Content uploaded by Rémi Berthon
Author content
All content in this area was uploaded by Rémi Berthon on Sep 11, 2017
Content may be subject to copyright.
Paléorient
On the Genesis of the Kura-Araxes phenomenon: New evidence
from Nakhchivan (Azerbaijan)
Catherine Marro, Veli Bakhshaliyev, Rémi Berthon
Citer ce document / Cite this document :
Marro Catherine, Bakhshaliyev Veli, Berthon Rémi. On the Genesis of the Kura-Araxes phenomenon: New evidence from
Nakhchivan (Azerbaijan). In: Paléorient, 2014, vol. 40, n°2. The Kura-Araxes culture from the Caucasus to Iran, Anatolia
and the Levant: Between unity and diversity. pp. 131-154;
doi : 10.3406/paleo.2014.5639
http://www.persee.fr/doc/paleo_0153-9345_2014_num_40_2_5639
Document généré le 23/05/2017
Abstract
The formative processes of the Kura-Araxes cultural complex, in particular the date and
circumstances of its rise, have been debated for decades. This article takes advantage of the
recent data retrieved from Ovçular Tepesi, a Late Chalcolithic settlement located in Nakhchivan by
the Arpaçay river, to demonstrate that the earliest manifestation of the Kura-Araxes phenomenon
should be dated back at least to the last quarter of the 5th millennium BC. More importantly, since
the early Kura-Araxes pottery from Ovçular Tepesi was found within the Late Chalcolithic habitat
together with buff-coloured Chaff-Faced Ware, we make use of technological and functional
evidence to argue that the coexistence of Kura-Araxes and Chaff-Faced vessels corresponds to a
multicultural settlement pattern, which involves two human communities that interact at the
regional level. We then compare this pattern with later examples of multiculturalism involving Kura-
Araxes societies in the Kura and Euphrates basins, but also in Palestine.
Résumé
Les processus de formation du complexe culturel Kura-Araxe, notamment la date et les
circonstances de son développement, font l’objet de débats depuis des décennies. Cet article y
contribue à travers la présentation de données récentes provenant d’Ovçular Tepesi,
établissement situé au Nakhchivan au bord de la rivière Arpaçay. pour démontrer que la plus
ancienne apparition du phénomène Kura-Araxe remonte au moins au dernier quart du 5e
millénaire av. J.-C. De façon plus signifi cative, étant donné que la céramique Kura-Araxe a été
retrouvée à Ovçular Tepesi au sein de l’habitat Chalcolithique récent, associée à de la vaisselle
Chaff-Faced Ware, nous démontrons, à travers l’introduction de paramètres technologiques et
fonctionnels, que la présence concomitante de ces deux types de poterie traduit la coexistence de
deux communautés humaines interagissant au niveau régional. Ces modes de fonctionnement
sont ensuite comparés à d’autres exemples de multi-culturalisme impliquant les sociétés Kura-
Araxe dans les bassins de la Kura et de l’Euphrate, mais aussi en Palestine.
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014 Manuscrit reçu le 13 décembre 2013, accepté le 19 juin 2014
ON THE GENESIS OF THE KURA-ARAXES
PHENOMENON: NEW EVIDENCE FROM
NAKHCHIVAN (AZERBAIJAN)
C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Abstract: The formative processes of the Kura-Araxes cultural complex, in particular the date and circumstances of its rise, have
been debated for decades. This article takes advantage of the recent data retrieved from Ovçular Tepesi, a Late Chalcolithic settlement
located in Nakhchivan by the Arpaçay river, to demonstrate that the earliest manifestation of the Kura-Araxes phenomenon should be
dated back at least to the last quarter of the 5th millennium BC. More importantly, since the early Kura-Araxes pottery from Ovçular
Tepesi was found within the L ate Chalcolithic habitat together with buff-coloured Chaf f-Faced Ware, we make use of technological and
functional evidence to argue that the coexistence of Kura-Araxes and Chaff-Faced vessels corresponds to a multicultural settlement
pattern, which involves two human communities that interact at the regional level. We then compare this pattern with later examples
of multiculturalism involving Kura-Araxes societies in the Kura and Euphrates basins, but also in Palestine.
Résumé : Les processus de formation du complexe culturel Kura-Araxe, notamment la date et les circonstances de son développement,
font l’objet de débats depuis des décennies. Cet article y contribue à travers la présentation de données récentes provenant d’Ovçular
Tepesi, établissement situé au Nakhchivan au bord de la rivière Arpaçay. pour démontrer que la plus ancienne apparition du
phénomène Kura-Araxe remonte au moins au dernier quart du 5e millénaire av. J.-C. De façon plus signifi cative, étant donné que la
céramique Kura-Araxe a été retrouvée à Ovçular Tepesi au sein de l’habitat Chalcolithique récent, associée à de la vaisselle Chaff-
Faced Ware, nous démontrons, à travers l’introduction de paramètres technologiques et fonctionnels, que la présence concomitante
de ces deux types de poterie traduit la coexistence de deux communautés humaines interagissant au niveau régional. Ces modes de
fonctionnement sont ensuite comparés à d’autres exemples de multi-culturalisme impliquant les sociétés Kura-Araxe dans les bassins
de la Kura et de l’Euphrate, mais aussi en Palestine.
Keywords: Kura-Araxes ceramics; Multi-cultural settlement patterns; Behavioural patterns; Caucasus; Anatolia.
Mots-clé s : Céramique Kura-Araxe ; Structures de peuplement multi-culturelles ; Modes de fonctionnement ; Caucase ; Anatolie.
A major phenomenon in the history of the Caucasus and the
Near East, the Kura-Araxes cultural complex, in particular its
formative processes, remains paradoxically fairly little known.
The circumstances of its rise and later expansion have been
hotly debated for decades, both in the former Soviet Union
and in the West, but the resulting scenarios tend to be rather
vague, since they mostly rest on speculation rather than per-
tinent data. Partly tainted by political considerations, arising
from the putative Indo-European origins ascribed by certain
authors to the Kura-Araxes cultures (Burney 1989 and 1993),
the debates have focussed on three major issues: the date and
place of their earliest manifestation, their original cultural
background, and the reasons behind their startling expansion.
This complex, which includes the Karaz, Red-Black
Burnished Ware (RBBW) or Khirbet Kerak facies, compares
in many ways in its duration and wide geographical scope with
other great cultural phenomena that are still being discussed
at length in Near Eastern archaeology. As with the Ubaid and
Uruk cultures for instance, the Kura-Araxes phenomenon fi rst
developed in a fairly circumscribed area and later expanded
over a wide region, which in the case of the Kura-Araxes,
encompasses the Caucasus, Eastern Anatolia, Western Iran
Livre_BAT.indb 131Livre_BAT.indb 131 04/12/14 12:3604/12/14 12:36
132 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
and Palestine. As with the Ubaid complex, it is characterized
by a relatively formal homogeneity that is coupled with dis-
tinct local features, as is evident from its ceramic repertoire, its
funerary traditions and architectural manifestations. Similarly,
the historical signifi cance of this long-lived phenomenon,
which according to recent data, developed over more than
2000 years, is still obscure. This situation is partly due to the
marked diversity expressed through its various regional com-
ponents, which raises many issues: are the cultural similarities
perceptible at a macro-regional level the outcome of organic
ties between its different components? If so, what are the cul-
tural dynamics behind these organic ties? Conversely, should
we ascribe cultural idiosyncrasies to regional parameters, or
are there other explanations? Since most of these questions
have not been tackled with adequate archaeological data, there
still is wide disagreement over the genesis, chronological evo-
lution and overall signifi cance of the Kura-Araxes cultures
over the years.
This paper will take advantage of the data recently retrieved
from the excavations of Ovçular Tepesi in Nakhchivan to focus
on the fi rst of these issues: the Kura-Araxes formative pro-
cesses and their signifi cance for apprehending south Caucasian
cultural dynamics in Late Prehistory.
STATE OF THE ART
Lately over the past ten years, a timid consensus had
seemingly been reached with more and more scholars sup-
porting the idea that Kura-Araxes cultures originated ca 3500
BC in the Middle Araxes basin from a local Late Chalcolithic
background.1 Only marginally have other hypotheses devel-
oped, suggesting that the formation of the so-called Kura-
Araxes cultures was a complex phenomenon, rooted both
in the Caucasus and in Central Anatolia (Palumbi 2008;
Frangipane and Palumbi 2007). On the whole, most interpre-
tations seemed to be at odds over the perplexing propensity of
these cultures to mix ‘typical’ Kura-Araxes traits with local
specifi cities.
Such was the state of the art when an unexpected discovery
was made in 2009 on the settlement of Ovçular Tepesi (fi g. 1)
where a Kura-Araxes potsherd-scatter was found in a well-pre-
served, sealed context in the middle of the Late Chalcolithic
habitat. This context, a house fl oor, has been dated through
relative and absolute methods to the last quarter of the 5th mil-
1. See for instance: Kushnareva 1997; Kiguradze and Sagona 2003; Kohl
2009; Ristvet et al. 2011; Gopnik and Rothman 2011.
lennium; the early dating of this fi nd was later confi rmed by
similar discoveries made in 2010 and 2011 in other parts of
the site.
After a presentation of the available data, this paper will
focus on the interpretation of these fi nds, testing a number of
explanatory hypotheses.
THE EARLIEST KURA-ARAXES ARTEFACTS
KNOWN TO DATE: THE EVIDENCE FROM
OVÇULAR TEPESI (CA 4300-4100 BC)
STRATIGRAPHY AND CONTEXT OF THE FINDS
The fi rst discovery of Kura-Araxes artefacts in a strati-
fi ed Late Chalcolithic context was made over the fl oor (locus
5194) of a domestic house. These fi nds consisted of a scat-
ter of potsherds, corresponding to a total of eight different
vessels (fi g. 2: 1-2, 6-7; fi g. 3: 3-4, 7-8). Interestingly, this
potsherd scatter was located next to a complete Chaff-Faced
Ware (CF W) bowl in situ (fi g. 4: 2), which was placed upside
down on the same fl oor. In the following years, other Kura-
Araxes potsherds were found in a pit (locus 11224) that ha d
been sandwiched between two work platforms in pisé (fi g. 2:
3-5; fi g. 3: 1-2, 5), as well as in a grave (locus 5313) that had
later been sealed by a Late Chalcolithic wall (fi g. 3: 6). The
work platforms, as well as the wall, belong to the Ovçular
facies.2
These fi nds came somewhat as a surprise, since this settle-
ment, which we started to investigate in 2006, had so far pre-
sented the ‘regular’ features of a small, possibly seasonal, south
Caucasian village, characterized by free-standing mud-brick
houses, plain chaff-tempered pottery, infant jar-burials and a
lithic industry dominated by obsidian. The pottery assemblage
in particular was extremely homogenous and had clear links
with the Chaff-Faced Ware at home in a wide region extending
from the Caucasus to Upper Mesopotamia in the second half of
the 5th millennium BC (Marro 2012). Occasional Kura-Araxes
potsherds had in fact been found here and there throughout the
Late Chalcolithic habitat before 2009, but since they usually
came from disturbed or unsealed loci, they had received little
attention. It is essential to note that the data we discuss in the
following pages comes exclusively from well-stratifi ed, sealed
2. The Ovçular facies may be describe d as a south Caucasian La te Chalcolithic
culture with Gawra overtones; for a more detailed description see Marro et
al. 2011; Gülçur and Marro 2012.
Livre_BAT.indb 132Livre_BAT.indb 132 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
133
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
contexts; the dating of which has been secured by the stratigra-
phy, pottery evidence and when possible, a series of 14C dates.
The fi rst discovery of Late Chalcolithic Kura-Araxes pot-
tery (LCKA)3 is thus the potsherd scatter that was found on
the fl oor of a domestic house, labelled House 5.3, excavated
in the north-west corner of Chantier 5. The context of these
sherds is quite clear, fi rst because they lay over a house fl oor
(locus 5194) and secondly because most of the scatter was
later sealed by a circular work platform (called ‘dial’ because
of its raised edge) made of pisé (locus 5167). In line with the
architectural tradition of Phase II,4 the far side of this house
3. For the sake of clarity, the Kura-Araxes pottery that was found in Late
Chalcolithic occupation layers has been labelled ‘LCKA’, as opposed to
‘EBKA’, which refers to the Early Bronze Age Kura-Araxes pottery.
4. The chronological sequence of the Late Chalcolithic occupation comprises
two phases, Phase I and Phase II. The architecture of Phase I is charac-
consisted of a stone terrace-wall (Wall 5132). The general lay-
out of House 5.3 may be inferred from the plan of earlier and
later buildings that lay respectively under or above Floor 5194,
since all the walls of House 5.3—except for Wall 5132—were
destroyed by later rebuilding activities.
Together with House 5.3, at least six architectural levels are
attested in the NW corner of Chantier 5 (fi g. 5); they all dis-
play a rectangular room, containing a dial or a circular hearth
located in the south-east side of the room. From Level 5.1
onwards (chronology speaking), the far side of the room was
delimited by the said terrace-wall, which was rebuilt at least
terized by semi-subterranean huts (Marro et al. 2011). For the dating of
these phases, see Footnote 8. After a hiatus of about 700 years, the Late
Ch alc olit hic level s of O vçula r wer e fol lowed by a n Ea rly Br onz e Age occ u-
pation (Phase III and IV), which is dated through 14 C dates to the 3200-
2400 BC timespan.
44°45'
44°45'
45°0'
45°0' 45°15'
45°15'
45°30'
45°30'
45°45'
45°45'
46°0'
46°0'
39°0'
39°0'
39°15'
39°15'
39°30'
39°30'
39°45'
39°45'
01020
km
OvOvOv
Ov
ul
u
ula
ulaula
ula
r
r
r
r
r
Ovçular
TT
T
T
T
T
T
T
T
ee
e
e
e
e
T
T
T
T
T
pe
pe
pe
pepe
pe
si
sisi
si
si
si
i
si
s
epesi
ARMENIA
IRAN
TURKEY
AZERBAIJAN
Nax
Nax
ax
i
Ni
e
v
a
n
n
Arpa
A
ay
ay
Arpaçay
Araxes river
Araxes river
NaxNaxNax
x
i
i
i
i
i
xi
x
eva
evaevaeva
v
nn
n
n
n
ay
ay
ay
a
ayay
y
ay
Naxçivançay
settlement
salt mine
Gemikaya
Batabat
pastureland
Nakhchivan
(AZERBAIJAN)
Kültepe II
Kültepe I
Culfa
Ordubat
Sarur
Duzdagi
Nakhchivan
Fig. 1 – Nakhchivan. Regional map (O. Barge – CNRS) .
Livre_BAT.indb 133Livre_BAT.indb 133 04/12/14 12:3604/12/14 12:36
134 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
Fig. 2 – Ovçular Tepesi. Late Chalcolithic Kura-Araxes pottery (I) (K. Alhamed).
three times (loci 5132/5152/5369). In Level 5.6 (the latest), this
terrace-wall is hardly visible, but the shape and the extension
of the room, which were roughly the same as in the previous
levels, could be made out from the presence of pisé benches, as
well as one perpendicular stone wall that marks the southern
side of the house.
The best preserved of these architectural levels is the
earliest (House 5.1), where the so-called ‘rectangular room’
Livre_BAT.indb 134Livre_BAT.indb 134 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
135
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
Fig. 3 – Ovçular Tepesi. Late Chalcolithic Kura-Araxes pottery (II) (K. Alhamed).
Livre_BAT.indb 135Livre_BAT.indb 135 04/12/14 12:3604/12/14 12:36
136 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
appears as part of a two-room building. House 5.1 displays
mud-brick walls, which sometimes rest on stone founda-
tions (Wall 5199); indoor features comprise a work basin in
pisé (locus 5342) and a storage jar (5333) found half-buried
in situ in Floor 5336 (fi g. 6 and table 1). The southern wall
built for House 5.1 (5199), as well as the inner mud-brick wall
(5300), have been reused in House 5.2 (fi g. 7). From House 5.3
onwards, the spatial layout is marked by some change, since the
‘rectangular room’ slightly shifts towards the south. House 5.4
is characterised by the building of a new mud-brick (?) parti-
tion wall (Wall 5235), which stands exactly at the same place
as Wall 5300 in Houses 5.1/5.2 (fi g. 8). In House 5.6, this mud
wall is replaced by a stone wall (5079), which is also built in
the same place as Walls 5300 and 5235 (fi g. 5).
The location of the dials tends to vary a little from level
to level, but they are always built in the south-eastern part of
Fig. 4 – Ovçular Tepesi. Chaff-Faced Ware (I) (K. Alhamed).
Livre_BAT.indb 136Livre_BAT.indb 136 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
137
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
the eastern room, except for House 5.1, where two dials (5296,
5329) were found in the adjacent room. The dials are invari-
ably made of mud, with a raised pisé wall around the edge.
In certain cases, this pisé wall has been reinforced by a stone
infrastructure (Dials 5259, 5177), which suggests that the func-
tion of these features may have varied from one to the other:
some of them (Dials 5161, 5177) bear marked traces of fi re;
they should however be distinguished from proper hearths,
which have a different shape.
Each building level has been modifi ed throug h tim e by sev-
eral additions (new pisé dials, hearths or storage jars), showing
that each level was home to a number of distinct occupation
phases (table 1); this is especially clear for House 5.2. But the
important information here is that the LCKA potsherds from
fl oor 5194 were found in the middle of a stratigraphic sequence
(fi g. 5) that is characterized by the repeated building of basi-
cally the same room, with similar domestic features that have
been built and rebuilt more or less in the same place over three
hundred years.
The second discovery of LCKA pottery was made in a plain-
earth tomb (locus 5313), which contained the remains of an
adult (fi g. 6). The body lay in its original position and was well
preserved, in spite of a Late Chalcolithic stone wall (Wall 5266)
that was later built over the tomb (fi g. 7). This wall belongs to
House 5.2, and marks the north-east limit of the ‘rectangular
room’. The tomb was void of funerary gifts but yielded four
black-burnished, Kura-Araxes body sherds (fi g. 3: 6).
It seems from the stratigraphy that tomb 5313 was wedged in
between House 5.1 and House 5.2 in the architectural sequence
of Chantier 5 (compare fi gs. 6 and 7). It did not damage any
domestic structure belonging to House 5.1, because Wall 5304,
which stratigraphically speaking may be considered as the
ancestor to Wall 5266, was built about 20 cm further to the
south of Wall 5266. Considering the regularity of the building
layout however, Tomb 5313 appears as an intrusive event, a
funerary one, within the stratigraphic sequence.
Lastly, another piece of evidence comes from Chantier
11, where several Kura-Araxes potsherds were found in a pit
(locus 11224) that was dug between two occupation phases in
a Late Chalcolithic building called House 11.1 (fi g. 6). As in
Chantier 5, this house, of which two rooms have been brought
to light, yielded a great number of dials in pisé, as well as sev-
eral hearths and storage jars in situ. In addition, House 11.1
also contained other domestic implements, among which a
mortar, a pestle and two door-sockets, which lay over a well-
preserved fl oor (locus 11260).
House 5.1 + House 5.2
(Wall
House 5.2
5152
5111
907.01 m
905.57 m
House 5.3
5235
5194
906.29 m
Terrace
Kura-Araxes potsherds
5167
House 5.4
5161
House 5.5
House 5.6 5079
5077 5118
907.25 m 907.28 m
906.83 m
5234
906.69 m
906.31 m
5300 )
+ House 5.1
House 5.1
(South-Eastern side
cut into the virgin soil)
AB
905.83 m
5199Wall
905.30 m
906.14 m
colluvium
pisé / mud brick
lling layer
0 1 2 m
Fig. 5 – Ovçular Tepesi, Chantier 5. The south-east ‘rectangular room’ stratigraphic sequence
(southern profi le + restitution) (B. Özgür, C. Jagoury and C. Marro).
Livre_BAT.indb 137Livre_BAT.indb 137 04/12/14 12:3604/12/14 12:36
138 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
In comparison with the houses excavated in Chantier 5,
the building sequence in Chantier 11 is characterized by fewer
architectural levels, but a greater number of occupation phases
within each level. Floor 11260, which probably belongs to the
second occupation phase, could easily be followed thanks to its
bluish colour and clayish texture. These occupation layers can be
made out mainly through the presence of pisé dials, which have
been constantly rebuilt on top of —and more often than not, cut
across by—each other. A series of at least eight superposed dials
has thus been evidenced on the northern side of Floor 11260,
next to a stone circular feature (locus 11093) that is so far inter-
preted as a silo. Most interestingly, the repeated rebuilding of
dials in this part of the room was on one occasion interrupted by
the digging of Pit 11224, which appears as an oddity within the
stratigraphic sequence. The pit was later sealed by another dial
(locus 11164), which is particularly well preserved.5
The digging of the pit is no doubt an intrusion in the archi-
tectural development of the room, characterized otherwise by
its regularity. The intrusive character of this pit is further illus-
trated by its content, with the presence, next to a handful of
5. For the location of Dial 11164 over Pit 11224, see Marro et al. 2011: Fig. 7.
Chantier 6
Chantier 1
Chantier 11
Chantier 5
905.48 m
905.32 m
905.30 m)
(Average oor elevation:
905.82 m
5304
905.49 m
905.04 m
5333
Jar
5342
905.68 m
905.31 m
905.77 m
905.23 m
Floor
5343
905.25 m)
(Average oor elevation :
5368
Grave
5313
Floor
11260
905.44 m
905.24 m
Dial
11238
905.55 m
904.92 m
Jar
Jar
11248
905.22 m
Door-socket
905.10 m
904.89 m
Mortar Hearth
11264
905.13 m
905.10 m (northern area)
905.28 m (southern area)
Door-socket
Burial jar
Burial jar
905.00 m
Floor
904.88 m
Door-socket
6420
Chantier 20
11093
905.68 m
Floor
5336
Pit
11224
20163 11220
052,5
m
Terrace cut into the virgin soil
Door-step
A
B
South-eastern
« rectangular room »
See Fig. 5
905.52 m
905.32 m
5329
5296
5305
5300
5199
5369
Terrace wall
905.40 m
All jars are half-buried in the oors
N
mud brick / pisé
Fig. 6 – Ovçular Tepesi, House 5.1 on top of House 11.1. House 5.1 = Floor 5336 + Floor 5343;
House 11. 1 = Floor 11260 + Floor 20163 (B. Özgür, C. Jagoury, P. Lebouteiller and C. Marro).
Livre_BAT.indb 138Livre_BAT.indb 138 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
139
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
Table 1 – Ovçular Tepesi. The architectural sequence of the NW corner of Chantier 5 (C. Marro).
Chantier 5 -NW
Architectural sequence Loci 14C dating
Sub-level 5.6-1
average floor alt. 907,25 m Stone wall 5079; Floor 5077 One 14C dating from lentils seeds found scattered over 5077
House 5.6
average floor alt. 907,10 m Stone walls (5079 + 5148); Benches (5112 + 5118); Floor 5111 No 14C dating available
Sub-level 5.5-1
average floor alt. 906,80 m Same wall features + Floor 5234 No 14C dating available
House 5.5
average floor alt. 906,70 m Terrace walls (5132 + 5152); Dial 5161; Floor 5212 Three 14C datings available from charcoal retrieved
respectively in 5137, 5161 and 5212
House 5.4
average floor alt. 906,40 m
Terrace walls (5132 + 5152); Mud-brick ( ?) wall 5235; Dial 5167;
Storage jars 5155 + 5162 + 5164); Floor 5168 No 14C dating available
House 5.3
average floor alt. 906,30 m
Terrace wall (5132); Dial 5177; Silo (?) 5193; Floor 5194
Kura-Araxes potsherd scatter One 14C dating from a piece of charcoal found in Floor 5194
Sub-level 5.2-3
average floor alt. 906,10 m Same wall features + Dial 5247 No 14C dating available
Sub-level 5.2-2
average floor alt. 905,95 m Same wall features + Dial 5268 No 14C dating available
Sub-level 5.2-1
average floor alt. 905,85 m Same wall features + Dial 5259 One 14C dating from a piece of charcoal found in 5259
House 5.2
average floor alt. 905,70 m
Mud-brick wall on stone foundations (5199); Mud-brick (?) walls
without stone foundations (5266 + 5300); Terrace wall (5369); Dial
5290; Storage jars (5315 + 5284); Work basin 5344; Floor 5306
No 14C dating available
House 5.1
average floor alt. 905,30 m
Mud-brick wall on stone foundations (5199); Terrace wall (5369);
Mud-brick walls without stone foundations (5300 + 5304 + 5305
+ 5368); Work basin 5342; Dials 5296 + 5329); Storage jar 5333;
Floors 5336 + 5343
One 14C dating from a piece of charcoal found within 5333
Chantier 6
Chantier 11
906.01 m
905.70 m
Terrace cut into the virgin soil
N
mud brick / pisé
OVCULAR TEPESI 2013
Chantier 05
House 5.2
Scale 1/25e
C. Jagoury, P. Lebouteiller
Jar
5284 905.73 m (oor-level)
Terrace wall
905.40 m
5369
Jar 5315 905.61 m
906.41 m
(905.70 m - average oor elevation)
Floor
5306
5266
5290 905.83 m
906.69 m
Dial
905.39 m
6359
906.08 m
905.63 m
906.01 m
905.29 m
5199
052,5
m
All jars are half-buried into the oors
906.14 m
A
B
(See g. 5)
(See g. 5)
905. 83 m
5300
Fig. 7 – Ovçular Tepesi, House 5.2 (B. Özgür, C. Jagoury, P. Lebouteiller and C. Marro).
Livre_BAT.indb 139Livre_BAT.indb 139 04/12/14 12:3604/12/14 12:36
140 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
typical CFW pottery, of seven Kura-Araxes sherds, most of
which are black-burnished. It is possible that each pisé dial,
which is usually separated from the previous one by a very thin
fi lling layer (a few centimetres at the most), corresponds to the
rebuilding of a key domestic feature by the Ovçular commu-
nity after a period of absence (a year?). At any rate, the pres-
ence in between two dials (11238 and 11164) of a pit containing
black-burnished pottery suggests that the site was visited by
Kura-Araxes groups when the usual inhabitants of the Ovçular
village were away.
THE FINDS
As evident from the above, all the Kura-Araxes artefacts
so far retrieved from the Late Chalcolithic contexts of Ovçular
Tepesi are ceramic potsherds: no complete vessel, nor any other
artefact that could be identifi ed as ‘typically Kura-Araxes’ is
attested with certainty within the Late Chalcolithic habitat.
Five Kura-Araxes andirons have so far been found but they
all come from Early Bronze Age loci belonging to Phase III
(Silo 12089, Pit 19052) or from disturbed contexts. Similarly,
no traces of Kura-Araxes domestic structures, apart from pits,
have been related with certainty to this period; the only other
locus is Grave 5313.
Some of these potsherds nonetheless piece together to form
fairly la rge vessels, as is the case of the pithos found over Floor
5194 (fi g. 2: 7 and fi g. 9). Half of its collar and upper body
was found in House 5.3: it is clear that the potsherds belong-
ing to this pithos may not be considered as stray sherds, even
if the presence of a fragmentary Kura-Araxes storage jar in
this house is far from being clear. The potsherds scattered
over Floor 5194 otherwise belong to various vessels: apart
from the pithos, there are fragments of two medium-sized jars
characterized by so-called Nakhchivan lugs, two jarlets (one
with traces of a Nakhchivan lug) and one bowl (fi g. 2: 1-2, 6;
fi g. 3: 7-8). Decoration is provided by the presence of a dimple
on two body fragments (fi g. 3: 3-4), whereas the bowl bears
a silver sheen that could be graphite. Interestingly enough,
one of the medium-sized jars is made of coarse, grit-tempered
Fig. 8 – Ovçular Tepesi, House 5.4 + Kura-Araxes potsherd scatter from House 5.3
(K. Dupinay, C. Jagoury, P. Lebouteiller and C. Marro).
5152
Terrace wall
907.01 m
906.40 m
Kura-Araxes
potsherds
906.30 m
906.38 m
906.52 m
906.31m
906.52 m
906.40 m
906.59 m
906.24 m
Jar
Jar 5155
906,28 m
Chantier 6
052,5
m
Floor
(Average oor elevation: 906.40 m)
906.30 m)
(House 5.3 = Floor 5194:
All jars are half-buried in the oors
N
mud brick / pisé
A(See g. 5)
B(See g. 5)
906.93 m
906.31 m
5132
Terrace wall
Terrace cut into the virgin soil
5167
Dial
Wall 5235
906.54 m
906.27 m
Jar 5162
5164
5168
Livre_BAT.indb 140Livre_BAT.indb 140 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
141
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
ware (Ware 59)6 displaying drab, unburnished surfaces. All
others display the typical black or dark grey burnishing that is
characteristic of the Kura-Araxes culture.
The pithos is of particular interest, since it displays the
black/red contrasting colour scheme that is one of the hall-
marks of Kura-Araxes pottery. But unlike EBKA examples,
the red colour does not result from elaborate fi ring processes.
In this particular case, the red surface of the vessel comes from
a coat of red paint that was applied over the interior of the ves-
sel (fi g. 10). On discovering this pithos, we fi rst thought that
this coat of red paint could be a chronological marker, sug-
gesting that paint was the initial technique employed by Kura-
Araxes potters for producing the famous contrasting colour
scheme. But then we discovered another bicoloured potsherd
in Pit 11224, admittedly fairly small, which displays black/red
contrasting surfaces without the use of paint. In the latter case,
the red colour was seemingly obtained through a process of
burnishing before fi ring.7
These fi nds raise the question of the specifi city of the
LCKA as opposed to the EBKA assemblages, since many of
the traits that are usually considered as typical of the ‘classi-
cal’ phase of the Kura-Araxes repertoire are already attested
at Ovçular Tepesi during the Late Chalcolithic period: this is
the case in particular of the rail-rims, dimples, dark-burnished
6. For the description of Ware groups, see the catalogue of the plates.
7. As far as exterior surfaces are concerned, six of the seven sherds retrieved
from Pit 11224 are black-burnished; one example is highly burnished,
while two of them bear traces of graphite. The fabric of the unbur nished jar
compares well with the jar potsherd found over fl oor 5194, except that it is
mixed-tempered (=Ware 58), it als o displays a ‘smok ed’ effect. De corat ion
is attested with two dimples impressed on the body.
exterior walls, and the colour contrast between exterior and
interior surfaces.
Judging from the available data, however, some traits
seem specifi c to the LCKA assemblage: bowls with simple
rims, which are attested at Ovçular Tepesi during the Late
Chalcolithic (fi g. 2: 5-6), are unknown during the Early
Bronze Age, where bowls occur with a beaded rim. Similarly,
LCKA jars tend to have simple rims (fi g. 2: 1-4; fi g. 3: 8),
while most EBKA jars display a beaded rim, sometimes with
a tapered edge. Otherwise, the general composition of the two
repertoires is certainly comparable (mainly jars and jarlets,
with occasional bowls). Elaborate decorative patterns, or the
use of incised or raised decoration, are absent from the LCKA,
as well as from the early EBKA assemblages.
The LCKA ceramic corpus of Ovçular is certainly too
small for us to proceed to a thorough comparative analysis
with its late 4th millennium counterpart, but the presence dur-
ing the Late Chalcolithic of features considered to embody
the spirit of Kura-Araxes aesthetics is certainly a fact that has
important historical implications.
RELATIVE AND ABSOLUTE DATING
The relative dating of these extraordinary fi nds is fi rst given
by their position in the stratigraphic sequence. All the LCKA
pottery from Ovçular Tepesi has so far been found in Phase II,
which has been dated through a coherent series of radiocar-
bon readings to the last quarter of the 5th millennium BC,8 ca
8. Marro et al. 2009 and 2011. The dating of Phase II was originally ascribed
Fig. 9 – The LCKA pithos from Locus 5167-5194
(© Mission Araxe) .
Fig. 10 – Inner scraped surface of the pithos
(© Mission Araxe) .
Livre_BAT.indb 141Livre_BAT.indb 141 04/12/14 12:3604/12/14 12:36
142 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
4300-4000 BC – (2 sigmas). No such fi nd has ever been made
in Phase I, which is characterized by a fl imsy settlement made
up of semi-subterranean huts. Judging by the settlement layout,
composed of individual houses that never encroach on each
other, Phase I is seemingly a short time period that has been
attributed through radiocarbon analyses to the ca 4350-4300
BC time span; but its real duration could be less than ten years.
Secondly, the dating of these fi nds is confi rmed by the
CFW pottery, usually whole vessels in situ, that are often found
together with the Kura-Araxes evidence: apart from the CFW
bowl found on Floor 5194 next to the LCKA potsherd-scatter,
a series of three storage jars (loci 5155, 5162 and 5164) were
brought to light in House 5.4, just above Floor 5194 (fi g. 8). In
Chantier 11, one storage jar (locus 11248) was found in situ
associated with Floor 11260; these jars were in use before the
digging of Pit 11224, but they certainly bear affi nities with the
jars from Chantier 5, which stratigraphically speaking, come
after Pit 11224. From a technological and a morphological
point of view, all these jars are typical of the Ovçular facies,
which is itself part of the Chaff-Faced Ware oikumene (Marro
2011 and 2012) at home in Late Chalcolithic Transcaucasia,
Eastern Anatolia and Upper Mesopotamia.
Last but not least, the LCKA fi nds have been dated by eight
14C analyses, sampled from sealed contexts located respec-
tively in House 5.1, House 5.2, House 5.3, House 5.5, House 5.6
and House 11.1 (tables 2-3; fi g. 11). The samples, which were
obtained from charred botanical remains, were prepared and
measured using AMS facilities. The dates were calibrated and
modelled with the OxCal v.4.2 programme (Bronk Ramsey
to the 4250-4000 BC timespan, but the 14 C dates recently obtained from
Chantier 5 and Chantier 11 suggest that its beginning should be dated back
to at least 4300 BC.
2009), using the calibration curve IntCal13 (Reimer et al.
2013). Thanks to the stratigraphic position of the samples, a
multiphase model could then be applied to the calibrated dates
in order to reduce their time range through the use of Bayesian
statistics.
Each house in this model corresponds to a phase for which
one or several 14C dates are available, with the exception of
House 5.4 (table 1), for which no appropriate sample could be
measured. According to the stratigraphy, House 11.1 is older
than House 5.1 but we do not know whether the two build-
ing phases are actually separated by a defi nite time gap. At
all events, we placed two boundaries between House 11.1 and
House 5.1 (Bronk Ramsey 2009: 348). On the other hand, the
different house phases in Chantier 5 seemingly follow each
other fairly continuously; accordingly, only one boundary was
placed after each phase. The model we propose here is fairly
accurate since its agreement index (Amodel) is over 60 (Bronk
Ramsey 2009: 356-357). If we look at the individual agree-
ment indices (A), only one sample (LTL13319A from locus
5137) does not clearly correspond to the model (table 3): this
sample is slightly older than expected, but its low agreement
index does not challenge the entire model. Finally, it should be
emphasized that the antiquity of these radiocarbon dates can-
not result from any kind of ‘old-wood effect’, since among the
samples retrieved from the latest building levels, one reading
(5077) was obtained from lentil seeds (House 5.6), and another
(5161) from tree branches (House 5.5) that were probably col-
lected for fi re-wood (table 2).
The radiocarbon dates obtained from the stratigraphic
sequence of Chantiers 5 and 11 certainly confi rm that the
LCKA fi nds date back to the last quarter of the 5th millennium
BC. Recalculation through Bayesian statistics even pushes this
dating further back, as shown by the data from locus 5194,
Table 2 – Ovçular Tepesi. Radiocarbone dates (BP) of the selec ted
samples by chronological order of the measured dates (R. Berthon) .
Sample ID Locus Radiocarbon
Age (BP)
δ13C (‰) Material
LTL13320A 5212 5200 ± 45 -28.3 ± 0.4 Wood charcoal
LTL5311A 5077 5210 ± 50 -31.9 ± 0.4 Seed (Pisum/Vicia)
LTL5312A 5161 5215 ± 50 -33.4 ±0.4 Wood charcoal
LTL8087A 5194 5364 ± 40 -28.7 ± 0.3 Wood charcoal
LTL13319A 5137 5389 ± 45 -23.6 ± 0.3 Wood charcoal
LTL13321A 5259 5450 ± 45 -21.4 ± 0.4 Wood charcoal
LTL12565A 5333 5600 ± 45 -20.9 ± 0.2 Wood charcoal
LTL13323A 11267 5635 ± 45 -24.1 ± 0.1 Wood charcoal
Table 3 – Unmodelled and modelled 14C calibrated age of the sam-
ples (cal. BC, 2 σ). The model agreement index (Amodel) is 71.8 (A:
agreement index). Duration of the sequence (span): 232-742 years
(R. Berthon).
Phase Sample ID Locus
Unmodelled
calibrated
date (2 σ)
Modelled
calibrated
date (2 σ)
A
House 5.6 LTL5311A 5077 4230-3947 4220-3949 108.8
House 5.5
LTL13320A
LTL5312A
LTL13319A
5212
5161
5137
4228-3849
4230-3951
4340-4059
4229-3986
4229-3991
4241-4041
66.5
80.8
39.9
House 5.3 LTL8087A 5194 4329-4056 4329-4122 111.4
House 5.2 LTL13321A 5259 4438-4176 4362-4252 108.2
House 5.1 LTL12565A 5333 4518-4350 4454-4344 105.1
House 11.1 LTL13323A 11267 4550-4358 4549-4378 107.2
Livre_BAT.indb 142Livre_BAT.indb 142 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
143
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
where the LCKA ceramic scatter was found: Bayesian statis-
tics suggest that the dating of 5194 could be restricted to the
4329-4122 cal. BC time span, instead of 4329-4056 cal. BC
(fi g. 11). More samples from each building phase would in fact
help to reduce this date range more effectively, but, thanks to
the available data, we can at least check the reliability of our
samples: the accuracy of the stratigraphic analysis, as well as
the absence of intrusion from later deposits, are certainly con-
fi rmed by the consistency between the measurements and the
model.9
9. This model also provides some information about the possible duration of
each phase: the recalculation of the available data shows that the architec-
tural sequence from House 11.1 to House 5.6 corresponds to an occupation
of at least some 230 years, suggesting a minimum of 30 years per house.
CONCLUSION
The attribution of the early Kura-Araxes ceramic fi nds
from Ovçular Tepesi to the Late Chalcolithic is thus unques-
tionable: the stratigraphy, the pottery evidence, as well as the
radiocarbon dates all converge to place these fi nds in the last
quarter of the 5th millennium, and possibly a little earlier.
These fi nds support Gadzhiyev’s hypothesis,10 according to
which the Kura-Araxes culture should be dated back to the
Late Chalcolithic, but this is the fi rst time that the antiquity of
this culture is actually proved by a series of reliable 14C dates.
More importantly, it must be stressed that the LCKA pottery
from Ovçular Tepesi does not correspond to some early stage
in the development of this ceramic trend, but to a phase where
10. Gadzhiyev 1983: 42. A similar hypothesis is also put forward by Kavta-
radze who marshalls evidence from the Upper Euphrates valley and the
Near East (Kavtaradze 1983: 82; 2004: 546 sq.).
Fig. 11 – Modelled dates (R. Berthon).
Livre_BAT.indb 143Livre_BAT.indb 143 04/12/14 12:3604/12/14 12:36
144 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
its main features appear already in full bloom: clearly, we are
not faced with a genesis here; and there is no reason to consider
the LCKA from Ovçular as some kind of ‘Proto-Kura-Araxes’
ware.
If the LCKA pottery from Ovçular Tepesi is certainly the
earliest known to date, it must be noted that this discovery has
in fact recently been echoed by similar fi nds in the cave of
Areni-1. As at Ovçular Tepesi, Kura-Araxes pottery has been
found at Areni in Late Chalcolithic loci that have been dated
by radiocarbon analyses to the very end of the 5th millennium
BC (Wilkinson et al. 2012). The interpretation of the fi nds
from Areni-1, however, is diffi cult because of the complexity
of the cave stratigraphy. Whether we are faced with a proper
Kura-Araxes occupation layer, or with a few Kura-Araxes arte-
facts within a CFW environment, as at Ovçular Tepesi, is still
not clear.
At all events, the discovery from Ovçular Tepesi certainly
questions the analytical framework so far proposed in the
interpretation of the Kura-Araxes formative processes: more
than a Bronze Age society replacing Late Chalcolithic com-
munities in a linear change, we may be dealing with a complex
settlement pattern, whose examination requires a complete
renewal of analytical paradigms. As for the fi nds from Areni-1,
they show that the presence of early Kura-Araxes components
at Ovçular Tepesi during the 5th millennium BC cannot be con-
sidered as an oddity, and suggest that similar fi nds should be
expected in Nakhchivan and beyond in the future.
INTERPRETATION: THE COEXISTENCE
VERSUS THE POLYMORPHISM HYPOTHESES
If the emergence of Kura-Araxes phenomenon can now be
dated back to the last quarter of the 5th millennium, the signifi -
cance, as well as the implications, of this important discovery
remains to be established. The interpretation of the LCKA
pottery from Ovçular Tepesi is certainly not straightforward:
are we faced with two distinct groups, the Ovçular versus
the Kura-Araxes communities, living together in the same
area? Or is the presence of Kura-Araxes potsherds in the Late
Chalcolithic habitat the result of stylistic polymorphism within
the Ovçular cultural facies?
Whether the evidence from Ovçular testifi es or not to some
kind of multiculturalism at this early period has now become
the issue at stake. This question will be tackled through a sys-
tematic analysis of the artefacts at hand with a view to assessing
the organic relationships between the two groups, or absence
thereof. Since the LCKA material assemblage is limited to
ceramics, the following analysis will focus on the comparative
study of the CFW versus the LCKA pottery repertoires, before
introducing other data related to the macro-region.
COMPARATIVE ANALYSIS OF THE POTTERY
ASSEMBLAGES
First of all, it must be noted that the comparative study of
the CFW and LCKA assemblages is hampered by the small
size of the LCKA corpus, which only comprises potsherds
belonging to about 18 different vessels (minimum number).
Yet, technological, morphological and even functional analy-
ses are feasible, thanks to the relatively good state of preserva-
tion of these vessels.
At fi rst sight, the examination of the LCKA and CFW assem-
blages conveys the impression that they have no relationship
with each other, whether it be in the shapes, the decoration, the
pastes or the surface treatment: we are seemingly faced with two
distinct manufacturing systems, the fi rst being characterized by
a heavy work and time investment, while the second reveals a
concern for rapidity and standardisation. Yet, one has to con-
sider the possibility, at least from a ‘theoretical’ point of view,
that these two assemblages were produced by the same potters
but used in specifi c contexts that may explain their widely dif-
ferent features. In theory, it is possible that the LCKA pottery,
which represents a minority in the overall Late Chalcolithic
ceramic corpus, corresponds to a particular usage within the
Ovçular culture, perhaps during a religious festival; just as spe-
cifi c tableware is used at Seder during the Jewish celebration of
Pessah. In the following lines, we will proceed to a systematic
analysis of the LCKA pottery by comparing it to the CFW, in
order to establish how far such as an interpretation is tenable.
Composition of the pottery repertoires:
the functional specifi city hypothesis
The analysis of the functional composition of the two groups
shows that the CFW corpus is basically divided into jars and
bowls. Bowls of different shapes amount to 46% of the total
CFW assemblage in Chantier 1, the pottery of which has been
studied in its entirety (Gülçur and Marro 2012). Most bowl
shapes are fairly basic and represented by simple-rim bowls with
a round bottom (fi g. 4: 1-2, 5); however, specifi c shapes such as
bowls with annular bases or hollowed bases, or fl ower-shaped
bowls also occur. Jars are usually wide-necked (about 25%) in
Chantier 1, but hole-mouth (fi g. 4: 6) and narrow-necked jars
Livre_BAT.indb 144Livre_BAT.indb 144 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
145
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
are also attested (about 3% in each case). Most necked jars
are decorated with an annular coil with or without decoration
(fi g. 12: 1-2); sometimes with an added fi gurative frieze of cap-
rines around the shoulder. Handles are usually found on narrow
necked-jars, which may have been pouring vessels.
Apart from jars and bowls, specifi c shapes such as pans,
probably used for the making of fl at bread or pancakes, as well
as mangaly, possibly used as a grill, are also frequent. Among
rare shapes that come in CFW, one can mention a cubic pot-
stand, mugs and cottage-cheese strainers.
LCKA pottery on the other hand is mostly restricted to col-
lared jars. These jars may be either narrow or wide-necked
(fi g. 2: 1-4 and fi g. 3: 8), the former being more frequent; hole-
mouth jars are unknown. All these jars are characterized by
the presence of one or two Nakhchivan lugs. Bowls are attested
Fig. 12 – Ovçular Tepesi. Chaff-Faced Ware (II) (K. Alhamed).
but rare; they display a basic half-spherical shape with a small,
fl at bottom.
As evident from the above, the CFW and the Kura-Araxes
repertoires are perceptibly different from each other; even
if the functional interpretation of these differences is not as
simple as it may seem. The CFW assemblage presents a fairly
balanced picture with jars and bowls amounting respectively
to ca 33% and 46% of the total repertoire in Chantier 1, while
the Kura-Araxes assemblage seems more specifi c, since most
of the vessels are jars. It must be noted however that these jars
come in very different sizes and probably have different func-
tions: from jarlets that could serve as drinking mugs to the red-
black burnished pithos, a variety of uses may be inferred from
the available collection, that include serving, pouring, cook-
ing11 and storing. Moreover, the rare representation of bowls or
basins could be explained by the use of wooden ware in these
once tree-covered regions: indeed, the rarity of bowls in the
LCKA repertoire may have other explanations than the puta-
tive functional specifi city of the Kura-Araxes assemblage.12
Last but not least, the specifi city hypothesis is not supported by
the context of the fi nds, which may be domestic (house fl oor,
pit) or funerary (tomb): as it stands, the LCKA repertoire from
Ovçular Tepesi seems, just as the CFW assemblage, to have
a multi-purpose function, with a standard production that is
rarely decorated but for an occasional dimple.
Morphological and technological analysis
A better insight into the relationship between the LCKA
and the CFW pottery is actually given by the morphologi-
cal and technological analysis of the fi nds, which has been
conducted here by following the methodology established by
V. Roux (Roux 2010 and 2011).
Morphologically-speaking, these two repertoires have very
few traits in common: the body of CFW jars is always globe-
or egg-shaped; they are characterized by their curved profi le,
whereas the body of LCKA jars tends to be carinated (fi g. 3:
2, 4), even if this carination is not as frequent as in the Early
Bronze Age. This is also true for the join between the collar and
the body, which is usually curved in CFW jars, but angular in
LCKA vessels, where this join is sometimes even emphasized
11. The use of Kura-Araxes pottery (either LCKA or EBKA) as cooking-pots
at Ovçular may be inferred from traces of fi re over the bottom part of
several vessels.
12. It must be noted that classical EBKA pottery repertoires from late 4th-3rd
millennia Caucasian settlements do not display a greater proportion of
open vessels, as if the use of terracotta bowls or basins in the Kura-Araxes
communities was rather limited.
Livre_BAT.indb 145Livre_BAT.indb 145 04/12/14 12:3604/12/14 12:36
146 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
by a groove (fi g. 3: 5). Lastly, the bottom of CFW jars is usually
round, suggesting that the jar was modelled upside-down, pos-
sibly moulded over the bottom of another jar.13 This contrasts
with the bottom part of Kura-Araxes pottery, which is always
fl at, showing that the vessels were modelled right side up.
In both cases, the potters modelled slabs about 1-2 cm thick
and 10-12 cm high to craft their vessels, but in the case of the
13. Alternatively, the bottom of these jars could be modelled right side up by
using some kind of rotative, spherical support: see Gosselain 2010: 672,
n. 13-16.
Fig. 13 – Ovçular Tepesi. Comparative crafting techniques
of the LCKA (1-2) and CFW (3-4) (K. Alhamed, C. Marro).
CFW jars, the vessel was shaped by adding paste astride the
previous slab; the paste was then stretched into a new slab by
working symmetrically on both sides of the body wall (fi g. 13:
3-4). LCKA jars, on the other hand, are crafted by adding a
piece of clay either from inside or outside the previous slab
(depending on the orientation of the wall, concave or convex).
The paste is then shaped into a new slab by stretching it up
from this particular side, even if this method of course also
involves making the two slabs stick together by working on
both sides of the wall (fi g. 13: 1-2).
This difference in technique has important implications
for the overall shape of the vessel, since the addition of a slab
astride the previous one involves a vertical pressure over the
wall that tends to distort the overall shape of the vessel. This
crafting method probably explains why large CFW vessels
are often slightly slanted, with uneven surfaces (fi g. 14).14 The
walls of LCKA vessels, on the other hand, are most of the time
remarkably even, which suggests greater craftsmanship in pot-
tery making.15
The tendency of CFW vessels to have uneven surfaces
probably explains why a strong emphasis is put on scrap-
ing. CFW medium to large vessels are always scraped, but
this technique is also commonly found on bowls; it is usu-
ally applied on bowls and jars alike with the help of a dented
comb (fi g. 15). Some form of scraping is also attested on
LCKA pottery but this occurs much less frequently. Unlike
the tool used in CFW scraping, the device applied for level-
ling LCKA jars, probably wood or fl int, seems fairly basic
and leaves uneven traces (fi g. 10).16 Lastly, when the scraping
technique occurs on LCKA pottery, it is usually confi ned to
the interior surface of the vessels, whereas comb-scraping on
CFW is frequently applied on exterior surfaces, and rarely
inside.17
14. The only way to prevent such deformation is by applying a new slab on
leather-dried paste, which in turn may jeopa rdize the solidity of the vessel
(information provided by Y. Dede, Istanbul University). Judging by the
number of CFW potsherds whose breaking point precisely corresponds
to the junction between two slabs, this seems however to be the technique
used by CFW potters.
15. It should also be noted that the addition of a new slab pasted sideways
from the interior or the exterior wall of the vessel, as is the case with
LCKA potters, leads to quicker vessel crafting : there is no need to wait
for the paste to dry: this method, too, reveals a better understanding of
pottery making than the CFW technique.
16. Work is in progress to determine the raw material used for the making of
the combs. Use-wear analyses have shown that the lithic tools available
at Ovçular Tepesi have not been used for ceramic scraping (C. Lemorini,
University of Rome).
17. It should be noted, however, that true comb-scraping on Kura-Araxes
ware is occasionally attested at Ovçular at the end of the 4th millennium
on EBKA pottery.
Livre_BAT.indb 146Livre_BAT.indb 146 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
147
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
Another major difference in the surface treatment between
the CFW and the LCKA repertoires is of course the use of bur-
nishing, which occurs on most LCKA vessels, whereas bur-
nished surfaces are very rare on CFW.
18 Burnished surfaces
on LCKA are usually dark (black, brown, drab or grey) on
the exterior, while the interior is slightly smoothed or simply
left plain.19 Examples of shiny black-burnished surfaces, which
were probably meant to give a metallic sheen to the pots, show
that burnishing processes had already been mastered by LCKA
potters during the Late Chalcolithic;20 no equivalent in quality
has been found in the CFW repertoire, where burnished sur-
faces tend to be uneven and fairly dull.
In short, if we compare the main surface treatments used
respectively by the CFW and the LCKA potters, differences
in craftsmanship are quite evident: burnishing is never car-
ried out to the point of obtaining an even, smooth surface in
CFW, and it never achieves the shiny appearance of their Kura-
18. Burnished pottery in the CF W assemblage from Ovçular starts to occur at
the very end of the occupation sequence; it usually comes from disturbed
fi lling layers in the upper levels of the site, and thus has little stratigraphic
value.
19. One example of a red-burnished interior surface has also been found in
Pit 11224; it appears as the fi rst example of true Black-Red Burnished
ware.
20. The methods used by Kura-Araxes potters to produce such an effect have
been the object of much speculation: experiments are currently in prog-
ress at Ovçular Tepesi under the supervision of Yüksel Dede (Istanbul
University).
Araxes counterparts. On the whole, the rare examples of bur-
nished CFW appear as a far cry from the LCKA models. As
for the Kura-Araxes pottery, the use of scraping is rare, and
mostly devoted to the levelling of interior surfaces, but never to
the fashioning of the overall shape, as is the case with certain
CFW jars: scraping devices are clearly an auxiliary tool in the
making of Kura-Araxes pottery, while they play a central role
in CFW craftsmanship.21
To conclude, it is clear that CFW and LCKA potters used
different crafting techniques: their common technological
background is fairly limited, while the gestures implied in
each chaîne opératoire are on the whole distinct: if the potters
sometimes used similar tools, we are certainly not faced with
a community of practices.
The pastes
Most of the pastes present in the LCKA corpus are coarse,
grit-tempered with grey, fl attish, angular inclusions (Ware 52,22
n=5 ; 23%). Variations in this group relate to the i nclusions, which
21. The comb, for instance, is also employed as a means of decoration, as in
the impression of oblique stitches along annular coils on jars: Marro et
al. 2011: Pl. VIII: 1. This is also the same technique used on the famous
comb-stamped ware: Marro et al. 2011: Pl. XIX: 3; Gülçur and Marro
2012: Fig. 7: 9-10.
22. For a description of the main ware families, see Gülçur and Marro 2012:
Table 2.
Fig. 14 – CFW jar (Locus 5121) (© Mission Araxe). Fig. 15 – Comb-scraped surface of CFW (© Mission Araxe).
Livre_BAT.indb 147Livre_BAT.indb 147 04/12/14 12:3604/12/14 12:36
148 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
may be purple (Ware 53, n=1), or purple and grey (Ware 54,
n=1). Similar groups are also attested with a smoked surface
(Ware 58, grit-tempered; and Ware 59, mixed-tempered). All
together, the coarse ware-groups comprise about 50% of the
fi nds. They display an uneven break that is due to the coarseness
of the inclusions which appear to be some kind of schist.
The second main group comprises fi ne, mixed-tempered
ceramics (Ware 41, n=5; 23%), while the third group corre-
sponds to ceramics with no visible temper, sometimes with
the occasional presence of very fi ne grits included in the clay
matrix (Ware 62, n=3; 15%).
By contrast, most of the CFW pastes are vegetal-tempered
(Wares 11 and 12); their texture is usually coarse or semi-
coarse; fi ne vegetal-tempered pastes are very rare. The veg-
etal temper is sometimes replaced by a mixture of straw and
medium-sized grits (Wares 21 and 22), but it should be noted
that all the CFW ware groups contain to a greater or a lesser
extent mineral inclusions that were present in the clay matrix,
indicating that the clay was poorly levigated. When grit inclu-
sions have been intentionally added to the vegetal temper, they
appear as sand grits of different sizes, resulting from mini-
mal sieving. The texture of the CFW vessels is always loose
because of the heterogeneity of the paste and the presence of
the numerous voids left by the straw imprints; by contrast,
the texture of the LCKA is often fi ne (Ware 41) or even dense
(Ware 62), with hardly any void visible.
Another technological difference between the LCKA and
the CFW assemblages is given by the geochemical analy-
ses carried out by F. Dessène: the analysis of six potsherds
retrieved from Floor 5194 suggests that the clay used for pro-
ducing LCKA vessels came from a single source (3ARI), which
is distinct from that exploited in the making of CFW vessels
(M2APV).23 It is thus clear that the clays used by CFW and
LBKA potters came from different sources; but more impor-
tantly, it also appears that they were cured differently. The clay
involved in the making of LCKA vessels is often (50%) care-
fully prepared before use, which is never the case with CFW.
Firing processes
As with the clays, the fi ring processes involved in the mak-
ing of CFW and LCKA have little in common. Unfortunately,
no kilns related to the making of CFW have been found with
certainty at Ovçular Tepesi: the bottom of two large circular
23. Frédéric Dessène (Univ. Laval, Québec), pers. comm. ; Dessène and
Boileau, poster presented at the EMAC 09: “Late 5th to Early 3rd
Transcaucasian Wares: a Technological and Petrographic Perspective”.
ovens was brought to light in Chantier 6, but they were empty
and no ceramic waste was found in the vicinity. On the other
hand, several superimposed open-air hearths containing a num-
ber of in situ Kura-Araxes vessels were found in the later levels
of Chantier 6; these hearths were dated through 14C readings to
the last quarter of the 4th and the beginning of the 3rd millennium
BC (= Phase III). These hearths, by their relative small size and
their light infrastructure, could be domestic pottery-fi ring facili-
ties, but they have little resemblance with an artisan’s kiln.
Thus, the fi ring processes in use at Ovçular Tepesi
can mainly be studied through the analysis of the pottery
itself. According to the study conducted on the corpus from
Chantier 1 (Gülçur and Marro 2012), about 25% of the CFW
pottery has grey to dark cores, testifying to the practice of pot-
tery fi ring in a reduced atmosphere. Firing conditions, how-
ever, are certainly variable, since an important proportion of
CFW displays fully oxidised cores. Considering the available
evidence, we have no means of establishing whether partly-
oxidised and fully-oxidised cores were retrieved from the
same, or from separate, ceramic batches.
The LCKA pottery on the other hand displays much more
regular features: most vessels (15 out of 18) display a bico-
loured break, corresponding to the contrasting colour scheme
described above. This effect mostly appears in a black/drab or
black/grey contrast, but black/buff and black/beige cores are
also attested. Vessels with a smoked surface (Wares 58 and 59)
also display a bicoloured break, but the lighter colour comes
fi rst: buff/drab or buff/dark grey.
Last but not least, it should be noted that the LCKA ceram-
ics are usually fairly dense, heavy and hard, whereas CFW are
lighter and more breakable. These specifi cities partly derive
from the temper and clay preparation, but also possibly from
the clay itself: experiments conducted by Y. Dede with differ-
ent kinds of clay sampled around Ovçular Tepesi and Duzdağı
has shown that the grey clay from the salt mine of Duzdağı
can be fi red to a fairly hard consistence at a temperature of
only 750°C.24 This clay, which displays the same macroscopic
specifi cities as Wares 58-59 (with inclusions of schist), is never
used in the making of CFW.
To conclude, it is clear from the comparison of the CFW
and LCKA manufacturing processes that the potters shared
a limited number of practices, mostly circumscribed to the
use of slabs and scraping tools.25 In each case, distinct craft-
24. Experiments conducted with an electric oven at the University of Istanbul
during the 2013-2014 winter season.
25. It must be noted however, that surface scraping was in fact widely used
in most Late Chalcolithic communities from the South Caucasus down to
Upper Mesopota mia: this tech nique clearly is part of the L ate Chalcolithic
Livre_BAT.indb 148Livre_BAT.indb 148 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
149
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
ing methods betray different traditions: the CFW and LCKA
corpuses not only differ in shapes,26 paste colours and surface
treatments, they also vary in the choice and preparation of the
clay, in the gestures involved in pottery making, and probably
in the fi ring methods and organisation of pottery production
altogether. In short, the chaînes opératoires involved in both
assemblages have little in common.
The question whether the CFW and LCKA ceramics grew
out of craft or domestic activities is still not answered, but the
available evidence shows that the CFW was produced with min-
imal preparation, as well as little time and work investment. The
production of CFW could be the work of the domestic house-
hold, but alternative models are also possible: the pottery kilns
found within the CFW oikumene, for example at Tell Kosak
Shamali in Syria (Nishiaki et al. 1999), clearly show that the
production of CFW on certain sites was in the hands of arti-
sans. The production of specifi c standardised shapes, such as
Coba-bowls or Wide Flower-pots, also suggests the existence of
a class of artisans, at least on major settlements. Since all the
sites within the CFW oikumene did not have the same function,
as for example with Ovçular, where occupation seems to be dis-
continuous, the organisation of the pottery production on one
site cannot be extrapolated for other sites. The available data
from Ovçular, which testify to small-scale variations in bowl
sizes or jar profi les (Gülçur and Marro 2012: 312-315), recall the
irregularities of domestic work more than craft production, but
these inferences must be accepted with caution.
On the other hand, the making of LCKA ceramics cer-
tainly required fi ne craftsmanship, with the mastering of spe-
cifi c surface treatments (burnished dark surfaces, sometimes
with a graphite sheen imitating metal) and elaborate fi ring
processes (the colour contrast). It also implies fi ne craftsman-
ship in the fashioning of the shapes, which are characterized
by their regularity and relative sophistication: the making of
Kura-Araxes pottery is evidently governed by strict opera-
tional rules, according to an aesthetic canon. Whether the
craftsmanship involved in the making of Kura-Araxes ceram-
ics did require a degree of specialisation, which would imply a
class of artisans, is still an open question.
Whatever the exact organisation of pottery productions at
Ovçular, it is clear from the above that the differences between
technological know-how during the second half of the 5th millennium
(LC 1 and the beginning of LC 2) and tends to disappears afterwards
(Marro 2012).
26. Very rare examples of Kura-Araxes shapes produced in CFW pastes
(Gülçur and Marro 2012: Photo 8) show that interactions between CFW
and LCKA potters probably existed, but nothing is known so far of the
context of these interactions.
the CFW and the LCKA assemblages are not mere variations
in style or the result of functional specifi cities: they refl ect
contrasting concepts in pottery crafting that relate to distinct
technological spheres.
INTERACTIONS BETWEEN COMMUNITIES
AND KURA-ARAXES BEHAVIOURAL PATTERNS
Since the chaînes opératoires at work in the CFW and
Kura-Araxes ceramic productions are conceptually dissimilar,
it now seems evident that the CFW and LCKA vessels could
no t have been prod uce d by t he sa me po tter s. We n eed, however,
to further check this hypothesis by studying the interactions at
work between the Ovçular and the Kura-Araxes components
on the local and the regional scale. In the followings lines, we
will fi rst examine these interactions by focussing on the spatial
distribution of the LCKA artefacts at Ovçular, while the evolu-
tion of settlement patterns over the macro-region will be dealt
with in a second step. The results from this multi-scalar analy-
sis are expected to give some insight into Kura-Araxes behav-
ioural patterns.
Local scale
The context of the LCKA fi nds is of special importance
for understanding the signifi cance of the Kura-Araxes compo-
nents at Ovçular Tepesi during the Late Chalcolithic period.
Again, the LCKA artefacts are very rare, even if we include
the ceramic potsherds that were found in unsealed loci. As
far as sealed contexts are concerned, two of them, Grave 5313
and Pit 11224, are clearly intrusive: they were dug within the
domestic habitat, both partly destroying Floor 11260. They
were both sealed by later domestic structures, respectively
by Wall 5266 and Dial 11164. The intrusive character of the
LCKA potsherds scatter over Floor 5194 on the other hand is
much less evident; its interpretation remains to be established,
especially as it is a single case.
The general impression, however, is that LCKA compo-
nents are not part of daily life at Ovçular Tepesi; they occur
sporadically, and in two cases out of three, they occur as dis-
carded matter (a dead body, some broken pottery in a pit). If
we add the fact that no Kura-Araxes artefact other than pottery
has been found in the Late Chalcolithic habitat, and that there
are no traces of LCKA buildings, the conclusion that leaps to
mind is that Kura-Araxes groups did not live on this settlement
together with the Ovçular community: they mostly frequented
the site when its other inhabitants were away. Thus, we may
Livre_BAT.indb 149Livre_BAT.indb 149 04/12/14 12:3604/12/14 12:36
150 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
venture to suggest that the presence of Ovçular and LCKA
groups on the site was mutually exclusive: we are most prob-
ably faced with two distinct communities that lived together
on the regional level, but not on the same site. This interpre-
tation is further supported by similar artefact combinations
attested in the fi rst half of the 4th millennium at Berikldeebi in
Georgia, as well as in the second half of the 4th millennium at
Arslantepe in Eastern Turkey.
Macro-regional scale
The signifi cance of the fi nds retrieved from Ovçular Tepesi
is highlighted when they are set against the evidence available
at the macro-regional level over the longue durée.
Indeed, there are at least two other cases where the coex-
istence of Kura-Araxes groups with local communities may
be inferred from the presence of Kura-Araxes pottery within
the Late Chalcolithic habitat. The fi rst is Berikldeebi in Shida
Kartli (Georgia), where Kura-Araxes potsherds have been
found in a number of pits mingled with CFW.27 One of these pits
(locus 172) has been dated to the early 4th millennium through
one 14 C reading; it belongs to the earliest Late Chalcolithic
occupation level (Level V.1) excavated on this site.28 The sec-
ond case concerns the site of Arslantepe in Eastern Anatolia,
which yielded a small number of Kura-Araxes vessels in Syro-
Mesopotamian-related contexts dated to the second half of the
4th millennium (Frangipane 2004; Palumbi 2008). It is inter-
esting to note that in both cases, the local Late Chalcolithic
pottery may be considered as an organic development of the
early CFW typical of the ‘Chaff-Faced-Ware’ oikumene, to
which the Ovçular facies also belongs.29
At Berikldeedbi, the Late Chalcolithic period is divided
into two occupation phases: the earliest (Level V.1) mostly
yielded pits, while a large, rectangular mud-brick building,
identifi ed by the excavators as a ‘temple’, was brought to light
in Level V.2 (Dzhavakhishvili 1998: Abb. 1; Kiguradze and
Sagona 2003: Fig. 3.2). This ‘temple’ is itself ridden with
pits belonging either to Level IV (Kura-Araxes) or Level III
(Bedeni), which suggests a period of rupture between Level V.2
and Level IV. The dating of Level IV is not specifi ed, but it
27. Dzhavakhishvil i 1998: 14, Abb. 3-6 and personal comm. from L. Glonti in
1998 (C. Marro).
28. Kiguradze and Sagona 2003: 44 and 93, note 1. Pit 172 is attributed to
Level V.1 in Dzhavakhishvili 1998: 11.
29. On the Chaff-Faced-Ware oikumene, see Marro 2010. The pottery from
Arslantepe Period VII clearly belongs to the Amuq F complex; this is
also the case with some of the vessels brought to light at Berikldeebi:
Dzhavakhishvili 1998: Abb. 3.
corresponds to the fi rst Kura-Araxes settlement (‘Bauphase’)
to be established after the ‘temple’ phase (Level V.2).
The case with Arslantepe is slightly different, since not
only potsherds but also a few whole vessels in situ have been
found within the Late Chalcolithic habitat, both in Period VII
(Amuq F) and Period VIA (Late Uruk-related).30 This situa-
tion is closer to the case recorded at Ovçular Tepesi, where
Late Chalcolithic Kura-Araxes potsherds are attested not just
in pits, but also within the Late Chalcolithic buildings. As in
Ovçular, it suggests that the inhabitants of Arslantepe had
close contacts, but were not actually living together with Kura-
Araxes groups, since no Transcaucasian artefacts apart from
ceramics were found in Periods VII or VIA: presumably, the
two communities were interacting, but not living in the same
place. As in Berikldeebi, this period of regional interaction is
followed at Arslantepe by the establishment of a proper Kura-
Araxes settlement (Period VIB1), after the disappearance of
the Late Chalcolithic component.31
Thus, in two cases out of three, a period of coexistence
between Late Chalcolithic and Kura-Araxes groups is fol-
lowed by the establishment of a Kura-Araxes settlement. At the
regional level, this period is followed by the hegemony of the
Kura-Araxes component in the three regions under study, but it
should be emphasized that this hegemony does not develop con-
comitantly in the Kura, Araxes and Euphrates basins. Following
the evidence from Nakhchivan, the appearance of the fi rst
Kura-Araxes ‘settlements’ in the Middle Araxes region prob-
ably occur red in the fi rst half of the 4th millennium, as suggested
by one unpublished early 14C reading from Kültepe II.32 But in
the Upper Kura basin, the fi rst occurrence of a Kura-Araxes site
dates to the second half of the 4th millennium (Berikldeebi IV),
while in the Upper Euphrates catchment proper Kura-Araxes
settlements are only attested from the beginning of the 3rd
millennium onwards, as illustrated by the sites of Arslantepe
(Period VIB1), Han Ibrahim Şah or Pulur-Sakyol (Marro 2011:
300). It follows from this analysis that the Kura-Araxes com-
ponent cannot be considered as part of the CFW facies during
the Late Chalcolithic, otherwise the disappearance of the latter
would have entailed the end of the former more or less at the
same time. The coexistence of CFW and LCKA pottery tradi-
30. See Frangipane 2004: 49, for the restitution of Temple B with a Kura-
Araxes pithos standing next to a number of Late Uruk related jars.
31. The data published in the preliminary excavation reports from Tepecik
suggest that this site, too, was occupied during the Late Chalcolithic by a
Late Uruk-related community interacting with Kura-Araxes groups at the
regional level. This Late Chalcolithic occupation was then followed by a
Kura-Araxes settlement (Marro 1997: 170-171).
32. Personal observation by Veli Bakhshaliyev.
Livre_BAT.indb 150Livre_BAT.indb 150 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
151
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
tions at Ovçular during the Late Chalcolithic therefore certainly
refl ects the existence of two different communities, one of which
(the Kura-Araxes) outlived the other.
This analysis also implies that the Kura-Araxes phenom-
enon cannot be considered as an organic ‘development’ of the
local CFW cultures. If the Kura-Araxes phenomenon had been
a ramifi cation of the Ovçular, Berikldeebi or Arslantepe facies,
one would again have expected the Kura-Araxes ceramic
trends to take over in these three regions more or less at the
same time, but not with a 500-year or even a 1000-year timelag.
To conclude, the coexistence of CFW and Kura-Araxes
ceramics during the Late Chalcolithic may not be analysed in
terms of cultural polymorphism within a given assemblage, but
should rather be regarded as the expression of some form of
coexistence between distinct cultural groups. In point of fact,
examples of coexistence between Kura-Araxes and various
local groups are not specifi c to the highlands, if one takes the
Palestinian case into consideration: recent research at Khirbet
Kerak (Bet Yerah) has shown how Kura-Araxes-related groups
settled on this site next to local Early Bronze Age communities.33
It is clear that coexistence with other human groups is a major
feature in Kura-Araxes behavioural patterns, even if actual cases
of cohabitation on the same site, as in Khirbet Kerak, are fairly
rare. In many cases however, interactions between Kura-Araxes
and other groups result in a fair degree of acculturation, which,
in ceramic making, is evident in the evolution of decoration and
even modelling techniques (Marro 1997: 95 sq.; 2011: 300).
CONCLUSION
The information given by the data recently retrieved from
Ovçular Tepesi is important for apprehending the formative
processes of the Kura-Araxes phenomenon mostly because it
refutes a number of deep-rooted assumptions: contrary to cur-
rent beliefs, the Kura-Araxes pottery is not an organic develop-
ment of the Late Chalcolithic chaff-tempered ware, whether it
is called Late Sioni or Chaff-Faced-Ware. In fact, as demon-
strated in this paper, the fi rst attestation of Kura-Araxes pot-
tery (the LCKA) at Ovçular Tepesi is contemporary with the
CFW manufactured by the Ovçular potters.
33. Greenberg 2007. In the case of Ovçular Tepesi, the scarcity of LCKA
pottery, as well as the absence of LCKA houses, precludes the possibil-
ity of cohabitation between Ovçular and Kura-Araxes groups on the site
itself. Considering the available evidence, the presence of LCKA pottery at
Ovçular can only be explained by some kind of interaction between the two
groups, which also implies a degree of mobility in the Ovçular community.
Since it appears from the spatial distribution of the fi nds that
no functional specifi city can be attributed to the LCKA vis-à-
vis the CFW, while the technological analysis of these ceramics
clearly points to distinct crafting traditions, it seems clear that
the CFW and LCKA corpuses refl ect the existence of distinct
human groups that live together within the Middle Araxes basin.
We now have suffi cient evidence to claim that the Kura-
Araxes and Ovçular cultures developed on two distinct cul-
tural substrata, each characterized by their own technological
universe. Interestingly enough, the LCKA attested at Ovçular
Tepesi appears in full bloom, without any trace of a genesis, a
fact that questions the very concept of “Proto-Kura-Araxes pot-
tery” that we have sometimes used in our previous work in refer-
ring to early 4th millennium vessels (Marro 2005; Bakhshaliyev
and Seyidov 2013). Considering the available data, the rare
examples of CFW pottery displaying Kura-Araxes traits (or vice-
versa) that we had thought to be ‘prototypes’ of Kura-Araxes
vessels, for example at Ziyaret Tepe (Marro and Özfi rat 2003:
391) in Eastern Anatolia, are in fact most probably the outcome
of hybridization brought about by constant interactions.34
Coming back to the question of the origins of the LCKA,
the existence of a few technological similarities between this
pottery and the CFW (the use of slabs and surface-scraping)
certainly points to a common sphere of development for these
pottery traditions, which probably was located in the high-
lands. But so far, the actual formative processes of the Kura-
Araxes complex itself, when and under which circumstances
they developed, remain as mysterious as ever.
Catherine MARRO
Rémi BERTHON
CNRS, UMR 5133 Archéorient
Maison de l’Orient et de la Méditerranée
7, rue Raulin – 69007 Lyon – FRANCE
catherine.marro@mom.fr
rberthon@mnhn.fr
Veli BAKHSHALIYEV
Azerbaijan National Academy of Sciences, Nakhchivan Branch
Heydar Aliyev Prospekti 76, 7000 Naxçivan – AZERBAIJAN
velibahshaliyev@mail.ru
34. Other cases of ‘Proto-Araxes pottery’ have also been described elsewhere
within the Kura-Araxes purview, as in Didube (Kiguradze and Sagona
2003) or Sos Höyük (Sagona 2000) for instance. The case of Didube is
particularly interesting, since this ‘Proto-Kura-Araxes’ pottery has been
found together with plain, buff-colored, chaff-tempered ceramics typical
of the CFW oikumene, as at Ovçular Tepesi, Areni-I or Berikldeebi. The
general appearance of the Didube ceramics is indeed different from the
LCKA ceramics found either at Ovçular or Areni-I, suggesting that the
formative processes of all these assemblages may be partly unrelated.
In short, it seems now necessary to redefi ne the concept of ‘Proto-Kura-
Araxes ware’, which may be relevant in certain cases but not in others.
Livre_BAT.indb 151Livre_BAT.indb 151 04/12/14 12:3604/12/14 12:36
152 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
APPENDIX - CATALOGUE
A. Surface colour; B. Break colour; C. Temper description, D. Sur-
face treatment, E. Decoration type (if any).
LCKA I (fi g. 2)
1. Locus 5167: A. Ext: black (collar), drab (lip); int: drab; B. Black /
drab; C. Fine mixed temper, almost invisible; D. Ext: irreg-
ular, glossy burnishing; int: burnished (upper collar), smoothed
(lower collar).
2. Locus 5167: A. Ext: black; int: grey; B. Grey/light grey; C. Fine
mixed temper, almost invisible; D. Ext: glossy burnished; int:
matte burnished (collar), smoothed (body).
3. Locus 11224: A. Ext: black; int: drab; B. Grey/light grey, W62;
C. Fine mixed temper, almost invisible; D. Ext: burnished; int:
burnished (collar), plain (body).
4. Locus 11224: A. Ext: drab; int: drab; B. Buff/drab, W58;
C. Coarse mineral temper; D. Ext: ‘smoked’ effect; int: plain.
5. Locus 11224: A. Ext: black; int: dark grey; B. Dark grey/light
grey, W52; C. Coarse mineral temper, with a few carbonate inclu-
sions; D. Ext: burnished; int: matte; E. Painted grey band along
the rim (inside).
6. Locus 5167: A. Ext: black; int: dark grey; B. Dark grey/grey C.
Medium mixed temper with a few white coarse grits; D. Ext: bur-
nished with traces of graphite; int: matte burnished; E. Traces of
graphite over exterior surface; grey painted band inside the bowl
– paint slightly fl aking.
7. Loci 5167 and 5194: A. Ext: black; int: red (inside collar, turns
brownish red inside the body); B. Dark grey/light buff, W52; C.
Fine to medium mineral temper with some fi ne vegetal inclu-
sions; D. Ext: glossy burnished; int: painted; E. Red paint,
scraping traces are visible inside the vessel under the paint.
For nos. 1-2, 6-7, see Marro et al. 2011: 85, pl. V-VI.
LCKA II (fi g. 3)
1. Locus 11224: A. Ext: black; int: dark-grey; B. Black/grey, W62;
C. Fine mixed temper, almost invisible; D. Ext : glossy bur-
nished; int: scraped; F. E. One dimple on the body.
2. Locus 11224: A. Ext: black; int: grey; B. Black, W41; C. Fine to
medium vegetal temper, with a few fi ne mineral inclusions; D.
Ext: glossy burnished; int: plain; E. One dimple on the body.
3. Locus 5167: A. Ext: grey-beige; int: grey? (damaged surface);
B. Grey C. Medium mineral temper with a few fi ne vegetal inclu-
sions; D. Ext: matte burnished with a few vegetal imprints; int:
damaged surface; E. Two dimples over the body.
4. Locus 5167: A. Ext: black; int: grey-beige; B. Black/grey C. Fine
mixed temper, almost invisible; D. Ext: burnished; int: plain;
E. One dimple over the body. For 3-4, see Marro et al. 2011: Pl. VI.
5. Locus 11224: A. Ext: black; int: pink beige; B. Drab/buff, W53;
C. Coarse mineral temper, with a few carbonate inclusions; D. Ext:
unevenly burnished; int: plain; E. One groove along the collar
base.
6. Locus 5313: A. Ext: black; int: drab; B. Dark grey/drab, W41;
C. Fine to medium vegetal temper, with a few fi ne mineral inclu-
sions; D. Ext: lightly burnished; int: unevenly burnished.
7. Locus 5167: A. Ext: ‘smoked pink’; int: beige; B. Grey-beige/
beige C. Medium to coarse mineral temper; D. Ext : smoothed;
int : plain. Marro et al. 2011: Pl. V, 5.
8. Locus 5167: A. Ext: black or brown: varies from potsherd to pot-
sherd; int: pinkish beige; B. Pinkish beige C. Fine to medium
mineral temper with some coarse mineral inclusions; D. Ext: bur-
nished; int: plain.
CFW I (fi g. 4)
1. Locus 1284: A. Ext: beige to buff, mottled; int: beige; B. Drab;
C. Medium vegetal temper with a few white coarse grits; D. Ext:
smoothed; int: plain.
2. Locus 5194: A. Ext: beige; int: beige; B. Drab; C. Medium veg-
etal temper with a few white coarse grits; D. Ext: chaff-faced and
smoothed; int: plain; E. Watery brown-painted festoon motif.
3. Locus 11067: A. Ext: beige, chaff-faced; int: beige; B. Buff;
C. Medium vegetal temper with fi ne to medium white mineral
inclusions; D. Ext: smoothed; int: smoothed. Marro et al. 2011:
Pl. IV, 2.
4. Locus 7017: A. Ext: beige, chaff-faced; int: beige; B. Buff;
C. Medium vegetal temper with fi ne to medium white mineral
inclusions; D. Ext: scraped; int: smoothed; E. Snake-shaped (?)
relief decoration around the collar.
5. Locus 5005: A. Ext: beige, chaff-faced; int : beige ? (covered
with salt concretions); B. Dark core; C. Medium vegetal temper
with some coarse mineral inclusions; D. Ext: comb-scraped; int:
?; E. Row of knobs below the rim.
6. Locus 7026: A. Ext: mottled buff and beige; int: buff; B. Buff,
Ware 2; C. Medium mixed temper; D. Ext: chaff-faced, traces of
vertical comb-scraping under the smoothed surface, int : irreg-
ular smoothed surface.
7. Locus 1204: A. Ext: beige; int: beige; B. Buff; C. Medium veg-
etal; D. Ext/int: chaff-faced, burnished.
CFW II (fi g. 12)
1. Locus 11247: A. Ext: beige, int. beige; B. Beige ; C. Medium
vegetal temper with a few white mineral inclusions, Ware 1;
D. Ext: chaff-faced, comb-scraped and smoothed; Int: smoothed;
E. Annular coil with comb-impressed decoration.
2. Locus 12052: A. Ext: cream to buff; int: buff; B. Brownish black
core; C. Medium vegetal temper with a few white mineral inclu-
sions; D. Ext: Chaff-faced, comb-scraped and smoothed; int:
plain, slightly damaged surface. Traces of fi re at the bottom;
E. Annular coil with impressed decoration (comb impressions);
discontinuous row of knobs around the shoulder. Marro et al.
2011: Pl. I.
Livre_BAT.indb 152Livre_BAT.indb 152 04/12/14 12:3604/12/14 12:36
On the Genesis of the Kura-Araxes phenomenon: New evidence from Nakhchivan (Azerbaijan)
153
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
BIBLIOGRAPHY
ASHUROV S.
2005 An introduction to Bronze Age sites in the Şarur plain.
Archäologische Mitteilungen aus Iran und Turan 37: 89-99.
BAKHSHALIYEV V. and SEYIDOV A.
2013 New findings from the settlement of Sadarak (Nakhchivan-
Azerbaijan). In: Anatolia Antiqua 21: 1-21.
BRONK RAMSEY C.
2009 Bayesian Analysis of Radiocarbon Dates. Radiocarbon 51,1:
337-360.
BURNEY C.
1989 Hurrians and Proto-Indo-Europeans: the ethnic context of
the Early Transcaucasian Culture. In: EMRE K., HROUDA B.,
MELLINK M. and ÖZGÜÇ N. (eds.), Anatolia and the Near East.
Studies in Honor of Tahsin Özgüç: 45-51. Ankara: Türk Tarih
Kurumu Basımevi.
1993 Arslantepe as a Gateway to the Highlands. A note on Periods
VIA-VID. In: FRANGIPANE M., HAUP TM AN N H., LIVERANI M.,
MATTHIA E P. and MELLINK M. (eds.), Between the Rivers and
over the Mountains. Archaeologica Anatolica et Mesopotamica
Alba Palmieri Dedicata: 311-318. Roma: Università di Roma
“La Sapienza”.
DZHAVAKHISHVILI A.
1998 Ausgrabu ngen in Beri kldeebi (Sh ida Kartl i). Georgica 21: 7-20.
FRANGIPANE M.
2004 Arslantepe. Alle origini del potere. Arslantepe, la collina dei
leoni. Milano: Electa.
FRANGIPANE M. and PALUMBI G.
2007 Red-Black ware, pastoralism, trade and Anatolian-Trans-
caucasian interactions in the 4th-3rd Millennium BC. In:
LYONNET B. (éd.), L es cultures du Caucase (VIe-IIIe millénaires
avant notre ère). Leurs relations avec le Proche-Orient : 233-
255. Paris : CNRS Éditions, ERC.
GADZHIEV M.G.
1983 Poseleniya gornogo Dagestana epokhi rannej Bronzy (Early
Bronze Age settlements in the mounta inous region of Dagestan).
In: OSMANOV M.-O. (ed.), Drevnie i srednevekovye poseleniya
Dagestana (Ancient and Middle Age settlements in Daghestan):
6-42. Makhachkala: Institute of History, Language and Lite-
rature of Ac. of Sc. USSR.
GOPNIK H. and ROTHMAN M.
2011 On the High Road: the History of Godin Tepe, Iran. Costa
Mesa, CA: Mazda publishers.
GOSSELAIN O.
2010 Ethnographie comparée des trousses à outils de potiers au sud
du Niger. Bulletin de la Société préhistorique française 107 :
667-689.
GREENBERG R.
2007 Transcaucasian Colors: Khirbet Kerak Ware at Khirbet Kerak
(Tel Bet Yerah). In: LYONNET B. (éd.), Les cultures du Caucase
(VIe-IIIe millénaires avant notre ère). Leurs relations avec le
Proche-Orient: 257-268. Paris : CNRS éditions, ERC.
GÜLÇUR S. and MARRO C.
2012 The view from the North: comparative analysis of the
Chalcolithic pottery assemblages from Norşuntepe and Ovçular
Tep es i. In: MARRO C. (ed.), After The Ubaid: Interpreting
Change from the Caucasus to Mesopotamia at the Dawn of
Urban Civilization (4500-3500 BC) . Proceedings of the Inter-
national Workshop Held at Fosseuse, (29th June-1st July
2009): 305-352. Paris: De Boccard (Varia Anatolica 27).
KAV TA R A D Z E G.
1983 K khronologii epokhi eneolita i bronzy Gruzii (The Chronology
of the Eneolithic and Bronze Age in Georgia). Tbilisi: Metsnie-
reba.
2004 The chronology of the Caucasus during the Early Metal Age:
observations from Central Trans-Caucasus. In: SAGONA A.
(ed.), A View from the Highlands. Archaeological Studies in
Honour of Charles Burney: 539-556. Leuven: Peeters (Ancient
Near Eastern Studies Suppl. 12).
KIGURADZE T. and SAGONA A.
2003 On the origins of the Kura-Araxes cultural complex. In:
SMITH A.T. and RUBINSON K.S. (eds.), Archaeology in the
Borderlands. Investigations in Caucasia and beyond: 38-94.
Los Angeles: University of California, The Cotsen Institute of
Archaeology (Monograph 47).
KOHL P.L.
2009 The Making of Bronze Age Eurasia. Cambridge, New York:
Cambridge University Press (Cambridge World Archaeology).
KUSHNAREVA K.
1997 The Southern Caucasus in Prehistory. Stages of Cultural and
Socio-economic development from the Eigth to the Second mil-
lennium BC. Philadelphia: The University Museum.
MARRO C.
1997 La culture du Haut-Euphrate au Bronze Ancien: essai d’inter-
prétation à partir de la céramique peinte de Keban (Turquie).
Paris : De Boccard et IFEA (Varia Anatolica VIII).
2005 Cultural duality in Eastern Anatolia and Transcaucasia in Late
Prehistory (ca. 4200-2800 BC). Archäologische Mitteilungen
aus Iran und Turan 37: 27-34.
2011 Eastern Anatol ia in the Early Bronze Age. In: STEADMAN S. and
MACMAHON G. (eds.), The Oxford Handbook of Anatolia 10,
000-323 BCE: 290-309. Oxford: Oxford University Press.
2012 Is there a Post-Ubaid Culture? Reflections on the Transition
from the Ubaid to the Uruk Periods along the Fertile Crescent
and Beyond. In: MARRO C. (ed.), After the Ubaid: Interpreting
Change from the Caucasus to Mesopotamia at the Dawn
of Urban Civilization (4500-3500 BC). Proceedings of the
International Workshop Held at Fosseuse, 29th June-1st July
2009: 13-38. Paris: De Boccard (Varia Anatolica 27).
MARRO C. and ÖZFIRAT A.
2003 Pre-classical Survey in Eastern Turkey. First preliminary
Report: the Ağrı Dağ (Mount Ararat) region. In: Anatolia
Antiqua XI: 385-422.
MARRO C., BAKHSHALIYEV V. and ASHUROV S.
2009 Excavations at Ovçular Tepesi (Nakhchivan, Azerbaijan).
First Preliminary Report: the 2006-2008 seasons. In: Anatolia
Antiqua 17: 31-87.
Livre_BAT.indb 153Livre_BAT.indb 153 04/12/14 12:3604/12/14 12:36
154 C. MARRO, V. BAKHSHALIYEV and R. BERTHON
Paléorient, vol. 40.2, p. 131-154 © CNRS ÉDITIONS 2014
2011 Excavations at Ovçular Tepesi (Nakhchivan, Azerbaijan).
Second Preliminary Report: the 2009-2010 seasons. In: Ana-
tolia Antiqua 19: 53-100.
NISHIAKI Y., KOIZUMI T., LE MIÈRE M. and OGUCHI T.
1999 Prehistoric occupa tions at Tell Kosaq Shama li, Upper Euphrates,
Syria. Akkadica 113: 13-68.
PALUMBI G.
2008 Mid-fourth Millennium Red-Black Burnished wares from
Anatolia: a cross-comparison. In: RUBINSON K. and SAGONA
A. (eds.), Ceramics in transitions: Chalcolithic through Iron
age in the Highlands of the Southern Caucasus and Anatolia:
39-58. Leuven: Peeters (Ancient Near Eastern Studies Suppl.
27).
REIMER P.J., BARD E., BAYLISS A. et al.
2013 Intcal13 and Marine13 Radiocarbon Age Calibration Curves,
0-50,000 Years Cal BP. Radiocarbon 55,4: 1869-1887.
RISTVET L., BAXŞELIYEV V. and AŞUROV S.
2011 Settlement and society in Naxçivan: 2006 excavations and
survey of the Naxçivan archaeological project. Iranica Antiqua
XLVI: 1-53.
ROUX V.
2010 Lecture anthropologique des assemblages céramiques. Fonde-
ments et mise en œuvre de l’analyse technologique. Les
Nouvelles de l’Archéologie 119: 4 -9.
2011 An anthropological interpretation of ceramic assemblages: the
foundations and implementations of technological analysis. In:
SCARCELLA S. (ed.), Archaeological Ceramics: A Review of
Current Research: 80-88. Oxford (BAR Int. Ser. 2193).
SAGONA A.
2000 Sos Höyük and the Erzurum Region in Late Prehistory: A
Provisional Chronology for Northeast Anatolia. In: MARRO C.
et HAUP TM AN N H. (éd.), Chronologie des Pays du Caucase
et de l’Euphrate aux IV
e-IIIe Millénaires. Actes du Colloque
d’Istanbul, 16-19 décembre 1998 : 329-373. Paris : De Boccard
(Varia Anatolica XI).
WILKINSON K., GASPARIAN B., PINHASI R., AVETISYAN P., HOVSEPYAN R.,
ZARDARYAN D., ARESHIAN G., BAR-OZ G. and SMITH A.
2012 Areni-1 Cave, Armenia: A Chalcolithic-Early Bronze Age sett-
lement and ritual site in Southern Caucasus. Journal of Field
Archaeology 37,1: 20-33.
Livre_BAT.indb 154Livre_BAT.indb 154 04/12/14 12:3604/12/14 12:36
