You are on page 1of 41

The Burnt House at Arpachiyah: A Reexamination Author(s): Stuart Campbell Reviewed work(s): Source: Bulletin of the American Schools

of Oriental Research, No. 318 (May, 2000), pp. 1-40 Published by: The American Schools of Oriental Research Stable URL: . Accessed: 27/01/2012 12:41
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact

The American Schools of Oriental Research is collaborating with JSTOR to digitize, preserve and extend access to Bulletin of the American Schools of Oriental Research.







STUART CAMPBELL Department of Art History and Archaeology University of Manchester Oxford Road Manchester M13 9PL Great Britain
Mallowan's excavations at Arpachiyah in northern Iraq in 1933 uncovered a Burnt House at the summit of the mound. Although the importance of the discovery has been recognized ever since as vital evidence for society in the late Halaf (mid-fifth millennium B.C.;ca. 5300 cal B.C.)in northern Mesopotamia, the basic data have never been reviewed and no comprehensive interpretation has been put forward. Using both published and unpublished material, this article attempts to review the evidence and propose a new interpretation,based on the current knowledge of the period. It is suggested that the Burnt House was the last in a series of structures where exchange-of obsidian, among other commodities-took place within a formalized social context, and it was through this role that Arpachiyah acted as a center of regional integration. The destruction of the Burnt House may have been a deliberate, ritual act.


cant tell now surroundedby a housing estate on the edge of a village just outside the urban sprawl of Mosul in northernIraq.It is an unlikely spot for one of the key prehistoric sites of the Near East. It was the first excavation of the British School of Archaeology in Iraq, directed by Max Mallowan in 1933 and published with remarkable speed two years later (Mallowan and Rose 1935). The discoveries of that excavation are regularly referredto and the artifactsrepublished in books and articles. There was also a small later excavation by Ismail Hijara in 1976 (Hijara, Watson, and Hubbard 1980; Hijara 1997). In many ways, though, the site has been rather neglected. Virtually no attempt has been made to reexamine what was actually found nor to fundamentally reassess or reinterpretthe more spectacular discoveries in the light of recent knowledge. Almost always, more recent interpretationshave been based on Mallowan's conclusions and usually illustrated from the original report.1

is a small,superficially rpachiyah insignifi-

The major period of occupation at Arpachiyah was between the Halaf lb (ca. 5000 b.c.; ca. 5800 cal B.C.,2Campbell 1992) and the Ubaid (ca. 4200 b.c.; ca. 5100 cal B.C.). In many ways, this is a critical phase in the development of complex society in the Near East. By the end of the fourth millennium, Mesopotamia had moved from having a developed Neolithic society to one on the brink of urbanism, with most of the key elements of state-level society already in place. Yet we know very little of the processes involved. In the Halaf, it has proved difficult to define the degree of regional integration present or the extent of centralized control or social hierarchies. Although it has been frequently suggested that Halaf society represents a chiefdom, the evidence used to support this is thin, and mechanisms that might articulate the meaning of that general description have barely been explored. After the earlier set of excavations, Tell Arpachiyah was by far the most important site for the interpretationof the Halaf culture, and to an extent it still is. In addition to its being, for a long time, the only stratified sequence available, it also was home

in Curtis b.c. andcali2All datesaregiven as bothuncalibrated 1The previouslyunpublished photographs brated B.C. (1981) areone notableexception. 1



to a unique discovery: a building from the end of the Halaf occupation which had burnt down, an event that preserved many of its contents. This assemblage of objects was immediately recognized as outstanding. In particular,the pottery sets standardsof excellence in prehistoricpotterythatare difficultto parallel anywhere else and, rightly, have a prominentplace in the development of ceramic art. Much of the rest of the assemblage is of equal quality.Despite this prominence, very little effort has been made to reexamine the role of the Burnt House, either within the settlement of Arpachiyah or within the wider context of fifth millennium society. The contradiction between the small size of the site and its possible importance has not been explored in any detail. Even the new evidence from Hijara's excavation is not fully set in the context of the previous excavations, and the interpretationbarely touches on the burnt building. There has never been an attempt to draw up a complete catalog of the material in the Burnt House nor to analyze this material as a single assemblage. It is hoped that such an analysis will contribute at least partially to our understandingof late Halaf society. This article originates in an attempt, conceived in the mid-1980s, to completely republish all the material from the site in the light of more recent discoveries and using the limited archive records.3That more ambitious study is suspended because of the difficulty of completing work on the materialin Iraq. This article has a more limited focus and concentrates on the TT6 Burnt House for which, unlike for the rest of the site, there is a substantialbody of material that can be tied to a particular architectural context. The reevaluationis based on a study of most of the artifacts from TT6, the original excavation records, and the publications of both Mallowan and Hijara. archivematerial consistsof a volumeof 3Theprimary notes(Mallowan, a significant numRecords), Arpachiyah ber of photographs Vol. I), and (Mallowan, Photographs Mallowan's Small originalsmall findsrecord(Mallowan, heldin the BritishMuseum FindsRecord) andthe original findsdivisionlist keptin the IraqMuseum. Thereis some additional materialin the BritishMuseum.The original theBritish Museum accessions andthe publication, catalog, not IraqMuseumaccessionscatalogincludeinformation

Tell Arpachiyah is a small site, not much more than 125 m in diameterand standing 5.5 m above the surrounding area. Mallowan excavated Arpachiyah for six weeks in the spring of 1933 (fig. 1; Mallowan and Rose 1935). He employed a huge team with minimal supervision:up to 180 workmen and at most two archaeological supervisors on site, although a few foremen with extensive experience were also employed. It is indicative of the potential problems, recognized even at that time, that work was started on the outer part of the mound to "enable the men, at least half of whom had not been employed on excavations before, to obtain some initial training" (Mallowan and Rose 1935: 8). It is not surprising that almost no structures were found in these outer areas. Deliberately, and fortunately,the center of the mound was excavated last, with ratherbetter trained workmen. Because of this and because of the unusual nature of the finds, it was better recorded as well. As far as a detailed analysis of the function of the site is concerned, only this central area can be used from the 1933 excavations. The site was reexcavated in 1976 by Hijara using a long and stepped slit trench running from the extant summit of the mound down to natural soil (fig. 2; Hijara, Watson, and Hubbard 1980; Hijara 1997). THE BURNT HOUSE The sixth level from the top of the mound was designated TT6 (TT stands for "Top of Tepe"). A building, partly destroyed by fire and containing a large number of objects, was the major structurein TT6. In the original records, this is generally referred to as the "TT6 Burnt House" or, sometimes, the "TT6 BurntRoom."Mallowan recordedover 150 objects of a wide variety of types lying on the floor of two rooms; depending on how the objects are counted, it would be easy to arrive at a much larger total. Although I will discuss the contents of the room in more detail below, the majorityof these objects are extremely and often uniquely fine, including the famous polychrome plates. Many of the objects have remainedeitherincompletely published,or completely unpublishedbecause, owing to various political difficulties, Mallowan was unable to reexamine the objects that were assigned to Iraq in the division of the finds.

present in the primary archive material, and it is assumed that this information,recordedclose to the time of excavation, is either based on records now lost or on firsthand Previous Interpretations knowledge. In many cases, there is additional information either writtendirectly on artifactsor on scrapsof paperkept The main interpretationsare quite restricted and with the artifacts. have rarely been developed in any detail. Mallowan



E.W Ext




Trench boundaries


Trench Name A Orignal

0 lom

- ..

Fig. 1. Approximateareas excavated by Mallowanin 1933 at Arpachiyah.

offered various different possibilities in the initial report (Mallowan and Rose 1935), but the most favored-and generally accepted-one is thatthe building representsa combined potter'sand stoneworker's workshop. LeBlanc and Watson have interpreted it simply as a chief's house but without elaboration (Watson and LeBlanc 1973). Additional suggestions, not elaborated in detail, are its function as a storeroom for a community's wealth or as the treasury of a local chief (Roaf 1990: 49). Maisels adds the hypothesis that, within a nonhierarchical network of settlements, the Burnt House might have served as

"a kind of 'central emporium' where settlements 'in credit'... could have drawn on resources coming from any part of the network" (Maisels 1999: 143). Its destruction was ascribed by Mallowan to invading peoples of the Ubaid culture(Mallowan and Rose 1935: 106), an idea that has almost universally been rejected since, although an alternative cause of destruction has rarely been specifically proposed. Certainly, it seems unlikely that any invaders would have been Ubaid as the pottery does not represent the final Halaf phase, but the possibility of destruction in some sortof hostilities has never been convincingly




Mallowan's Trench


Trench Hijara's

/ " .


Fig. 2. Hijara's1976 contour plan and trench, with probable area of Mallowan'sexcavations overlaid.

disproved. The most innovative, and perhaps most interesting, theory has been advanced by Munchaev, who suggested that it was a ritualcremation,although Mallowan specifically rejected this (see Mallowan 1977: 96 for reference to Munchaev'stheory; see also Oates 1978: 119). There are a number of problems associated with these interpretations.The most basic is that they are not based on the nature of the building, the actual finds, or the detailed context of deposition; rather they are based on Mallowan's interpretationof the evidence as presented and emphasized in the original report. Secondly, the implications of the interpretationof the Burnt House have not been followed through and integrated with what we know of the Halaf in general. The Data It should be emphasized that the evidence is less than ideal. Other studies of more recently excavated

materialfrom contexts with large quantitiesof in situ materialand possible ritual associations have emphasized the value of complete and detailed recording of the objects themselves, their context and relationship with one another,and the natureof the surrounding context in which they were set (e.g., Peltenburg 1991; Stevanovid 1997; Verhoeoven 1999). Very little is known of the exact distribution of the objects within the Burnt House. Insufficient information is available on the nature of the burning and the architecture of the building. There are no faunal or botanical remains. It also seems certain that not all the artifacts originally in the Burnt House were located. Many small objects must have been missed without systematic retrieval methods. Indeed, many small finds from the site are simply recordedas being found on the spoil heap. It seems likely that some of the material supposedly from TT5 (e.g., A1004; fig. 7:2) may have originated in the upper fill of TT6 before its significance was realized (see, in particular, discussion of bowl A1004 below). Conversely,where



. .. . . .. . . . . .:. . .:.: ? . . . .: : :. . .. . : . . . .. . . .1. .. .. . ....:. .. . . . .... .. .:: .: :::::: - ..": ..: . . .. ... . . : . :. ... . . .. . . . . . . . .. . . ... . . . . .. . . ... .. . . . . ...:. ... .. .. . .. ... .1 .? :. ... . . .. ..... .. .: . . , . .. ... . ..: .. . . . .. . ,S.- ...N. .. .. . .. .. . .... . .. . .. .I .. . . . . . . .. . . . ......: . . ... . : . .: . :. .... N . .. . ... i. .:: . . .. .:: . . . . .. NE -.. .. .. .. . . .. .:..... ... .. .. :. . ...: N..- ... ... . ..: :-:: .. .. -,..: .. ... .... . .. . .. -.: .. ..... . . .... . -. . . .v .. I . ... ...:.... .. . .... .. :i. . . .. .. . . . . .. ... . . . ... . ....... .::-.: -:H .... . . . .. ... .... . . . I . . . .: ii-i:: .. .: .... . . .: . . .. . . . .. .. : . . . . .. .. ... . - . ...-. ... .. : . .-:' ... ...'... .: : . ...: .. .. . : - . . -.:-.: . . .. .. ..:.. .. ... . .... .. . ::: Hs;: . i. -. .. .: : :::. .. ?- . ::f: . .. ... . ... .. . . .. ..: . ..' .. ..N . . -" - .: :-. . ..:-1 ... .% . ,. .... .:, . ." . .. ..... .. .. .. ..... ..:. . ... ........ ..:. .. .. .:. .... . : .: : ... . -.:..,..::.: ... :.. .. ,:.. .. ... .: .: ... ,.., .. . .:.. -... . .. .. .. . .% .. .. .. .: . .. . : .. . : :.. . ... . .. : : . .. . :. . . ...:.. ... . ,... .... --. : ... : . .. ..... : ... ..:-. .:.. .. .::::-:.,%, 1 . .. : ..... .. . .. . .. . ... .. ... .... .. .: -: .. .. .. . ... .... . .... .: .. . . ...:. :.. ..... .I .. :. f:: :. . ..::.: .14" ....... ... ... .. .. .. .. ... . .. ... " . . .....:. . . ..:..:: .. . :.. .. . ... : . :.i.. . . -. . .. . .. :.. . :. .... .,.. ..... : --. : .. .... .. .::f. : :: -: .: ... .... ..... ..: : :.... -: ..... :.. .... ..... .. :: -.'.%::. ...... ... . ..... . . . . .. ..... ... .. .. . : .. . .. . ., ... .. ... -: I . ...: :,.1 . . . ... -: . .: :.:........ .... .. ... . ...... .. ..1. N :.? .... .... .... .. . ...... -. .. . ,. . ...:... ... . . .. .: .. ... . .. -: .: . .... . . .... .. . . :, . . , - . ..:.. . . .. .. : . I .. . .. . ... . . . :. ,. .: ..: :: : . . .. ?:: . ..-:: .-: : : ... :. -.:-:::-:.. , '...: .:..:..:.:: ...-:-:-::. -- .:-:--:.--.. . ...-.. .. .. . , .... :.. . .1 .... ,: :.:-. . ... . . ::... ,.-f '., -, ..-. - . . ..: . .: : :.. :. .. :..::: ..:::% . :-. ':... :'. ... .. :. .. .: .. ..S , . .:.- . :...: :.. ... .: .. , . .: ,.. ..... .. .. ..... :.:.: -:? . . .. ..::.. .. ...:. ....: .: .. . ....... ...... .. ., . . . :.-.. :.: ... .... . .. ..:, .... ..:: ..:....: .: ... :--:.--.: ..: .. -:. I . . ... .. ..... .. ,..: . .- %: . .. ..;.:.. ... . ... ....... :. . N ... . ...... .. . .:,--.. . .. . I.. .. .. ..: ...... :-: ? .-. ..X ... ::. -.... :". :-. I ,i .1 :. .....: .. . : - :.I-X :: .....-.. .-M .:::.. . .-.. . .. . .. lV-..-%:.::. -:. . :% --..% : .N. ... ... ..... .. ..... -:::..% .. . ....... .... .... .-.-. ..... .- . . . .... . i :. :..::;.% - .-:::..:,%%. ,... , :.:-. I.. .. . ..% .... .. .-:.:.:-, . :-1-::.:.-.:%s-%. . ... :,, .....-.V. .... ,.., . . ... :::::::::::::.:. .1 .Z.:.. .:. I . 4. - ... ..... .. .-:::::::i:::.,. :it.. .......:-:.:, . . . . ..... :. .. ...... ::::: . .:.::::::: ?, . ..... .. . ... :.. . .. . . ... .: -? . .. :::.: .:. : . . -:: : ........ S,;. . ::::::: :.... ..X ....:::::.:.. ,;,--: : ,.... ... :s% .: .1 Z,: ??Ki . .. . . . . .. . .::.-:':..., ... . I %. -.-: .%-,--.,::f:.:.I . ,. :?! .i -.74 Z -: , .;,-!:::::::f:, .. .....111. .1 . :: : . 1 : ?: : . ? :X::: . :. : . .... . .:.. .... , .I :. .:..:Sit : :I.:.. . :' I,--,I - .. -? .

: .. ...


. . . . ..., . .?:?? . .. ..???? : . .. .. ..: .... . . . ...? : . . .. . .. .?? : 1. .. .. . . : . ... . .. . .. ....... ....i.. . . ... ..

.: ::

-. : . .. .:-

..I. ..- .... N .:. ... .......:.

.. .

. ..

.. . . 1. ..l.;" .

. :: T . .: : ,: :;%, I:? : . . ::.!% .. :

?iM ?:i? ?i ,::0. ....,%.:::,., :::.. ::: :: N : 4;: .x?

::: ::Y ::! ;?:: -j :?!iiiiii~ i?

X :.,: -:::::;::M ,,.-.,%? :?,-,:: : .::: ?D j `z x : i~

. : : M' !?? ...... ,N? . .,.:

::::::::::. .: .1.... ? -:?X x ::?::::,? .. ......

-1. 'It ,444" s ;::::::: BE .; ... :::? .11 .... ?:.. ..

.: .:.:::-::.: :-..":. ..: .. . .. . .:.--::. .... . .. . . .: -I :.. ... :. .. -% . .--: .:. . . . . .: ,:; s!%,!! :f: :4:. :. .. ..:? -I.. . . . ..:.::.. .:.. . ..:: ..
.::: -:. ...:.: ..: :. :.. ...::,% .... . . . .: .. ... .. ::::....:".

? ?' ? ?:' I

.,::4: IN . .. . ::. :: . N. --, .:...' . ....

Fig. 3. 3D projectionof the TT6 BurntHouse.

The Burnt House is in a very prominent position at the summit of what was by that date a low mound. There seem to have been paths made from close packed pebbles, some of which are almost certainly contemporarywith TT6, leading to the center of the site (Mallowanand Rose 1935: 21, pl. 1:b).These further emphasize the nature of the Burnt House as a 4lndeed,given the eventsof recentyearsit is doubtful focal point of the settlement. The layout of the TT6 if all the material, some of which had been dispersed to building is rambling with no immediately obvious survives and still museums, plan (fig. 3). This may be complicated by an excava(Baker,Matthews, provincial tion technique that may have failed to distinguish Postgate1993).

an artifact is recorded as simply coming from TT6, we cannot be certain that it comes from the Burnt House-it may come from elsewhere in that level, although that remains much the most likely location. Damage, cleaning, and restorationthat may have occurred during fifty years in museum storage hinders detailed examination of the objects. It was particularly difficult to examine some of the material that remained in Iraq, much of which had been distributed among variousregional museums.4Furthermore, incomplete records were kept of the distribution of the material that was exported-objects are distributed among multiple institutions in Britain, New Zealand, Canada, and Australia, as well as given to

a few private individuals. Although most of the artifacts have been traced, a few have not yet been found and examined. Nonetheless, a core data set can be isolated from which we may hope to draw plausible conclusions.

The Structures



between adjacentunits. Partsof the building are of a significant size; the long room that contained many of the objects is approximately 10 m by 3 m and had walls 0.5 m thick. In the levels below TT6 there was a series of circular buildings-the structures that encouraged Mallowan to use the misleading name of "tholos." Accordingly, TT6, with its rectangulararchitecture, has traditionallybeen seen as a radical change in the architecturaltraditions of the site. In TT8, a single tholos was excavated with a 10 m external diameter and 1-m-wide walls built on a foundation of river boulders. This was reconstructed on the same plan in TT7, and part of a second massive tholos at this level was also found. Mallowan argues that because of their size and position, "it may thereforebe taken for granted that they [the tholoi] were buildings of outstanding importance"(Mallowan and Rose 1935: 34). He suggests, furthermore,that the absence of later disturbance, the presence of two rich graves, and the discovery of several female figurines in the vicinity all point to the sanctity of the buildings. Hijara subsequently followed this argument, adding his suggestion of a central tholos precinct (discussed below), more unusual burials, and the observation that the fill of the tholoi seems to be of clean red clay rather than more usual cultural debris (Hijara 1978: 127-28). He further argues that the investment of labor in the construction of these foundations would have required a significant mobilization of labor (Hijara 1978). In contrast, Akkermans has argued against this interpretation,pointing out that the figurines and burials are not directly associated with the tholoi and that the tholos precinct enclosure and the clean fill were only seen in very restricted areas (Akkermans 1993: 300-302). In particularhe argues, convincingly, that the stone foundations of the tholoi, probably carried from the river Khosr about 3 km west of the site, need not have involved such a major investment of labor. Thereremains, however, a previously unremarked but possibly more convincing reason to suggest that these tholoi may have had some particularstatus. If one superimposes the plan of the TT6 buildings on that of TT7, it is clear that the long, northernroom of TT6 is on exactly the same plan as the antechamber of the TT7 tholos (fig. 4). TT6 is perhaps better thought of as a reconstruction of TT7 on a slightly different plan rather than as a radical change in architecture. This runs against traditionalassumptions in that the part of the TT7 structurethat shows con-

tinuity, and may therefore be the most important room, is not the tholos but its rectangular attachment. Mallowan states that "the foundations of the potter's shop [the Burnt House] rested directly upon those of the older level of destroyed buildings in TT7" (Mallowan and Rose 1935: 17; contra Akkermans 1993: 302). Indeed, some unpublished photographs in the British Museum and the section in the original publication (Mallowan and Rose 1935: fig. 4) indicate that some of the walls of the TT6 building may have been below the tops of the TT7 tholos foundations, which suggests that the later building may even have been an alterationof a standingbuilding rather than a reconstruction. It may not be entirely irrelevant that the concept of reconstructing a building on a preexisting ground plan is familiar from later Mesopotamian archaeology, particularly for religious buildings. The Burnt House must have been only the last in a series of importantstructures with a shared significance. It is possible that the contents of these earlierbuildings were comparablewith those of the TT6 Burnt House but were simply not preserved. Hijara's Interpretation Hijara has argued that the TT6 building, and the central area of the site in the preceding levels, constituted a distinct area within the settlement of Arpachiyah, marked off by a boundary wall and with little or no building activity in the outer part of the site (Hijara 1978: 127; 1997; Hijara, Watson, and Hubbard 1980: 134). Given the small size of Hijara'strench, projecting a boundarywall aroundthe top of the tell seems very conjectural. If the section of Hijara'strench is studied (Hijara, Watson, and Hubbard 1980: figs. 3-5), the so-called enclosure wall is easily seen between the points marked 30 m and 35 m and is recorded as being a thick mass of tauf, or packed mud. While it is undeniable that it could be a wall, the trench in which it was found is only 2.5 m wide which is scarcely wide enough to argue that it is a wall rather than a mound of tauf of a differentnature.Moreover, the reason why Hijara detected little or no later inhabitation in the outer part of the site is probably simple: it had already been removed in the 1933 excavations. The point at which there is a break in habitation is at about the 40 m mark on Hijara'spublished section (Hijara, Watson, and Hubbard 1980: figs. 4-5). It is clear from a combination of Hijara's





Housesuperimposed on the planof the TT7tholoi. Fig. 4. Planof TT6Burnt is certainly the large, long room on the northeastern side of the building. The other is not exactly specified, but may be the adjacent room. Although the term "full room" suggests that it contained most of the objects, this may be incorrect. Most of the objects for which there is any indication of location are specified as coming from the long room. It is notable that this room may be the primary one and is also the room that preserves the plan, and perhaps the significance, of the preceding buildings. The finds were scattered around the rooms, some close to the walls, others not; some individually and some in groups. Some of the artifacts were found on THE DISTRIBUTION OF THE ARTIFACTS charcoal, which Mallowan plausibly interpretsto inThe objects seem to have been concentrated in dicate that they had been placed on tables or shelves two rooms of the Burnt House, one referredto as the aroundthe room. The few photographsshow objects "long room,"the other as the "full room."The former that are lying in groups but do not appearto be close to a wall (Mallowan and Rose 1935: pl. 21:b; Mallowan, Photographs I: photos M412, M516, and Ar- M520 in particular).Unfortunately it is now impos5This is clear in the archivevolume (Mallowan, Records: andfromdepths sible to reconstruct the original locations of more 178, 188,in particular) pachiyah than a few of the objects. recorded on sherds. overall plan, Mallowan and Rose's overall plan, and several published and unpublishedaerialphotographs that this is almost exactly the edge of one of the trenches of the 1933 excavations (fig. 2). Mallowan's unpublished notes in the British Museum5 make it clear that up to 4 m of deposit was removed in this area, which probably accounts for all of the phases of occupation apparentlymissing on the outer partof the tell. Therefore, I suggest that we have to assume that the outer areas of the site were occupied when the TT6 building was in use.


? i-:!:x::,: IKZW


!-s is, :MI ?.::?:::.i~i IMi~l li~ii'I Z?;i ili?:?.; iii~liiiiriii':~liC,~illlll~~re"ssItIist ?:::"M tits: At ;':: ::~i??::: ??;: "l~ci:jiii~i~iiia'i.F~i:isil; :~::::? int:iilI!z :~??:i::iiIiif.: ?i :?ol ? NN-4,1 T~:?: -fl i XV ril;;;I: I' If,: f. ????,iiii~j~i~iil'';it E ?Z.iii f. it:

i K i 01i: fF:;1:ii iiiiiii. f -'s tsii,W.

ayvs -ri

ils ijii,

iW ;WI, lu f


Ro -;s~' AA:i:'

. ......... -4 i.,M :Mmrix iis.; N't" I....jiii i F i:z 47 "R i F i.... iiijj .... .... . 'Eiiiim RZI; i ...... . ... ":? .f,.;.!rii liiii~iiAgi::li'Yld M '?T iiil


. -


.::-ii r-4~~r RP iilixiii .-gt1~ wiil Iiiiii ?us:: i is it ii -siiesi ~i; -iissil"Z: zz I; i

Is -R?Aii-.


Wa, ms~liis, -,.I A :::: 'sf z:: -.:J ?ss if~i ;g MOMFr:illCil illit.H.,:: t::1NNO I:.f?; 41. zi;CiRzi " Kiigl:*:":i5,og :t~ s. iti U;,4 ...............: iiiil~dU/U;-:Itil. htl diii E 1:: sit s?: .4 :sti : Ili's::f fsf ........ ....... SWiii~iiiii'eiri~u:921.l'i .... ... .iinns ... .. . .... ?I:Y?: ? si-m.:,::; ii.: ii?. -qirg W sifszz f;:f: .,.. -:::T,? Role, .......::?:?.~ R ii, . IF .I :?:?r:;.; ?;;:-?:::;?:?;; ??::?:::;r:. :r?;: l?::l::.:;. .;!jr.:r~r.:::::?:t; i~i6I::i'ii'iiilil ?'?4 Vili:B~iiliii:;P::...... Ii:::ili~ 0 ,4Rill~liiii$il Ili! l~:?i E .r :'':'? :i:l:: ??` ?l ir:;;rr :I: ?''~"?: ' i~ill' i .

Fig. 5. Polychrome plate A528.

Not all of the building was equally burnt. Mallowan states that the southern portion suffered the most (Mallowan and Rose 1935: 106). This contrast is extremely markedin the condition of the artifacts: some are heavily burnt but others are completely unmarked. This may indicate that the burning was only partial and that the fire may have been halted by the collapse of the roof of the building, either accidentally or deliberately. NATURE OF THE OBJECTS The finds can be discussed in rough groups. Much of the most important information is not included in the original publication, and some categories of finds are omitted entirely there.

The best-known artifacts are the highly decorated pottery plates. Among them are thirteen polychrome plates, manufactured and decorated with exceptional care and skill. Remarkably, not all of these are published in the original report (figs. 5-6). Together with another polychrome example, recorded from TT5 but which probably originates in the Burnt House (A1004; fig. 7:2), and six fine monochrome plates, these plates form a distinct group. Most of the polychrome plates are decorated in only two colors, but four examples have additional white paint. The quality of this group is unparalleled at any other Halaf site in terms of form, fabric, and decoration.





plateA529. Fig.6. Polychrome courtesy of the (Photograph British Museum.)

There is a considerable degree of homogeneity within this groupof open plates. The forms areclearly related, although there are specific variations and a considerable range in the dimensions. Plates A748 and A749 have more complex profiles (Mallowan and Rose 1935: pl. 19:1, fig. 54:4). Some have much wider bases and steeper angles on the walls (e.g., A745 and A751; Mallowan and Rose 1935: pls. 17:b, 18). However, the similarity in dimensions of the largest plates is striking. Plates A746, A753, A750, A749, A748, A752, and A528 form a very tight group in terms of diameter, especially if A515 is considered as being a slightly different shape and therefore set aside (Mallowan and Rose 1935: fig. 56:2). If it does come from the Burnt House, A1004 would also fit in this group. In terms of the height of the vessels, these pieces are also very similar, with the possible exception of A528, for which the height had to be estimated from a photograph and may be too low. An alternative, looser grouping might be those

plates with a diameter of over about 240 mm. This would add to the group A529 (fig. 6), a trichrome plate; perhaps A751 (Mallowan and Rose 1935: pl. 22:10), a ratherlow bichrome plate; and A754 (Mallowan and Rose 1935: pl. 19:4), a fine monochrome plate. The general scheme of decoration is identical in all cases. The interior is the main focus and very little of it is left without decoration; indeed in most cases very little of the original vessel surface is left unpaintedat all. With the exception of A528 (fig. 5), there is a strong central motif which is usually some form of rosette but may be what Mallowan refers to as a "Maltese Square" (A751; Mallowan and Rose 1935: pl. 22:10) or a group of bukrania(A515, which is a slightly different shape from the other examples; Mallowan and Rose 1935: fig. 56:2). The walls of the interiors are covered with tight horizontal bands of densely painted and visually complementary motifs. This unit of composition is unusual.




In other Halaf examples of plates, the decoration, although mainly in the interior, is somewhat sparser and more open. The exteriors of the plates are more plainly decorated, either with parallel bands or a bichrome design of triangles, as with A528 (fig. 5) and A752 (Mallowan and Rose 1935: pl. 15). There are certain designs that recur in the decoration of the plates. Two of the plates are identical (A742 and A743; Mallowan and Rose 1935: fig. 57:1). Stippled circles are frequent, occurring on three vessels as the uppermostband of interior decoration. In two cases this is on vessels decorated in three colors, associated with the use of small rosettes in the next horizontal band and with a bichrome, triangular design on the exterior of the vessel. Negative designs occur in three cases (A745, A750, and A751; Mallowan and Rose 1935: pls. 17b, 18); each is associated with the same external motif and, in the first two cases, with bands of cross-hatching made up of horizontal and vertical lines. Rosettes are commonly used in this group, not just as a central motif, but also as a component in the checkerboardingand bands. None of these motifs are unique within the Halaf ceramic tradition, but their frequency here is unusually high. Therefore, this group of plates has a strong degree of internal similarity. It can be suggested, with some degree of confidence, that at least a subgroup of these plates may have been made by a single potter of great technical skill. This subgroup would include A748, A753, A746, A750, A752, A528, A529, A1004, and, possibly, A749 and A745. Others of the group of fine plates may also have originated in this group, but they lack the very close similarities in shape and form that occur within it. This core subgroup of eight to ten large plates probablyhad a single use. The quality of the plates and their context suggest that this function was high status. The general lack of use-wear suggests that they were only used on specific occasions, most probably in formal social rituals. A relatively small group of motifs was employed in a consistent and interrelated manner on these vessels. This homogeneity may be simply stylistic, but the decorative motifs may also have carried quite specific meanings, perhaps conveying messages relating to the contexts in which they were used. Most of the plates are broken, some of them very extensively and one into 76 pieces (Mallowan and Rose 1935: 107). These plates must always have been fragile and, in the destructionof the build-

ing, would almost inevitably have been damaged. That this damage sometimes happened in situ as a result of the pressure from the surrounding deposits is shown by the only illustration of in situ finds (Mallowan and Rose 1935: pl. 21:b): three plates can be seen, clearly extensively broken but undisturbed. However, some plates seem to have been broken and their pieces scattered before being burnt. This is particularlyclear in the cases of A529, A750, A752, A753, and A754 where conjoining sherds have been burnt to radically different temperaturesand in different atmospheres (e.g., A529 in fig. 6). This might have occurred because a plate was dropped from a table or shelf or even shatteredduringthe burning of the building. Mallowan, however, states that some of the sherds from a single vessel were found at opposite ends of the room (Mallowan and Rose 1935: 106). If we accept the accuracy of Mallowan's observations, he must be correct that these pieces were broken and scattered deliberately. Such deliberate breakage must have been selective ratherthan comprehensive as not all artifacts or even all plates were affected. One fragmentary plate (A1004; fig. 7:2) reconstructed from a number of burnt sherds is almost certainly from the Burnt House but is recorded as being found in TT5. It was probably excavated before the rest of the Burnt House had been identified, either through inaccuracy of recording or because it was lying on the top of the TT6 deposit. It differs from other plates in that the edges of the individual sherds are very worn. The most likely explanation is that, after the destruction of the Burnt House, this area of the site was abandoned. The sherds may have lain on the surface or eroded out, and the edges may have become worn through exposure. A gap in occupation after TT6 has been suggested previously, although not from the same or from such conclusive evidence (Davidson and McKerrell 1980: 164; Curtis 1981: 33). Other Pottery In addition to these fine plates, there was a considerable quantity of other pottery. Some of this was also extremely fine. The two bichrome jars (A512, A739; Mallowan and Rose 1935: pl. 20:a-b) are particularly notable. There were also two fragmentary champagne vases, probably A526 and A527 (Mallowan and Rose 1935: 136). The former is recorded in the small finds list as being decoratedwith stippled




Scale 1:2
Fig. 7. Plates probablyfrom TT6. 1: A 1003; 2: A 1004 (hatched paint is red-brown).




circles but has not been traced. A fragmentary and unnumberedbow-rim jar in the Institute of Archaeology, London, again decorated with stippled circles, probably came from the Burnt House as well. The stippled circles on extremely well-made pottery are reminiscent of the decoration on several of the polychrome plates and emphasize the stylistic links within the assemblage. In addition to this fine pottery, there were also a considerable number of vessels that, although by no means crude, are more typical of a Halaf ceramic assemblage. These include a considerable number of small pots and bowls. Two very similarjars are noteworthy (figs. 8:2-3). From the differential burning visible on the exteriors, both must have been sitting at an angle when they were burntratherthan on their flat bases. Was the Burnt House in some disarraybefore the fire took place? Both these jars and several of the other, non-fine ware vessels have clear signs of wear, particularly chips around the rims. These broken areas are themselves worn and must represent the results of wear and tear in the normal life of a vessel. These plain, more heavily used pots may represent the typical component of a normal Halaf dwelling. A contrast can be drawnbetween the often extensive wear on these plainer vessels and the absence of such on the fine plates. A few of the smaller and less fine plates have traces of wear, especially scraping in the center of the vessel, presumably where food was scooped up. The finest and largest of the plates appearto have minimal or no wear visible in this area.


Large quantities of items that could loosely be considered jewelry were found. These include large numbers of simple beads (A877, A878, A879, and A880; Mallowan and Rose 1935: pl. 6:b). Very few of these have been traced in museums. Therefore the identificationssuggested by Mallowan-that they contain examples in exotic materials (steatite, carnelian, and serpentine are all mentioned)-must remain unconfirmedand unquantified.However, there are certainly also some more exotic items. Among them is a necklace of cowrie shells and obsidian links apparently found lying in the order that they have been restrung (fig. 9). Mallowan notes that the shells were originally filled in with red ocher. However, the deposits that remain in the shells give no indication of this but seem to be a mixture of reddish ash and tauf which may reflect the deposits that surrounded them. The obsidian links in the necklace are technologically paralleled by the considerable numberof other obsidian links found in a group (A905 and A906; fig. 10; Mallowan and Rose 1935: pl. 11:b). Once again, these indicate a technology of obsidian polishing and grinding that is very different from standard Halaf lithic manufacture. The sizes of the blades from which these links were made (at least 60 mm long and 30 mm wide) are very unusual in the Neolithic of northernMesopotamia. The value of such links is emphasized by their absence from any other contexts at Arpachiyah and the scarcity of parallels within the Halaf. Specific parallels occur at Banahilk (Watson 1983: 573, fig. 10.4) and Yarim Tepe II (Munchaev, Merpert, and Bader 1984: 38, Stone Vessels fig. 5:4) but these are all isolated finds. The cowrie Nine stone vessels were present. Stone vessels shells in the necklace are clearly exotic, long-distance generally occur consistently but in small numbers imports. in Halaf sites. This rarity may be due to their being strongerthan pottery vessels and thereforediscarded Seals less often, but it may also reflect their greater value. In any case, nowhere else has such a large number At least eleven seals were found (13 if A43 and been found together. Of particularnote is the obsid- A893b are included, althoughthey may not be seals), ian jar (A411; Mallowan and Rose 1935: fig. 44:15). but there were probably more; six more have the amThis is an exceptional object, although there are now biguous location of TT6-7 and two come from TT5. Halaf parallels for obsidian vessels from Domuz- The seals are typical Halaf stamp seals with linear tepe in southeast Turkey (Campbell et al. 1999: fig. incised designs, but it should be noted that this is 16). Mallowan states that the vessel may have been by far the largest concentration found in any Halaf unfinished or broken in manufacture. There is no site (cf. von Wickede 1990). All of the seals that I specific indication of this; certainly its two frag- have been able to study have signs of long use, usuments were found lying together, suggesting that ally visible as a very high polish on raised areas.Two they were simply broken in the destruction of the are of particularinterest. One is broken (A867; Mallowan and Rose 1935: fig. 51:11), but the breakitself building (Mallowan and Rose 1935: pl. 22:b).






Scale 1:2
Fig. 8. Jars: 1: A525 (TT5);2: A516 (dotted line indicates angle at which it was sitting when burnt;3: A517 (dotted line indicates angle at which it was sitting when burnt);4: A535. 5: Stone vessel, A413. 6: Small jar, A493. 7: Stone vessel, A414.




Scale 1:2
Fig. 9. Examples of groundand polished obsidian and shell beads from necklace A909. 1:A909a, stone; 2: A909m, stone; 3: A909p, obsidian;4: A909b, obsidian; 5: A909f, obsidian (hatchingindicates heavy grinding);6: A909u, shell; 7: A909q, shell.

is heavily worn. The seal certainly continued in use for a considerable period after being damaged, suggesting that these seals were of considerable inherent importance. A second seal had been deliberately defaced (A874; fig. 11:3): 1-2 mm of stone has been ground away from its face, leaving only faint traces of the original incised patternsaroundthe edge. The small size of the seal seems to rule out secondary use. Instead, it has been deliberately taken out of use. Once again, this seems to emphasize the significance given to such seals. It was not enough for it simply to fall out of use; it also had to be destroyed.

Seals had probably also been manufactured in or around the Burnt House. There is one seal with no design carved on it (A866; fig. 11:2), and two further unfinished seals were found in spoil from TT6 or TT7 (A41 and A578; Mallowan and Rose 1935: pls. 8:a, 7:a). Almost all Halaf seals have holes for suspension and wear marks consistent with rubbing for a long period, suggesting that they may have been worn about a person. They were probably associated with an individual or with an institution that that person represented.Typically they are found singly, and no




Scale 1:2

Fig. 10. Groundand polished obsidianlinks. 1:A906b;2: A906c; 3: A906f;4:A906g; 5: A906i;6: A9061;7:A906m;8: A906o; 9: A906p.




Scale 1:1



6 L!
Scale 1:2

Fig. 11. Seals: 1: A575; 2: A866; 3: A874. Sealings: 4: A619a; 5: A619b; 6: A620b; 7: A620a.




other significant grouping has ever been found. It is tempting to suggest that an institution based in the Burnt House may have had some specific control over the use and availability of seals-manufacturing them and taking them out of circulation. Charvait has emphasized the possible personal associations of the use of Halaf seals and, in particular,a possible specific role in gift exchange (Charvait1994: 13). At least one possibility is that the Burnt House had a central position in mediating such transactions. Sealings Twenty-seven burnt clay sealings were found but more may well have been missed. The stamp seals themselves are well known, but Halaf sealings are much less so, probably owing to accidents of survival and excavation. The discovery of some 300 sealings at Tell Sabi Abyad from the late sixth millennium (Akkermans and Duistermaat 1996) suggests that their use was much more widespread than suggested by the small number from Arpachiyah, Tepe Gawra (Tobler 1950), and KhirbetDerak (Breniquet 1996: pls. 56-60). In the Burnt House, they form two functional groups. The first type is a group of oval sealings wrappedaroundstrings (the 19 sealings in group A619). This type seems to have been used in a standardized manner with impressions of a single seal evenly distributed,often in pairs, around the circumference of the sealing. The sealings of the second type are characteristicallyflat discs with the seal impressionmade into the uppersurface (the eight sealings in group A620). The reverse usually has no marks to indicate what it was pressed onto. Von Wickede suggests these were pot lids-although as he points out, Halaf jars with sufficiently narrow necks are extremely rare-or control markers or tokens (von Wickede 1990: 97). Fourteen of the sealings were impressed with the same hand-shapedseal (von Wickede 1990: 95). These include examples of both the first type of sealing (e.g., A619a-c; fig. 11: 4-5) and the second (e.g., A620d-e; von Wickede 1990: nos. 58-59). The impression of this seal is never found in combination with the impression of any other seals. There is no indication that the sealings were associated with other objects that they might have sealed. Indeed, the number of sealings of the first type, impressed around cords, is much greater than the number of jars that might plausibly have been

sealed. This suggests that the sealings were discarded after use, or even kept as an accounting system or, much more likely, that perishable goods were also kept in the Burnt House but did not survive. There are no sealings that match the patterns on any of the seals found in the Burnt House. Sealings were clearly used in a standardizedand recognized manner to identify objects or goods. In a very general sense, this suggests the concept of ownership and the beginnings of bureaucraticprocedures has argued that "the to mark and control it. Charvait conditions under which a practice of sealing may be introduced exist only if the commodities sealed are transferred into different competence spheres" (Charvait1988: 57). If so, a rather complex system was in operation, with material being passed to the Burnt House at Arpachiyah from other people or locations which in turn possessed some degree of control. The status and location of the owner(s) of the hand-shaped seal, which was the source of the most common impressions, are unknown but potentially critical. The social context in which the transfer between different spheres took place is unclear. It may have been asymmetric or symmetric exchange, or it may have been a form of central social storage; in either case formal or informal gift exchange was most probably involved. Certainly, the contexts in which seals and sealings were used are likely to have been embedded in the other social rituals that seem to have taken place in the Burnt House. Stone Axes Among the finds from the Burnt House are six stone axes, the largest concentration at a single find spot in the Halaf. The four that have been studied are all true axes with symmetrical profiles of the cutting edge. Two of these are heavily used with chipped and resharpened edges. Two more, which have not been traced, seem from Mallowan's small finds record to have been equally heavily worn. One ax is only very lightly worn but is otherwise unremarkable.The final example is more unusual (A666; fig. 12:4). It is small and very well made from an unusual green stone. There is minimal sign of wear on the cutting edge, and the stone of which it was made is not hard enough to be suited to extensive use. It therefore seems probable that this had some prestige or ceremonial significance rather than being strictly functional.




Li,1 ))?.




/:?I ?

/// "







I I Scle1:

Scle 1:

Fig. 12. Stone tools. 1: Chert implement,A1001a. Groundstone axes: 2: A664 (hatching indicates polish from probable hafting);3: A663; 4: A666. 5: Stone palette, A901.






89"10 ,

Scale 1:2
Fig. 13. 1: Stylized anthropomorph,A904a. Knucklebones: 2: A922b; 3: A922c. Pottery discs and spindle whorls: 4: A913a; 5: A913b; 6: A913c; 7: A913d; 8: A913e; 9: A913f; 10: A913g.

Miscellaneous Objects

trough (A923; Mallowan and Rose 1935: pl. 10:a). This group is enigmatic; it may be suggestive of There was a group of objects found together at ritual or cult but this can only be speculation. It the one end of the Long Room (Mallowan and Rose may be significant that one figurine of each sex was 1935: 99, pl. 10:a). The objects include the figurines present, despite the fact that distinctively male figufrom the Burnt House (A921, A941; Mallowan and rines are very rare in the Halaf. Not associated with Rose 1935: pl. 10:a, fig. 46:3), the collection of real this group were three (in the small finds catalog) or and stone knuckle bones (A992; fig. 13:2-3; Mal- four (in the publication) stone objects, at least one lowan and Rose 1935: fig. 52), and a small "steatite" of which is a very stylized anthropomorph(A904a:




Scale :2

Fig. 14. Bone tubes. 1: A915a; 2: A915b; 3: A915c.

fig. 13:1). The one that is not in Baghdad is definitely of pumice, and two of the others may be. Their context and function are unclear. A second type of object that can usefully be considered here is the four bone tubes (A915a-d; fig. 14:1-3). Two of these were found together at the west end of the Long Room, lying on charredwood (Mallowan, Arpachiyah Records: 243). The location of the other two is unknown. Only one is preserved to its full length, and it is possible but unlikely that two of the others may be opposite ends of the same object. These bone tubes have some characteristics in common. All are highly polished. This may well be through handling and, together with the heavily worn ends, may suggest a long period of use. Each has a small area of wear close to the end, where the outer surface of the bone is completely worn through. Groups of notches occur on one side of the two longer pieces and around the midpoint on the complete piece, although it is possible that these are a natural characteristic of the bone caused by stress (Nicola Murray, personal communication 1991). A

stone mouthpiece, found as part of this group, may have a related use (A915e). A palette was found (A901; fig. 12:5), broken, probablyin antiquity.Although this object is far from unique, with parallels at KharabehShattani (Campbell 1995a: fig. 60:2) and Yarim Tepe (Merpert and Munchaev 1984: fig. 13:1), it is very well made from an extremely fine and colorful piece of banded limestone. Mallowan recordedit as being associated with a lump of red ocher which he interpretsas evidence that pottery was painted in the building. This appears to be highly conjectural.There is little definitive evidence to associate such a palette with potterypainting, nor is there any other evidence of pottery manufacture in the Burnt House. A conical lump of lead, 42 mm high, was also recorded (A918). However, it was not possible to examine this in Baghdad. There are no other examples of lead in a Halaf context, although there is a lead bracelet from Level 12 at Yarim Tepe I dating to the earlier sixth millennium (Merpert and Munchaev 1993: 114). It must be taken as an unusual,






-TIffig If

Y O"A m
K-4 7, A W?. JIM" M Q': ug'


N9 lie -.3,

ZZA R:01 r?A

-x; I'M .,w7f LEI,

41-Ex -IM

ly ?.j e 41 ef

.. ..... ...
N ?.k Oll - ---

.. ... .....

ff M


wO, 0 IP


House. cores (A911)fromthe Burnt Fig. 15. Obsidian

and thereby potentially valuable, object as well as a rare piece of evidence of Halaf metallurgy.

obsidian knapping took place in the Burnt House as well. There is a general absence or extreme rarity of obsidian cores at most other Halaf sites. The two usual explanations for this are either that obsidian Chipped Stone was imported in the form of blades or that the knapMallowan records "thousands of pieces of flint ping occurred in unexcavated areas of the sites. The and obsidian cores and chips" in the Burnt House presence of cores in the Burnt House at Arpachiyah (Mallowan and Rose 1935: 105). It is very unfortu- is significant. It is the only material that does not nate that this component, numerically the most sig- fall into the categories of luxury, prestige goods, or nificant find of all from the Burnt House, is virtually normal domestic equipment. It is almost the only unrecorded and that little of it can be traced in material for which there is real evidence for manumuseums. Small finds A907, A908, A910, and A911 facture,and the large quantitiespresentare totally out (fig. 15) can only representa minute proportionof the of proportion with our expectations for this size of total. Three additionalobjects, unusualtabular"chop- exposure. There is a large quantity of obsidian from Arpers,"exist in the Institute of Archaeology, London, and are marked as being from the Burnt House pachiyah in the Institute of Archaeology, London. (A1001; fig. 12:1). Along with them are quantities of All of this material is technically unstratified. In flakes of an identical flint, at least one of which can view of the areas excavated by Mallowan, almost be joined to one of the "choppers."These objects, at all must come from Halaf levels. It may, in theory, least, were made in the building. The obsidian cores come from any partof the long sequence, but it seems of A910 and A911 may well indicate that primary almost certainthat much (and possibly all) came from




TABLE 1. Colors of Obsidian from Arpachiyah

Green Gray Green% Gray% Ratio
Core preparation Blades Bladelets 210 457 64 11 109 50 95% 81% 56% 5% 19% 44% 19:1 4:1 1.3:1

the TT6 Burnt House. When examined in 1983, it was stored as though it came from a single collection assembled in the field, and it was stored together with A1001, which is known to have come from the Burnt House. It is very probable that if Mallowan was keeping any lithics, he would have kept them from here and, indeed, the only stratified lithics known from Arpachiyahcome from the Burnt House. In a preliminary analysis,6 a sample of the assemblage was divided into blades or bladelets (and tools on retouched blades) and debris from core preparation. This latter group consists of primary and secondary core preparation,some of the former retainingthe outer surface of obsidian nodules (patinated, smoothed, and abraded). Each piece was examined in transmittednaturallight, and the color was categorized by the presence or absence of a green tinge. Almost all of those that are not green are gray, but a few are colorless or opaque. There are a small number of cores but it was impossible to examine their color systematically due to the thickness of the obsidian, althoughboth green and gray obsidiancores are present. Although trace element analysis is generally used to determine the source of obsidian, color analysis allows a basic and rapid division between peralkaline and calc-alkaline obsidian as only the former produces green obsidian (Cann 1983: 234). Comparisons between obsidian color and chemical sourcing in Anatolia suggest that this correlation is very high (Campbell 1992: 141-42; Healey, in prep.), and we know that the only peralkaline sources are Bing6l and Nemrut Dag near Lake Van. The gray obsidian from Arpachiyah could have come from several locations but sources ig (the calc-alkaline sources at Bing61l)and 3a (Zernaki), both to the north of Lake Van, are attested in chemical analysis of material from the site (Renfrew, Dixon, and Cann 1966 and 1968 for the analyses; Chataigner 1998 for the lo6A more detailed study has been undertaken recently by Elizabeth Healey (Healey, in prep.) which confirms the general accuracy of this rough analysis.

cation of the sources). There has been no chemical analysis of green obsidian from Arpachiyah.The obsidian from Arpachiyah shows a clear and marked pattern. Among the core preparation sample, green obsidian outnumbersgray by a ratio of 19:1. For the finished blades the equivalent ratio is only 4:1. Among bladelets, gray obsidian is almost as common as green. Green obsidian from Bing61lor Nemrut Dag was being imported and worked on site, at least some of it from raw nodules. Gray obsidian was being imported largely as blades or as prepared blade cores. This seems to reflect a complex exchange system in which Arpachiyah had a key position. At least some of the green obsidian was being taken from the source to the site as raw or minimally prepared nodules, either by direct access to the sources or through an exchange network to which Arpachiyah had preferential access. The material was processed at Arpachiyahinto blade or blade core form and distributed to surroundingsites. Another site, or sites, must be postulated which was fulfilling the same role for gray obsidian-importing it, processing it, and trading it on. The system included multiple primary centers of importation and distribution. Arpachiyah seems to have participatedboth as a primary distribution center and as a recipient of blades from elsewhere in the system. Control of distribution of the green obsidian must have given Arpachiyah an important position within an exchange network, both as an economic resource and, possibly, as controlling the distributionof a status item. The presence of the gray obsidian, though, suggests that the direct economic importance was not necessarily critical. If it was cheaper in crude economic terms to obtain raw material, why also acquire manufacturedblades made at another site? It is unlikely that exchange systems functioned on a purely economic basis in the Halaf. The nature of the commodity may have been important and its economic value relevant to some extent, but it may have been the act of exchange itself that was most meaningful-a social rather than an economic transaction. It may be that the Burnt House acquired a significant proportion of its status as a setting for such exchanges by providing a social and a ritual context. "Domestic" Objects Along with these more exceptional objects, there were a number of items ignored in the publication




that are somewhat more domestic in nature, including spindle whorls and pierced sherds (A913; fig. 13:4-10). Some of the objects considered in the groups above, such as some of the plainer pottery and all but one of the stone axes, also probablyought to be considered here. These objects might perhaps be found in any Halaf building where material has been preserved in situ. Although they are of good quality, they are not exceptional, and several of them show signs of long use. It is unfortunatelyimpossible to know whethermost of these objects came from the Long Room and Full Room, along with the high prestige objects, or from the other parts of the building. Therefore, we do not know if they represent a domestic context that coexisted in a different part of the same building as the more exceptional elements or whether they were used alongside them. It does seem clear, however, that if the richer deposits in the Burnt House have any status, ritual, or cultic associations, they had not yet been completely separated from a more domestic context. Furthermore, the status items may have been restricted to use at certain times and in specific social contexts. THE DESTRUCTION OF THE BUILDING We should consider the destruction of the building as distinct from its function. The burning of the Burnt House could be accounted for by accident, by hostile invaders, or by an attack from a rival central site, and it is difficult to discount these possibilities entirely. However, there are some indications of a more deliberatedestruction.Some of the finest plates appear to have been deliberately smashed and scattered to the extent that pieces were found at opposite ends of the Long Room prior to the burning. Mallowan, assuming that the destructionwas in war, suggests that this is the natural behavior of an invader (Mallowan and Rose 1935: 106). Yet there may be another explanation. If the quantity and exceptional fineness of undamaged objects left in the building are an indication, there is little evidence of extensive looting or attempted recovery of the objects. As considerable portions of the building were not extensively burnt,it seems unlikely that collapse of the roof in the unburntareas was total or immediate. Attempts could have been made to recover at least the less fragile objects, such as the stone bowls, the obsidian links, and the seals. It is not impossible, of course, that such an effort was made and that what we are left with is what escaped such a salvage

operation. It seems more likely, however, that no such salvage took place. Therefore, we may be looking at a building in which there were large numbers of very valuable objects, some of which may have been deliberately destroyed before the building was burnt and which were left deliberately unrecovered following that fire. It is difficult to avoid at least a tentative conclusion that the end of the building was a ritual one. The destruction was followed by a significant gap in occupation. The subsequent building on the site, TT5, seems much smaller and has a different building plan. Any ritual destruction would have marked the end of the role of the TT6 building and the individual or institution it may have housed. It may also have been a key formal act in a conscious abandonmentof the site. When comparanda are sought elsewhere in the Halaf culture, there are significant parallels on a much smaller scale in the Halaf levels of YarimTepe II. There are three characteristics of the destruction of the TT6 Burnt House that are significant here. * It is a collection of specifically the finest objects. * Some objects may well have been deliberately smashed. * Burning was involved. There are close similarities between these factors and the characteristics of other nonarchitecturaldeposits known at Yarim Tepe II, where extremely fine objects (although not in as large concentrations as at the Burnt House) were smashed, deliberately according to the excavators, and associated with burning (Campbell 1995b: 33). These deposits are often associated with human remains but not always-the presence of human remains seems optional rather than central to the ritual-so this is no barrierto this interpretation.Thus, an interpretationof the end of the Burnt House as a ritual destruction on a grand scale, somewhat similar to that suggested by Munchaev (Mallowan 1977: 96), would fit very closely to what we know of Halaf ritual practice. DISCUSSION It may be useful to summarize some of the conclusions made above. The Burnt House succeeded previous complexes which may have had related functions, although without comparable sets of artifacts from them it is difficult to consider them in more detail. While it is conceivable that some or all of the objects found in the Burnt House may have




been brought to the building only at the very end of its life, it will be assumed in this discussion that they relate primarily to the use of the building. The objects in the building are overwhelmingly, but not entirely, high quality and probably high status. Evidence for formal social ritual, associated with consumption and probably taking place only on certain occasions, is suggested strongly by the polychrome plates. Cultic ritual is much more weakly indicated. There is evidence of large quantities of imported obsidian. Storage of material is furtherindicated by the sealings present. The large number of seals is probably significant, and there are hints that some degree of control over seals may have been exercised. Finally, the building was abandoned,arguably in an act of formal and deliberate destruction, and this area of the site was not reoccupied for a significant period of time. If the Burnt House was a workshop, as has commonly been suggested, one would expect more evidence of manufacture-perhaps wasters, waste from the manufactureof the ground stone vessels, and so on. In particular,one might expect raw materials and partially completed objects to be present and less of a preponderanceof finished, very high quality objects. Only for flint and obsidian is there any compelling evidence for manufacture within the building. The only other potential raw material seems to have been a lump of red ocher that may or may not have been used for paint. We must reject the idea that it was a potter's workshop, although lithics manufacture does seem to be present. Instead, there was a collection of the finest objects of their kind found at any Halaf site. This implies a great degree of control over the products of the best craftsmen by an individual or institution. It seems unlikely that the finest potter or potters and the finest stoneworkers, in various mediums, would all be local to the relatively small site of Arpachiyah. Therefore, it is likely either that the products of fine craftsmen were gathered from a number of settlements or that the craftsmen themselves were persuadedto relocate to Arpachiyahbecause of economic or social reasons. Both cases suggest that Arpachiyah exercised influence over the highest quality, highest status material on a regional basis, although the size of the region is conjectural. The sealings supportthis. Standardizedand organized methods of marking property and controlling ownership, in whatever form, are only likely where significant numbers of people participatein the sys-

tem, that is, when abstract symbols are needed to supplement personal knowledge. This implies considerable integrationand control of society, probably over a significant geographical area. It also indicates the existence of spheres of control other than that of the Burnt House, among which sealings helped to regulate the movement of goods. This in turn implies subsidiary settlements from which sealed goods were transferredto the Burnt House (an essentially unilinear transferof goods), and/or centers equivalent to Arpachiyah that exchanged goods (a bidirectional transfer of goods). If the earlier large tholoi in TT7 and TT8 shared some of the function as well as the plan of the Burnt House, these relationships would have persisted over at least three building phases. This suggests an institutionalized control associated with either a series of individuals or a central institution that could sustain itself over a long period, perhaps in partthrough some forms of quasi-bureaucracy. It seems very likely that obsidian had a key position in the importance of the Burnt House. As an item only accessible throughlong-distance exchange mechanisms, it would have been ideally suited for manipulation by an elite institution in building and reinforcing a position of regional importance. There is additional support for Arpachiyah, although not necessarily the Burnt House, having a key role in the manufacture and distribution of artifacts. Davidson and McKerrell'sneutron activation analysis of Halaf pottery from Tepe Gawra and Arpachiyah suggests that the former site received substantialquantities of pottery from Arpachiyah during the late Halaf, over a distance of some 25 km (Davidson and McKerrell 1980). It is entirely possible that other materialsmay have been involved but have not survived in the archaeological record. The critical importance of the distribution network that may have centered on Arpachiyah was probably not economic or even crudely political but social. It has been argued above that, with the obsidian, it may have been the act of exchange that was meaningful. Arpachiyah may have been able to center part of that exchange activity on the Burnt House owing to its preferential access to one or more particular sources, but it also participated in the exchange systems of obsidian from other sources. It also seems clear that exchange was embedded in a particularsocial context with secular and possibly sacred rituals providing a formal setting in which relationships could be negotiated and reconfirmed.




Given the possible evidence for control over the seals which may have been used as partof a proto-bureaucratic system associated with the BurntHouse, it may be that the key role of the individual or institution within the Burnt House was as a mediatorregulating exchange transactions. Given the traditional approaches to settlement hierarchy in the Near East, where size is equated to centrality, it is surprising that Arpachiyah is so small. It may have been at the center of a wide-ranging economic and probably political network, but it was a center that did not have a large concentration of population. It is equally interesting that there is not an obvious social hierarchy, supposedly the inevitable accompaniment of this degree of social complexity in a traditional chiefdom setting. The only clues to such a hierarchyare indirect. Certainly, there is a major concentration of wealth in the Burnt House, but we need not assume that it is associated with an individual rather than an institution. Similarly the wealth in some burials at Yarim Tepe, with the single possible exception of the outstandingburial on Yarim Tepe I (with the skull of a huge bull and about 200 astragalus bones along with pots and ground stone objects; Merpert and Munchaev 1971: 17), is probably not associated with the deceased individual; ratherit hints at an attemptto gain prestige by conspicuous consumption (Campbell 1995b). Finally, the possible restrictionof polychromepotteryto specific, high-status contexts suggests an attempt to symbolize power by specific stylistic devices. While all this does perhaps imply individual status, it is also compatible with power being concentrated in an institution. While it is difficult to apply the more traditional definitions of chiefdoms to this situation, Kristiansen's proposalsarepossibly useful (Kristiansen1991). He sees a new level of "decentralized, stratified society" as a vital jump from a tribe-based society on its way to a state. Chiefdoms are related to the tribal level, where kin is the vital means of linking people, ratherthan Kristiansen'sproposed new level. "Decentralized, stratified society" is seen as having decentralized subsistence patterns, formalized ownership of land, and no towns. Instead there is specialized craft production,connected to elites, and control and taxation can play a significant economic role in development. Sahlins had previously highlighted the importance of the shift from kinship to property as being a crucial development (Sahlins 1972: 92-94). It would be foolish to accept this as

a new category into which late Halaf society in northernIraq could be slotted without qualification, but Kristiansen'smodel is a surprisinglyclose fit and may be a useful starting point for future work. This model for Halaf society contrasts strongly with the picture emerging in southeast Turkeywhere Halaf sites such as Domuztepe (Campbell et al. 1999) and Kazane (Coursey, Bernbeck, and Pollock 1998) are now known to reach up to about 20 ha in size. There, large population centers seem to be demonstrable, and there are strong argumentsthat they are acting as social and economic centers as well. It is certainly possible that there were similar large Halaf sites in northernIraq-they may well be buried under later urban tells such as Nineveh and Erbil-but we cannot, at present, point to any clear examples despite the very considerable amount of survey that has been carried out. Instead, our only good candidate for a central settlement is Arpachiyah and it is undeniably small. It is too early to construct anything more than a suggestion, but it may be that there are two very different types of regional integration within the Halaf by the mid-fifth millennium. In north Syria and southeast Turkeythere are major population centers at the center of the system while, perhaps, in northeast Syria and northern Iraq, regional stratification does not lead to population concentration. It is certainly possible that beneath the traditional regularity of the Halaf lurk radically different types of social complexity in different regions. Undoubtedly, we will not understand the developmentof highly complex societies in Mesopotamia during the fifth and fourth millennia unless we gain a more sophisticated understandingof the social processes in the earlier part of that period when key prerequisites of state and urbanism may have been emerging.

THE CATALOG OF OBJECTS IN TT6 The catalog of objects has been compiled from several sources. The basic record is taken from Mallowan's original small finds records (now in the British Museum) and the list of the division of objects between Baghdad and Mallowan (kept in the Iraq Museum in Baghdad). Some additionalfacts are only given in the publication (Mallowan and Rose 1935). Where any information is recorded on the artifacts (usually in pencil), it has been taken as being more




accuratethan the small finds recordsince it was probably added in the field. The original scheme of small finds numbering has been retained: a capital A followed by a number. Where Mallowan assigned a single number to a group of several objects, they have been designated with Mallowan's number,followed by a letter to provide unique numbers; e.g., A913a to A913g. Where Mallowan did not allot a small finds number to an artifact, new numbershave been assigned starting at A1001. Wherever possible the Arpachiyah small finds have been located in museums, recorded in detail and, where necessary, redrawn. However, in cases where no museum is listed, the objects have not yet been traced. Also a considerable number of the objects in the Iraq Museum were not available for

examination or had been dispersed to regional museums. In both these cases, Mallowan'soriginal small finds record has been used as the primary source of details, supplemented by the publication of the site. In several cases also, in the IraqMuseum and in Iraqi provincial museums, an object could be seen in a display case but not examined in detail. This is intended to be a provisional summarycatalog. When the remainder of the objects have been examined, it is hoped that a complete and comprehensive catalog can be published. The location contains all the information in Mallowan's notes, even where the information is difficult to interpret. Illustrationsare only published with this article where the artifact has not been previously illustrated or where the original illustration is defective. Ideally this article should be read with the original report at hand.

Number Museum Illustration Location Description

Polychrome Pottery

A512 A528 A529 A530 A739 A745 A746 A748

IM ? IM 14734 BM 127504 IM 14765 IM 14736 BM 127585 IM 17836 IM 17837

M andR pl. 20:a fig.5 fig.6

in Fragsscattered TT5 andTT6 BurntHouse BurntHouse BurntHouse

M andR pl. 20:b M andR pl. 17:b M andR pl. 16:a M andR frontispiece, pl. 19:1 M andR fig. 54:4 M andR pl. 18

BurntHouse BurntHouse(?) BurntRoom BurntRoom Khallaf'sroom BurntHouse

Jarwithhorizontal incisionsoverpainted withred andblackpaint. Platewithred,black,andwhitepaint. Burnt. Platewithred,black,andwhitepaint. Burnt. Bow-rim jar withred andblackpaint. Burnt. Jarwithblackandredpaint.Undamaged by fire. Platewithredandblackpaint.Slight burning. Platewithbrownandredpaint. Platewithred,black,andwhitepaint. Slightburning. Platewithred andblackpaint.Heavily burnt. Platewithredandblackpaint.Burnt.

A749 A750

BM 127502 BM 127507

A751 A752 A753 A755

IM 14733 BM 127508 IM 14753 IM 14724

M andR pl. 18, pl. 22:10 M andR pl. 15 M andR pl. 14 M andR pl. 17:a

BurntHouse Eastend of LongRoom on northside of TT6 BurntRoom Platewithredandblackpaint.Heavybut variable burning. BurntHouse Platewithblack,red,andwhitepaint. BurntHouse Platewithredandblackpaint. BurntHouse Platewithred andblackpaint.




CATALOG-(Continued) Number Museum Illustration

M and R pl. 5:a M and R fig. 69:5



Vessels OtherCeramic IM 14751 A99a and 14977 BM 127529 A289

A487 A491 A492

IM 14832 BM 127630 BM 127633

M and R fig. 41:10 M and R fig. 42:11 M and R fig. 43:12

A493a A 505b A515 A516

IM 14852 IM 14849 BM 127511 BM 127554

M and R fig. 43:10; fig. 8:6 M and R fig. 42:12 M and R fig. 56:2 M and R fig. 59:5; fig. 8:2

A517 A518 A521 A526

I of A 53/300 M and R fig. 64:4; fig. 8:3 IM 14826 M and R fig. 7:1 M and R fig. 58:2 IM ? I of A?

A527 A535 A740 A741 A742 A743 A744 A747 A754 A763 A768 A769 A802

unknown I of A 53/??? fig. 8:4 BM 127530 IM 15338 BM 127583 IM 14726 IM 14760 IM 14741 IM 14720 M and R fig. 60:5 M and R fig. 60:3 M and R fig. 57:1 M and R fig. 64.2, pl. 19.8 M and R fig. 53:1

M and R fig. 55, pl. 19:4 IM 14762 M and R fig. 58:1 BM 127564 M and R fig. 59:1 I of A 53/304 M and R fig. 60:4 IM 14821 M and R fig. 79:6

In BurntHouse or adjoin- Zoomorphic vessel in the form of a pig or hedgehog. Red paint. ing Burnt House Bowl with red paint. Broken and mended Full Room in antiquity with signs of extensive wear. Unpainted cup with open spout. Burnt House Small jar with red paint. Burnt House Roughly made small jar with possibly accidental traces of brown paint on the exterior. Unburnt. Small unpaintedjar with a chipped rim. Burnt House Burnt. Small jar with red paint. Burnt House Bowl with orange to brown paint. Burnt. Burnt House Jar with dark gray paint. Very heavily burntand sitting at an angle at the time of burning. Burnt House Jar. Heavily burnt and sitting at an angle at the time of burning. Bowl with brown paint. Heavily burnt. Burnt House Bowl with black paint. Burnt champagne vase fragment with black paint recorded in small finds catalog, but the identification with sherds in the I of A is tentative. Burnt House Fragment of a vessel. Details unknown. Burnt House Jar, probably painted but too heavily burnt to be sure. Burnt House Bowl with dark red paint. Heavily burnt. Burnt House Bowl with red-brown paint. Burnt House Plate with black and red paint. Burnt Room Smaller duplicate of A742 Burnt Room Jar with two lug handles and red-black paint Burnt Room; same end Plate with light red paint. Little burnt. of room as A748 Burnt Room Plate with red paint. Burnt Room Burnt House Burnt House Burnt House Probable lid with red paint. Jar with dark red paint. Bowl with red paint. Unpainted oval bowl with two ledge handles. Reported stolen from Basra Museum (Baker, Matthews, and Postgate 1993: 67).




Number Museum Illustration Location Description Pink and white mottled limestone bowl. Burnt House Obsidian jar. The exterior has been pecked to shape but not smoothed, leaving a rough surface with many glinted facets. The interior appears to have been bored and the rim has been ground flat. Black and white diorite bowl. Finely made and highly polished stone bowl. Bowl, almost certainly stone but so burnt and encrusted that it is difficult to be certain. Chipped aroundrim during use. Burnt matter in interior is probably original contents. Dark gray limestone bowl. Black steatite bowl. Gray and cream mottled limestone champagne vase fragment. The pedestal base is circular but the upperbowl is oval. Small trough of black steatite.

StoneVessels IM 14889 A409 A411 IM 14860

M and R pl. 5:d, fig. 44:16 M and R pl. 5:c, fig. 44:15

A412 A413 A414

? BM 127793 BM 127790

M and R pl. 10:b, fig. 44:13 M and R fig. 44:2; Burnt Room fig. 8:5 M and R fig. 44: 1; Burnt Room fig. 8:7

A415 A420 A912

IM 14823 ? IM 14858

M and R fig. 44:3 M and R fig. 44:6 Burnt House


IM 15032

M and R pl. 10:a

Burnt House

StoneAxes IM 14883 A663 A664 IM 14884

M and R pl. 12:b; Burnt House fig. 12:3 M and R pl. 12:b; Burnt House fig. 12:2


Fitzwilliam E.216.1934 BM 127765

M and R pl. 12:b

Burnt House


M and R pl. 12:b; fig. 12:4

Burnt House

Dark gray stone ax. Working edge is lightly chipped through use. Black stone ax. Heavy use along working edge, with probable resharpening. Probable hafting traces on the butt and traces of what seems to be bitumen. This is probably the ax found with a carbonized shaft and illustrated in M and R fig. 51:12. Dark gray stone ax. Working edge is heavily worn and chipped. Probable hafting traces around butt. Small stone ax made from a hard, very fine grained stone, dark green in color, but where thin, a translucent midgreen. Very finely made with sharp corners and edge. Almost exactly symmetrical and it is very highly polished. There is a light chipping on the butt and the cutting edge is fractionally chipped (at 8x magnification only).




CATALOG-(Continued) Number
A667 A668

? ?

M and R pl. 12:b M and R pl. 12:b

Burnt House Burnt House

Ax made from black stone. Ax made from black stone.

A920 A921 A941 A942 Beads A587 A891

BM 127717 IM 15015 IM 15050 BM 127710

M and R pl. 10:a, fig. 52:3 M and R pl. 10:a M and R fig. 46:3 M and R fig. 47:24

Burnt House Burnt House Burnt House Loose in soil of TT6

Female figurinemade from gray limestone. Male figurine made from alabaster. Ceramic bird figurine with black paint. Sun-dried clay figurine of a bird.

? BM 127677 M and R pl. 6:b, fig. 51:5 M and R pl. 6:b M and R pl. 6:b M and R pl. 6:b M and R pl. 6:b M and R pl. 6:b, fig. 51:7 M and R pl. 6:b, fig. 51:7 M and R pl. 6:b M and R pl. 6:b M and R fig. 51:8 M and R pl. 6:b, fig. 51:1 M and R pl. 6:b M and R pl. 6:b M and R pl. 6:b

Burnt Room"Girgis' room" Burnt House

Black steatite figurine of a duck. Black steatite double ax bead. Partly broken in antiquitybut the smoothing of the break suggests a use after breaking. Black steatite double ax bead. Dark gray steatite double ax bead. Dark gray steatite double ax bead. Dark gray steatite double ax bead. Black steatite duck bead. Black steatite duck bead. Black steatite duck bead. Black steatite duck bead. Black steatite duck bead. Lozenge bead of black steatite with groove along one side. Ring bead of dark gray stone. Burnished, black ceramic conical bead. Mostly black and gray steatite. Five illustrated from a total of 14. Lentoid, lozenge, ring, flattened double conoid beads noted. 13 black and gray steatite beads. Barrel lozenge, ring, circular beads noted. 17 obsidian links with groundandpolished surface and biconical perforations. Two perforations at one end and only one at the other. Green obsidian. One perforation at both ends. Green obsidian (100 mm long). Two perforations at each end. Green obsidian. One perforationat both ends. Red obsidian.

A862 A863 A864 A865 A870a

IM 15004 IM 15005 BM 127676 IM 15006 BM 12767

Burnt Room Burnt Room Burnt House Burnt Room Burnt Room Burnt Room Burnt Room Burnt Room Burnt House Burnt Room Burnt House Burnt House

A870b-d IM 15036 A871 A872 A873 A875 A877 A878 A879

IM 15040 IM 15044 BM 127818 IM 15045

A880 A905 A905a A905b A905c A905d

IM 15088 BM ?

M and R pl. 6:b M and R pl. 11:b

Burnt Room At the end of a single room




Number A905e A905f A905g A905h A905i A905j A905k A9051 A905m A905n A905o A905p A905q A906 IM 1505615075 M and R pl. 11:b; Burnt House; all at one Anon. 1985: 353 end of single room Museum Illustration Location Description Two perforations at one end and only one at the other. Green obsidian. Two perforations at each end. Gray obsidian. Two perforations at each end. Gray obsidian. One perforation at both ends. Green obsidian. One perforation at both ends. Green obsidian. One perforation at both ends. Gray obsidian. One perforation at both ends. Gray obsidian. Two perforations at one end and only one at the other. Green obsidian. One perforation at both ends. Green obsidian. Two perforations at each end. Green obsidian. Two perforations at one end and only one at the other. Green obsidian. One perforation at both ends. Gray obsidian. Two perforations at each end. Green obsidian. 19 rectangularobsidian links with ground and polished surface and biconical perforations. The links are all made from long, flat blades. The unillustrated links are very similar to the illustrated ones or were not available for study. Two perforations at one end and only one at the other. Green obsidian. Two perforations at each end. Green obsidian. Green obsidian. One perforation at both ends. Green obsidian. One perforation at both ends. Green obsidian. One perforation at both ends. Green obsidian. Gray obsidian. Green obsidian. IM 15068 fig. 10:6 One perforation at both ends. Green obsidian.

A906b A906c A906e A906f A906g A906i A906j A906k A9061

IM 15058 IM 15059

fig. 10:1 fig. 10:2

IM 15062 IM 15063 IM 15065

fig. 10:3 fig. 10:4 fig. 10:5




CATALOG-(Continued) Number
A906m A906n A906o A906p A909

IM 15069 IM 15070 IM 15071 IM 15072 BM 127814

fig. 10:7 fig. 10:8


One perforation at both ends. Green obsidian. Green obsidian. Two perforations at one end and only one at the other. Green obsidian. One perforation at both ends. Green obsidian. Obsidian and cowrie shell necklace. See M and R, p. 97 for description. Five obsidian, 3 pebble, and 16 cowrie beads. Burnt House. Numbering below is taken as strung in the BM, starting with the stone pendent to the top and running clockwise. Stone pendent. Obsidian. Cowrie shell bead. Cowrie shell bead. Cowrie shell bead. Green obsidian link with ground and polished surface. Cowrie shell bead. Cowrie shell bead. Cowrie shell bead. Green obsidian link with ground and polished surface. Cowrie shell bead. Cowrie shell bead. Gray stone similar to A909a. It has been pierced along its length with a cylindrical hole (i.e., as reconstructed, it is strung differently from the other beads in the necklace). Both ends are damaged, probably in antiquity. Cowrie shell bead. Cowrie shell bead. Ground and polished obsidian link. Cowrie shell bead. Cowrie shell bead. Cowrie shell bead. Green obsidian link with ground and polished surface. Cowrie shell bead. Cowrie shell bead. Cowrie shell bead. Green obsidian link with ground and polished surface.

fig. 10:9 M and R pl. 11:a Burnt House

A909a A909b A909c A909d A909e A909f A909g A909h A909i A909j A909k A9091 A909m

fig. 9:1 fig. 9:4

fig. 9:5

fig. 9:2

A909n A909o A909p A909q A909r A909s A909t A909u A909v A909w A909x fig. 9:6 fig. 9:3 fig. 9:7




Number A914 A919 Museum IM 15090 ? Illustration Location Burnt House Burnt House Description Mostly small frit ring beads. Mixed group of beads with small frit ring beads, some shell, one reddish brown stamp sherd. Limestone cylinder bead, neatly made and finished with a cylindrical hole through the center. Cream to very pale brown. One of three beads lying in A414 and, if associated with it, from the Burnt House. A bead cut and ground down from a larger piece of pottery-probably a sherd of Halaf fabric. Cylindrical clay bead, broken at one end.





As A1002a

A1002c Seals A43 A554


As A1002a

BM 127648 IM 15003

M and R fig. 50:12, pl. 8:a M and R pl. 8:a

Outside tholos(?)


I of A 53/473 M and R pl. 7:a; fig. 11:1 ? BM 127680 BM 127667 M and R pl. 6:b M and R pl. 6:b; fig. 11:2 M and R pl. 6:b, fig. 51:11 Burnt House Burnt House Burnt House

A860 A866 A867

A868 A869 A874

BM 127663 IM 15036 BM 127660

M and R pl. 6:b M and R pl. 6:b M and R pl. 6:b, fig. 51:12; fig. 11:3

Burnt House Burnt Room Burnt House


IM ?

M and R pl. 6:b

Burnt Room

Pottery. Triangularwith linear markings and two perforations. Greenish gray limestone. Flat, dropshaped. Linear markings with perforation in the middle. Black steatite. Circular with perforated handle. Linear markings on both sides. Broken. Black steatite. Double ax-shaped seal. Black steatite, circular disc with a knob handle. Probable unfinished seal. Seal of a white marble, probably a fragment from a broken crescent. It is pierced at one end and is incised on one side. The end away from the hole is broken but the broken area is also polished, suggesting extended use after breakage. Black steatite. Shield-shaped with linear markings on the flat side. Black steatite. Lozenge-shaped. Brownish gray steatite seal. The incisions are deep and neat. In the center of the incised face all of the raised, central part of the design has been deliberately abraded away; approximately 1-2 mm of material must have been removed. This must be deliberate, done by rubbing the seal facedown on an abrasive material. Black steatite lozenge-shaped seal.




CATALOG-(Continued) Number
A890 ?


M and R pl. 7:b

TT6 or F 0.9

Dark gray (greenish) serpentine(?) or stone. Flat, double conoid with linear markings on one side. Black steatite drop-shaped seal. Incised bead/seal made from yellowbrown stone, perforated along its length. Its surface is smooth, and polished on the raised portions-notably along the ridge on the reverse and the ridges of the decoration, as though polished through wearing and long use. The decoration is deeply incised (up to 3 mm).

A892 A893b

BM 127659 BM 127815 or M and R pl. 7:b BM 127817

Burnt House

A619 IM 15185 M and R pl. 9:b Black clay. 15 are poorly illustrated in M and R, and 10 out of the total of 19 are in Baghdad. Double conoid, perforated at both ends and containing string marks. See also von Wickede 1990. Dark brown to black clay pressed around a string or strings (2 strands enter but only 1 exits) in a double conoid. It has 10 seal impressions, all from the same seal, all oriented the same way, and evenly spaced in pairs around the circumference. Very similar to A619a except that the seal was pressed arounda two- or three-strandstring. There are eight impressions of the same seal as A619a evenly spaced in pairs aroundthe circumference. Burnt House, BurntRoom Black clay. Double conoid, perforatedat both ends and containing string marks. Same impressions as A619a and A619b. Burnt House, Burnt Room Half of a sealing wrapped around string impressions. Unclear impressions. Burnt House, BurntRoom Dark brown clay. A two-strand string runs through the center. It has been impressed nine times around its circumference by the same seal. The design is indistinct, the only visible part being a line running down the center. It may be the impression of what could be a bead seal with a single linear groove (cf. A875). Burnt Room


BM 127693

M and R pl. 9:b; fig. 11:4


BM 127696

M and R pl. 9:b; fig. 11:5


BM 127701

M and R pl. 9:b

A619d A619e

I of A 53/461 I of A 53/967




Number A619f Museum I of A 53/467 Illustration Location Description


I of A 53/467


I of A 53/619

A619i A620

IM 15185

von Wickede 1990: no. 61 M and R pl. 9:b

Burnt House, Burnt Room Dark gray-brown clay. Formed around a two-strand string. Only one surface is preserved with the rest broken off. There is a single seal impression, too faint to make out the design apartfrom two short lines. BurntHouse, BurntRoom Lump of darkgray-brown clay, broken at one end. It is not formed around a string. There are two distinct chaff impressions and two possible seal impressions. These latter are very indistinct and may not be from seals. Lump of dark gray clay shaped around a three(?)-strandstring and impressed by four seals, two probably of the same seal. The other two are different and may both originate from a second seal; however, they are very indistinct. Half the sealing is broken in antiquity. Sealing. Black sun-driedclay. Six illustrated from the total of eight. Disk-shaped clay lumps stamped with Halaf seals. Dark brown to black, subcircular, flattened lump of clay. There are five deep seal impressions in the clay. They are probably from two different seals: one for the upper three impressions, the other for the lower two. The back is relatively smooth with no indication that it was used to seal anything. Brown clay sealing with four circular seal impressions, all from the same seal. As with A620a there are no markings on the back. Remains of a circular sealing. There are two impressions on the upper surface. Both are shallow but are clearly Halafstyle impressions and almost certainly from the same seal. Sealing with same impressions as on A619a-c. Sealing with same impressions as on A619a-c. Sealing with same impressions as on A619a-c.


BM 127697

M and R pl. 9:b; fig. 11:7


BM 127698

M and R pl. 9:b; fig. 11:6


I of A 53/468 M and R pl. 9:b

A620d A620e A620f

IM 15184 IM 15184 IM 15184

von Wickede 1990: no. 58 von Wickede 1990: no. 59 von Wickede 1990: no. 62




CATALOG-(Continued) Number Museum Illustration Location Description

Stone palette made from strikingly colored stone; very colorful with gray to dark gray veins, orange to red-brown veins, and white veins running throughout the stone. Originally associated with a lump of red ocher. One end is broken. Blue-gray veined limestone mace head. Pumice and basalt or sandstone objects. Two pumice, one basalt (M and R, p. x) or sandstone (M and R, p. iv). Object carved from pumice. The lower part is D-shaped in plan, and it tapers slightly to a spheroid "head."The base is flat and the object sits steadily on it. There is a slight ridge running vertically down the front of the head that may suggest a nose. One of A904b and A904c is pumice; the other is either sandstone or basalt.

Miscellaneous BM 127786 A901

M and R fig. 52:4; Burnt House fig. 12:5

A903 A904


IM 14866 IM 14890 and 14891; BM 127728 BM127728

M and R pl. 10:c Full Room, Burnt House M and R pl. 10:d-g Full Room, Burnt House

M and R pl. 10:e; fig. 13:1

Full Room, Burnt House

A904b A904c A907 A908 A910 A911

IM ? IM ?

M and R pl. 10:d or g M and R pl. 10:d or g

Full Room, Burnt House Full Room, Burnt House Burnt House Burnt House

IM 1507615080 IM 15082 IM 15083

M and R pl. 12:b fig. 15

Burnt House Burnt House


IM 14900

Burnt House

A913a A913b

IM 14900 IM 14900

fig. 13:4 fig. 13:5

Burnt House Burnt House

A913c A913d A913e A913f A913g

IM 14900 IM 14900 IM 14900 IM 14900 IM 14900

fig. 13:6 fig. 13:7 fig. 13:8 fig. 13:9 fig. 13:10

Burnt House Burnt House Burnt House Burnt House Burnt House

Group of complete flint knives. Several obsidian blades, including one translucent. Some also exported. Obsidian chips and cores. Some also exported. Obsidian cores. Three are mentioned in the small finds catalog but four are present in the Iraq Museum. Pottery discs and spindle whorls. Seven objects (numbered as A913a-g) in Baghdad. Unknown numbers were also exported. Pierced pot disc carefully cut from a sherd. Pierced pot disc cut from a sherd with polish on the edge probably caused by extensive handling. Pierced pot disc roughly chipped from a sherd. Pierced pot disc, carefully cut from a sherd. Ring bead possibly cut from a sherd and highly smoothed. Spindle whorl. Spindle whorl.




Number A915a Museum Illustration Location Burnt Room Description Long bone tube, discolored through burning. The surface is highly burnished. One end has been cut and smoothed; there are clear abrasion marks across it. There are also deep abrasions and some notches on the surface around the center. One end of a tube, similar to A915a. The bone is burnt, with a very high polish. The surviving end is very worn but has clearly been cut to shape and then extensively worn during use. One end of a bone tube, apparentlysimilar to A915a. It is burnt with a highly burnished surface. There is a small area at one end where the polish has been worn through to a rougher finish. At around the point of the break there are three notches similar to those on A915a. One end of a bone tube, apparentlysimilar to A915a. It is very similar in detail to A915c although with greater abrasion to the surface and with no notches. Trumpet-shapedgray marble object. Large lump of red paint. Also exported. Conical lead lump. Burnt bone "knuckle bone," heavily burnt. Both ends have been worn to some degree, especially around the edges. Light gray stone "knuckle bone." Well made and polished with a D-shaped section. There are grooves in the shape of a cross on the base. It will not stand upright. There is little sign of wear or sign of fire damage. Stone "knuckle" bone. Very similar to A992b. It will stand upright unsupported. There is a single groove running across the base. Stone "knuckle bones." Slabs or parts of slabs of a dark gray to purple tabular flint and the flakes struck from it. There are two other sizable chunks with no retouch and many smaller pieces (20-30), at least one of which conjoins the first retouched piece described below.

I of A 53/374 fig. 14:1


I of A 53/374 fig. 14:2

Burnt Room


I of A 53/374 fig. 14:3

Burnt Room


I of A 53/374

Burnt Room

A915e A916 A918 A992a

BM 127739 IM 14966 IM 14864 BM 127718

M and R pl. 10:a, fig. 52:1-2

Burnt Room Burnt House Burnt House Burnt House


BM 127719

M and R pl. 10:a, fig. 52:1-2; fig. 13:2

Burnt House


BM 127720

M and R pl. 10:a, fig. 52:1-2; fig. 13:3 M and R pl. 10:a, fig 52:1-2

Burnt House

A992d-f A1001

IM 15048 I of A

Burnt House TT6 BR




CATALOG-(Continued) Number
A1001a A1001b A1001c

I of A I of A I of A

fig. 12:1


Roughly triangularwith retouch. Areas of cortex on one side. Subrectangularwith a retouched edge. Irregularwith retouch along one edge, with some cortex remaining.

ProbableTT6BurntHouseArtifacts A1003 I of A 53/337 fig. 7:1


Ashmolean 1934.106

fig. 7:2

Marked TT5

Plate with orange to dark brown paint. Partially burnt. The burning, the vessel shape, and decoration suggest that this may have originated from the Burnt House, but there is no marking on the plate to indicate this. Fragmentaryplate of which 15 sherds are preserved. The paint is red-brown and brown-black. Shape and decoration very similar to plates from the Burnt House. Two of the sherds have evidence of burningwhich does not occur on conjoining sherds-similar to the plates known to have come from the Burnt House. The edges of most of the sherds are clearly worn, unlike any of the observed sherds from TT6. Signs of repeated wear on the base.

recorded Artifacts from TT7and TT5whichmighthavecomefrom TT6 A41 BM 127657 M and R pl. 8:a TT6-7 South side A336 A339 A348 A419 A488 A511 A524 A525 A532 A557 A559 A578 IM 14879 BM 127758 IM 15156 BM 127634 I of A 53/290 Louvre 1973 IM 15702 BM 127555 I of A 53/302 I of A 53/458 M and R pl. 8:a IM 15001 M and R pl. 8:a IM 15029 M and R pl. 7:a M and R pl. 8:b M and R pl. 8:b M and R pl. 8:b M and R fig. 44:8 M and R fig. 41:4 M and R pl. 16:b M and R pl. 19:6 fig. 8:1 TT5 TT5-6 by kiln TT5 TT5-6 TT5 TT5 TT7, outside tholos

Unfinished seal. A few linear markings on the flat side. Dark green stone. Dark greenish gray stone ax.


IM 15028

M and R pl. 7:a

Gray stone ax. Light green stone ax. Gray limestone vessel. Small unpaintedpot. Pot with red and dark orange paint. Near-complete pot with brown paint and buff fabric. TT5 Near-complete jar with black and red paint. TT5-6, edge of Halaf ter- Pot with red-brown paint. race and 50 cm below it TT6-7 Flat seal with rough linear markings. TT7 Flat, drop-shaped,black steatite bead/seal. TT6-7 loose soil Drop-shaped, black steatite seal/bead; unmarkedand, therefore, unfinished. Convex one side. TT6-7 loose soil Drop-shaped amethystine quartz seal; linear markings on the flat side.




Number A580 A581 A584 A761 A881 A917 A924 Museum IM 15002 BM 127663 BM 127668 IM 14752 BM 127669 IM 14863 IM 14842 Illustration M and R pl. 7:a M and R fig. 50:10, pl. 7:a M and R fig. 50:23, pl. 7:a M and R fig. 64:1 M and R pl. 7:b, fig. 50:26 M and R fig. 45:10 Location TT5 loose soil TT7 Variously attributedto TT7 and TT6 TT5-6 AK. W.C. TT5 TT5-6 Description Ovoid black steatite/stone seal; linear markings on flat side. Gray limestone seal; horizontally ribbed with linear markings. Triangularblack steatite seal with linear markings. Pot with red-brown paint. Crescent-shaped black steatite seal with linear markings. Female figurine with red paint on light drab fabric. Pot disc in situ in jar neck stuck with bitumen. Disc cut from bowl base with faint trace of rosette. Sun-dried clay figurine with black paint.

M and R fig. 49:23 TT7

A940 A1005

BM 127725 M and R fig. 45:16 TT5 I of A 53/287 TT5. TT AK WE House 5 Unpainted pot.

Abbreviations M andR: Mallowan andRose 1935 I of A: Institute of Archaeology, London BM:BritishMuseum IM:IraqMuseum in boldindicateillustrations in this article. NOTE: Figurereferences

ACKNOWLEDGMENTS This article has its distant origins in a paper given at the British Association for Near Eastern Archaeology in 1988 and later in a chapter of the writer's Ph.D. dissertation (Campbell 1992) although it has been almost entirely rewritten. Comments on that original version from Trevor Watkins, Edgar Peltenburg, and Peter Akkermans were gratefully received. The original excavation records are in the British Museum, and many thanks are due to John Curtis, in particular, for access to them. The artifacts from Arpachiyah are widely scattered, and the following institutions and individuals are gratefully acknowledged: The IraqMuseum, The British Museum (John Curtis),The Ashmolean Museum (RogerMoorey), Instituteof Archaeology, London (Peter Parr and Harriet Crawford), Birmingham (Philip Watson), The Fitzwilliam Museum, Cambridge, The Royal National Museum of Scotland (Elizabeth Goring). The original drawings are by the writer and were sometimes drawn in difficult circumstances; any errorsare my responsibility. Many thanks are due to Bronwen Campbell for the inked drawings.

REFERENCES Akkermans, P. M. M. G. 1993 Villages in the Steppe: Later Neolithic Settlement and Subsistence in the Balikh Valley, Northern Syria. Archaeological Series 5. Ann Arbor:InternationalMonographsin Prehistory. Akkermans, P. M. M. G., and Duistermaat, K. Of Storage and Nomads: The Sealings from 1996 Late Neolithic Sabi Abyad, Syria. Paleorient 22, no. 2: 17-44. Anonymous 1985 The Land between TwoRivers: TwentyYearsof Italian Archaeology in the Middle East: The Treasuresof Mesopotamia. Turin:II Quadronte Edizioni. Baker, H. D.; Matthews, R. J.; and Postgate, J. N. 1993 Lost Heritage: Antiquities Stolen from Iraq's Regional Museums, Fasc. 2. London: British School of Archaeology in Iraq.




Breniquet, C. La Disparition de la Culture de Halaf: les 1996 Origines de la Cultured'Obeid dans le Nord de la Mesopotamie. Bibliothbque de la D616gation arch ologique frangaise en Iraq 9. Paris: Editions Recherche sur les Civilisations. Campbell, S. 1992 Chronology, Culture and Change in the Late Neolithic of North Mesopotamia. Ph.D. Dissertation, University of Edinburgh. 1995a The Small Finds. Pp. 147-60 in Excavations at Kharabeh Shattani, Vol. 2, eds. D. Baird, S. Campbell, and T. Watkins. Edinburgh:Department of Archaeology, University of Edinburgh. 1995b Death for the Living in the Late Neolithic in North Mesopotamia. Pp. 29-34 in TheArchaeology of Death in the Ancient Near East, eds. S. Campbell and A. Green. Oxford: Oxbow. Campbell, S.; Carter, E.; Healey, E.; Anderson, S.; Kennedy, A.; and Whitcher, S. 1999 Emerging Complexity on the Kahramanmara? Plain, Turkey:The Domuztepe Project, 19951997. American Journal of Archaeology 103: 395-418. Cann, J. R. 1983 Petrology of Obsidian Artefacts. Pp. 227-55 in The Petrology of Archaeological Artefacts, ed. D. R. C. Kempe and A. P. Harvey. New York: Oxford. Charvit, P. 1988 Archaeology and Social History: The Susa Sealings Ca. 4000-2340 B.C. Paleorient 14, no. 1: 57-63. 1994 The Seals and Their Functions in the Halaf- and Ubaid-Culture(A Case Study of Materialsfrom Tell Arpachiyahand Nineveh 2-3). Pp. 9-15 in Handwerkund TechnologieimAlten Orient:Ein Beitrag zur Geschichte der TechnikimAltertum, Internationale Tagung Nerli, 12.-15. Miirz 1991, ed. R.-B. Wartke.Mainz: von Zabern. Chataigner, C. 1998 Sources des artefacts du Proche Orient d'apres leur caract6risationg ochimique. Pp. 273-324 in L'Obsidienne au Proche et Moyen Orient, eds. M.-C. Cauvin, A. Gourgaud, B. Gratuze, N. Arnaud, G. Poupeau, J.-L. Poidevin, and C. Chataigner. BAR International Series 738. Oxford: B.A.R. Coursey, C.; Bernbeck, R.; and Pollock, S. 1998 Excavations of the Halaf Occupationat Kazane Hdytik, 1998. Neo-Lithics 3/98: 6-7. Curtis, J. 1981 Arpachiyah. Pp. 30-36 in Fifty Yearsof Mesopotamian Discovery, ed. J. Curtis. London: British School of Archaeology in Iraq.

Davidson, T. E., and McKerrell, H. 1980 The Neutron Activation Analysis of Halaf and cUbaid Pottery from Tell Arpachiyah and Tepe Gawra. Iraq 42: 155-67. Healey, E. In prep. The Role of Obsidian in the Halaf. Ph.D. Dissertation, University of Manchester. Hijara, I. 1978 Three New Graves at Arpachiyah. World Archaeology 10: 125-28. 1997 The Halaf Period in Northern Mesopotamia. London: Nabu. Hijara, I.; Watson, J. P. N.; and Hubbard, R. N. L. B. 1980 Arpachiyah 1976. Iraq 42: 131-54. Kristiansen, K. 1991 Chiefdoms, States, and Systems of Social Evolution. Pp. 16-43 in Chiefdoms: Power, Economy and Ideology, ed. T. K. Earle. Cambridge: CambridgeUniversity. Maisels, C. K. 1999 Early Civilizations of the Old World. London: Routledge. Mallowan, M. 1977 Mallowan's Memoirs. London: Collins. n.d. Arpachiyah Records. Notes from the excavation. London: British Museum. n.d. Photographs, Vol. 1. (BM archives, vol. 196.7). London: British Museum. n.d. Small Finds Record from Arpachiyah. London: British Museum. Mallowan, M. E. L., and Rose, J. C. 1935 Excavations at Tell Arpachiyah, 1933. Iraq 2: 1-178. Merpert, N. I., and Munchaev, R. M. 1971 Excavations at Yarim Tepe 1970: Second Preliminary Report. Sumer 27: 9-22. 1984 Soviet Expedition's Research at Yarim Tepe III: Settlement in Northwestern Iraq, 19781979. Sumer 43: 54-68. 1993 Yarim Tepe I. Pp. 73-114 in Early Stages in the Evolution of Mesopotamian Civilization: Soviet Excavations in Northern Iraq, eds. N. Yoffee and J. L. Clarke. Tucson: University of Arizona. Munchaev, R. M.; Merpert, N. I.; and Bader, N. O. 1984 Archaeological Studies in the Sinjar Valley, 1980. Sumer 43: 32-53. Oates, J. 1978 Religion and Ritual in Sixth-Millennium B.C. Mesopotamia. WorldArchaeology 10: 117-24. Peltenburg, E. J. 1991 Lemba Archaeological Project Volume, Vol. 2.2: A CeremonialArea at Kissonerga. Studies in MediterraneanArchaeology 70:3. G6teborg: Astroim.




Renfrew, C.; Dixon, J. E.; and Cann, J. R. 1966 Obsidian and Early Cultural Contact in the Near East. Proceedings of the Prehistoric Society 32: 30-72. Further Analysis of Near Eastern Obsidians. 1968 Proceedings of the Prehistoric Society 34: 319-33. Roaf, M. 1990 CulturalAtlas of Mesopotamia and the Ancient Near East. New York: Facts on File. Sahlins, M. 1972 Stone Age Economics. Chicago: AldineAtherton. Stevanovi(, M. 1997 The Age of Clay: The Social Dynamics of House Destruction. Journal ofAnthropological Archaeology 16: 334-95. Tobler, A. L. 1950 Excavations at Tepe Gawra, Vol. 2: Levels IXXX. Philadelphia: University of Pennsylvania. Verhoeven, M. 1999 An Archaeological Ethnographyof a Neolithic Community:Space, Place and Social Relations

in the Burnt Village at Tell Sabi Ayad, Syria. Uitgaven van het Nederlands Historisch-Archaeologisch Instituut te Istanbul 83. Leiden: Nederlands Historisch-Archaeologisch Instituut te Istanbul. von Wickede, A. 1990 Prdhistoriche Stempelglyptik in Vorderasien. Miinchener Vorderasiatische Studien 6. Munich: Profil. Watson, P. J. 1983 The Soundings at Banahilk. Pp. 545-613 in Prehistoric Archaeology along the Zagros Flanks, eds. L. S. Braidwood, R. J. Braidwood, B. Howe, C. A. Reed, and P. J. Watson. Oriental InstitutePublications 105. Chicago: University of Chicago. Watson, P. J., and LeBlanc, S. A. 1973 Excavation and Analysis of Halafian Materials from SoutheasternTurkey:The HalafianPeriod Re-examined. Unpublished conference paper.