careless burnishing. The clay is often heterogeneous, which results from hasty kneading. The vessels were crafted with coils that were added from the top—that is, not from the side—as shown by the shallow U-shaped groove sometimes visible along the lower break of the sherds. This usually produced uneven walls and somewhat crooked vessels, which were then clumsily straightened with a bat. Interestingly enough, the vessels retrieved from Level 1 were crafted more carefully than those of Level 2. This is particularly visible in the rims and the surface treatment of the early vessels, which are more regular than those occurring after the Level 1-Level 2 break. A general decline in craftsmanship is perceptible in Level 2, with the frequent occurrence of cylindrical pots displaying evident structural weaknesses: the vessel’s base is thinner than the side walls (fig. 13.3-4), which usually results in a break at the join between the bottom and the walls.

33Another contrast between Level 1 and Level 2 is the frequent presence of cream-slipped vessels in Level 1, while this type of surface treatment sharply declines in Level 2: buff, light buff and brown-coloured vessels then become predominant. Lastly, the pottery repertoire of Level 2 is characterised by the occasional presence of lugs or an annular band applied around the body of certain pots (fig. 15.1 and 4-5): this trait heralds the beginning of the relief decoration technique, so typical of the Caucasian Neolithic and Chalcolithic pottery. It should be emphasised, however, that this type of decoration is fairly rare at Kültepe I, especially in comparison with the pottery assemblages of Shomu-Shulaveris sites (Alekbarov 2018: 28), whether it be in the Kura or Araxes basins. Among the odd finds, the fragment of a strangely decorated jar with two vertical bands of knobs along the neck was brought to light in G-104, in the last levels of the sequence (fig. 15.2): this is a unique case at Kültepe I, but very similar examples have been found at Aruchlo (Sagona 2018: fig. 3.9.1).

Fig. 15 – Relief-decorated ceramics from the Neolithic period

Fig. 15 – Relief-decorated ceramics from the Neolithic period

1. Locus KT 16-G-076: (a) E: cream, I: buff; (b) greyish core, medium vegetal-tempered; (c) E: cream slipped, mat burnished, I: scraped and unevenly burnished; both sides chaff-faced; (d) simple annular band around the neck; (e) none; (f) Neolithic, Level 2.
2. Locus KT 16-G-104: (a) E: drab, I: light brown; (b) red core, medium straw-tempered ware with a few coarse grits; (c) E: unevenly burnished, flaked-off surface due to heavy; concentration of chaff; I (neck): unevenly burnished, (body) smoothed; both sides chaff-faced; (d) two vertical bands of knobs along the neck; (e) heterogenous paste from poor kneading; coil-crafting visible in shallow groove bordering the lower part of the sherd; (f) Neolithic, Level 2.
3. Locus KT 16-G-118: (a) E: dark brown, I: reddish brown; (b) dark grey/brown; medium straw-tempered ware with a few medium grits, including quartz, as well as some crushed bones; (c) E: evenly burnished? Flaked-off surface; I: evenly burnished, soft surface; (d) loop in relief bordering vessel rim; (e) none; (f) Neolithic, Level 2.
4. Locus KT 16-F-110: (a) E+I: light beige; (b) light core, medium straw-tempered ware with a few coarse grits; (c) E+I: unevenly burnished, matt surface; (d) simple annular band around the upper part of the body; (e) none; (f) Neolithic, Level 2.
5. Locus KT 16-E-234: (a) E: cream, I: buff; (b) beige/buff core; medium straw-tempered ware with a few coarse grits, including quartz; (c) E: cream slipped, evenly burnished? (stained potsherd); I: unevenly burnished; (d) simple annular band around the body; (e) wavy, uneven surface inside corresponding to the coils used for crafting the vessel; (f) Neolithic, Level 2.

CAD K. Alhamed, F. Quliyeva and T. Tiriş

34Painted pottery is conspicuous by its near absence. The sherds of two cream-slipped, crudely-painted vessels were collected in the earliest Neolithic levels in Chantier E (fig. 14.5-6), but they bear no relationship whatsoever with the two elegant Halaf jarlets brought to light in a similar phase by the Soviet team.

The lithic industry

  • 13 About 3500 pieces of obsidian and flint have been registered so far, 300 of which have been drawn.

35The lithic industry is mostly composed of obsidian artefacts and debris, with a small minority of flint pieces, which were examined during two study seasons in 2015 and 2018. This study has focussed on the assemblages from Chantiers A and E, with the view to covering the complete Neolithic sequence and studying its full technological and typological developments. Although this work is still in progress13 it is possible to describe a few trends.

Primary industry: the flaking technology

36The primary industry is represented by a significant quantity of core fragments, such as core flanks, core splinters, tips, or pieces with flaking platform residues (fig. 16.1-4).

Fig. 16 – The lithic industry

Fig. 16 – The lithic industry

CAD J. Thomalsky

37The latter is often described as “splintered pieces”, and regarded as a specific tool group of thicker flakes with wedge-like working edges. In fact, this category can be described as side-blow blade-flake (e.g., Nishiaki 1996). Here, the blanks were not flaked along the core face (vertically), but horizontally-flaked along the core platform—like core tablets. This method leaves specific traces on the blade flakes, such as residual core-platform parts. The ventral face often shows traces of bi-directional flaking (bipolar), as exemplified by double-bulb or double-flaking scars.

38However, 90% of the obsidian material from Kültepe I may be considered as knapping debris, especially core splinters that include parts of the knapping platform and flaking face. Another group—often regarded as flakes—again comprises core fragments, which display a flaking face with the negative contours of bladelets (fig. 16.5). This pattern indicates a certain percentage of mis-chipping accidents, during which the cores broke into several fragments while being flaked off. This happens mainly when the knapper aims at producing a secondary flaking platform (platform refreshment). Indeed, we observe a certain percentage of platform rejuvenation flakes, such as core tablets. Only a few cores or core residues show a regular conical shape; most others are amorphous pieces with multi-directional flaking patterns. No crested flakes have been found that could suggest a more elaborate core preparation. On the other hand, the knapping debris are characterised by “cape-like” thick conical flakes, which were at first knapped from the core with the aim of producing a core platform. All the first steps of tool production were performed by direct percussion, leaving large platform residues on the basal ends of the flakes and the blades. The fragmented cores obviously result from this knapping method. The following tool modifications are limited to evening the faces from which the blanks were flaked-off. At this point, an indirect hammering-method was used. As regards the size or the dimension of the primary industry, we are faced with a medium-sized blade-oriented technology, since bladelets and related cores are very rare (bladelets per definition have widths < 1 cm). In some cases, the blades exhibit traces of direct percussion—for instance in the form of splintered flaking platform residues on the basal ends.

39All in all, one may presume a high intensity of primary flaking at the site, which left a vast number of different categories of debris. The flaking method is not very standardised, as clearly shown by the high quantity of split core fragments. The cores seem to have been often rejuvenated during the knapping sequence.

40The blades—as the result of this flaking method—all appear fairly dissimilar, often with rather irregular long-edges; the short ends—particularly the distal ends—are often broad and flat or epsilon-shaped. Moreover, in the following blade modification, the irregular parts of the edges were straightened with the help of a micro-burin or simply retouched. The basal face is also often flattened by the removal of the bulb.

41The flint industry displays different characteristics. Though small in number, the technology used for flint entailed the preparation of the core by removing the cortex: a platform would then be created by removing the flakes in a circular pattern. All the working steps are attested in the assemblage; they indicate a complete production on-site. However, no real blade-core has so far been found. Abundant evidence of flint-flaking activity was found in Chantier D, where larger material pieces, amorphous cores and core residues along with flint flakes appear to represent the debris of a flint-knapping workshop. No flint tool was found within this context.

Tools

42The Neolithic tool assemblage is dominated by denticulated and notched pieces, on both flakes and blade flanks. Tools made from blades predominate. In many cases, the long-edges of the rather irregular flaked blades were further evened up by the micro-burin technique. This technique was also used to segment the irregular edges. A similar technique used for edge-straightening has been described for the Neolithic assemblages of the Khabur region (Nishiaki 1990). For the first time, we may confirm the existence of this technique in the South Caucasus as well.

43Another dominating tool group is composed of burins. Flakes and blades alike show burin blows along the edges, mostly along the shorter ends (oblique distal ends), but also along the long sides of the blades. It seems that a large number of blades with burin blows have in fact served as cores for producing the burin spall. However, the thin, needle-like burin-spalls which resulted from this operation, are noticeably under-represented in the whole assemblage.

44Obsidian blades often exhibit traces of heavy use, with traces of abrasion or broad notches visible along the working edges. It is often not easy to distinguish between explicit retouching meant to provide a working edge and usage-induced modifications. But it is clear that the first tool category results from a specific operation whereby the tool was used for smoothing or levelling surfaces, such as leather or bone surfaces. Generally, the visible use-wear suggests leather- and/or bone working. Another tool group that can be assigned to this category are the above-mentioned blades or bladelets with abraded edges, which in the following work sequences were re-sharpened through retouching. The result of this edge modification can most clearly be described as fine, deep, nibbling traces, which at times may have turned into notches (fig. 17.11-12).

Fig. 17 – The lithic industry

Fig. 17 – The lithic industry

45CAD J. Thomalsky

46A distinctive pattern that comes along with the nibbling-retouch technique is an alternating edge modification (fig. 17.13-16). This pattern also develops through specific tool movements during the working processes.

47It is important to note that all the specificities described above (burins, nibbled edges, notches)—may clearly be identified as “Neolithic”; they are common to all the Neolithic assemblages of Southwestern Asia from the 8th millennium BC onwards—whether it be traces of usage or intentional edge-modifications.

  • 14 Thomalsky J. (forthcoming), Obsidian tool production in the South Caucasus of the 5th-4th millenniu (...)

48Denticulated edges, albeit rare, are attested (fig. 17.17-19). In general, blades exhibit irregular edge retouches. Several of these blades show secondary traces of a burin, which clearly indicates the transformation of the blades into tools, which were used to extract burins from the pre-shaped edges (fig. 17.20). Significant exceptions are invasive retouched blades with fairly regular lamellar modifications (fig. 17.21-22). Since the pieces presented here come from the upper levels of Chantier E, it seems that we are faced with an early technological development of the characteristic broad lamellar retouches known, for example, from the Late Chalcolithic layers of Ovçular Tepesi.14

  • 15 Thomalsky J. (forthcoming), Lithic networks in Iran and Caucasus. In: Hansen S. and Lordkipanidze D (...)

49Very few geometric pieces were found next to these major tool types. Microlithic geometrics such as small trapezes or lunates are absent from the Neolithic assemblage, but a few pieces were found in disturbed contexts (fig. 17.5). Interestingly enough, several pieces retrieved from the Neolithic layers belong to a specific group that has been identified as the “Trialetian technological complex” (Kozlowski 1996; Aurenche and Kozlowski 2005), which is composed of large geometric pieces; this complex is deemed to be typical of the Caucasus (fig. 17.6). However, the very existence of this group is questionable since its definition relied on a type of tool whose technological and functional aspects remain elusive.15

50Fairly irregular flakes display a great number of edge retouches that are meant to shape them into large triangular or trapezoid forms. It is not clear yet whether these specimens are real tool types or result from “coincidental” flaking. However, some triangular-shaped pieces—made of chert (!)—exhibit “sickle” gloss along their long edges. Judging by the presence of complete sickle beams, these geometric triangles may have been used as end-pieces placed inside a sickle beam. As to the possibility that these obsidian pieces may belong to composite sickles, we need more use-wear analyses to answer this question. Use-wear analyses are also necessary to trace back the possible use of obsidian blades as sickle inlays.

51Thus, the Neolithic assemblage from Kültepe I appears to be fairly heterogeneous with a wide variety of debris. It is an obsidian-dominated, non-standardised blade industry, whose knapping technology was adapted to the physical properties of obsidian. On the whole, the tool inventory of Kültepe I is fairly restricted: it comprises burins, retouched blades, which includes abraded/smoothed and nibbled edges, as well as notched and denticulated pieces. There are almost no true scraping tools, but the blades with nibbled edges and abrasion traces may nonetheless fall into this category.

The Chalcolithic and Kura-Araxes occupation levels (ca. 3800-2500 cal. BC)

52The Kura-Araxes occupation layers had once been a major part of the tepe, but we soon realised that these layers had been almost entirely removed by the Soviet excavations: scanty remains were nonetheless brought to light in Chantier G in the southern part of the site, together with traces of heavy modern terracing caused by bulldozers: for this reason, it seemed pointless to extend the excavation area any further under the former car park. As for the Chalcolithic period, elusive traces showed that Kültepe I was briefly frequented at the beginning of the 4th millennium BC.

Stratigraphy and Architecture

53Work in Chantier G proved to be useful for describing the transition between the Neolithic and the Kura-Araxes occupation layers, which according to Abibullayev, was marked by a 15 to 40 cm-thick hiatus void of archaeological remains (Abibullayev 1982: 80).

54This was not the case in Chantier G, however, where a Kura-Araxes floor (G-008) that had once possibly belonged to a house was found sitting directly over the filling layer of a Neolithic house. Moreover, two graves were found to the north of this Neolithic house (fig. 18). Judging from a jarlet placed as a funerary gift (fig. 14.3 and 19), one of these graves (G-107), which contained the remains of a juvenile, may belong to the Late Chalcolithic period. A few potsherds pointing to the first half of the 4th millennium BC have also been found in a disturbed context not far from this grave (fig. 14.4).

Fig. 18 – Chantier G, Late Chalcolithic and Kura-Araxes level

Fig. 18 – Chantier G, Late Chalcolithic and Kura-Araxes level

CAD G. Gadebois

Fig. 19 – Chantier G, Late Chalcolithic tomb (G-107)

Fig. 19 – Chantier G, Late Chalcolithic tomb (G-107)

Photo N. Mattana, MBA

  • 16 Şorsu and Zirinçlik are two camp-sites located respectively at the foot and in the piedmonts of the (...)

55These finds are interesting because they show that another occupation period, however brief, should be introduced in the stratigraphy set up by the Soviet team, between their Stratum 1 and Stratum 2: the remains of this period, seemingly only present in a restricted area of the mound, may correspond to the visit of mobile groups, whose existence is otherwise attested in the region at Şorsu and Zirinçlik, where similar pottery has been found.16 But too little is left of the Chalcolithic occupation of Kültepe I, which may have just been restricted to graves, to tell us much about it.

56The situation is not much better as concerns the remaining Kura-Araxes levels, which only yielded a collective grave with three individuals (G-069), a fragmentary stone wall (G-027) and a fragmentary floor (G-008). Judging by the pottery and two 14C readings (LTL16016A and LTL16018A), these features belong to the last quarter of the 4th millennium BC, which indicates the beginning of the Kura-Araxes occupation at Kültepe I.

57Some coring was carried out to the west of the settlement to see whether there was any hope of finding Kura-Araxes layers along the outskirts of the main mound (fig. 2). But the results obtained from the radiocarbon analyses of two charcoal samples retrieved from this core were not coherent: as in Chantier B, the upper sample was dated to an earlier period than the lower sample, an anomaly probably resulting again from terracing activities.

The ceramic and lithic assemblages

  • 17 Personal observation by C. Marro, who has studied the Soviet pottery collection from Kültepe I stor (...)

58The Kura-Araxes pottery assemblage is fairly limited, as it was collected over a small area. However, it comprises the usual set of ceramic vessels attested on Kura-Araxes settlements, namely a majority of black or grey-burnished jars and jarlets with a few simple bowl, andiron and lid fragments (fig. 14.2 and 20). A scraper made from an oval potsherd was found in Tomb G-069. Decorative features are very rare, in comparison to the assemblage of Ovçular Tepesi: a few examples of very simple relief-decorated ware are attested (fig. 14.1), but incised, grooved or impressed decoration has not been found. Dimples, in particular, which are so frequent at Ovçular, are seemingly absent from Kültepe I. This could be explained by the small size of the assemblage retrieved in Chantier G, but the absence of dimple-like decoration is also notable in the ceramics collected by the Soviet team.17

Fig. 20 – Ceramics from the Kura-Araxes period

Fig. 20 – Ceramics from the Kura-Araxes period

1. Locus KT 15-G-018: (a) E: black (lip + upper wall), grey (lower wall), I: black; (b) black, fine mixed-tempered ware; (c) E: highly burnished surface, I: lightly burnished surface; (d) none; (e) none; (f) Early Bronze Age.
2. Locus KT 15-G-018: (a) E: light drab (lip + upper wall), black (lower wall); I: cream; (b) dark/grey bicoloured core; no visible temper but sparse, fine vegetal and mineral inclusions; (c) E+I: highly burnished surface; (d) none; (e) none; (f) Early Bronze Age.
3. Locus KT 15-G-030: (a) E: mottled dark grey and grey; I: grey; (b) dark grey; fine mixed-tempered ware; (c) E+I: highly burnished surface; (d) none; (e) none; (f) Early Bronze Age.
4. Locus KT 15-G-018: (a) E: grey, I: black; (b) black, fine mixed-tempered ware; (c) E: highly burnished surface; I: lightly burnished surface; (d) noe; (e) none; (f) Early Bronze Age.
5. Locus KT 15-G-018: (a) E+I: grey; (b) greyish core, fine mixed-tempered ware; (c) E: lightly burnished surface, I: smoothed; (d) none; (e) none; (f) Early Bronze Age.
6. Locus KT 15-G-018: (a) E: light grey (lip + neck, upper part; grey (neck, lower part + body); I: light grey; (b) dark core, fine mixed-tempered ware; (c) E+I: lightly burnished surface; (d) none; (e) none; (f) Early Bronze Age.
7. Locus KT 15-G-010: (a) E: buff, I: beige; (b) buff, fine mixed-tempered ware; (c) E: lightly burnished surface; I: unevenly burnished surface; (d) none; (e) none; (f) Early Bronze Age.

Fizze Quliyeva; CAD T. Tiriş

59A limited quantity of artefacts pertaining to the lithic industry was retrieved from the Kura-Araxes levels during the 2012-2018 excavations but we may at least say that this assemblage, which is again mainly composed of obsidian, exhibits a certain level of standardisation: its regular blade industry, related cores and knapping debris thus strongly contrast with the heterogeneous character of the Neolithic productions.

Subsistence strategies

The Neolithic

60On account of the differences observed in the sedimentation and the architecture between Level 1 and Level 2, the faunal and the botanical assemblages of each level will be discussed separately. This preliminary study will be focused on the assemblage collected from Chantier E, since this is the only area which encompasses the whole Neolithic sequence.

61According to a preliminary analysis carried out on wood charcoals (Decaix 2016), the inhabitants of Kültepe I settled in an open landscape dominated by juniper trees. The riparian forest was also exploited, possibly along the Naxçıvançay river. It appears that cereals, including naked wheat (Triticum aestivum/durum), barley (Hordeum vulgare) and emmer (Triticum turgidum subsp. dicoccon), were the main cultivated crops, followed by lentil (Lens culinaris): all the taxa that have been identified so far are domesticates, but it should be noted that the botanical study is still in progress.

62The faunal spectrum of the earlier occupation (Level 1) is rather limited with only nine different species identified (table 3). The new settlers almost exclusively exploited domestic sheep and goats. Beside caprines, which represent 85% of the identified specimens, domestic cattle are attested with only 5% of the identified specimens. Among the domestic mammals, dogs are also present. Wild animal remains (wild boar, red deer, fox, hare, and tortoise) are rare, which suggests that their role in Neolithic subsistence strategies was anecdotal. One notable exception is the discovery in Level 1 of a large heap of male deer skulls with unshed antlers (fig. 7, Sub-level 1B). This feature appears to be different from regular consumption refuse because of the absence of post-cranial remains in the heap. This peculiar deposit, whose function is not clear yet, will be analysed in detail elsewhere.

Table 3 – Composition of the faunal assemblage from Level 1 in number of identified remains (NISP)

Table 3 – Composition of the faunal assemblage from Level 1 in number of identified remains (NISP)

6 fragments of deer antler were not included in the table

R. Berthon

63The quantitative importance of domestic mammals, together with the presence of domestic cereals from the earliest occupation onwards, indicate that the groups which first settled at Kültepe I had a full productive economy. This of course raises the question of the genesis of this economy, since no traces of a domestication process are visible: did the first settlers of Kültepe I come from another region together with their Neolithic tool set? Or are we faced with local populations, who somehow managed to acquire domesticated plants and livestock, together with Neolithic agro-pastoral know-how?

  • 18 Berthon R., Giblin J., Balasse M., Fiorillo D. and Bellefroid E. (forthcoming), The role of herding (...)

64The pastoral productions of the first inhabitants of Kültepe I were focused on the exploitation of sheep and goats. Their kill-off pattern suggests a focus on meat exploitation since the animals were mainly slaughtered when they had reached their optimal weight, aged 6 months or older (fig. 21). The absence of animals slaughtered before the age of 6 months, however, is difficult to interpret. In the case of mobile forms of pastoralism for instance, the lambs and the kids would be living in highland pastures, not in lowland settlements, during the spring and the summer seasons: the absence of young animal remains at Kültepe I could also be explained by the practice of vertical pastoralism—we are currently working on this hypothesis.18

Fig. 21 – Kill-off pattern based on sheep and goat teeth from Level 1 (n=17)

Fig. 21 – Kill-off pattern based on sheep and goat teeth from Level 1 (n=17)

CAD R. Berthon

  • 19 See previous footnote.

65A similar agro-pastoral way of life is also attested in Level 2. From the preliminary results of the archaeobotanical analysis, domestic cereals are again the main cultivated crops, followed by lentil (Decaix 2016). Domestic ruminants are still the most frequent animals consumed by the villagers (table 4). The rarity of wild mammals (wild boar, red deer, wild goat, gazelle, fox, bear, mustelid together with birds and tortoise remains), reflect the minor role of wild animal resources in the subsistence strategies. The only difference with Level 1 is the (slight) increase in domestic cattle remains. Although domestic sheep and goats represent 80% of the identified specimens, the ratio of cattle to caprines rose from 1:15 in Level 1 to 1:5 in Level 2. It is interesting to note that this slight increase in cattle remains mirrors the evolution evidenced at Ovçular Tepesi between the two Chalcolithic layers.19 In both cases the transition from semi-subterranean housing to mud-slab or mud-brick architecture is accompanied by an increase in cattle remains. Sheep and goats, however, are still the main sources of animal products in Level 2. In Level 2, caprines were usually slaughtered at an earlier age than in Level 1 (fig. 22), between 6 months and 1 year old. This shows that production was focused on tender meat. This is the case for both species, but it must be said that at the other end of the spectrum, between the age of 4 and 6 years old, more goats than sheep were slaughtered: this could mean that goats were exploited for the production of secondary products such as milk or hair.

Table 4 – Composition of the faunal assemblage from Level 2 in number of identified remains (NISP)

Table 4 – Composition of the faunal assemblage from Level 2 in number of identified remains (NISP)

6 fragments of deer antler were not included in the table

R. Berthon

Fig. 22 – Kill-off pattern based on (A) sheep (Ovis aries, n=26) and (B) goat (Capra hircus, n=29) teeth from Level 2

Fig. 22 – Kill-off pattern based on (A) sheep (Ovis aries, n=26) and (B) goat (Capra hircus, n=29) teeth from Level 2

CAD R. Berthon

Kura-Araxes subsistence strategies

66As stated above, Kura-Araxes occupation layers were heavily disturbed and could only be exposed over a very small area. Only 108 faunal remains were collected by the French-Azerbaijani excavations, of which 28 were identified as far as the specific or the generic level (table 5). This does not give us much information about Kura-Araxes subsistence strategies at Kültepe I, but at least, we may say that the available information does not contradict our current knowledge on Kura-Araxes animal resource exploitation in the region (Davoudi et al. 2018). This is also true for the botanical evidence. The analysis of a few samples collected by the Soviet excavations led to the identification of barley, naked wheat, Fabaceae, goatgrasses, capperbushes, and Panicoideae (Decaix 2016).

Table 5 – Composition of the faunal assemblage from the Kura-Araxes occupation at Kültepe I in number of identified remains (NISP)

Table 5 – Composition of the faunal assemblage from the Kura-Araxes occupation at Kültepe I in number of identified remains (NISP)

R. Berthon

Regional parallels

67The comparison of Kültepe I with other Caucasian sites sheds some new light on two important issues: the advent of a Neolithic way of life in the Caucasus and the inner dynamics of the Kura-Araxes phenomenon.

The Neolithic assemblage

68Kültepe I is one of the rare Caucasian sites, together with Hacı Elamxanlı, to be dated to the end of the 7th millennium BC. Kültepe might even be a bit earlier than Hacı Elamxanlı, since the latter seems to straddle the 7th-6th millennium BC transition. The first Neolithic occupation on most other excavated sites, which belong to the so-called “Shomu-Shulaveris complex”, only starts ca. 5900-5800 cal. BC, as at Aratashen in the Araxes basin, or at Aruchlo, Göytepe, and Mentesh Tepe in the Kura basin. The Neolithic occupation of Göytepe has even been recently reduced to the 5650-5450 cal. BC time span, after applying Bayesian statistics to 14C dates (Nishiaki et al. 2018). As for the Neolithic sites of the Mil steppe, which display marked differences with the Shomu-Shulaveris complex, they seem to have been occupied between 5650 and 5200 BC cal. (Ricci et al. 2018).

  • 20 The stratigraphy of Aknashen was previously divided into five main stages (Badalyan et al. 2010: 18 (...)
  • 21 Haruntyunian A., Badalyan R., Chabot J., Chataigner C., Christidou R. and Hovsepyan R. (2019), The (...)

69An interesting case, however, is that of Aknashen, where an early occupation level labelled Horizon VII has recently been found under a 40 cm-thick layer of sterile soil located below Horizon V, thus revealing a hiatus in the stratigraphic sequence of this site.20 Horizon VII has yielded a small collection of brown-on-beige painted pottery, while the grit-tempered ware typical of the Aratashen-Shomu-Shulaveris group is conspicuous by its near absence (Badalyan and Harutyunian 2014: 166). According to the excavators,21 Horizon VII has been radiocarbon-dated to the very end of the 7th millennium BC, which suggests that Aknashen should in fact probably be added to the Kültepe I/Hacı Elamxanlı group.

70These three sites are certainly of particular interest for examining the processes involved in the adoption of a productive economy in the Caucasus. However, since the details pertaining to Aknashen VII have not been published yet, this analysis will mostly focus on the comparison of Kültepe I with Hacı Elamxanlı.

71Kültepe I and Hacı Elamxanlı are characterised by a number of features that at first sight appear to be broadly similar. Most importantly for the study of Neolithisation processes, for instance, they have both yielded vegetal and animal remains that show a predominance of fully domesticated taxa right from the onset, with an interesting emphasis on caprine herding. The presence of domestic pigs at Hacı Elamxanlı Tepe, albeit rare, is significant since pig herding has been evidenced on the Late Neolithic settlements of the Kura basin, but not in the Araxes basin.

72No traces of a domestication genesis have been found on either site, as if the Neolithic way of life had arrived in the Caucasus as a ready-made package. Since the predominance of caprines in the faunal assemblages recalls the herding profiles attested in the Zagros during the Neolithic (Hesse 1978; Daujat et al. 2016), it is possible that the human groups who settled in the Kura and Araxes basins some time at the turn of the 7th-6th millennia BC actually came from Iran.

  • 22 See previous footnote.

73But the overall evidence is in fact ambivalent, as shown by the pottery assemblages: the bulk of the Kültepe I pottery, for instance, displays few links in shapes, technical processes or decoration with known Neolithic ceramic repertoires from Iran (Adams 1983; Niknami et al. 2013; Mortensen 2014), apart from certain carinated vessels with traces of a red paint (fig. 14.5-6), whose shape and colour recall some of the painted vessels from Hajji Firuz (Voigt 1983: 190-192). The evidence from Hacı Elamxanlı is more difficult to assess, since very few potsherds (21 pieces only) were collected in the excavated area (Nishiaki et al. 2013, 2015), in marked contrast with the large Kültepe assemblage. Moreover, the sherd collection from Hacı Elamxanlı shows a great variability in production techniques, whether it be the preparation of the paste, surface treatment, decoration processes, and firing methods: both chaff and grit-tempered productions are attested, even if mineral-tempered ware is more common. Apart from two brown-on-beige painted potsherds that could be related to Samarrean pottery (Nishiaki et al. 2013: 13 and fig. 19), it is clear that the small ceramic collection of Hacı Elamxanlı does not compare with southern crafting traditions any more than does the Kültepe assemblage. It should be noted that Aknashen VII seems to stand apart from the other two sites here, since the ceramic assemblage from Horizon VII has strong connections with the South, especially with the Mesopotamian world. These connections are all the more interesting since virtually no pottery other than painted ceramics has been found at Aknashen VII.22

74Beyond the similarities in the faunal assemblage and the presence of domesticated species at Kültepe I and Hacı Elamxanlı, significant differences are nonetheless perceptible. It seems wise not to make any comment about the comparative use of cultivated plants, since the study of plant remains from Kültepe I is still in progress. However, if hulled wheat was seemingly the norm at Hacı Elamxanlı (Nishiaki et al. 2015: 20 and table 7; Akashi et al. 2018: table 2), naked wheat was apparently cultivated on a wider basis at Kültepe I, as was the case on many Late Neolithic sites belonging to the Shomu-Shulaveris culture (Arimura et al. 2010: 82). This pattern will be confirmed or invalidated by further analyses.

75As concerns the architecture, it is remarkable that the first occupation layers of Kültepe I (Level 1) are seemingly void of mud-built houses. By contrast, the occupation levels of Hacı Elamxanlı are characterised by circular mud-buildings from the onset. Houses similar to those of Hacı Elamxanlı develop at Kültepe I from Level 2 onwards, but they are not exactly the same: the buildings of Kültepe I are single-roomed, whereas the mud-buildings of Hacı Elamxanlı are composed of two adjoining circular rooms that recall the “snowman-shaped” houses described at Aruchlo, Shulaveris Gora or Imiris Gora in Western Georgia (Kushnareva 1997: fig. 11.2-3 and 5-6; Hansen and Mirtskhulava 2012: fig. 80-81; Nishiaki et al. 2015: fig. 9-10). A few pits and storage bins have also been found at Hacı Elamxanlı (Nishiaki et al. 2015: fig. 9), which are conspicuous by their absence at Kültepe I. A feature that Kültepe I and Hacı Elamxanlı have in common, however, is the presence of caches of small pebbles (ca. 5 cm in length) in different places in the settlement, which were probably used as sling balls (Nishiaki et al. 2013: fig. 9). Hearths are also common inside and outside the buildings on both sites, but those of Kültepe I are paved with pebbles or larger river-stones.

  • 23 See footnote 15.

76Differences between the two sites are also evident in the lithic industry: the lithic tool assemblage is composed of flint and obsidian at Hacı Elamxanlı (Nishiaki et al. 2013: 16; Nishiaki et al. 2015: fig. 13.1-21), whereas it is clearly dominated by obsidian at Kültepe I, as is the case at Late Göytepe or Mentesh Tepe (Guliyev and Nishiaki 2012; Lyonnet et al. 2016). The excavations of Hacı Elamxanlı have yielded a small collection of trapezes (5%), next to round scrapers, blades, and bladelets (Kadowaki et al. 2016: 718 and table 1). By contrast, microlithics are absent from the neolithic assemblage of Kültepe I, although the filling layers sampled from pits, hearths and floors have been systematically wet-sieved: in fact, no links with Mesolithic traditions are perceptible in the Kültepe I repertoire. The obsidian industry from Kültepe I actually finds its best parallels in the assemblages from Göytepe and Aruchlo.23 The knapping method used at Kültepe I indicates a lowly-standardised technology, which focused almost exclusively on the use of obsidian: this produced a high quantity of knapping debris, as well as tools such as notches, burins and blades, but no specific geometric or microlithic component has been found. The full “chaîne opératoire is attested on the site itself, including the chipping waste and the mis-chipping events. We are seemingly faced with an opportunistic and domestic production, which produced chipping waste that was not cleaned up but left in situ.

77Thus, the overall differences between Kültepe I and Hacı Elamxanlı are at least as significant as the similarities. Do these differences reflect a fundamental contrast between the cultural environments of the Kura and the Araxes basins? This is a difficult question since chronologically comparable sites are rare. But if we turn to Late Neolithic settlements, those belonging to the so-called Shomu-Shulevaris complex, it appears that the sites located in the Araxes river basin, namely Aknashen (Horizons V-II), Aratashen (Levels IId-IIa) and Kültepe I (Level 2), are also marked by clear differences between them, whereas they are supposed to be all contemporaneous.

  • 24 Note that no architectural data are so far available for Horizon VII.
  • 25 Please note that the construction techniques described for these two sites simply indicate the use (...)

78At Aratashen for instance, building traditions are characterised by circular mud-buildings from the onset (Level IId), where they appear together with a few ovens located in the courtyards. Circular mud-houses are also attested at Aknashen,24 but the available data suggests the existence of walled enclosures linking several circular buildings together around a domestic compound: this kind of enclosure is also present at Aruchlo between 5800 and 5400 BC, for instance, where mud walls link together “snowman shaped” houses (Hansen and Mirtskhulava 2012: fig. 80-81). At Kültepe I, houses are usually built of “möhre, as in Aknashen and Aratashen,25 but walled enclosures, just like “snowman shaped” houses, seem to be unknown altogether.

79If we turn to the ceramic assemblages, we are faced with the same heterogeneity in almost every aspect of this particular field: the earliest level of Aratashen is mostly devoid of pottery (Arimura et al. 2010: 81), which suggests the existence of a Pre-Pottery period, which clearly does not exist at Kültepe I. When pottery finally develops at Aratashen-Level IIb, the vessels are mainly grit-tempered and remain scarce (Badalyan et al. 2007: 43), whereas the ceramic production of Kültepe I is vegetal-tempered, chaff-faced and just as abundant in Level 2 as in Level 1.

80A similar case may be described for Aknashen: the morphological repertoire of the later levels (Horizon V-II), which mostly comprises simple bowls and cylindrical vessels of various heights, is very similar to that of Kültepe I, but the vessels from Aknashen are grit-tempered (Badalyan et al. 2010: fig. 7 and 9.1.3; Badalyan and Haruntunyan 2014), which again contrasts with the pottery of Kültepe I.

81Heterogeneity is also perceptible in the ceramic industries of the Kura valley: the morphological repertoire from Mentesh Tepe (Period I) for instance, which displays strong similarities with that of Aknashen and Kültepe I, is mostly vegetal-tempered, which matches the repertoire of Kültepe I, but contrasts with the pottery of most other Neolithic sites in the Kura valley. For this reason, the Neolithic assemblage of Mentesh Tepe had first been considered to belong to yet another, unknown, cultural horizon by B. Lyonnet (2017: 141-143) when it was brought to light. As for the repertoire of Aruchlo, which is grit-tempered like the assemblages of most other sites in the Kura valley, it differs from them all through its morphological specificities (jars with a fairly narrow base and a large mouth).

82The variety perceptible in the material assemblages of the Kura and Araxes valleys questions the very concept of a “Shomu-Shulaveris culture”, and even more so of an “Aratashen-Shomu-Shulaveris culture” (Chataigner et al. 2014: 12) currently used in Caucasian archaeology, since in most fields differences appear to be at least as numerous as similarities: what do we actually mean when we refer to the Shomu-Shulaveris culture?

83The heterogeneity of Caucasian Neolithic cultures probably reflects the composite substratum on which the Neolithic “way of life” developed in this mountainous region—a mixture of local hunters-gatherers converted to agriculture and/or herding, together with small groups of migrants from different horizons—but more investigation is needed if we are to break down the interaction patterns underlying the different facets of the Shomu-Shulaveris culture before we can decide if this term is useful at all.

  • 26 The applied decoration attested on a narrow-necked jar from Kültepe I appears to be somewhat differ (...)

84From all these examples, it seems clear that regional parameters alone are not relevant to explain the many differences observed in the ceramic repertoires of the Kura and Araxes valley-sites during the Neolithic. The complexity of the regional dynamics perceptible during the 6th millennium BC in Southwestern Asia may be further illustrated by the occasional presence of a “Caucasian” type of relief decoration, i.e., parallel rows of pellets, on the pottery from three sites of the Tigris and Euphrates basins: Çayönü, Mezraa Teleilat and Yayvan Tepe (Özdogan 2013: fig. 34.2a-c). The presence of applied decoration on Late Neolithic sites located in Upper Mesopotamia clearly testifies to the existence of links between the northern fringes of the Mesopotamian world and the Caucasus, but strangely enough, these links seem to have been established with the Kura, not with the Araxes basin: parallel rows of applied pellets are so far unknown in the latter.26 This may appear somewhat paradoxical, since the Araxes valley lies fairly close to Upper Mesopotamia, whereas the Kura valley is located much further north, nestled between the Lesser Caucasus and the Greater Caucasus mountain ranges. Whatever lies under these interactions, it seems clear that distance and geographic proximity are not among the main criteria that may explain the links perceptible between the Neolithic groups of the two regions.

  • 27 For the Mil steppe, D’Anna 2012: 38-44. The painted pottery from the Mil plain displays strong conn (...)

85Lastly, mention should be made here of the Neolithic sites of the Mil and Mugan steppes: from the available data, these settlements display marked differences in most fields with the sites located in the Kura and Araxes basins: at Kamil Tepe for instance, the presence of mud-brick, multi-cellular and orthogonal buildings, of a lithic industry with a strong flint component, of a few carinated ceramic shapes, as well as clay tokens (D’Anna 2012: fig.49; Helwing and Aliyev 2012: 10, fig. 13; Guilbeau et al. 2017; Helwing et Aliyev 2017: fig. 10), point to the Neolithic “faciès of Hajji Firuz in Northwestern Iran, rather than to the Shomu-Shulaveris culture. This is also true for the painted pottery trends attested both in the Mil and the Mugan steppes.27 Interestingly enough, however, most of the repertoire from Kamil Tepe comprises shapes that are typical of the ceramic assemblages of Kültepe I and Aknashen, while technology-speaking, the pottery from Kamil Tepe is also very comparable to that of Kültepe I (vegetal-tempered chaff-faced pottery that is occasionally cream-slipped and burnished). This is probably why certain scholars have placed the site of Kültepe I within the Mil-Mugan horizon, rather than associate it with the Shomu-Shulaveris culture (Munchaev 1982: 130). This interpretation should be examined in detail, a task which is beyond the scope of the present article; but in any case, this challenging idea should not be discarded too hastily.

The Kura-Araxes assemblage

86The Kura-Araxes pottery from Kültepe I is of special interest, since it does not follow the chromatic evolution expected from Kura-Araxes ceramic assemblages belonging to the 3200-2500 BC period. Drawing on the ceramic repertoires from many sites in Eastern Anatolia and the South Caucasus, a number of scholars expect these assemblages to evolve from a black-burnished (sometimes also grey or brown) into a black-red-burnished repertoire: they posit the appearance of black-burnished vessels with a red, beige or pink-coloured interior sometime towards the end of the 4th millennium BC (Palumbi 2008: 205; Sagona 2018: 256).

  • 28 The presence of black-red-burnished pottery at Kültepe II was further checked by C. Marro in August (...)

87But the pottery from Kültepe I is invariably black or grey-burnished on both sides, a specificity that is also attested in the ceramic repertoire brought to light by the Soviet excavations. The absence of black-red-burnished vessels at Kültepe I is all the more interesting since this chromatic contrast is nonetheless attested in the Kura-Araxes levels of Kültepe II, which is located further north at only 5 km from Kültepe I (Ristvet et al. 2011: 11).28 Moreover, black-red-burnished potsherds have also been found on two sites located to the south of Kültepe I, at Taze Kend and Külün, during the survey we carried out in 2018 in the Naxçıvançay region: Kültepe I, with its invariably monochrome black-or-grey burnished ceramic assemblage, seems to be standing in the middle of black-red-burnished-ware producing sites.

88Thus, it appears from the evidence retrieved in Central Nakhchivan that the absence (or presence) of black-red-burnished ware on a Kura-Araxes site should not be system-atically used as a chronological, or as a regional marker. This fact was first evidenced in 2009 at Ovçular Tepesi in Western Nakhchivan, where the potsherds of a black-red-burnished pithos were found lying over the floor of a Late Chalcolithic house dated to the end of the 5th millennium BC (Marro et al. 2014, 2015). The monochrome repertoire of Kültepe I is probably linked to the Kura-Araxes traditions of Northwestern Iran, where the black-red chromatic contrast is seemingly unknown (Summers 1982: 45 sq.), but the fact that this settlement stands in the middle of other Kura-Araxes sites where black-red-burnished-ware is commonly used strongly suggests that the black-red chromatic contrast should rather be interpreted as a marker of identity among Kura-Araxes groups: if Kura-Araxes groups from Iran migrated to Nakhchivan ca. 3200 BC (if this was indeed the case), they did not change their ways of producing pottery once they settled in Kültepe I, although they went on living next to black-red-burnished-ware producing villages for the next 800 years. This suggests that the black-red colour contrast (or absence thereof) is a key-element in the social construction of Kura-Araxes groups; and not the mere reflection of fashion, which is bound to change through time and influence.

89Rather than a regional or a chronological marker, we therefore posit that the black-red colour contrast, but also other aesthetic specificities such as the presence of dimples on one site, and relief decoration on the other, relate to socio-symbolical structures or the cosmogony of the Kura-Araxes world, about which we still know very little.

Conclusions

90The information retrieved from Kültepe I by the 2012-2018 excavations have produced a wealth of new data that will first and foremost contribute to our understanding of the Neolithic period. The data obtained on the inner dynamics of the Kura-Araxes phenomenon are certainly significant, but they can only be interpreted in the light of the material exhumed by the Soviet team, which will be done in a second step. Judging by the comparison of the ceramic assemblage of Kültepe I with those of surrounding Kura-Araxes sites, we surmise that the community which settled at Kültepe I in the last third of the 4th millennium BC came from Northwestern Iran. Whatever the origin of this group, it is clear that it maintained a specific ceramic repertoire, different from the pottery traditions of its immediate neighbours, throughout the Early Bronze Age.

91The new data obtained on the Neolithic period, on the other hand, are particularly important since they cast a new light on the formation of the Neolithic period in the Caucasus: Kültepe I now appears as one of the earliest Caucasian Neolithic settlements known to date (if not the earliest). The fact that this site is located near the Urmiah region, which opens onto the Mesopotamian world via the Solduz valley (Voigt 1983: 324), is probably not a coincidence: if the impetus that set the Neolithisation processes into motion in the Caucasus originally came from Iran—or else, from Mesopotamia—the earliest Caucasian sites to display a Neolithic “faciès should indeed be located in the middle Araxes valley. But the available data remain ambivalent at best.

92In any case, it is now clear that the adoption of a productive economy in the Caucasus was not a uniform trend: if we compare the settlements of the Araxes with those of the Kura basin, especially Kültepe I and Hacı Elamxanlı, the formation of the Caucasian Neolithic clearly stands as a multi-faceted phenomenon. On the one side, a link with the Middle East in undeniable, through the rare, but repeated, occurrence of Mesopotamian or Iranian-related painted pottery on several sites of both the Araxes and the Kura basins. A further testimony of this link is given by the constant presence of cream-slipped pottery in the earlier Neolithic levels of Kültepe I (Level 1), which could be interpreted as some imitation of the light-coloured wares frequent in Late Neolithic Mesopotamia. The marked prevalence of caprines in the faunal assemblages of both Kültepe I and Hacı Elamxanlı is also a trait that may be compared with the specialised herding strategies of the pastoral groups living in the Zagros.

93But all these links remain elusive and do not account for the striking differences attested between the earliest Neolithic settlements so far known in the South Caucasus. The comparison of the ceramic repertoires of Kültepe and Hacı Elamxanlı is a good illustration of this puzzle: how can we explain the virtual absence of pottery at Hacı Elamxanlı set against the profusion of ceramics found at Kültepe I in the same period?

94It is clear from this example that the trajectory of the Early Neolithic groups that settled in the South Caucasus between 6200 and 5900 BC differs from one case to the other; and this is also indicated by the different types of painted pottery attested on these sites: if the painted jarlets found by the Soviet team at Kültepe I clearly point to the Halaf horizon, those attested at Aknashen (Horizon VII) and Hacı Elamxanlı are closer to the Samarra-Hassuna repertoire (Nishiaki et al. 2013: 13; Badalyan and Harutyunyan 2014: fig. 8, p. 175), whereas the colour of the red-painted vessels found in Chantier E at Kültepe I recalls the painted ceramic productions from Hajji Firuz in Iran. Apart from their light-coloured paste, one should note that all these painted-pottery trends have altogether very little in common, whether it be in shape or decoration.

95Thus, we surmise the existence of separate, irregular and multi-form relationships between South Caucasian and Near/Middle Eastern groups in the early steps of the Caucasian neolithisation process. These relationships may at times have been paralleled by a few migration episodes of peoples from Mesopotamia or Iran. Migrants from the South may indeed have been responsible for the diffusion of herding north of the Araxes, as suggested by the practice of specialised husbandry at both Kültepe I and Hacı Elamxanlı, which is reminiscent of the subsistence patterns attested in the Zagros. This possibility remains to be demonstrated, but the case of Aknashen VII and its all-Mesopotamian painted-ware assemblage could be an indication that small groups of people had migrated from the Middle East to the Caucasus by the end of the 7th millennium BC. At any rate, the discrepancy between the cultural assemblages of excavated South Caucasian Neolithic sites works against any hypothesis positing a single impulse, let alone a single factor, that would account for the rise of the Neolithic way of life in the Caucasus.

96The relative heterogeneity of the cultural context in which a Neolithic way of life developed in the Caucasus is perceptible throughout the 6th millennium BC, as evidenced by the Late Neolithic settlements of the so-called Shomu-Shulaveris culture: this cultural complex seems to be much less homogeneous than what is generally implied by the use of this label. A complete reassessment of known Caucasian Neolithic cultures, including the outliers located in the Mil and the Mugan steppes, is needed before we may break down the dynamics underlying the formation of the Shomu-Shulaveris culture, and decide what is exactly meant by this term. This should be a collective work, involving all the archaeological teams presently working in the Caucasus.

Haut de page

Bibliographie

Abedi A., Shahidi H. K., Chataigner C., Eskandari N., Kazempour M., Niknami K., Piromohammadi A., Hosseinzadeh J. and Ebrahimi G.
2014 – Excavation at Kul Tepe (Hadishahr), North-West Iran, 2010: First preliminary report, Ancient Near Eastern Studies 51: 33-167.

Abibullayev O.A
1959 – Kultəpədə arxeolozhi gazyntylar [Archaeological excavations at Kültepe]. Baku: Elm (in Azeri Turkish).

Abibullayev O.A
1982 – Eneolit i Bronza na territorii Nakhichevanskoy ASSR [The Eneolithic and Bronze Ages within the territory of the Nakhchivan Autonomous SSR]. Baku: Elm (in Russian).

Adams R.M.
1983 – The Jarmo stone and pottery vessel industries. In: Braidwood L.S., Braidwood R.J., Howe B., Reed C.A. and Watson P.J. (eds.), Prehistoric Archaeology along the Zagros flanks: 209-232. Chicago: University of Chicago Press (Oriental Institute Publications 105).

Akashi C., Tanno F., Goluyev F and Nishiaki Y
2018 – Neolithisation processes of the South Caucasus: As viewed from macro-botanical analyses at Hacı Elamxanlı Tepe, West Azerbaijan. Paléorient 44,2: 75-89.

Alekbarov V.A.
2018 – Technological development of the Neolithic pottery at Göytepe (West Azerbaijan). Archaeology, ethnology and anthropology of Eurasia 46,3: 22-31.

Arimura M., Badalyan R., Gasparyan B. and Chataigner C.
2010 – Current Neolithic research in Armenia. Neo-Lithics 1/10: 77-85.

Aurenche O. and Kozlowski S.K.
2005 – Territories, boundaries and cultures in the Neolithic Near East. Oxford: Archaeopress (BAR Int. Ser. 1362).

Badalyan R. and Harutyunyan A.
2014 – Aknashen – The Late Neolithic settlement of the Ararat valley: Main results and prospects for the research. In: Gasparyan B. and Arimura M. (eds.), Stone Age of Armenia. A guide-book to the Stone Age archaeology in the Republic of Armenia: 161-176. Kazanawa: Kanazawa University.

Badalyan R., Harutyunyan A., Chataigner C., Le Mort F., Chabot J., Brochier J.E., Bălăşescu A., Radu V. and Hovsepyan R.
2010 – The settlement of Aknashen-Khatunarkh, a Neolithic site in the Ararat Plain (Armenia): Excavation results 2004-2009. TÜBA-AR 13: 185-218.

Badalyan R., Lombard P., Avetisyan P., Chataigner C., Chabot J., Vila E., Hovsepyan R., Willcox G. and Pessin H.
2007 – New data on the late prehistory of the Southern Caucasus. The excavations at Aratashen (Armenia): Preliminary report. In: Lyonnet B. (ed.), Les cultures du Caucase (VI-IIIe millénaires av. n.è.). Leurs relations avec le Proche-Orient: 37-62. Paris: CNRS Éditions.

Bronk Ramsey C.
2009 – Bayesian analysis of radiocarbon dates. Radiocarbon 51,1: 337-360.

Burney C. and Lang D.M.
1971 – The peoples of the hills. Ancient Ararat and Caucasus. London: Weidenfeld and Nicolson.

Chataigner C., Badalyan R. and Arimura M.
2014 – The Neolithic of the Caucasus. Oxford Handbooks online [DOI: 10.1093/oxfordhb/9780199935413.013.13].

Cucchi T., Kovács Z.E., Berthon R., Orth A., Bonhomme F., Evin A., Siahsarvie R., Darvish J., Bakhshaliyev V. and Marro C.
2013 – On the trail of Neolithic mice and men towards Transcaucasia: Zooarchaeological clues from Nakhchivan (Azerbaijan). Biological Journal of the Linnean Society 108,4: 917-928.

D’Anna M.B.
2012 – The pottery of Kamil Tepe (MPS 1). AMIT 44: 39-47.

Daujat J., Mashkour M., Emery-Barbier A., Neef R. and Bernbeck R.
2016 – Qale Rostam: Reconsidering the “rise of a highland way of life”. An integrated bioarchaeological analysis. In: Roustaei K. and Mashkour M. (eds.), The Neolithic of the Iranian plateau. Recent research: 107-136. Berlin: ex Oriente (SENEPSE 18).

Davoudi H., Berthon R., Mohaseb A., Sheikhi S., Abedi A. and Mashkour M.
2018 – Kura-Araxes exploitation of animal resources in Northwestern Iran and Nakhchivan. In: Çakırlar C., Chahoud J., Berthon R. and Pilaar Birch S. (eds.), Archaeozoology of the Near East XII. Proceedings of the 12th international symposium of the ICAZ archaeozoology of Southwest Asia and adjacent areas working group, Groningen, 14th-15th June 2015: 91-108. Eelde: Barkhuis; Groningen: University of Groningen.

Decaix A.
2016 – Origine et évolution des économies agricoles dans le Sud du Caucase. Unpublished PhD thesis. Paris: Université Paris 1 Panthéon-Sorbonne.

Garelli P.
1969 – Le Proche-Orient asiatique : des origines aux invasions des peuples de la mer. Paris: Presses universitaires de France.

Guilbeau D., Astruc L. and Samzun A.
2017 – Chipped stone industries from the Mil Plain (Kamiltepe) and the Middle Kura Valley. In: Helwing B., Aliyev T., Lyonnet B., Guliyev F., Hansen S. and Mirtkhulava G. (eds.), The Kura Projects: New research on the Later Prehistory of the Southern Caucasus: 385-398. Berlin: Dietrich Reimer Verlag (Archäologie in Iran und Turan 16).

Guliyev F. and Nishiaki Y.
2012 – Excavations at the Neolithic settlement of Göytepe, the Middle Kura Valley, Azerbaijan, 2008-2009. In: Matthews R. and Curtis J. (eds.), Proceedings of the 7th ICAANE, London,12th April-16th April 2010: 71-84. Wiesbaden: Harrassowitz Verlag.

Guliyev F. and Nishiaki Y.
2014 – Excavations at the Neolithic settlement of Göytepe, West Azerbaijan, 2010-2011. In: Bielinski P., Gawlikowski M., Kolinski R., Lawecka D., Soltysiak A. and Wygnanska Z. (eds.), Proceedings of the 8th ICAANE, Warsaw, 30th April-4th May 2012. Vol. II: Excavation and progress reports. Posters: 3-16. Wiesbaden: Harrassowitz Verlag.

Hamon C.
2016 – Salt mining tools and techniques from Duzdağı (Nakhchivan, Azerbaijan) in the 5th to 3rd millennium BC. Journal of Field Archaeology 41,4: 510-528.

Hamon C., Jalabadze, M., Agapishvili, T., Baudouin E., Koridze I. and Messager E.
2016 – Gadachrili Gora: Architecture and organisation of a Neolithic settlement in the Middle Kura Valley (6th millennium BC, Georgia). Quaternary International 395: 154-169.

Hansen S. and Mirtskhulava G.
2012 – The Neolithic settlement of Aruchlo. Report on the excavations in 2009-2011. AMIT 44: 58-86.

Helwing B. and Aliyev T.
2012 – Fieldwork in the Mil plain: The 2010-2011 expedition. AMIT 44: 4-17.

Helwing B. and Aliyev T.
2017 – Excavations in the Mil Plain sites, 2012-2014. In: Helwing B. Aliyev T., Lyonnet B., Guliyev F., Hansen S. and Mirtkhulava G. (eds.), The Kura Projects: New research on the Later Prehistory of the Southern Caucasus: 11-42. Berlin: Dietrich Reimer Verlag (Archäologie in Iran und Turan 16).

Hesse B.C.
1978 – Evidence for husbandry from the Early Neolithic site of Ganj Dareh in Western Iran. Unpublished PhD thesis. New York: Columbia University.

Kadowaki S. Guliyev F. and Nishiaki Y.
2016 – Chipped stone technology of the earliest agricultural village in the southern Caucasus: Hacı Elamxanlı Tepe (the beginning of the 6th millennium BC). In: Stucky R.A., Kaelin O. and Mathys H.-P. (eds.), Proceedings of the 9th ICAANE, Basel, 9th-13th July, vol. 3: 709-722. Wiesbaden: Harrassowitz Verlag.

Kozlowski S.K.
1996 – The Trialetian “Mesolithic” industry of the Caucasus, Transcaspia, Eastern Anatolia, and the Iranian Plateau. In: Kozlowski S.K. and Gebel H.G.K. (eds.), Neolithic chipped stone industries of the Fertile Crescent, and their contemporaries in adjacent regions: 161-174. Berlin: ex Oriente (SENEPSE 3).

Kushnareva K.K.
1997 – The Southern Caucasus in Prehistory. Stages of cultural and socio-economic development from the eighth to the second millennium BC. Philadelphia: The University Museum.

Lyonnet B., Guliyev F., Bouquet L., Bruley-Chabot G., Samzun A., Pecqueur L., Jovenet E., Baudouin E., Fontugne M., Raymond P., Degorre E., Astruc L., Guilbeau D., Le Dosseu G., Benecke N., Hamon C., Poulmarc’h M. and Courcier A.
2016 – Mentesh Tepe, an early settlement of the Shomu-Shulaveri culture in Azerbaijan. Quaternary International 395: 170-183.

Lyonnet B.
2017 – Mentesh Tepe 2012-2014: The pottery. In: Helwing B., Aliyev T., Lyonnet B., Guliyev F., Hansen S. and Mirtkhulava G. (eds.), The Kura Projects: New research on the Later Prehistory of the Southern Caucasus: 141-151. Berlin: Dietrich Reimer Verlag (Archäologie in Iran und Turan 16).

Marro C., Bakhshaliyev V. and Ashurov S.
2009 – Excavations at Ovçular Tepesi (Nakhchivan, Azerbaijan). First preliminary report: the 2006-2008 seasons. Anatolia Antiqua 17: 31-87.

Marro C., Bakhshaliyev V. and Ashurov S.
2011 – Excavations at Ovçular Tepesi (Nakhchivan, Azerbaijan). Second preliminary report: the 2009-2010 seasons. Anatolia Antiqua 19: 53-100.

Marro C., Bakhshaliyev V. and Berthon R.
2014 – On the genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan). Paléorient 40,2: 131-154.

Marro C., Bakhshaliyev V. and Berthon R.
2015 – A reply to G. Palumbi and C. Chataigner. Paléorient 41,2: 157-162.

Mortensen P.
2014 – Excavations at Tepe Guran: The Neolithic Period. Leuwen: Peeters (Acta Iranica 55).

Munchaev R.M.
1982 – Pamiatniki kul’tury eneolita Kavkaza. In: Masson V.M. and Merpert N.I. (eds), Arkheologiia SSSR: Eneolit SSSR: 100-164. Moscow: Nauka.

Niknami K.A, Niksad M. and Alibaigi S.
2013 – Neolithic settlement patterns of the Sarfirouz Abad Plain, Central West Zagros. In: Matthews R. and Fazeli Nashli H. (eds.), The Neolithisation of Iran. The formation of new societies: 35-47. Oxford: Oxbow Books (Themes from the Ancient Near East BANEA Publication Series 3).

Nishiaki Y.
1990 – Corner-thinned blades: A new obsidian tool type from a Pottery Neolithic mound in the Khabur basin, Syria. BASOR 280: 5-14.

Nishiaki Y.
1996 – Side-blow blade-flakes from Tell Kashkashok II, Syria: A technological study. In: Kozlowski S.K. and Gebel H.G.K. (eds.), Neolithic chipped stone industries of the Fertile Crescent and their contemporaries in adjacent regions. Proceedings of the 2nd workshop on PPN chipped lithic industries, Warsaw, 3rd-7th April 1995: 311-325. Berlin: ex Oriente (SENEPSE 3).

Nishiaki Y., Guliyev F., Kadowaki S., Arimatsu M., Hayakawa Y., Shimogama K., Miki T., Akashi C., Arai S. and Salimbeyov S.
2013 – Hacı Elamxanlı Tepe: Excavations of the earliest Pottery Neolithic occupations on the Middle Kura, Azerbaijan, 2012. AMIT 45: 1-25.

Nishiaki Y., Guliyev F., Kadowaki S., Alakbarov V., Takehiro M., Salimbayov S., Akashi C. and Arai S.
2015 – Investigating cultural and socioeconomic change at the beginning of the Pottery Neolithic in the Southern Caucasus: The 2013 Excavations at Hacı Elamxanlı Tepe, Azerbaijan. BASOR 374: 1-28.

Nishiaki Y., Guliyev F., Kadowaki S. and Omori T.
2018 – Neolithic residential patterns in the southern Caucasus: Radiocarbon analysis of rebuilding cycles of mudbrick architecture at Göytepe, West Azerbaijan. Quaternary International 474,B: 119-130.

Özdogan M.
2013 – Reconsidering the Late Neolithic period in South-Eastern Turkey: A regional perspective. In: Nieuwenhuyse O., Bernbeck R., Akkermans P. and Rogasch J. (eds.), Interpreting the Late Neolithic of Upper-Mesopotamia: 377-386. Turnhout: Brepols (Papers on Archaeology from the Leiden Museum of Antiquities 9).

Palumbi G.
2008 – The red and black. Social and cultural interaction between the Upper Euphrates and Southern Caucasus communities in the fourth and third millennium BC. Roma: Università di Roma “La Sapienza” (Studi di preistoria orientale 2).

Payne S.
1973 – Kill-off patterns in sheep and goats: The mandibles from Aşvan Kale. Anatolian Studies 23: 281-303.

Reimer P.J., Bard E., Bayliss A., Beck J.W., Blackwell P.G., Bronk Ramsey C., Buck C.E., Cheng H., Edwards R.L., Friedrich M., Grootes P.M., Guilderson T.P., Haflidason H., Hajdas I., Hatté C., Heaton T.J., Hoffman D.L., Hogg A.G., Hughen K.A., Kaiser K.F., Kromer B., Manning S.W., Niu M., Reimer R.W., Richards D.A., Scott E.M., Southon J.R., Staff R.A., Turney C.S.M. and van der Plicht J.
2013 – IntCal13 and Marine13 radiocarbon age calibration curves 0-50,000 years cal BP. Radiocarbon 55,4: 1869-1887.

Ricci A., D’Anna M.B., Lawrence D., Helwing B. and Aliyev T.
2018 – Human mobility and early sedentism: the Late Neolothic landscape of southern Azerbaijan. Antiquity 92,366: 1445-1461.

Ristvet L., Baxşəliyev V. and Aşurov S.
2011 – Settlement and society in Naxçıvan: 2006 excavations and survey of the Naxçıvan archaeological project. Iranica Antiqua 46: 1-53.

Sagona A.G.
2018 – The archaeology of the Caucasus. From earliest settlement to the Iron Age. Cambridge: Cambridge University Press (Cambridge World Archaeology).

Seyidov A.
2003 – Nаxçıvan e. ə. VII-II minilliklərdə (Nakhchivan between the 7th and the 2nd millennium BC). Elm: Baкı (in Azeri Turkish).

Summers G.D.
1982 – A study of the architecture, pottery and other material from Yanik Tepe, Haftavan Tepe VIII and related sites. Unpublished PhD thesis. Manchester: University of Manchester.

Voigt M.
1983 – Hajji Firuz Tepe, Iran. The Neolithic settlement. Hasanlu excavations reports I. Philadelphia: University Museum (University Museum Monograph 50).

Haut de page

Notes

1 A few caravansarays located in the villages bordering the foothills, as at Qarabağlar, testify to the existence of this route. All these buildings are now in ruins.

2 Bakhshaliyev V. (forthcoming), Archaeological excavations at Nakhchivan Tepe. In: Marro C. and Stöllner T. (eds), On salt, copper and gold: The beginning of early mining and metallurgy in the Caucasus. Proceedings of the international conference of Tbilisi, 16th-19th June 2016. Lyon: Maison de l’Orient et de la Méditerranée.

3 This work was part of a research programme funded by The Shelby White and Leon Levy Program for Archaeological Publications (PI: R. Berthon), aiming at reassessing the results of the 1951-1964 excavations. This task entailed the restudy of the archaeological material stored in the History Museum in Baku, and the re-opening of former Soviet trenches. Work is still in progress and will be published in a separate volume.

4 This work was supported by the French Ministry of Foreign Affairs (“Mission archéologique du bassin de l’Araxe”, dir. C. Marro) and the French National Science Foundation (ANR) within the framework of the “Mines” project (ANR FRAAL-0012). We would like here to thank the Azerbaijani National Academy of Sciences, and especially I. Haciyev for his constant help and support.

5 Or “Eneolithic” in the Russian terminology.

6 Abibullayev’s second report was published in 1982 but since he died in 1970, this report was presumably written in the late 1960s.

7 Uzdurum M. (2013), Aşıklı Höyük yerleşmesinde ateş yerleri ve kullanımı [Fire places and their function at Aşıklı Höyük]. Unpublished Master thesis. Istanbul: University of Istanbul (see p. 44 and 127 sq.).

8 As suggested by the size of the hearth (1 m × 0,8 m preserved over a height of 0.8 m) and the presence of potsherds belonging to “crudely modelled vessels”.

9 See table 1: LTL16015A, 6473±45 BP, 5520-5330 cal. BC 2σ.

10 Chips have not been included in this count. A potsherd is considered to be a chip when at least one of the wall surfaces is missing.

11 This refers to the filling layers located between 943.2 m and 944.8 m asl., in the northern half of the deep sounding of Chantier E, which was excavated in 2018. The southern half was excavated in 2016: for this reason, the number given to the loci does not always follow the order of the stratigraphic units (see table 2).

12 See table 1; please note that LTL16905A and LTL18620A have been excluded from the dating of Level 1, since preservation issues have cast doubt on their reliability. A new set of analyses will be carried out on the charcoal sampled from Locus E-264 before we may validate these two readings.

13 About 3500 pieces of obsidian and flint have been registered so far, 300 of which have been drawn.

14 Thomalsky J. (forthcoming), Obsidian tool production in the South Caucasus of the 5th-4th millennium BC: Technology, typology and socio-cultural implications. In: Marro C. and Stöllner T. (eds)On salt, copper & gold: The beginning of early mining and metallurgy in the Caucasus. Proceedings of the international conference of Tbilisi, 16th-19th June 2016. Lyon: Maison de l’Orient et de la Méditerranée.

15 Thomalsky J. (forthcoming), Lithic networks in Iran and Caucasus. In: Hansen S. and Lordkipanidze D. (eds.), New research on the Neolithic in the circumcaspian regions. Proceedings of the conference DAI of Tbilisi, 25th September-1st October 2011.

16 Şorsu and Zirinçlik are two camp-sites located respectively at the foot and in the piedmonts of the Sirab mountains. Zirinçlik was excavated in 2014 under the responsibility of N. Gailhard (post-doc, CNRS UMR 5133), and Şorsu in 2014-2017, under the responsibility of S. Sarıaltun (doctoral student, University of Paris Nanterre).

17 Personal observation by C. Marro, who has studied the Soviet pottery collection from Kültepe I stored in the History Museum of Baku.

18 Berthon R., Giblin J., Balasse M., Fiorillo D. and Bellefroid E. (forthcoming), The role of herding strategies in the exploitation of natural resources by early mining communities in the Caucasus. In: Marro C. and Stöllner T. (eds)On salt, copper & gold: The beginning of early mining and metallurgy in the Caucasus. Proceedings of the international conference of Tbilisi, 16th-19th June 2016. Lyon: Maison de l’Orient et de la Méditerranée.

19 See previous footnote.

20 The stratigraphy of Aknashen was previously divided into five main stages (Badalyan et al. 2010: 188-189): Horizon I (Chalcolithic) and Horizons II-V (Neolithic). Horizon V was itself divided into a number of sub-levels (V1-V5). The stratigraphy of Aknashen has been partly re-labelled after the unexpected discovery of this early occupation level in a 1 m × 1 m sounding opened in 2013: Horizon V-3 became Horizon VI; hence Horizon VII probably refers to former sub-levels V-4 and V-5, although this is not clearly specified (Badalyan and Haruntyunian 2014: 165).

21 Haruntyunian A., Badalyan R., Chabot J., Chataigner C., Christidou R. and Hovsepyan R. (2019), The first farmers of the Araks valley: The formative stage of the “Aratashen-Shulaveri-Shomutepe” culture. Paper given at the international conference held in Lyon, 14th-15th May 2019: The Araxes river in Late Prehistory: Bridge or border? (dir. A. Abedi and C. Marro).

22 See previous footnote.

23 See footnote 15.

24 Note that no architectural data are so far available for Horizon VII.

25 Please note that the construction techniques described for these two sites simply indicate the use of “cob” (Chataigner et al. 2014: 15), which is not necessarily the same as “möhre”: the latter entails the shaping of cob into large bricks on the wall itself—“möhre” is the Azeri Turkish equivalent of the “bauge technique”.

26 The applied decoration attested on a narrow-necked jar from Kültepe I appears to be somewhat different from the examples attested at Aruchlo or Göytepe, for instance, since we are faced with two rows of knobs rather than pellets; moreover, it must be emphasised that this jar is a unique case at Kültepe I (see fig. 15.1).

27 For the Mil steppe, D’Anna 2012: 38-44. The painted pottery from the Mil plain displays strong connections with that of Alikemek Tepesi in the Mugan steppe (pers. observations by C. Marro, who has had the opportunity to examine the ceramics from Alikemek Tepesi stored at the Institute of Archaeology and Ethnography in Baku in 2005. C. Marro would like to thank T. Akhundov for showing her the pottery from Alikemek Tepesi).

28 The presence of black-red-burnished pottery at Kültepe II was further checked by C. Marro in August 2018 in the framework of a survey carried out at Kültepe II as part of the PAST-OBS programme (dir. F.-X. Le Bourdonnec, University of Bordeaux-Montaigne).

Haut de page

Table des illustrations

Titre Fig. 1 – Regional map with sites mentioned in the text
Crédits Map O. Barge and C. Marro
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-1.jpg
Fichier image/jpeg, 898k
Titre Table 1 – Radiocarbon dates, presented in stratigraphic order for each excavation area
Légende The dates were calibrated with the OxCal programme (Bronk Ramsey 2009), using the calibration curve IntCal13 (Reimer et al. 2013)
Crédits R. Berthon
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-2.jpg
Fichier image/jpeg, 959k
Titre Fig. 2 – 2012-2018 excavation areas at Kültepe I
Crédits CAD P. Lebouteiller and G. Gadebois
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-3.jpg
Fichier image/jpeg, 650k
Titre Fig. 3 – Location of former Soviet “Area 5”
Crédits Photo C. Marro, MBA
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-4.jpg
Fichier image/jpeg, 810k
Titre Fig. 4 – Chantier E, deep sounding (north section)
Crédits CAD G. Gadebois
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-5.jpg
Fichier image/jpeg, 485k
Titre Fig. 5 – Chantier E, deep sounding (South section)
Crédits CAD G. Gadebois
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-6.jpg
Fichier image/jpeg, 595k
Titre Fig. 6 – Chantier E, deep sounding (East section)
Crédits CAD G. Gadebois
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-7.jpg
Fichier image/jpeg, 469k
Titre Fig. 7 – Chantier E, with Sub-level 2B (main square), Sub-level 2A and Sub-level 1B (deep sounding)
Crédits CAD G. Gadebois
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-8.jpg
Fichier image/jpeg, 347k
Titre Fig. 8 – Chantier E, deep sounding (south-western angle), Neolithic Level 1
Légende Washed-down wall of a semi-subterranean dwelling carved into the yellowish silt (E-294)
Crédits Photo C. Marro, MBA
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-9.jpg
Fichier image/jpeg, 1,5M
Titre Fig. 9 – Chantier E, deep sounding, Neolithic Level 1
Légende Post-holes (E-358 and E-360) are visible in the north section
Crédits Photo C. Marro, MBA
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-10.jpg
Fichier image/jpeg, 541k
Titre Fig. 10 – Chantier E, deep sounding, Neolithic Level 1
Légende Fire-pit (E-270)
Crédits Photo C. Marro, MBA
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-11.jpg
Fichier image/jpeg, 817k
Titre Fig. 11 – Chantier A, large hearth (A-167) lined with river-stones, Neolithic Level 2
Crédits Photo C. Marro, MBA
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-12.jpg
Fichier image/jpeg, 1,1M
Titre Fig. 12 – Storage jar (E-312) from House E-242, Neolithic Level 2
Crédits Photo N. Gailhard, MBA
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-13.jpg
Fichier image/jpeg, 502k
Titre Fig. 13 – Plain ceramics from the Neolithic period
Légende (a) Vessel colour; (b) Paste colour and temper; (c) Surface treatment; (d) Decoration if any; (e) Specific observations if any (f) Periodisation/stratigraphic attribution. E: exterior surface; I: inside surface1. Locus KT 15-F-085: (a) E: mottled beige and buff; I: beige (b) buff, medium straw-tempered; (c) E: plain, I: slightly smoothed; both sides chaff-faced; (d) none; (e) heterogenous paste from poor clay-kneading; (f) Neolithic, Level 2.2. Locus KT 15-F-031: (a) E: buff; I: buff; (b) black core; medium straw-tempered with a few mineral inclusions (e.g., quartz); (c) E: buff, evenly burnished; I: unevenly burnished (upper body) or smoothed (lower body); both sides chaff-faced; (d) none; (e) none; (f) Neolithic, Level 2.3. Locus KT 14-A-040: (a) E+I: light buff; (b) buff with grey core, medium straw-tempered with mineral inclusions (c) E+I unevenly burnished; matt burnishing; both sides chaff-faced; (d) none; (e) heterogenous paste from poor clay-kneading; (f) Neolithic, Level 2.4. Locus KT 14-A-090: (a) E: cream, I: buff; (b) darkish core, medium straw-tempered; (c) E: cream wash, evenly burnished, matt burnishing; I: unevenly burnished; both sides chaff-faced; (d) none; (e) none; (f) Neolithic, Level 2.5. Locus KT 14-A-077: (a) E: mottled cream and buff; I: light buff; (b) brown, medium straw-tempered with a few mineral inclusions; (c) E: cream slipped; E+I: light burnishing? Matt surface. Both sides chaff-faced; (d) none; (e) none; (f) Neolithic, Level 2.6. Locus KT 15-A-161: (a) E: mottled cream and buff; I: beige (neck) and drab (body); both sides chaff-faced; (b) dark core; medium mixed-tempered ware with a few coarse grits; (c) E+I (neck): unevenly cream- slipped, unevenly burnished, matt surface. I (body): drab, smoothed. (d) none; (e) Vessel shaped with a bat; (f) Neolithic, Level 2.7. Locus KT 17-E-312: (a) E+I: buff; (b) buff, medium straw-tempered; (c) E: unevenly burnished, chaff-faced; I: (worn out surface); (d) none; (e) use of coil technique visible through a shallow groove; (f) Neolithic, Level 2.
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-14.jpg
Fichier image/jpeg, 491k
Titre Fig. 14 – Ceramics from various periods (1-2) Kura-Araxes; (3-4) Late Chalcolithic; (5-6) Neolithic Level 1
Légende 1. Locus KT 15-G-030: (a) E: black (upper wall) and grey (lower wall); I: grey; (b) dark core, fine mixed-tempered ware; (c) E: lightly burnished, I:  smoothed; (d) capsule-shaped relief decoration; (e) none; (f) Early Bronze Age.2. Locus KT 15-G-008: (a) E+I: cream; (b) buff, fine vegetal-tempered ware; (c) E+I: cream-slipped; (d) anthropomorphic face in relief over upper end; (e) none; (f) Early Bronze Age.3. Locus KT 16-G-107: (a) E+I: beige; (b) dark core, fine mixed-tempered/ mineral temper: fine black grits; (c) E+I: slipped and smoothed; (d) finger traces of brown paint (e) none; (f) Late Chalcolithic.4. Locus KT 15-G-026+G033: (a) E+I: beige; (b) beige core, medium straw-tempered ware, (c) E+I: scraped and smoothed; (d) none; (e) none; (f) Late Chalcolithic.5. Locus KT 18-E-358: (a) E+I: cream; (b) light buff, medium vegetal-tempered ware with a few mica inclusions; (c) E: cream-slipped, burnished, I: cream-slipped, smoothed; (d) E: red-brown painted geometric motif; (e) use of watery paint; (f) Neolithic, Level 1.6. Locus KT 18-E-352: (a) E: cream, I: buff, plain; (b) buff core, medium vegetal-tempered ware with a few mica inclusions; (c) E: cream-slipped, burnished; I: buff, plain; (d) red-painted geometric motif, possibly finger-painted (uneven coat of paint); (f) Neolithic, Level 1.
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-15.jpg
Fichier image/jpeg, 293k
Titre Table 2 – Evolution of the ceramic assemblage during the Neolithic at Kültepe I (ca. 6200-5000 BC)
Crédits C. Marro
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-16.jpg
Fichier image/jpeg, 630k
Titre Fig. 15 – Relief-decorated ceramics from the Neolithic period
Légende 1. Locus KT 16-G-076: (a) E: cream, I: buff; (b) greyish core, medium vegetal-tempered; (c) E: cream slipped, mat burnished, I: scraped and unevenly burnished; both sides chaff-faced; (d) simple annular band around the neck; (e) none; (f) Neolithic, Level 2.2. Locus KT 16-G-104: (a) E: drab, I: light brown; (b) red core, medium straw-tempered ware with a few coarse grits; (c) E: unevenly burnished, flaked-off surface due to heavy; concentration of chaff; I (neck): unevenly burnished, (body) smoothed; both sides chaff-faced; (d) two vertical bands of knobs along the neck; (e) heterogenous paste from poor kneading; coil-crafting visible in shallow groove bordering the lower part of the sherd; (f) Neolithic, Level 2.3. Locus KT 16-G-118: (a) E: dark brown, I: reddish brown; (b) dark grey/brown; medium straw-tempered ware with a few medium grits, including quartz, as well as some crushed bones; (c) E: evenly burnished? Flaked-off surface; I: evenly burnished, soft surface; (d) loop in relief bordering vessel rim; (e) none; (f) Neolithic, Level 2.4. Locus KT 16-F-110: (a) E+I: light beige; (b) light core, medium straw-tempered ware with a few coarse grits; (c) E+I: unevenly burnished, matt surface; (d) simple annular band around the upper part of the body; (e) none; (f) Neolithic, Level 2.5. Locus KT 16-E-234: (a) E: cream, I: buff; (b) beige/buff core; medium straw-tempered ware with a few coarse grits, including quartz; (c) E: cream slipped, evenly burnished? (stained potsherd); I: unevenly burnished; (d) simple annular band around the body; (e) wavy, uneven surface inside corresponding to the coils used for crafting the vessel; (f) Neolithic, Level 2.
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-17.jpg
Fichier image/jpeg, 295k
Titre Fig. 16 – The lithic industry
Crédits CAD J. Thomalsky
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-18.jpg
Fichier image/jpeg, 786k
Titre Fig. 17 – The lithic industry
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-19.jpg
Fichier image/jpeg, 995k
Titre Fig. 18 – Chantier G, Late Chalcolithic and Kura-Araxes level
Crédits CAD G. Gadebois
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-20.jpg
Fichier image/jpeg, 354k
Titre Fig. 19 – Chantier G, Late Chalcolithic tomb (G-107)
Crédits Photo N. Mattana, MBA
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-21.jpg
Fichier image/jpeg, 856k
Titre Fig. 20 – Ceramics from the Kura-Araxes period
Légende 1. Locus KT 15-G-018: (a) E: black (lip + upper wall), grey (lower wall), I: black; (b) black, fine mixed-tempered ware; (c) E: highly burnished surface, I: lightly burnished surface; (d) none; (e) none; (f) Early Bronze Age.2. Locus KT 15-G-018: (a) E: light drab (lip + upper wall), black (lower wall); I: cream; (b) dark/grey bicoloured core; no visible temper but sparse, fine vegetal and mineral inclusions; (c) E+I: highly burnished surface; (d) none; (e) none; (f) Early Bronze Age.3. Locus KT 15-G-030: (a) E: mottled dark grey and grey; I: grey; (b) dark grey; fine mixed-tempered ware; (c) E+I: highly burnished surface; (d) none; (e) none; (f) Early Bronze Age.4. Locus KT 15-G-018: (a) E: grey, I: black; (b) black, fine mixed-tempered ware; (c) E: highly burnished surface; I: lightly burnished surface; (d) noe; (e) none; (f) Early Bronze Age.5. Locus KT 15-G-018: (a) E+I: grey; (b) greyish core, fine mixed-tempered ware; (c) E: lightly burnished surface, I: smoothed; (d) none; (e) none; (f) Early Bronze Age.6. Locus KT 15-G-018: (a) E: light grey (lip + neck, upper part; grey (neck, lower part + body); I: light grey; (b) dark core, fine mixed-tempered ware; (c) E+I: lightly burnished surface; (d) none; (e) none; (f) Early Bronze Age.7. Locus KT 15-G-010: (a) E: buff, I: beige; (b) buff, fine mixed-tempered ware; (c) E: lightly burnished surface; I: unevenly burnished surface; (d) none; (e) none; (f) Early Bronze Age.
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-22.jpg
Fichier image/jpeg, 276k
Titre Table 3 – Composition of the faunal assemblage from Level 1 in number of identified remains (NISP)
Légende 6 fragments of deer antler were not included in the table
Crédits R. Berthon
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-23.jpg
Fichier image/jpeg, 224k
Titre Fig. 21 – Kill-off pattern based on sheep and goat teeth from Level 1 (n=17)
Crédits CAD R. Berthon
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-24.jpg
Fichier image/jpeg, 39k
Titre Table 4 – Composition of the faunal assemblage from Level 2 in number of identified remains (NISP)
Légende 6 fragments of deer antler were not included in the table
Crédits R. Berthon
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-25.jpg
Fichier image/jpeg, 274k
Titre Fig. 22 – Kill-off pattern based on (A) sheep (Ovis aries, n=26) and (B) goat (Capra hircus, n=29) teeth from Level 2
Crédits CAD R. Berthon
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-26.jpg
Fichier image/jpeg, 69k
Titre Table 5 – Composition of the faunal assemblage from the Kura-Araxes occupation at Kültepe I in number of identified remains (NISP)
Crédits R. Berthon
URL http://journals.openedition.org/paleorient/docannexe/image/589/img-27.jpg
Fichier image/jpeg, 158k
Haut de page

Pour citer cet article

Référence papier

Catherine Marro, Veli Bakhshaliyev, Rémi Berthon et Judith Thomalsky, « New light on the Late Prehistory of the South Caucasus: Data from the recent excavation campaigns at Kültepe I in Nakhchivan, Azerbaijan (2012-2018) »Paléorient, 45-1 | 2019, 81-113.

Référence électronique

Catherine Marro, Veli Bakhshaliyev, Rémi Berthon et Judith Thomalsky, « New light on the Late Prehistory of the South Caucasus: Data from the recent excavation campaigns at Kültepe I in Nakhchivan, Azerbaijan (2012-2018) »Paléorient [En ligne], 45-1 | 2019, mis en ligne le 08 septembre 2021, consulté le 21 juillet 2024. URL : http://journals.openedition.org/paleorient/589 ; DOI : https://doi.org/10.4000/paleorient.589

Haut de page

Auteurs

Catherine Marro

CNRS, UMR 5133 Archéorient Environnements et sociétés de l’Orient ancien, 7 rue Raulin, 69007 Lyon – France

Articles du même auteur

Veli Bakhshaliyev

Azerbaijan National Academy of Sciences, Nakhchivan Branch, 76 Haydar Aliyev Prospekti, 7000 Nakhchivan – Azerbaijan

Rémi Berthon

Muséum national d’Histoire naturelle, UMR 7209 Archéozoologie, archéobotanique : sociétés, pratiques et environnements, 55 rue Buffon, 75005 Paris – France

Judith Thomalsky

German Archaeological Institute, Tehran Branch, Khiabane Shahid Hasan Akbari 7, PO Box 3894, 19639 Tehran Elahiyeh – Iran

Haut de page

Droits d’auteur

CC-BY-4.0

Le texte seul est utilisable sous licence CC BY 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search