Old and New Perspectives
on the Origins of the Halaf Culture
Peter M.M.G. Akkermans
Introduction
The Halaf culture is one of the most intriguing phenomena of the later prehistoric Near East. The broad
outlines of Halaf material culture are well-known by now and it is renowned for its elaborately decorated
pottery and round buildings (the so-called "tholoi"). Other aspects of Halaf society, however, are still poor-
ly understood. The chronology and the origin of Halaf society is a matter of much debate, and likewise our
insights into Halaf economic or social organization are generally based on sheer speculation. The Halaf
culture is said to have appeared rather suddenly somewhere in the later sixth millennium BC. subsequent-
ly spreading over an extensive region (from the Zagros in the east to the Taurus in the west) within a pre-
sumably rather short period of time and in a more or less uniform manner.
The unparalleled wide spread of Halaf material culture is for some regions attributed to an invasion or
migration of Halaf populations (e.g. Davidson 1977), whereas for other areas the role of trade or the 'dif-
fusion* of Halafian traits is emphasized (e.g. Perkins 1949: 44-45: Braidwood and Braidwood 1960: 137ff;
Thuesen 1988: 187). Watson and Leblanc (1973) have associated the extensive distribution of Halaf fea-
tures with the rise of strictly hierarchically organized chiefdoms, which required a considerable amount of
communication between elite groups of different communities and encouraged the sharing and exchange
of status goods of all kinds.
Generally, an increasingly complex social interaction and organization is assumed to have taken defi-
nite shape with the rise of the Halaf culture, and Halaf is considered to have been of crucial importance as
an intermediate stage between village economies on the one hand and early state formation on the other.
In the early decades of this century the Halaf culture attracted considerable attention, as shown by the
large-scale excavations at sites like Nineveh, Arpachiyah and Tepe Gawra, but attention flagged for some
reason or other in the late 1930s, with hardly any investigation into the Halaf until the end of the 1960s.
More recently there has been a renewed interest in the Halaf period, reflected in a series of excavations and
survey projects in Iraq, Syria and Turkey. So far, the emphasis of research has been on the later, mature
stages of the Halaf culture, with only little attention devoted to its rise, an event still largely shrouded in
mystery.
The first evidence of the Halaf culture was found in the plains of Syria, at the sites of Tell Halaf and
Yunus near Carchemish, at the beginning of this century. Nevertheless, the Syrian Jezirah has never been
seriously considered to have formed part of the Halaf homelands. The presence of Halaf in Syria has con-
sistently been treated as a derivation of cultural development in either Anatolia or Mesopotamia: for some,
the Halafians were originally hill people descending from the mountains of southeastern Turkey and invad-
ing the lowland of Syria and Iraq, whereas others reasoned the other way round and regarded the Halafians
as herdsmen of the North Mesopotamian plain who migrated into the neighbouring regions somewhere in
the early fifth millennium or whose cultural traits spread due to trade and 'diffusion'. Undoubtedly, this
preference is largely based upon the state of knowledge; the more is known about a specific region, the
more one is, unconsciously, inclined to emphasize this area's importance in cultural development vis-a-vis
regions less intensively investigated.
Recent discoveries in the Balikh valley of north-central Syria, in the heart of the Syrian Jezirah, have
yielded fascinating new information about the earliest stages of the Halaf culture. Particularly the excava-
tions at Tell Sabi Abyad, the "Mound of the White Boy", have produced new insights and perspectives,
which change the prevailing views on late sixth millennium cultural development in Syria and adjacent
regions. Detailed accounts of the Early Halaf period at Sabi Abyad have been presented elsewhere (see e.g.
Akkermans ed. 1989; Akkermans 1993; Akkermans - Le Miere 1992); here a much lesser-known period
will be shortly commented on, i.e. the so-calied transitional period, intermediate between the pre-HaJaf
Neolithic and the Early Halaf. This transitional period is of crucial importance in understanding the rise of
the Halaf culture (cf. Akkermans - Verhoeven 1995).
43
Subanu VII
The transitional settlement at Tell Sabi Abyad
Tell Sabi Abyad, one of the larger prehistoric sites in the upper Balikh valley of northern Syria, com-
prises an area of over four hectares and rises about ten metres above the surrounding fields. The site is one
of the focal points of excavation of a regionally oriented research project, aimed at clarifying the chronol-
ogy, settlement organization and ecology of later Neolithic society in the Syrian Jezirah. Settlement at Tell
Sabi Abyad seems to have started around 5800/5700 BC (uncalibrated) and lasted for about 800 years. At
present, eleven main building levels are distinguished, which give evidence of a continuous sequence,
without any major hiati (except perhaps between the lower levels 11 and 10). The lower levels 11 to 7 are
all part of the pre-Halaf Pottery Neolithic, the subsequent levels 6 to 4 are considered to represent a tran-
sitional period, whereas the topmost levels 3 to 1 are definitely Early Halaf.
So far, the earliest levels 11 to 7 have only been exposed in a limited area and have yielded sparse
remains of rectangular structures built of pise. The pottery from level 11. directly above virgin soil at a
depth of four metres below the level of the surrounding fields, mainly consisted of coarse vessels, occa-
sionally with loop handles, which closely resemble the ceramics from nearby Tell Assouad and Tell
Damishliyya (cf. Cauvin 1972; Le Miere 1979: Akkermans 1986/87). These ceramics are among the ear-
liest pottery known in Syria and date to the very beginning of the sixth millennium BC. The pottery of the
subsequent levels 10-7. too. mainly consist of irregularly shaped, thick-walled and often burnished ceram-
ics with little variety in shape. Most common are simple plain-rim bowls, hole-mouth pols and jars with
low necks, some of which show incised or impressed patterns of crosshatching. oblique lines or herring-
bones, or bands of dark-red paint. Huskingtrays with shallow ridges or finger-impressed pits upon their
interior bases have been found as well as red-burnished pottery and small quantities of imported, Levantine
Dark-Faced Burnished Ware.
TELL SABI ABYAD
- Fig. 1: Isometric reconstruction of the so-called Burnt Village at Tell Sabi Abyad. SB. 5200 BC.
44
La Djezire el VEuphrate syriens
- Fig. 2: Level 6 (Ihe Burnt Village) house remains at Tell Sabi Abyad.
The topmost levels 6 to 1 have been excavated in a considerable area, i.e. up to 900 square metres.
Earlier excavations (see e.g. Akkermans 1987. 1993: Akkermans ed. 1989: Akkermans - Le Miere 1992)
exposed a series of Early Halaf occupation levels (3-1) which can all be placed in the later sixth millenni-
um, i.e. about 5100-5000 BC (Akkermans 1993. 1991). These strata showed evidence of a loose aggrega-
tion of large rectangular buildings on top. surrounded along the slopes by rectangular annexes and numer-
ous circular structures {tholoi). A most interesting find was the fact that these Early Halaf strata were
directly preceeded by earlier Neolithic occupation levels at the site. No gap in occupation is indicated and
Halaf at Sabi Abyad was apparently the result of a gradual and continuous local process of cultural change.
There can be little doubt that the Syrian Jezirah forms part of the Halaf homelands from its earliest days.
So far. no other excavated site has provided a similar transitional sequence, as all other investigated sites
were newly founded somewhere in the Halaf period.
Interestingly enough, the transition was associated with the appearance of small quantities of finely
painted ceramics, figurines, etc.. which have clear counterparts in the Samarra and. perhaps, the Hassuna
cultural assemblages of north-central Iraq (see e.g. Akkermans 1993: 73). Apparently, some kind of inter-
action seems to have existed between the local late sixth millennium communities of northern Syria and
those located in the regions further east.
Three levels of occupation (6-4) document this period of transition and important innovations, the most
interesting of which is level 6, the so-called Burnt Village at Tell Sabi Abyad (fig. 1). The level 6 settle-
ment was built in terraces along the slope of the mound and consisted of dense clusters of large rectangu-
lar houses surrounded by smaller circular structures, which were all laid in ashes in a violent fire around
5200 BC. The houses were of pise and very regular in layout (fig. 2), although all kinds of renovations and
reconstructions took place in the course of time. Basically, the oblong structures seem to have been divid-
ed in three rows or wings, each of which consisted of a series of small rooms. Some of these houses must
have had 15 or more rooms, all very small and varying in size between three and five square metres. Some
of the rooms had normal doorways, others had entrances of such restricted size that one had to crawl
through them on hands and knees. Occasionally no passage at all was present; apparently these rooms were
accessible from the roof only. Ovens and fireplaces were virtually absent in the houses. These features were
mainly found in the open courtyards surrounding the buildings or constructed in small auxiliary structures.
One of the ovens was beehive-shaped, and stood to a man's height with the roof almost completely intact.
45
Subartu VII
In the open areas surrounding the houses, circular structures up to five metres in diameter were found.
The buildings were covered with white plaster and had small doorways. The largest of these circular build-
ings was divided in a series of smaller compartments, some of which had bins standing on the floor. Tholoi
have always been considered to be one of the main characteristics or Leitfossile of Halaf society, but these
circular structures were definitely not a Halafian 'invention'. Excavations at sites like Tell Hassuna and
Yarim Tepe I already yielded tholoi in early sixth millennium, Hassunan contexts and our work at Sabi
Abyad entirely confirms their root in a pre-Halaf Neolithic tradition. It cannot be denied, however, that for
one reason or another tholoi became extremely important in the Halaf period.
Pottery was locally produced in bulk and mainly consisted of coarse vessels, often oval in shape and
manufactured rapidly with little care. The more carefully finished examples were sometimes incised or
painted. Locally manufactured gray or black-bumished pottery and imported Dark-Faced Burnished Ware
were found in small quantities, as were Samarra imports or site-made Samarra cognates. In addition to the
pottery, thousands of other objects were recovered in situ from the houses of the Burnt Village, like ground-
stone equipment, flint and obsidian tools, clay figurines of both women and animals, clay sealings, stone
labrets which most likely served as lip or ear ornaments, small perforated discs made of sherds, stone axes,
bone implements and ornaments. In some houses these items seem to have been largely concentrated in
one room only, which may have served as a kind of archive. The most exciting finds from the 'archive
rooms' were the clay sealings with or without stamp-seal impressions, of which 300 examples have been
found so far, and the abundantly occurring, simply geometrically-shaped tokens. These sealings and tokens
point towards the existence of a well-developed system of recording and accounting in the second half of
the sixth millennium BC. So far. the earliest sealings in clay were known from the later Halaf period, i.e.
the early fifth millennium BC, as shown by sites like Arpachiyah. Tepe Gawra and Khirbet Derak
(Mallowan - Rose 1935: 98ff; Tobler 1950: 177; Breniquet 1990: 165; Von Wickede 1990: 94ff), but the
Sabi Abyad sealings date five or six hundreds years earlier. In general the Ubaid in the north of Iraq has
received most attention as being the origin of true sealings but this appears to be incorrect.
The sealings (fig. 3a-b) consisted of lumps of clay originally placed on the fastenings of all kinds of
containers, such as ceramic storage vessels, stone bowls, baskets and sacks, or completely covering the
opening of these containers. One of the sealings fitted a small, oval stone bowl; both were found in the
same house but in different rooms. The majority of the sealings was provided with one or more stamp-seal
impressions, depicting a wide variety of designs. Until now, dozens of different motifs have been recog-
nised, most of which are geometric (zigzag lines, triangles, concentric circles, diamonds, crosshatching,
etc.) but naturalistic representations are found as well. Some designs occur only once or twice, others are
found in massive numbers such as the billy-goats or gazelles, depicted in a lively manner with long, curv-
ing horns and with great attention for detail.
- Fig. 3a: Clay sealing with impressions of gazelles or goats, Tell Sabi Abyad, ca. 5200 BC (Max. Length: 4.5 cm);
46
La Djezire el i'Euphrate syriens
- Fig. 3b: Clay sealing with impressions of bukrania (?), Tell Sabi Abyad. ca. 5200 BC (Max. Length: 3.8 cm).
[t appeared that over 60 stamp seals of different shapes and sizes must have been used for sealing pur-
poses. However, not a single one of these stamp seals was found in the houses of the Burnt Village (so far
they have appeared only in later levels at Tell Sabi Abyad). Most likely, the actual sealing was not done at
Tell Sabi Abyad but carried out somehere else; in this view, the sealings themselves arrived at Sabi Abyad
as parts of trade products. The numerous clay tokens found in association with the sealings seem to sup-
port this view. The tokens were all very small and had simple geometric shapes (balls, cylinders, discs,
cones). Most likely they acted as counting devices expressing the quantities of objects exchanged or oth-
erwise deployed.
The existence of extensive exchange networks in the late sixth millennium BC seems to be beyond
doubt: in the case of Sabi Abyad it appeared that, apart from various products locally won, Samarra-like
pottery was brought in from eastern Syria or north-central Iraq, that so-called Dark-Faced Burnished Ware
came from western Syria and that obsidian, basalt and other precious stones were obtained from Anatolia.
The Burnt Village seems to have been part of extensive networks of long and short-distance exchange.
Moreover, it seems that the inhabitants of the Burnt Village were able to mobilize and exploit these exter-
nal resources regularly and to a considerable extent, beyond the level of incidental transactions. It is not
yet clear what caused the final destruction of the level 6 settlement at Sabi Abyad. Perhaps it was merely
a tragic accident but it may also have been due to a deliberate act of violence imposed by outsiders
(although there is at present no evidence pointing in this direction). The Burnt Village was soon replaced
by a new settlement partly founded upon the still standing remains of the Burnt Village. However, occu-
pation at Sabi Abyad now seems largely to have lost its aggregative character and instead gradually became
more open and dispersed in nature, with finally one or two rectangular buildings on the top of the mound,
surrounded by auxiliary structures on the slopes. The architecture ascribed to the Early Halaf level 3 was
monumental in outline (fig. 4), with extensive stone constructions and a large, probably two-storey build-
ing supplied with buttresses and niches along the exterior facade (see e.g. Akkermans - Le Miere 1992).
This building, which stood in an isolated and almost fortified manner, has little in common with the very
regularly built houses of the Burnt Village. The Early Halaf settlement most likely consisted of four or five
such extended house-complexes, and, despite the considerable size of the mound, may have comprised a
population of estimatedly 30 to 50 individuals only (cf. Akkermans 1993: 186ff).
47
Subartu VII
- Fig. 4: Early Halaf building at Tell Sabi Abyad. i u. 5200.
Other major changes in local Neolithic society can be associated with the full onset of the Halaf. such
as an increase in population (after a millennium or so of population decline) and a twofold development
of both subsistence strategies and settlement organization. On the one hand we now find relatively large
and permanently occupied agricultural settlements of long duration, whereas on the other hand there is an
increasing number of short-term, basically pastoral sites of a more mobile nature (ibid.). At the same time
finely painted ceramics are found in rapidly increasing numbers. Al Sabi Abyad fine-ware pottery was first
introduced in minute quantities in the transitional level 6. But whereas this pottery hardly comprises 6%
of the pottery assemblage in these transitional times, it appears that up to 809c of the ceramic inventory is
busily painted in the Early Halaf period. The present stratigraphic evidence and the available radiocarbon
dates from Sabi Abyad suggest that the various changes took place within a remarkably short period of
time, i.e. a few generations only (ca. 5200-5100 v.Chr.).
The beginnings of Halaf in a wider perspective
In the light of the above, it is clear that Tell Sabi Abyad provided exciting new data on the origins and
development of the Halaf culture. The excavations gave evidence of an important series of Early Halaf
occupation levels, which, as argued in detail elsewhere (Akkermans 1993: 128ff), is definitely not the same
as the 'Early Halaf found before at sites like Arpachiyah or Tell Aqab; the Sabi Abyad Early Halaf pre-
cedes, at least partially, that of the other excavated sites. Halaf society at Sabi Abyad gradually emerged
out of an earlier Neolithic complex, most likely related to the so-called Altmonochrome of Tell Halaf.
around 5200/5150-5100 BC.
48
La Djezire et I'Euphraie syriens
Commonly Halaf is said lo begin around the middle of the sixth millennium but this view appears to
be incorrect. The series of radiocarbon dates available now from Tell Sabi Abyad makes it clear that the
often-mentioned date of 7570 ± 35 BP from Tell Halaf, said to date the transition from the pre-Halaf
Aitmonochrone-ttaiditioa to Halaf, is much too old and has to be rejected (Akkermans 1991). At present the
stratigraphic sequence at Sabi Abyad shows no break, and the material culture of the series of occupation
levels gives evidence of close coherence as well. In this respect, it is strongly felt that Halaf in the Balikh
basin (and in northern Syria in general) is the ultimate result of a long and locally founded sequence of
Neolithic development. A similar conclusion can be reached from the survey results in the Balikh region
(cf. Akkermans 1993: 138ff). The present survey data suggest that the early half of the sixth millennium
BC was characterised by a trend towards site abandonment and contraction of occupation to a few larger
mounds, but that the lands earlier deserted were re-colonised around the middle of the sixth millennium or
shortly afterwards. Our excavations at Tell Sabi Abyad suggested that local Neolithic society now incor-
porated more and more external, foreign influences and sought access to wider, interregional modes of
communication and exchange of goods, beliefs and ideas. It is recalled that Sabi Abyad was undoubtedly
part of a wide network of exchange relationships and that products were brought in from far-away regions,
such as Dark-Faced Burnished Ware from western Syria or Cilicia and obsidian from central Anatolia. The
appearance of Halaf is part (or. perhaps, the result) of this trend towards intensive interregional contact.
In the early days of Near Eastern prehistory. Mallowan's suggestion that the Halaf culture originated
from the mountainous regions of southeastern Turkey found wide acceptance, this on the basis of some
assumed relationships in decoration between Halaf ceramics and the pattern-burnished wares found along
the Taurus range (Mallowan 1936: 4; Mallowan - Rose 1935: 177). An Anatolian origin is also favoured
by Mellaart (1970) and Bogoslavskaja (1972), who both stress the resemblances between the material cul-
ture found at Syro-Mesopotamian Halaf sites and at the early sixth millennium settlements of fatal Hiiyuk.
Hacilar and Can Hasan in south-central Anatolia. Mellaart (1970: 281) points out that "There are many fea-
tures that seem to be Anatolian rather than Mesopotamian: the emphasis on bull, ram, leopards, birds in
association with a Mother Goddess, the prominence of obsidian, metalworking and weaving, the natural-
istic animal figures, the 'St Andrew's' crosses and the 'double-axe' motifs in pottery and stone, the high
quality and the burnishing, the use of lustrous paint, etc.. are all features that we have come to recognize
as familiar from southern Anatolia since the neolithic period in the seventh millennium". However, apart
from some marginal or incidental parallels, some of which are even not limited to Anatolia either in time
nor in place, it seems clear nowadays that an origin of Halaf society west of the Taurus range finds no sup-
port. The material assemblage of. for example. Can Hasan, which must have been partially contemporary
with Halaf (Can Hasan 2B was destroyed by fire at around 4900 BC), is of a wholly different nature than
that of Halaf communities in Syria or Iraq. The same holds for fatal Hiiyuk West or Hacilar. each of which
represent regionally bound Anatolian communities which have little or nothing to do with the Halaf.
Actually, Mellaart revoked his view in the mid-1970s, noting that it was based upon a faulty chronology
and suggesting that parallels were perhaps based upon similarities in basic subsistence strategies (Mellaart
1975: 169).
Nevertheless, the general idea that the Halafians were originally hill people was still held, although
their region of origin was now situated somewhere betw een the mountains of Lake Van and the north Iraqi
steppe. Mellaart took up Diane Kirkbride's suggestion that the burnished pottery of early sixth millennium
Umm Dabaghiyah represented an early stage of Halaf (Kirkbride 1972) and suggested a symbiosis with the
Hassuna settlements of northern Mesopotamia, thereby trying to link sites like Umm Dabaghiyah or
Hassuna itself with an obsidian trade controlled by Halafians.
Unfortunately, little is known about the sixth millennium occupation in the vast piedmont of south-
eastern Anatolia. So far, most of the Neolithic sites that have been excavated, like Cayonii, Hayaz Huyiik
or Cafer Hiiyuk. are part of the Pre-Pottery Neolithic B cultural tradition and were either wholly abandoned
towards the end of the seventh millennium BC or gave evidence of poorly preserved Pottery-Neolithic set-
tlement, separated from the lower aceramic phases through a hiatus in occupation. Generally the pottery
assemblages of sixth millennium southeastern Turkey, like the one found at Kumartepe on the Euphrates
north of Urfa (Roodenberg et at. 1984), are considered to be local variants of the Levantine Dark-Faced
Burnished Ware tradition. Actually, regional variation and largely independent adaptation, even on a micro-
scale, may easily have developed in the often remote and barely accessible valleys and plains of eastern
Anatolia. On the other hand it seems clear that the Anatolian lands were not wholly outside the mainstream
of later sixth and early fifth millennium cultural development. Numerous Halaf settlements can be found
in the region, indicating that Anatolia, too, formed pan of this widespread tradition. Two observations
should, however, be taken into account. First, the present survey evidence (e.g. Dbnmez - Brice 1949;
Burney 1958; Brown 1967; Ozdogan 1977; Russell 1980; Algaze 1989) and the excavations at sites like
Gerikihaciyan (Watson and Leblanc 1973, 1990), Cavi Tarlasi (e.g. Von Wickede - Herbordt 1988) and
Tilki Tepe (Korfmann 1982) suggest that the Anatolian sites all fit into the later stages of the Halaf period;
49
Subartu VII
Halaf at these sites apparently arrived in a fully developed, mature state (see also Davidson 1977). So far,
the sole exception may be Sakce Gbzii west of modern Gaziantep, where levels I-Il have yielded a few
Early Halaf and Samarra-like ceramics among the vast bulk of plain, grey-black burnished and incised pot-
tery (cf. Garstang 1908; Garstang et at. 1937; Du Plat Taylor el a\. 1950). These finds suggest that the ear-
liest Halaf is not completely absent from the Anatolian piedmont. Second, it appears that at least some of
the sites in the area do not represent true Halaf settlements but should be considered as peripheral com-
munities which adopted Halaf traits to a limited extent. Girikihaciyan and Tilkitepe, and definitely the sites
of the Elazig-Keban area, are each Halaf-influenced or Halaf-derived communities on the boundaries of
the Halaf region of distribution.
Similarly, Halaf seems to have been an intrusive feature in the coastal parts of Anatolia and the lands
up to Gaziantep; here Halaf 'presence' is solely indicated by ceramics, some of which seem to represent
genuine import products, others local imitations (e.g. Braidwood - Braidwood I960). In these Mediterranean
areas the Dark-Faced Burnished Ware tradition had a firm footing. Radiocarbon samples from Ramad III
and Labwe I1A suggest that the earliest known Dark-Faced Burnished Ware (Amuq A) dates from around
5900 BC, and this pottery remained in use (although in a continuously changing form) for almost a mil-
lennium to come. Le Miere and Picon (1987) have brought forward some evidence that there was wide-
spread traffic in Dark-Faced Burnished Ware already by the early sixth millennium BC. this pottery being
found as far as Bouqras in eastern Syria. More recently true Dark-Faced Burnished Ware pottery was found
in a sixth millennium context at Tell Sabi Abyad in north-central Syria, these ceramics undoubtedly repre-
senting imports into this region as well (cf. Le Miere 1989).
Mallowan's idea of Anatolia, or perhaps even coastal Syria, as the heartland of Halaf society lost its
attraction to most scholars in the late 1940s, following the publication of Ann Perkins' The Comparative
Archaeology of Early Mesopotamia. Perkins (1949: 43-45) argued that Halaf most likely originated from
the Mosul region of northern Iraq, and spread into neighbouring regions in a diffusionist and ever-weak-
ening manner. Perkins based her conclusions on the supposedly more coherent and well-developed nature
of the Mosul ceramics (Arpachiyah, Tepe Gawra) when compared with those found in the west (Tell Halaf,
Chagar Bazar, Tell Brak). Moreover, she suggested that this Mosul pottery was found in a 'logical'
sequence, with a constant progression in technical qualities, forms and decoration.
The Mesopotamia-oriented approach has dominated research into the Halaf problem up to the present
day; Halaf is virtually always reviewed in association with the Mesopotamian Hassuna and Samarra cul-
tural traditions (see e.g. Mellaart 1975; Davidson 1977; Watkins - Campbell 1987). Actually, the Samarra
culture does seem to have affected the development of Halaf to a considerable extent. True Samarra ceram-
ics or Samarra cognates are rather commonly found within the Sabi Abyad sequence. The present data sug-
gest that Samarra partially precedes the rise of Halaf but the Middle or Classical Samarra (Tell as-Sawwan
levels IIIA-V), to be dated around 5300-5000 BC. is contemporary with the transitional and Early Halaf
levels 6 to 1 at Sabi Abyad. Interestingly, Samarra traits at Sabi Abyad are recognisable only in the ceram-
ic inventory of the transitional levels 6 to 4 and seem to disappear with the occurrence of true Halaf at the
site. In this respect it seems that the Samarra impact on the late sixth millennium communities of northern
Syria, or the arrival of Samarra derivations and 'impulses' in these peripheral regions, was of restricted
duration (the transitional period at Tell Sabi Abyad lasted hardly one hundred years or so, ca. 5200-5100
BC). Moreover, it appears that the rise of Halaf entirely replaced the earlier system of interregional inter-
action; indigenous Halaf traits now superseded the 'foreign' Samarra features. Evidently this development
is not merely a matter of stylistic change of ceramics and the like, but represents an era of profound
changes in the organization of space and society. The Halaf ideological world may have had little in com-
mon with the Samarra universe; considerable differences in material culture may reflect different modes of
perception and traditional beliefs. For example, the architecture found at the Halaf sites of Arpachiyah,
Cavi Tarlasi or Tell Sabi Abyad strongly deviates from the Samarra building traditions found at sites like
Tell as-Sawwan, Chogha Mami or Tell Songor. Generally, the Samarra architecture seems to be much more
uniform and built according to strict, tradition-bound rules. Whereas the maintenance of the Halaf or
Samarra tradition of house building will to some extent merely have been due to custom, for another part
certain structuring principles must have been at work which serve to organize and bind together a society
by creating a past shared by all members of a society. By acting within the traditional frame, one acknowl-
edges and pays respect to the past and ultimately to the present, or, in other words: one accepts the ruling
ideological system (see e.g. Miller - Tilley eds. 1984). Thus the houses were not simply raised according
to the wishes of their individual builders but according to the rules of society; they express part of these
societies' world-view (cf. Fritz 1978; Kus 1981; Hillier - Hanson 1984; Kent ed. 1990).
Whereas the Samarra is directly relevant to the rise of Halaf society, this is not yet clear for the Hassuna
culture. Hassuna is mainly found in northern Iraq but recent Japanese excavations at Tell Kashkashok II,
situated northwest of Hassake in Syria, seem to have yielded a Hassunan site with many of the character-
istics of its Mesopotamian counterparts (Matsutani 1991). Earlier, Hassuna-like materials were reported
50
La Djezire el i'Euphrale syriens
from Bouqras in eastern Syria (Le Miere 1983. 1986). Pottery at Bouqras appeared in mixed deposits in
the 'virgin-soil squares* from level 7 onwards but the larger part of it was found in the topmost south-
western part of the site (Le Miere 1986). A number of radiocarbon dates suggest that the pottery found in
the upper levels 7-1 of the 'virgin-soil squares' can be dated at around 6100-5900 BC, whereas the ceram-
ics from the southwestern area are of considerably later date, i.e. around 5600/5500 BC (ibid.: 120-21,253-
57). It has repeatedly been tried to link Halaf and Hassuna in Mesopotamia but any such proposed rela-
tionships are based on incorrect chronological assignments. Watkins and Campbell (1987: 433ff) have
made it clear that, so far, none of the sites excavated in northern Mesopotamia has shown evidence of a
contemporaneous appearance of both cultural complexes, or even of a transition between Hassuna and
Halaf. Actually, at all the excavated sites yielding both Hassuna and Halaf levels of occupation, like Tell
Hassuna. Hajjiluk. Tell Azzo. Nineveh, Yarim Tepe I and Kharabeh Shattani, it appeared that the Halaf stra-
ta belonged to the later stages of the Halaf period, clearly indicating a hiatus in occupation (ibid.: 434-36).
Evidently this does not necessarily imply that the earliest Halaf, as known from Sabi Abyad, is wholly
absent in northern Iraq but simply that it has not yet been found in a stratified context. A noteworthy excep-
tion may be the recently discovered site of Khirbet Garsour, which is said to have yielded "pottery dating
to the very end of the Hassuna period with clear signs of the developing Halaf style" (Campbell 1992: 183;
see also Wilkinson 1990). Another Iraqi site. NJP 72, revealed pottery which seems to be a somewhat later
development of the Khirbet Garsour ceramics and which closely resembles the ceramics of our Tell Sabi
Abyad (ibid.).
The identification of these transitional wares over an apparently large area suggests a widespread shar-
ing of cultural traits and a considerable degree of interregional communication and interaction, prior to the
full onset of the Halaf around 5100/5000 BC. The Sabi Abyad sequence has already made it clear that the
prevailing view that the North Mesopotamian plain is the heartland of Halaf society and that Halaf subse-
quently spread from northern Iraq into the surrounding regions (including Syria) in the early fifth millen-
nium can no longer be maintained. The occurrence of transitional ceramics in northern Iraq confirms this
picture and underlines that there is no such east-west movement of Halafian traits or people; all ingredi-
ents leading towards the Halaf culture seem to have been locally present already over much of what
became the Halaf range of distribution (cf. Campbell 1992: 183).
However, if indeed the origins of the Halaf culture have to be sought in various localised Neolithic cul
tures, new issues are raised. In particular it seems very questionable whether various distinct material cul-
tures, like the Altmonochrome-tradition in the west, which underlies the Sabi Abyad sequence (Akkermans
1993), and the Hassuna-tradition in the east, which seems to underlie the Khirbet Garsour/NJP 72
sequence, could both independently lead to the same outcome, i.e. Halaf society. Most likely each of these
traditions can be associated with different societal beliefs, ideas and regulations; is Halaf then merely a
kind of cultural umbrella or fashion package affecting numerous ethnic and social groups? A major and
unsolved problem in this respect is how to define our Halaf culture', apart from a summary description of
patterns of spatial and chronological variation in artefacts. More and more a distinct regional variation in
Halaf material culture is recognised (e.g. Davidson 1977 - Hijara 1980, contra the unitarian claims of
Leblanc and Watson 1973); instead of a neatly defined Halaf culture we find an increasing variety of cross-
cutting patterns and features and. most likely, overlapping social networks of varying intensity (cf. Mann
1986; Shennan 1989). It then seems clear that there is no such thing as "Hassuna", "Altmonochrome" or
"Halaf" in the sense of monothethic, independent entities firmly existing beside each other, but a series of
interwoven networks of often temporary and fluctuating groupings and alliances. Fried (1967, 1968) has
already argued that it was not until the rise of complex society, with its expanding empires, that these
alliances were treated and administered as 'ethnic' entities. Evidently ethnicity does not simply equal cul-
ture or people but, in the words of Stephen Shennan (1989: 14), refers to "a self-conscious identification
with a particular social group at least partially based on a specific locality or origin". At present, it cannot
be established with certainty whether or not different ethnic groups were involved in the Halaf phenome-
non, but it seems clear that social interaction at the end of the sixth millennium BC occurred on a much
larger scale than in previous times and that there was an ongoing trend towards interdependency. The Burnt
Village at Sabi Abyad. with its abundant 'foreign' traits and indicators of both short and long-distance
exchange, provides ample evidence in this direction. Actually it is precisely this increasing regional coop-
eration and, perhaps, social integration, associated with the emergence of extensive exchange networks,
which may largely explain the rapid rise of Halaf as a distinct, widespread style in ceramics and other mate-
rial-culture traits; if used over large distances and perhaps affecting numerous ethnic groups, the accep-
tance of the Halaf style may have been a perfect instrument to transmit clear symbolic messages about con-
scious cultural affiliation or social identity, beyond the individual. Style in this emblemic sense (Wiessner
1983) mainly serves in highly constrained and competitive contexts; in non-competitive and symbiotic
group relationships, there is little need to emphasize social group identities and boundaries by stylistic
behaviour (Hodder 1979: 446ff, 1981: 82ff). Significantly, the widespread success of the Halaf followed a
period of considerable instability at least in northern Syria lasting for some centuries, as indicated by a
51
Subartu VII
widespread abandonment of settlements and depopulation of almost entire regions (the so-called hiatus
palestinien; see e.g. Mellaart 1975: 67-69; Moore 1983: 99: Kohler-Rollefson 1988: 87-93 and the various
contributions in Bottema et at. 1990). Acceptance of the Halaf style, i.e. affiliation and conformation to the
new cultural norm, may have yielded considerable social or economic benefit to numerous communities
over a large part of the northern Fertile Crescent.
Bibliography
Akkermans 1986/87: Akkermans P.M.M.G., "Excavations at Tell Damishliyya, A Neolithic Settlement
in the Balikh Valley. Northern Syria", Annates Archeoiogiques Arabes
Syriennes 36/37: 40-66.
Akkermans 1987: Akkermans P.M.M.G., "A Late Neolithic and Early Halaf Village at Sabi Abyad,
Northern Syria", Paleorient 13: 23-40.
Akkermans 1993: Akkermans P.M.M.G., Villages in the Steppe- Later Neolithic Settlement
and Subsistence in the Balikh Valley, northern Syria. Ann Arbor.
International Monographs in Prehistory.
Akkermans 1991: Akkermans P.M.M.G., "New Radiocarbon Dates for the Later Neolithic of Northern
Syria", Paleorient 17/1: 121-125.
Akkermans 1989: Akkermans P.M.M.G. (ed.), Excavations at Tell Sabi Abyad - Prehistoric
Investigations in the Balikh Valley, Northern Syria, Oxford, BAR Int. Series
468.
Akkermans - Le Miere 1992: Akkermans P.M.M.G. & Le Miere M., "The 1988 Excavations
at Tell Sabi Abyad, a Later Neolithic Village in Northern Syria".
American Journal of Archaeology 96: 1-22.
Akkermans - Verhoeven 1995: Akkermans P.M.M.G. & Verhoeven M.. "An Image of Complexity:
The Burnt Village at Late Neolithic Sabi Abyad, Syria", American Journal
of Archaeology 99: 5-32.
Algaze 1989: Algaze G., "A New Frontier: First Results of the Tigris-Euphrates Archaeological
Reconnaissance Project, 1988", Journal of Near Eastern Studies 48:
241-281.
Bogoslvskaja 1972: Bogoslvskaja N.F., "On the Problem of the Origin of the Halaf Culture". Sovetskaa
Arkheologia 2: 3-16 (in Russian).
Bottema et al. 1990: Bottema S. et al, (eds.), Man's Role in the Shaping of the Eastern Mediterranean
Landscape, Rotterdam/Brookfield, A.A. Balkema.
Braidwood - Braidwood 1960: Braidwood R.J. & Braidwood L.S. Excavations in the Plain of Antioch I,
Chicago. The University of Chicago.
Breniquet 1990: Breniquet C, La disparition de la culture de Halaf ou les origines de la culture d'Obeid
dans le nord de la Mesopotamia Paris, Universite de Paris I.
Brown 1967: Brown G.H., "Prehistoric Pottery from the Antitaurus", Anatolian Studies 17: 123-164.
Burney 1958: Burney C.A., "Eastern Anatolia in the Chalcolithic and Early Bronze Age", Anatolian
Studies 8: 157-209.
Cauvin 1972: Cauvin ]., "Sondage a Tell Assouad (Djezireh, Syrie)", Annates Archeoiogiques Arabes
Syriennes 22: 85-88.
Campbell 1992: Campbell S.. "The Halaf Period in Iraq: Old Sites and New". Biblical Archaeologist 55:
182-187.
Davidson 1977: Davidson T.E., Regional Variation w ithin the Halaf Ceramic Tradition, Edinburgh,
University of Edinburgh.
Donmez - Brice 1949: Donmez A. & Brice W.C., "The Distribution of Some Varieties of Early Pottery
in South-Eastern Turkey", Iraq 11: 44-54.
Du Plat Taylor et al. 1950: Du Plat Taylor J. et al., "The Excavations at Sakce Gbzii", Iraq 12: 53-138.
Fried 1967: Fried M.H., The Evolution of Political Society, New York, Random House.
52
La Djezire et I'Euphrate syriens
Fried 1968: Fried M.H.. "On the Concepts of 'Tribe' and 'Tribal Society'", in Helm J. (ed.). Essays
on the Problem of Tribe, Seattle. University of Washington Press: 3-22.
Fritz 1978: Fritz J.M.. "Paleopsychology Today: Ideational Systems and Human Adaptation
in Prehistory", in Redman C.L. et al. (eds.). Social Archeology. Beyond
Subsistence and Dating. New York, Academic Press: 37-59.
Garstang 1908: Garstang J., "Excavations at Sakje-Geuzi in North Syria: Preliminary Report for 1908",
Annals of Archaeology and Anthropology 1: 97-117.
Garstang et al. 1937: Garstang J. et aL, "Third Report on the Excavations at Sakje-Geuzi. 1908-1911",
Annals of Archaeology and Anthropology 24: 119-140.
Hijara 1980: Hijara I.. The Halaf Period in Northern Mesopotamia, London. University of London.
Hillier - Hanson 1984: Hillier W. & Hanson J., The Social Logic of Space. Cambridge, Cambridge
University Press.
Hodder 1979: Hodder I., "Social and Economic Stress and Material Culture Patterning", American
Antiquity 44: 446-454.
Hodder 1981: Hodder I.. "Society. Economy and Culture: An Ethnographic Case Study amongst the
Lozi", in Hodder I. et al. (eds.), Patterns of the Past, Cambridge. Cambridge
University Press: 67-95.
Kent 1990: Kent S. (ed.), Domestic Architecture and the Use of Space, Cambridge. Cambridge
University Press.
Kirkbride 1972: Kirkbride D., "Umm Dabaghiyah 1971: a Preliminary Report", Iraq 34: 3-15.
Kohler-Rollefson 1988: Kohler-Rollefson I„ "The Aftermath of the Levantine Neolithic Revolution in
the Light of Ecological and Ethnographic Evidence". Paleorienl 14: 87-93.
Korfmann 1982: Korfmann M., Tilkitepe: die ersten Ansdtze prdhistorischer Forsclmng in der ostlichen
Tiirkei. Tubingen, Verlag Ernst Wasmuth.
Kus 1981: Kus S.. "The Context of Complexity", in van der Leeuw S.E. (ed.). Archaeological
Approaches to the Study of Complexity. Amsterdam, Universiteit van
Amsterdam: 197-227.
Leblanc - Watson 1973: Leblanc S.A. & Watson P.J., "A Comparative Statistical Analysis of Painted
Pottery from Seven Halafian Sites", Paleorienl 1: 117-133.
Le Miere 1979: Le Miere M.. "La ceramique prehistorique de Tell Assouad, Djezireh, Syrie",
Cahiers de I'Euphrate 2: 4-76.
Le Miere 1983: Le Miere M.. "Pottery and White Ware", in Akkermans P.A. et al., "Bouqras Revisited:
Preliminary Report on a Project in Eastern Syria", Proceedings of the
Prehistoric Society 49: 351-354.
Le Miere 1986: Le Miere M.. Les premieres ceramiques du Moyen-Euphrale. Lyon. Universite Lumiere-
Lyon 2.
Le Miere 1989: Le Miere M., "Clay Analyses of the Prehistoric Pottery: First Results", in Akkermans
P.M.M.G. (ed.). Excavations at Tell SabiAbvad. Oxford. BAR Int. Series
468: 233-235.
Le Miere - Picon 1987: Le Miere M. & Picon M., "Productions locales et circulation des ceramiques
au VIeme millenaire. au Proche-Orient". Paleorient 13/2: 137-151.
Mallowan 1936: Mallowan M.E.L., "The Excavations at Tell Chagar Bazar and an Archaeological
Survey of the Habur Region 1934-5", Iraq 3: 1-87.
Mallowan - Rose 1935: Mallowan M.E.L. & Rose J.C., "Excavations at Tell Arpachiyah, 1933", Iraq 2:
1-178.
Mann 1986: Mann M„ The Sources of Social Power, Cambridge, Cambridge University Press.
Matsutani 1991: Matsutani T. (ed.). Tell Kaskashok - The Excavations at Tell No. II, Tokyo,
The University of Tokyo.
Mellaart 1970: Mellaart J., "The Earliest Settlements in Western Asia from the Ninth to the End
of the Fifth Millennium BC", Cambridge Ancient History I: 248-303.
53
Subartu VII
Mellaart 1975: Mellaart J., The Neolithic of the Near East, London, Thames and Hudson.
Miller - Tilley 1984: Miller D. & Tilley C. (eds.). Ideology, Power and Prehistory, Cambridge,
Cambridge University Press.
Moore 1983: Moore A.M.T., "The First Farmers in the Levant", in Cuyler Young T. Jr. et al. (eds.J,
The Hilly Flanks and Beyond, Chicago, The University of Chicago: 91-111.
Ozdogan 1977: Ozdogan M., Lower Euphrates Basin 1977 Survey, Istanbul, Lower Euphrates Project
Publications.
Perkins 1949: Perkins A.L.. The Comparative Archaeology of Early Mesopotamia. Chicago,
The University of Chicago.
Roodenberg et al. 1984: Roodenberg J.J. et al., "Surveys and Soundings at Kumartepe, An Interim
Report", Anatolica 11: 1-16.
Russel 1980: Russel H.F., Pre-Classical Pottery of Eastern Anatolia. Oxford. BAR Int. Series 85.
Sherman 1989: Shennan S.J, "Introduction: Archaeological Approaches to Cultural Identity", in
Shennan S. (ed.), Archaeological Approaches to Cultural Identity, London.
Unwin Hyman: 1-32.
Thuesen 1988: Thuesen L, Hama: The Pre- and Protohistoric Periods, Copenhagen. Nationalmuseet.
Tobler 1950: Tobler A.J., Excavations at Tepe Gawra (Vol. 2), Philadelphia, University of Pennsylvania
Press.
von Wickede 1990: von Wickede A., Prahistorische Stempeiglyptik in Vorderasien. Miinchen. Profil
Verlag.
von Wickede - Herbordt 1988: von Wickede A. und Herbordt S., "Cavi Tarlasi: Bericht iiber die
Ausgrabungskampagnen 1983-1984", Istanbuler Milteilungen 38: 5-35.
Watkins - Campbell 1987: Watkins T. & Campbell S., "The Chronology of the Halaf Culture".
in Aurenche O. et al. (eds.). Chronologies in the Near East. Oxford.
BAR Int. Series 379: 427-464.
Watson - Leblanc 1973: Watson P.J. & Leblanc S.A., "Excavation and Analysis of Halafian Materials
from Southeastern Turkey: the Halafian Period Reexamined", Paper
presented at the 72nd Annual Meeting of the American Anthropological
Association, New Orleans, Nov.28 - Dec.2, 1973.
Watson - Leblanc 1990: Watson P.J. & Leblanc S.A., Gerikihaciyan: a Halafian Site in Southeastern
Turkey, Los Angeles, UCLA.
Wiessner 1983: Wiessner P., "Style and Social Information in Kalahari San Projectile Points". American
Antiquity 48: 253-276.
Wilkinson 1990: Wilkinson T.J., "The Development of Settlement in the North Jezira between the 7th
and the 1st Millennia BC", Iraq 52: 49-62.
54
PETER M M G AKKERMANS