• Aucun résultat trouvé

Obsidian consumption at Qdeir 1, a Final Pre-Pottery Neolithic site in Syria: An integrated characterisation study

N/A
N/A
Protected

Academic year: 2021

Partager "Obsidian consumption at Qdeir 1, a Final Pre-Pottery Neolithic site in Syria: An integrated characterisation study"

Copied!
16
0
0

Texte intégral

(1)

HAL Id: hal-02077604

https://hal.archives-ouvertes.fr/hal-02077604

Submitted on 23 Mar 2019

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Obsidian consumption at Qdeir 1, a Final Pre-Pottery Neolithic site in Syria: An integrated characterisation

study

Marie Orange, Tristan Carter, François-Xavier Le Bourdonnec

To cite this version:

Marie Orange, Tristan Carter, François-Xavier Le Bourdonnec. Obsidian consumption at Qdeir 1,

a Final Pre-Pottery Neolithic site in Syria: An integrated characterisation study. Comptes Rendus

Palevol, Elsevier, 2019, 18 (2), pp.268-282. �10.1016/j.crpv.2018.08.002�. �hal-02077604�

(2)

ContentslistsavailableatScienceDirect

Comptes Rendus Palevol

ww w . sc i e n c e d i r e c t . c o m

Human Palaeontology and Prehistory (Prehistoric Archaeology)

Obsidian consumption at Qdeir 1, a Final Pre-Pottery

Neolithic site in Syria: An integrated characterisation study

Économie de l’obsidienne à Qdeir 1, site du Néolithique précéramique final en Syrie : étude de caractérisation intégrée

Marie Orange

a,b,∗

, Tristan Carter

c

, Franc¸ ois-Xavier Le Bourdonnec

a

aIRAMAT-CRP2A,UMR5060CNRS–UniversitéBordeauxMontaigne,Maisondel’Archéologie,EsplanadedesAntilles,33607Pessac, France

bArchaeologyandPalaeoanthropology,SchoolofHumanities,ArtsandSocialSciences,UniversityofNewEngland,Armidale,NSW2351, Australia

cDepartmentofAnthropology/McMasterArchaeologicalXRFLab,CNH524,McMasterUniversity,1280MainStreet,LS84L9Hamilton, Ontario,Canada

a r t i c l e i n f o

Articlehistory:

Received10August2018

Acceptedafterrevision29August2018 Availableonline11December2018 HandledbyMarcelOtte

Keywords:

Obsidian

Geochemicalcharacterisation Lithicstudy

Syria Neolithic ED-XRF SEM–EDS

a b s t ra c t

Thispaperdetailsanintegratedcharacterisationstudyofasubstantialassemblageofobsid- ianartefacts(n=519)fromtheSyrianNeolithicsiteofQdeir1(ElKowmoasis).Theresults ofthechemicalcharacterisation(usingED-XRFandSEM–EDS)havebeencoupledwiththe typo-technologicaldata.Suchanapproachhasallowedus(i)toidentifyfourrawmaterials intheassemblage,namelyBingölAandBingölBfromeasternAnatolia,plusGöllüDa˘gand NeneziDa˘gfromcentralAnatolia,(ii)tospecificallysourcethedistinctivegreenperalkaline obsidiantoBingölA(ratherthan‘BingölAand/orNemrutDa˘g’),(iii)toobservethatthese fourrawmaterialswereconsumedinanigh-identicalmanner,probablyworkedlocallyby specialistcraftspeopletoproducefinepressureflakedblades,and(iv)tohypothesisethat thepeopleofQdeir1mayhaveplayedakeyredistributiveroleinthecirculationofobsidian tools,likelysupplyingtheneighbouringvillageofElKowm.

©2018Acad ´emiedessciences.PublishedbyElsevierMassonSAS.Thisisanopenaccess articleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.

0/).

Motsclés: Obsidienne

Caractérisationgéochimique Étudelithique

Syrie Néolithique ED-XRF SEM–EDS

ré s u m é

Cetarticleprésenteuneétudedecaractérisationintégréed’unassemblaged’artefactsen obsidienne(n=519)provenantdusiteNéolithiquedeQdeir1(Syrie,oasisd’ElKowm).

Lesrésultatsdelacaractérisationchimique,menéeparED-XRFetSEM–EDS,ontétécou- plésauxdonnéesdelalecturetypo-technologique.Unetelleapprochenousapermis(i) d’identifierl’utilisationdequatrematièrespremièresauseindel’assemblage,soitBingölA etBingölB(Anatolieorientale),ainsiqueGöllüDa˘getNeneziDa˘g(Anatoliecentrale),(ii) despécifierl’originedesartefactsdecompositionperalcaline(BingölA),(iii)demontrer

Correspondingauthor.ArchaeologyandPalaeoanthropology,SchoolofHumanities,ArtsandSocialSciences,UniversityofNewEngland,Armidale, NSW2351,Australia

E-mailaddresses:morange@une.edu.au,marie.orange@u-bordeaux-montaigne.fr(M.Orange),stringy@mcmaster.ca(T.Carter), Francois-Xavier.Le-Bourdonnec@u-bordeaux-montaigne.fr(F.-X.LeBourdonnec).

https://doi.org/10.1016/j.crpv.2018.08.002

1631-0683/©2018Acad ´emiedessciences.PublishedbyElsevierMassonSAS.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://

creativecommons.org/licenses/by-nc-nd/4.0/).

(3)

quelesmatièrespremièresenquestionontétéconsomméesd’unemanièresimilaire,prob- ablementtravailléeslocalementpardestailleurs«spécialisés»pourproduiredeslames débitéesparpression,(iv)d’argumenterqueleshabitantsdeQdeir1ontpujouerunrôle danslaredistributiondesoutilsenobsidienne,certainementenapprovisionnantlevillage voisind’ElKowm.

©2018Acad ´emiedessciences.Publi ´eparElsevierMassonSAS.Cetarticleestpubli ´een OpenAccesssouslicenceCCBY-NC-ND(http://creativecommons.org/licenses/by-nc-nd/

4.0/).

1. Introduction

Obsidiancharacterisationstudieshaveenjoyedarecent resurgence, the work being employed to engage with a range ofmajorquestions in anthropologicalarchaeol- ogy (Freund, 2013). Arguably obsidian sourcing is one of archaeological science’s greatest success stories, the uniquechemicalprofileofeachgeologicalsourceenabling archaeometerstoprovenancean artefact’sraw material through matching their elemental signatures (Glascock etal.,1998).Inthispaper,weemployanobsidiancharac- terisationmethodtoreconstructthenature,direction,and intensityofcontactbetweenprehistoricpopulationsinthe NearEasternNeolithic(cf.Chataigner,1994).

Methodologicallywefollowarecentturninthedevel- opmentofmoreintegrated characterisationstudiesthat trytomaximisethepotentialofbothaspectsofanarte- fact’s‘character’,i.e.workingwithina ‘chaîneopératoire’

frameworkthatconsidersbothelementalcompositionand typo-technological form (following suchwork as Briois et al., 1997; Carter et al., 2006; Le Bourdonnec, 2007;

Poupeauetal.,2007).Thusinsteadoftalkingabout‘sam- ples’,wereintroducetheterminologyofthearchaeologist andtechnologist,discussinghowanobsidianfrom‘source X’mayhavearrivedatthesiteintheformof‘ready-made unipolarpressureblades’,whiletheobsidianfrom‘source Y’wasprocuredintheformof‘rawnodules,thenknapped onsiteusinganopposedplatformpercussivebladetech- nology’interalia.Therefore,theintegrationofthesetwo typesofstudiesleadsustoconsiderthe‘consumption’of obsidianinamuchmoreholisticfashion(itsoverall‘econ- omy’;seeOrangeetal.,2017).

Anotherimportantaspectofthisstudyisthesample size, as untilquite recentlymost characterisation stud- iestendtoberestrictedtoahandfulofartefactspersite (cf.Renfrewetal.,1966),afactorthatultimatelylimitsa project’sinterpretativepotential.Recentlytherehasbeen aconsciousdevelopmentinobsidiansourcingtocharac- teriselargerdatasets(‘law oflargenumbers’,Shackley, 2005)toenabletheanalysttogainamorerepresentative sample oftheassemblageand aclearer insightintothe natureofprocurementandproduction(seealsoMillsetal., 2013;Reepmeyeretal.,2011).

In thisstudy,we aimedto(i)compare twodifferent methods for the provenance analysis: an energy dis- persiveX-rayfluorescencespectrometer[ED-XRF]anda scanningelectronmicroscopewithanenergy dispersive spectrometer[SEM–EDS],(ii)tolimitissuesofsampling representativenessbyanalysingthevastmajorityofthe Qdeir1obsidianassemblage(Fig.1),and(iii)tointegrate

raw material provenance datawith theartefacts’ typo- technologicalcharacteristics.

Apreviouspaper(Orange etal.,2013)presented the methodologicalimplicationsofourstudy,discussingthe significanceofacombinedED-XRF/SEM–EDSanalysisona largequantityofartefactsfromQdeir1andanotherSyr- ianNeolithicsite, Tell Aswad(Fig.1).In this paper,we nowfocusontheresultsfromQdeir1,integratingtypo- technologicalobservationswiththesourcingdatatoenable arichercharacterisationofhowobsidianwasconsumedat thesite(the‘obsidianeconomy’).Inturn,wethenlocateour Qdeir1dataandconclusionswithinabroadersynchronic regionalcontext,thusallowingustobetterunderstandthe roleofthiscommunitywithinlargernetworksofobsidian exchangeanddissemination,particularlywithregardtoits neighbouringsiteofElKowm.Byextent,ourstudytackles thequestionoftheimportanceofnomadicgroupsduring thefinalNeolithicperiod(cf.Cauvin,1994).

2. Qdeir1:aFinalPPNBnomadicsettlement

ThesiteofQdeir1(Fig.2)issituatedintheElKowm oasisin northernSyria(Fig.1), and wasfirstlocatedin 1980by OlivierAurenche, withsubsequentexcavations bythe‘Missiond’ElKowm-Mureybet’directed byJacques Cauvin (1981),revealing a FinalPre-PotteryNeolithic B [PPNB]settlementdated7100–5720calBC(Stordeur,1993, 2000;Table1).Aurenche’ssurveydemonstratedthattheEl Kowmoasiscontainsanumberofprehistoricsettlements (Fig.1),includingElKowm,Qdeir1(Stordeur,2000)and Umm-el-Tlel(Molistetal.,1992).A larger-scaleexcava- tionwasinitiatedin thelate1980s,initiallydirected by DanielleStordeur(in1989,1991and 1993), andsubse- quentlybyFrédéricAbbès(in1999,and2001–2003).The site is estimated tobec. 2000m2, of which some200 m2 havebeenexcavated(Abbès,2015).WhileElKowm andQdeir—8kmapart—werebothoccupiedduringthe FinalPPNB,theydisplayedsufficientoccupationaldiffer- encestowarrantattributiontotwodistinctculturalfacies (Cauvin,1994).WhileElKowmwasavillagewithalong andcomplexsettlementhistory,withatleastfivedifferent levelsofoccupationspanning7100–4230calBC(Stordeur, 2000),Qdeir1comprisedanexceptionallywellpreserved nomadiccamp,consistingofonlyafewarchitecturalstruc- tures,a semi-permanentcamp, and a workshopfor the manufactureofflintimplements(Abbès,2015;Besanc¸on etal.,1982;StordeurandWattez, 1998).Atotal offour occupation phases have been defined (Ibid.; Gourichon, 2004),twowitharchitecturalstructures(isolatedhouses1

(4)

Fig.1.Mapshowingthemainarchaeologicalsitesandobsidiansourcesmentionedinthetext.

Fig.1. Cartedesprincipauxsitesarchéologiquesetsourcesd’obsidiennementionnésdansletexte.

and2;phasesIIandIV),andtwowithout(phasesIandIII).

ThesecharacteristicsledQdeir1tobeclassifiedasbelong- ingtosomespecificfaciesofthePPNB,the‘desertPPNB’.

Whiledistinctinsize,location,andnatureofoccupation, thesethreesiteshaveallproducedevidencefortheuseof obsidian,anexoticrawmaterialwhoseclosestsourceslay hundredsofkilometrestothenorthincentralandeast- ernAnatolia(Fig.1);thismaterialwasmainlyrecovered intheformofpressurebladesandbladelets(Cauvinand Chataigner,1998).

Part of the Qdeir 1 chipped stone assemblage was studiedandpublishedin the1980s,albeit withmodest attentiontotheobsidianitself,duetothelimitednum- berofartefactsfoundatthispoint(Calley,1986;seealso Cauvin,2000).

3. TheQdeir1chippedstoneassemblageand samplingprotocols

Ourstudyinvolved theanalysisofeach artefactthat wasavailabletous(theremainingpiecesbeinginSyria), i.e. the519 artefacts comingfrom the excavationcam- paignsof1979–1980,1989,1991,1993,and2000.While exactfigureswereunfortunatelynotavailabletous(nor likelytobeaccessibleinthenearfuturegiventhepolit- icalsituationinSyriaatthetimeofwriting),therelative proportionofobsidiantoflintwithinthechippedstone

assemblageisknowntobequitelow(F.Abbès,pers.comm.).

The volume of worked flint per cubic meter has been estimatedat >12,000pieces,equivalentto22kgof raw material(Calley,1986).Mostofthisflintappearstohave beenprocuredfromanoutcrop4kmnorth-westofQdeir, aneasilyaccessiblesourcethatprovidedrawnodulesof excellentquality(Calley,1986).Theflintwasusedprimar- ilyforthemanufactureofbi-directionalbladesknapped fromopposedplatform/naviformcores(Calley,1986;see also Abbès,2003), a particularlyproductive technology, withestimatesof26/27bladesand13/14bladeletspercore (Calley,1986).Perhapsunsurprisingly,giventheproxim- ityofthesource,theflintassemblagecontainsasignificant quantityofearly-stagereductionknappingdebris,witha relativelyhighproportionofcorticalwaste(29%,Ibid.).As wedocumentbelow,itisinstarkcontrasttotheobsid- ianassemblages,which are dominatedby end-products andlate-stagemanufacturingpieces.Therelativepropor- tion and range of modified flintimplements is greater thanthatwitnessedintheobsidianassemblage,with31 arrowheads,156burins,108scrapers,andtensickleblades reportedfromthe17,642artefactsoftheGYandGXsurveys sectorsfromthe1980excavation(AurencheandCauvin, 1982).Flintandobsidianseemtobetheonlyrawmate- rialsemployedbythecommunitytomanufactureflaked stoneimplements,withtheexceptionofonechiselblade executedongreenstone(AurencheandCauvin,1982).

(5)

Fig.2.GeneralmapoftheQdeir1tell.CourtesyofF.Abbès.

Fig.2. PlangénéraldutelldeQdeir1.AveclapermissiondeF.Abbès.

Table1

Radiocarbondates obtainedfrom thesiteofQdeir1. Datafromthe BANADORAwebsite[http://www.arar.mom.fr/banadora/]andStordeur, 1993.

Tableau1

DatesradiocarboneobtenuespourlesitedeQdeir1,provenantdelabase dedonnéesenligneBANADORA[http://www.arar.mom.fr/banadora/]

ainsiquedeStordeur,1993.

Datingreference Samplelocation Calibrateddate Lyon-2338 CAR.EB96-US3 7175–6700BC Lyon-2339 CAR.EL76-Sond.E 7297–6840BC Lyon-2340 CAR.EL76-Sond.E 7176–6702BC

Ly-2578 nd 7100–5720BC

4. Obsidiansourcing

Inourstudy,weusedtwonon-destructivegeochem- ical characterisation techniques that have been widely employedforobsidiansourcingintheMediterraneanand NearEast,namelySEM–EDS(cf.Acquafreddaetal.,1999;

Le Bourdonnecet al., 2010) and ED-XRF(cf. Carter and

Shackley,2007;Carteretal.,2013;Poupeauetal.,2010).

Theoperatingconditionsforeachmethodaredetailedinan earlierpaper(Orangeetal.,2013);wepresenthereinstead adiscussionontherepeatabilityandaccuracyofourmea- surements(conceptsdiscussedinFrahm,2011).

4.1. Obsidiancharacterisation:UsingSEM–EDS 4.1.1. Repeatabilityandaccuracy

ToassesstherepeatabilityofourSEM–EDSanalyses, thegeologicalsampleKOMC5(Kömürcüobsidiansource, Göllü Da˘g massif) was analysed during every session.

Twenty-twomeasurements,correspondingtothemeanof 3to5measurementpoints,wereobtainedduringthispro- cess;thesedataarereportedinTable2,alongwiththe average,standarddeviation,andcoefficientofvariation(%).

Themeasuredvaluesneithersignificantlyvarythroughout asingleday,norfromonedaytoanother.

Concerning the accuracy of the SEM–EDS, the val- uesobtainedfor several geologicalstandards (KOM C5,

(6)

Table2

Average(wt%),standarddeviation,andvariationcoefficient(%)foreach elementassayedbySEM-EDSattheCRP2AonKOMC5onthe22mea- surements.ComparisonofAl,Si,Ca,andFecontents(wt%)oftheKOM C5geologicalsamplemeasuredeachmorningandeveningthoroughsix completedaysofanalysis.

Tableau2

Moyenne(%masse),écart-typeetcoefficientdevariation(%)dechaque élémentmesurépar MEB-EDSau CRP2Apourl’échantillon KOM-C5 (22mesures).ComparaisondesteneursenNa2O,Al2O3,SiO2,K2O,CaO etFe2O3(%masse)pourl’échantillongéologiqueKOM-C5mesuréchaque matinetchaquesoiraulongdesixjourscompletsd’analyse.

Allruns(n=22) Na2O Al2O3 SiO2 K2O CaO Fe2O3

18/01/12 4.4 13.1 76.8 4.6 0.45 0.61

19/01/12 Run1 4.0 12.9 77.3 4.8 0.43 0.60 Run2 4.1 12.8 77.1 4.8 0.49 0.67 20/01/12 Run1 4.1 12.9 77.2 4.7 0.50 0.61 Run2 4.6 13.1 76.8 4.4 0.46 0.61

23/01/12 4.4 13.0 76.8 4.6 0.43 0.70

24/01/12 Run1 4.1 12.7 76.9 4.8 0.62 0.70 Run2 4.0 12.9 77.0 4.8 0.50 0.74

25/01/12 4.7 13.1 76.7 4.4 0.47 0.58

30/01/12 Run1 4.3 13.0 76.9 4.6 0.43 0.65 Run2 4.6 13.4 76.5 4.4 0.53 0.58

31/01/12 4.3 13.5 76.3 4.6 0.61 0.66

01/02/12 4.2 13.0 77.0 4.7 0.42 0.64

02/02/12 4.4 13.0 77.0 4.6 0.45 0.62

03/02/12 4.1 12.9 77.1 4.8 0.47 0.71

06/02/12 4.2 12.9 77.0 4.7 0.46 0.68

02/07/12 4.5 13.1 76.8 4.4 0.44 0.61

04/07/12 Run1 4.1 13.0 77.1 4.7 0.43 0.67 Run2 4.3 13.0 77.0 4.6 0.48 0.61

05/07/12 3.9 12.8 77.2 4.9 0.45 0.67

09/07/12 Run1 4.0 12.8 77.0 4.8 0.48 0.74 Run2 4.1 12.9 77.0 4.7 0.52 0.67 Average(wt%) 4.2 13.0 76.9 4.7 0.48 0.65

Stand.Dev. 0.2 0.2 0.2 0.2 0.05 0.05

VC(%) 5.0 1.5 0.3 3.4 11.09 7.41

Contentsinoxidesareinweightpercent(wt%).Eachmeasureofnpre- sentedcorrespondstothemeanof3to5measurementpointsonKOM C5.

P35-18B1)wereshowntobeconsistentwithpreviously publishedvalues for these standards obtainedby PIXE, SEM–EDS(Poupeauetal.,2010),and/orICP-AES(Bellot- Gurlet, 1998).The accuracy of our SEM–EDS is further detailedinDelerue(2007),LeBourdonnecetal.(2010),and Poupeauetal.(2010).

4.1.2. Results

A total of 178 artefacts was analysed by SEM–EDS, togetherwith60geologicalsamplesfromvariousAnato- lianobsidiansources,specificallythosemostlikelytobe presentatthesitebasedonprioranalysesofmaterialfrom Qdeir1andpertinentsitesfromthesurroundingarea(see Chataigner,1998;Delerue,2007;Gratuzeetal.,1993inter alia).

ThemajorelementaldatageneratedbytheSEM–EDS revealedtheexistenceofthreedistinctrawmaterialgroups (onedominant), asclearlyshownbyFe2O3 vs. SiO2 and CaOvs. SiO2 binary graphs(Fig.3).Thesethree groups, easilydiscriminatedby theirFeand Cacontents, corre- spondtotheobsidiansources ofBingöl Band BingölA (easternAnatolia)andGöllüDa˘g(Cappadocia).Onearte- fact(KQ80.GX47.62,seeFig.3)isclosetothe90%normal densityellipseofNemrutDa˘g; however,ifwelookonly attheironcontent(3.34%),itcanclearlybeattributedto

Fig.3.SEM–EDSratioplotsofSiO2vs.Fe2O3andSiO2vs.CaOforQdeir1 artefactsandgeologicalsourcesamples(NNZD:NeneziDa˘g;GD:Göllü Da˘g;BB:BingölB;MDD:MeydanDa˘g;SD:SuphanDa˘g;ACGL:Acıgöl;

NMRT:NemrutDa˘g).Contentsinweightpercent.90%normaldensity ellipses.

Fig.3. DiagrammesbinairesdesrapportsSiO2vs.Fe2O3etSiO2vs.CaO obtenusparMEB-EDSpourlesartefactsdeQdeir1etleséchantillons géologiques(NNZD:NeneziDa˘g;GD:GöllüDa˘g;BB:BingölB;MDD: MeydanDa˘g;SD:SuphanDa˘g;ACGL:Acıgöl;NMRT:NemrutDa˘g).

Teneursexpriméesenpourcentagedemasse.Ellipsesdedensitéà90%.

theBingölAsource(seeSupplementarydata).Thewide spread oftheartefacts’ compositionaldataobserved for the sources of Bingöl A and Bingöl Bcan be explained bythesmallnumberofgeologicalsamplesanalysedfrom theselocations(n=7andn=5respectively).Alongsidethe productsofthesethreemainchemicalgroupswasasingle artefactwithacompositionof0.99%inCaOand1.04%in Fe2O3,whichcanbeattributedtotheNeneziDa˘gsource.

4.1.3. Problematicresults

Onlyoneartefact(KQ80.GY46.33)wasexcludedfrom the statistical analysis because of inconsistent results, mainlyobservedonsodium,potassium,calcium,andiron contents(standarddeviation<1%betweendifferentmea- suring points). This kind of phenomenon, reported in previousstudies(Luglièetal.,2007,2008;Poupeauetal., 2010),isoftenuniquetothearchaeologicalsurfaceitself,

(7)

Table3

Average(ppm),standarddeviationandvariationcoefficient(%)foreachelementassayedbyED-XRFattheMAXLabonRGM-2(34measures)andcomparison withUSGSrecommendedvalues(ppm).Comparisonofallassayedelements(ppm)ofRGM-2duringsixdaysofanalysis(tworunsperday).

Tableau3

Moyenne(ppm),écart-typeetcoefficientdevariation(%)dechaqueélémentmesuréparED-XRFauseinduMAXLabpourl’échantillonRGM-2(34mesures) etcomparaisonauxvaleursrecommandéesparl’USGS(ppm).Comparaisondetouslesélémentsdosés(ppm)pourRGM-2aucoursdesixjoursd’analyse (deuxexécutionsparjour).

Allruns(n=36) Ti Mn Fe Ni Cu Zn Ga Rb Sr Y Zr Nb Ba Pb Th

01/03/12 1512 306 14,015 6 12 33 17 146 105 26 234 14 825 18 13

14/03/12 1575 294 14,007 2 17 38 17 146 104 25 229 12 828 25 6

15/03/12 Run1 1599 318 14,023 11 14 35 17 147 105 22 234 5 856 20 10

Run2 1538 330 13,986 13 22 38 18 148 104 24 238 14 826 23 20

16/03/12 Run1 1541 332 13,953 22 39 18 142 108 25 229 9 803 28 17

Run2 1552 325 13,944 8 6 41 15 147 108 23 233 8 822 23 20

19/03/12 Run1 1554 344 13,941 10 39 15 144 105 25 232 7 791 18 11

Run2 1575 320 13,941 5 12 44 17 146 99 25 232 11 792 23 15

27/03/12 Run1 1570 314 14,001 1 5 36 18 150 104 22 238 13 822 19 17

Run2 1528 303 14,039 7 18 40 15 144 103 25 234 11 825 22 10

28/03/12 Run1 1534 314 13,986 12 37 12 147 103 24 231 9 832 17 12

Run2 1499 330 14,052 1 7 41 14 148 102 24 231 11 845 27 12

29/03/12 Run1 1662 335 13,991 2 12 45 12 145 104 27 232 10 821 17 7

Run2 1495 338 13,942 6 11 42 21 143 102 25 231 12 787 21 13

Run3 1609 321 13,969 7 16 40 19 142 105 25 237 12 792 22 15

04/03/12 Run1 1513 320 13,967 10 37 14 146 101 26 231 13 814 21 15

Run2 1552 332 13,963 22 38 18 144 103 25 231 7 825 21 13

Run3 1520 295 14,043 5 36 11 143 106 22 228 15 803 24 12

05/04/12 Run1 1565 315 13,978 12 39 19 146 106 23 231 9 809 19 12

Run2 1520 287 13,963 13 12 44 20 146 105 25 228 13 791 25 13

Run3 1480 314 13,942 12 37 16 148 104 23 231 11 819 21 10

07/04/12 Run1 1540 311 13,952 3 15 43 19 144 104 26 231 12 806 24 14

Run2 1510 312 13,983 9 17 39 13 144 106 24 234 12 794 25 16

09/04/12 Run1 1505 325 13,997 1 14 34 15 145 103 24 230 10 815 23 18

Run2 1594 312 13,965 1 17 36 16 146 107 22 228 11 810 21 8

11/04/12 Run1 1533 320 13,905 8 12 39 16 144 106 21 230 11 792 25 7

Run2 1537 330 13,932 5 16 36 19 149 104 21 226 7 769 22 15

Run3 1538 309 13,987 10 37 21 146 107 23 226 8 778 22 11

12/04/12 1524 325 13,953 8 41 18 140 105 25 229 8 782 21 13

13/04/12 Run1 1643 324 13,966 10 37 16 145 103 23 226 10 794 25 11

Run2 1486 358 13,938 11 35 15 144 109 23 229 13 787 18 9

Run3 1599 323 13,925 11 14 43 17 145 105 27 228 10 783 17 8

14/04/12 Run1 1572 291 13,972 8 35 19 148 104 23 224 11 772 24 1

Run2 1557 293 13,999 5 12 39 16 144 107 24 232 14 824 23 8

Run3 1559 320 14,014 7 16 39 17 145 102 23 229 11 793 23 6

17/04/12 1598 313 13,991 16 17 41 16 148 109 25 231 11 807 24 7

Average(wt%) 1550 318 13,976 6 13 39 17 145 105 24 231 11 806 22 12

Stand.Dev. 42 15 35 4 4 3 2 2 2 2 3 2 20 3 4

VC(%) 3 5 0 54 34 7 15 1 2 6 1 21 3 13 36

USGSvalues 1499±38 273±8 13,010±280 4 9.8±0.8 33±2 16±1 147±5 108±5 24±2 222±17 9 842±35 20±1 15±1 Elementcontentsareinppm.EachmeasureofnpresentedcorrespondstoonemeasurementpointonRGM-2.USGSvalues:recommendedvalues, informationvalues.±representsonestandardvariation.

frequentlypresentinganalterationofthechemicalcom- positioninthesurfacelayersoftheglass,duetopartial hydrationviathesoil.

4.2. Obsidiancharacterisation:UsingED-XRF 4.2.1. Repeatabilityandaccuracy

ThereproducibilityofourED-XRFmeasurementshas beenevaluated usingthe RGM-2internationalstandard producedbytheU.S.GeologicalSurvey(USGS),forwhich theelementalcompositionwasdeterminedoneachofthe 36analytical runs,i.e.36measurementpoints(Table3).

Theestimatedrelativestandarddeviation(%)showsthat therangeofmeasurementdeviationmostlyremainsunder 20%anddropstolessthan10%fortheelementsofgreatest discriminatorypower,namelyRb,Sr,Y,Zr,andBa,indi- catingasuitablereproducibilityofouranalyses.Wecan,

however,notethattheNi,Cu,andThelements,whichare notusedinourstatisticalanalysis,displayhigherdisper- sions(especiallyforNi,superiorto50%).Asshownbythe standarddeviationandthevariationcoefficient(%),wecan concludethat ourmethodhasa goodrepeatabilityover time,byday/week/month.

RGM-2beinganinternationalgeologicalstandard,has allowed us to compare our results with USGS recom- mendedvalues(Table3).Whiletheinstrumentgenerated problematic resultsfor Fe, Zn, Ni,Cu, and Th, we have excellentresultsforthose elementsmostusefulfordis- criminatingthemajorAnatolianobsidiansourceproducts, namelyGa,Rb,Sr,Y,Zr,andBa.

4.2.2. Results

Forthe517artefactsanalysedbyED-XRF,asetoffifteen majorandtraceelementswereassayed(Ti,Mn,Fe,Ni,Cu,

(8)

Fig.4. ED-XRFratioplotofSr/Pbvs.Zr/NbforQdeir1artefactsandsource samples.SourceabbreviationsasinFig.2.

Fig.4. DiagrammebinairedesrapportsSr/Pbvs.Zr/NbobtenusparED- XRFpourlesartefactsdeQdeir1etleséchantillonsgéologiques.Voir abréviationsdessourcessurlaFig.2.

Zn,Ga,Rb,Sr,Y,Zr,Nb,Ba,Pb,andTh).The561geological sourcesamplesemployedforthisstudywerepreviously analysedbytheMAXLabteamemployingthesameana- lyticalconditionsasthoseusedforourartefacts(cf.Carter etal.,2013).

As with the SEM–EDS analyses, the ED-XRF results indicatedtheexistenceof fourrawmaterials,asshown bya ratiodiagram (Fig.4)involvingZr, Nb,Sr,and Pb, fourofthemostusefulelementsforthediscriminationof obsidiansourceproductsintheNearEast(Poupeauetal., 2010).Mostoftheartefactshadachemicalsignaturethat matchedsourceproductsfromBingöl B(n=230),while otherscorrespondedtotheelementalprofilesoftheBingöl A/NemrutDa˘g,GöllüDa˘g,andNeneziDa˘gsources(n=102, 143,andfourrespectively).Todiscriminatebetweenthe BingölAandNemrutDa˘gsources(seeChataigner,1994;

Frahm,2012; Orange et al., 2013, regarding this topic), we plotted Y/Nb and Zr/Pbratios (Fig. 5); this demon- stratedthatall102ofQdeir1’sartefactsmadeofgreen peralkalineobsidiancamefromBingölA.Thefinalbinary diagram,comparingZrandSrcontents(Fig.6),allowedus todemonstratethattheremainingartefactsweremadeof CappadocianobsidianfromGöllüDa˘gandNeneziDa˘g.

4.2.3. Problematicresults

Amongstthe517artefactsanalysedbyED-XRF,38pre- sentedirregularvalues,evenafteranumberofrepeatruns.

AswiththeSEM–EDSanalysis,theseapparentlyproblem- aticvalues might be theresult ofsurface irregularities, soilcontamination,orpartialhydration. Butin thecase oftheED-XRF analyses, thesizeoftheartefactappears tobethe mostsignificantcriterion (Davis et al.,2011), withthe unattributed pieces tending to be the assem- blage’sthinnestand/ornarrowestpieces(seeFig.7).The ED-XRFanalysisusedanon-collimatedbeam,wherebythe diameteroftheseartefactswassmallerthantheincident beam,thusleadingtoproblematicresults.Unfortunately,

Fig.5. ED-XRFratioplotofY/Nbvs.Zr/PbforQdeir1artefactsandsource samplesfromBingölAandNemrutDa˘g.SourceabbreviationsasinFig.2.

Fig.5. DiagrammebinairedesrapportsY/Nbvs.Zr/PbobtenusparED-XRF pourlesartefactsdeQdeir1etleséchantillonsgéologiquesdessources deBingölAetNemrutDa˘g.VoirabréviationsdessourcessurlaFig.2.

Fig.6.Zrvs.SrcontentsdeterminedbyED-XRFforQdeir1artefactsand sourcesamplesfromGöllüDa˘g,SuphanDa˘g,Acıgöl,andNeneziDa˘g.

SourceabbreviationsasinFig.2.

Fig.6. DiagrammebinairecomparantlesteneursenZretSrobtenuspar ED-XRFpourlesartefactsdeQdeir1etleséchantillonsgéologiquesde GöllüDa˘g,SuphanDa˘g,AcıgöletNeneziDa˘g.Voirabréviationsdessources surlaFig.2.

lackoftimepreventedusinstallinganewcollimatorwith anarrowerbeamtore-runthese38artefacts;allbutone (KQ80.GY46.33)werehoweversuccessfullycharacterised bySEM–EDS.

4.3. Discussiononthesourcingtechniques

Thisstudy’saimwastocomparetwonon-destructive archaeometrictechniques.WeobservedthatED-XRFhas numerousbenefits,notleastitseaseofuse,automation, and speed (only 3h30 toanalyse 19 samples plus the

(9)

Fig.7. Histogramsrepresentingthewidthandthicknessmeasures(inmm)ofBingölA,BingölB,andGöllüDa˘gartefacts.Datageneratedfrommesial segmentsandwholeblades.

Fig.7. Histogrammesdesmesuresdelargeuretd’épaisseur(enmm)desartefactsattribuésauxsourcesdeBingölA,BingölBetGöllüDa˘g.Données obtenuessurpartiesmésialesdelamesetsurlamesentières.

geological standard), while SEM–EDS requires the con- stantpresenceoftheanalystandtakesaround10hours toanalyse33samples.Moreover,ED-XRFcansuccessfully discriminatebetweenbothperalkalineproductsofBingöl A andNemrut Da˘gand calco-alkaline productsofGöllü

Da˘gandAcıgöl inCappadocia,whileSEM–EDScanonly distinguish theperalkaline products(see Orange et al., 2013).That said, SEM–EDS also has its strengths, as it too can beused non-destructively, while its controlled beampermitstheanalysisoftheverysmallartefacts(few

(10)

micrometres)whichcancauseproblemswithED-XRF.In turn,whiletherequiredconstantpresenceoftheanalyst istime-consuming,itconverselyenableseachartefactto becarefullystagedtoprovidethebest(flattest)analytical surfacepossible.

5. Thetypo-technologicalapproach

5.1. Methodology

Theartefactswerefirstsortedbyprovenance,andthen byblanktype(flakes,blades,etc.)andpart(whole,prox- imal, medial, distal). A variety of attributes were then recorded on the blades, particularly the proximal seg- ments,asameansofobtainingfurtherinformationonhow theyweremade(e.g.,platformtype,lip,blankregularity, knappingrhythms,etc.),togetherwithavarietyofmetric informationsuchaslength,width,andthickness.Alldata wererecordedinanExcel®database.

5.2. Results

OurSEM–EDSandED-XRFanalysesdemonstratedthe presence of obsidian from four sources at Qdeir 1, all

locatedover 500km fromthesite (Fig.1).Cumulatively overtwothirdsoftheartefactsweremadeofrawmateri- alsthatcamefromeasternAnatolia,with245characterised asBingölBproducts(47%)and123asBingölA(24%).The restoftheartefactsweremadeofrawmaterialsfromthe centralAnatolian(southCappadocian)sourcesofGöllüDa˘g (n=144,28%)andNeneziDa˘g(n=4,<1%).

5.2.1. BingölB

The 245artefactsmade ofBingöl Bobsidian(47%of thetotal assemblage)arecomprisedprimarilyofblades that appeartohavebeenknappedfromsingleplatform (unipolar)coresbyapressuretechnique,asevidencedby the implements’distinctive parallel margins and dorsal ridges,consistentlongitudinalthickness,diffusebulbs,and smallplatforms(Fig.8a,b).Analysisoftheseblades’width and thickness(Fig.7)wasundertaken toexaminetheir standardisation(anothercharacteristicofpressureblades).

Mostofthedatafollowedanormaldistributionindicat- ingtherepetitiveproductionof fineprismaticbladesof muchthesameformandsize;however,ascatterinthe resultsindicatesthepresenceofwiderandthickerblades.

Oninspectionthesepieces(KQ89.GYNW82.17i.a.)hadthe

Fig.8.PhotographsoftheartefactsfromtheQdeir1archaeologicalsitediscussedinthetext(a–m).

Fig.8. PhotographiesdesartefactsdeQdeir1discutésdansletexte(a–m).

(11)

Fig.9. MapsshowingthesiteswithproofofuseoftheBingölA,BingölB,GöllüDa˘g,andNeneziDa˘gobsidianduringtheFinalPPNBperiod.

Fig.9. Cartedessitesayantdémontrél’utilisationd’obsidiennesprovenantdessourcesdeBingölA,BingölB,GöllüDa˘getNeneziDa˘gdurantlePPNBfinal.

characteristicsofbeingknappedbysoftstonepercussion (againfromunipolarcores),i.e.averythinplatform.

Concerningthe‘knappingrhythm’,i.e.thesequencein whichthepressurebladeswereflakedfromthecore(as evidencedbythestratigraphicrelationshipoftheblades’

dorsalscars[cf.Binder,1984]),themostfrequentlyviewed schemeis2-1-2’ asfortheBingölAandGöllüDa˘gprod- ucts.IfwerefertothestudyofthelithicassemblageofTell SabiAbyadII,abroadlycontemporarysitesome150km fromQdeir1(Astrucetal.,2007;seeFig.9),itwasargued that thedominanceof thistype ofblades(2-1-2’),with onlyafewtriangularsectionblades(1-2or2-1rhythm) wasindicativeofaspecificcoremanagementstrategy.This knappingsequenceallowsanoptimalproductionof2-1-2’

blades(inessencethemostregularprismaticproducts), togetherwithapproximately20to30%ofnecessarylateral

‘technicalblades’(i.e.“bladeswithtransversalscarsand/or corticalremnants”thatrelatetothemodeofbladeinitiation, Astrucetal.,2007:9).Onlythecrestedbladesaregenerally missingfromtheQdeir1obsidianassemblage(allsources included);onlyonewasfoundintheBingölBassemblage (KQ89.GYNW81.6).

Of the179examples of trueprismatic end-products (pleindebitage),51wereproximal,122weremedial,and

6weredistal;nonewerewhole.Basedonthenumberof proximalsections(and differentblade-types within the reductionsequence)thisassemblageofBingölBpressure bladesrepresentsaminimumof51implementsoriginally.

Thelongestexamplesmeasuredbetween52.0and55.4mm inlength.Theproximalsegmentsofthepressurebladesare distinctinthattheyhavetheiroriginalplatformoverhangs removedbytinyflakesand/orabrasion.Theplatformitself isalmostalwaysverythin(<1mm),i.e.linear,orplainin form.Theprofileofthemedialsegmentsisoftenstraight, whilethedistalfragmentsdisplayaslightcurvatureoreven atorsion,andalmostalwayshaveapointedtermination.

Theblade cores madeof BingölBobsidian thusappear tohaveoriginallybeenc.10–15cmlong,rathernarrow (hencethetwistedprofileofsomedistalparts),andconi- calinform,i.e.thewell-known‘bulletcore’type.Thistype ofnucleusisattestedthroughouttheregioninthelaterPPN periods(10th–8thmillenniumBC),asforexampleattested atAbuHureyra,TellHalula,andBouqras(Fig.9),orcloser toQdeir,atElKowm,whereanexamplewasrecycledas apendant. Obsidian‘bulletcores’haveinfactpreviously beennotedfromQdeir1(KozlowskiandAurenche,2005:

144),butnonewereavailableforouranalyses.Thegeneral absenceofcorticalknappingdebris,and bladeinitiation

(12)

blankssuchascrestedpieces,suggestsstronglythatthe communityprocuredBingölBobsidianintheformofpre- formed,andpart-workedpressurebladecores.Thereafter themajorepisodesofbladeremovaloccurredon-site,as attestedbythepresenceofend-products,afewmainte- nanceflakes,andaquantityofmicrochips(F.Abbès,pers.

comm.).Onlyoneflakerelatedtotheinitialcreationofa coreplatformwasfoundintheassemblage (KQ93.CS.1;

Fig.8c),apiecethatseemstohavebeenimportedtothe siteasatool(itwasretouchedintoascraper),ratherthan evidencefortheshaping ofcores locally,a stageof the manufacturingprocessthatwefeellikelyoccurredatthe quarries,oranintermediary site.Onecorticalflakewas reportedatQdeir(F.Abbès,pers.comm.),butthepiecewas unavailableforanalysis.

Ingeneral,modifiedpieceswerequiterare,suggesting thatthepressurebladesmayhavebeenemployedprimar- ilyin theirnatural, razor-sharpstate.Ofthe179Bingöl Bblades,only20hadbeenretouched(c.11%),including 18withsimplelinearretouch,plustwoexamplesofthe distinctive‘corner-thinned’ bladesmadeonmedialseg- ments.Asstatedabove,whilemostoftheBingölBblades wereproducedbypressureflaking,someappeartohave beenknappedbysoftstonepercussion(KQ89.GYNW82.17;

Fig.8d),whileonefinalpiece(KQ89.GYNW80.49;Fig.8e) drivesfromathirdtechnology,namelyaskilledpercus- sivetechniquefromanopposedplatformcore.Thispiece, previouslyattested intheQdeir1 flintindustry(Calley, 1986), wasalmost certainly procured ready-made from elsewhere.Whileopposedplatformbladesmadeofobsid- ianaregenerallyquiterareintheNearEast,thistechnical traditionisknowntohavebeenperformedattheKaletepe workshopatoptheGöllüDa˘gsourceinsouthernCappado- cia(BinderandBalkan-Atlı,2001).

5.2.2. GöllüDa˘g

Atotal of144artefactshad chemical signaturesthat matchedthoseofthesouthernCappadociansourceofGöllü Da˘g(asdefinedbyBinderetal.,2011),i.e.28%oftheQdeir1 assemblage.Thenatureoftheseproductsisverysimilar tothosemadeofBingölBobsidian,i.e.unipolarpressure blades(Fig. 8f, g), withthe same type of platform and overhangremoval.Asidefromthetrueend-products(plein débitage), there are a few ‘technical blades’ (asdefined above)and plungingblades.However, comparedto the BingölBproductstheGöllüDa˘gpressurebladesareslightly lessstandardised(whenoneconsidersthemetricaldata frommedial andwholeexamples),thoughmostremain verythin(1–2.5mm;Fig.7).Asmallerproportionofthe bladesweremodifiedthanamongsttheBingölBproducts, 11intotal(8%),thepiecesmainlymodifiedbysimplelinear retouch,togetherwithtwocorner-thinnedblades,which wereagainproducedonmedialfragments.Followingthe logicoutlinedabove,thisassemblagerepresentsamini- mumof32GöllüDa˘gpressureblades.

Theassemblagealsoincludesaflakewithanindirect abruptretouchonitsconvexdistalend, likelyascraper (KQ00.EB86.EF86.1;Fig.8h).

5.2.3. BingölA

Onecanmakemuchthesameobservationsaboutthe Bingöl A obsidian at Qdeir 1, as we made about the BingölBandGöllüDa˘gproducts,withthematerialdomi- natedbyunipolarpressureblades(KQ00.EB81.CS.5;Fig.8i), with thesame type of platform and overhang removal (KQ00.Extnrd.EG107.Z218.5; Fig. 8j). Further insight to the specifics of this technology are provided by a core rejuvenationflake(KQ91.GYNW83.17;Fig.8k;acuteplat- formangle),probablydeliberatelybrokentoobtainanew strikingplatformwithacorrectedangle;itsentirecircum- ferencehadbeenexploited.

A unique tool is observed within the peralkaline products of Bingöl A: a borer, produced on a regular prismaticbladefragment,withconvergentbackededges (KQ93.DW86.DZ98.10; Fig. 8l). Once again only a small proportionoftheBingölAobsidianartefactsweremodi- fied(n=8,7%),includingfivecorner-thinnedblades(e.g., Fig.8j).Themetricaldata(Fig.7)showthattheseblades derivefromasingleknappingtechnique(thesameasused toworktheotherrawmaterials), withthicknessesthat rangesfrom0.5and3.5mm,withtheanalysedassemblage comprisingaminimumof39BingölApressureblades.

5.2.4. NeneziDa˘g

Only four artefacts weremade of obsidian fromthe southern Cappadocian sourceof NeneziDa˘g, all medial segments of unipolar pressure blades (2 finished prod- ucts and 2 technical blades). One has a little tongue (KQ89.GYNE65.3; Fig. 8m), likeseveral blade fragments amongstthewholeQdeir1assemblage,revealingadeliber- atepost-manufacturefragmentationofthebladetoremove theproximalpart.Thedimensionsofthebladesaresimilar fromoneanother(averagewidtharound9mmandaverage thicknessaround2.5mm[seeFig.7]).

6. TheQdeir1resultsintheirbroadercontext

The Qdeir1obsidianassemblage thuscontains arte- factsmadeofrawmaterialsfrombothCappadocia(Göllü Da˘g,NeneziDa˘g)andeasternAnatolia(BingölB,BingölA [Fig.9]),withthemajoritycomingfromtheslightlycloser Bingölregion(c.450km,comparedtoc.520kmtothecen- tralAnatoliansources).This‘dualaccess’tobothcentral andAnatolianrawmaterialsissomethingwehavecome toassociatewithlaterPPNBcommunitiesofthenorthern Levant/UpperMesopotamia(CauvinandChataigner,1998;

Delerue,2007).

Whiletoanextentwehavereaffirmedpreviousideas concerningFinalPPNBobsidian‘diffusionzones’,wealso providenewimportantinformationastotheforminwhich theserawmaterialscirculatedandhowtheyweresubse- quentlyconsumed.Ourstudyshowsthatthesefourraw materialswereworkedin essentiallythesamemanner, usedtomakefineprismaticbladesbya skilledpressure technique,withthesamemodesofcorepreparationand reduction,leadingtotheend-productshavingmuchthe samesizeandform.OnlywiththeBingölBrawmaterial dowehaveevidenceformorethanonemodeofproduc- tion,withthepressuretechnologyaugmentedbyahandful ofsofthammerpercussionblades,andasinglelargerblade

(13)

fromaskilledpercussionopposedplatformtechnique.The standardisednatureof theunipolar pressureblades(all materials)isattestedbythegreatregularityofthefinished products,andthehighratioof2-1-2’blades,indicativeof ahighlyskilledpracticeinvolvingcarefulcontrolofcore preparationandthesubsequentknappingsequence.Ashas beensuggestedatthenearby,butslightlyearlier(middle PPNB)siteofTellSabiAbyadII(Fig.9),wemightalsobe dealingwiththeproductsofspecialistcraftspeople(Astruc et al., 2007), although hereat Qdeir 1we do not have evidence fortheuseof crutchandsmall leverpressure debitage (Abbès,2013).Asnoted previously,onlya few obsidiancorepreparationelementswerefoundinQdeir (allmaterialsconsidered),andonlyonecorticalflakewas reported.Thus,thefirststagesoftheknappingsequence appeartohavebeenundertakenelsewhere,whichleadsus towonderwhichaspectofthisindustrywouldhavebeen mostvaluabletothesepeople—theexoticrawmaterial (obsidian),ortherestrictedtechnicalknow-how(pressure technique)withwhichtoworkit?Insum,whiletheQdeir1 obsidianassemblagecontainsfourdifferentrawmaterials, thereisnoreasonwhatsoevertobelievethattheywere procuredviafourdistinctexchangenetworks.Instead,it appearsfarmorelikelythatanintermediarypopulation, perhapsintheMiddleEuphratesorfurthertothenorth- west(discussedbelow),wereresponsibleforprocuringthe differentrawmaterials,andthenconsumingthemwithin thesamepressure-bladetradition.Byextent,whomever wasresponsibleforworkingtheobsidianatQdeir1,bethat avisitingskilledknapper,oralocalcommunitymember, theytoothencontinuedtoreducethesecoresinthesame manner.

Ourfirstpointofcomparisoniswithotherassemblages fromthesurroundingsitesintheElKowmbasin,namely theElKowm2-CaracolvillageandUmm-el-Tlel2.Atthe formersite,wealsohaveamixofobsidianfrombotheast- ernandcentralAnatoliansources,mostlyintheformof end-products,againunipolarpressurebladesandbladelets (includingcorner-thinnedexamples),plussomeknapping debrisandasinglecorewhichappearssimilartotheexam- plefromQdeir1(Cauvin,2000).AtUmm-el-Tlel2(Borrell etal.,2013;MolistandCauvin,1990),somepressureblades werefoundbutnocores;withonlyBingölBobsidianthus farattestedfromthesite(Gratuzeetal.,1993).Thefact thatElKowm—thevillagesite—appearstohaveamore restrictedquantityandrangeofmaterial(nocorticaldebris for example) thanat the nomadiccampsite ofQdeir 1, mightbetakenasevidencetosupportthetheoryespoused by V. GordonChilde in the1930s that nomadic groups mayhave‘gravitated’aroundsedentaryvillages,supply- ingthemwithvariousnon-localproducts(Childe,1934).

Thatsaid,itshouldbenotedthatexcavationsatElKowm havethusfarbeenlimitedinnature,andthattheknown flintworkshopshavenotyetbeenexcavated.Moreover,in Qdeirtheexcavationtechniqueallowedallmicrodebitage tobecollected,whereasinElKowmthiswasnotthecase (F.Abbès,pers.comm.).

IfwethenconsidertheQdeir1datainitsFinalPPNB broadercontext,then onecanpointtowhatappearsto beverysimilarsetsofmaterial(rawmaterialandtechnol- ogy)fromnotonlyElKowmandUmm-el-Tlel,butalsoAbu

Hureyra(pers.obs.)andtheslightlyearlier(MiddlePPNB) TellSabiAbyadIItothenorth(Astrucetal.,2007).With obsidianrepresentingonlyaminoritycomponentofeachof thesesites’chippedstoneassemblages,andthefactthatthe materialwasoftenbeingconsumedinadistinctmannerto howflintwasbeingworked(atQdeir1theflintassemblage essentiallyrelatestoanaviform/opposedplatformpercus- sionbladetechnology[Calley,1986]),onehastoconsider theideathattheseobsidianassemblagescouldhavebeen producedbyoneortwonon-residentialitinerantspecial- ists(cf.Perlès,1990).Onemightarguethattheseitinerants wereactuallymembersofthenomadicgroupwhocamped atQdeir1,perhapsoccasionallyleavingthegrouptotravel northtoaccesstheobsidianquarriesofAnatolia,and/or interactwiththoseindividualswhocouldaccessthedistant materialsforthem.

ThefactthatatQdeir1—anomadiccampsite—wehave evidenceforagreaterrangeofknappingdebristhanitsnear neighbourssuggeststhatthismobilecommunitymayhave representedthemeansbywhichobsidianwascirculatedin thislocalarea.WhileElKowmhasacertainamountofpro- ductiondebris,theUmm-el-Tlelassemblageiscomprised exclusivelyofend-products.Herewemightclaimthatthe formercommunitymayhaverepresentedamore‘attrac- tive’localefortheitinerantspecialiststovisit,whereasthe latterpopulaceonlyreceivedtheirtoolssecond-hand,with perhapsa reflectionof localpowerrelations and domi- nanceinthisarrangement.In theorising‘attractiveness’, onemight think ofthepolitical capital tobegained by anindividual/sodalityintheirabilitytoattractacharacter suchastheitinerantobsidianworkertotheirvillage,the latterpotentiallyanexoticfigureofknow-howandexpe- rience(basedontheirskilledcraftingabilities andtheir regionaltravels[cf.Carter,2007:100–102;Helms,1993]).

Theseskilledknappersmightbeviewedastheminstrelsor rhapsodesoftheirtime,theteller-of-tales,relayingcultural traditionsasmuchastheyweretool-makers.Theirappear- anceatavillagemightbeseenintermsofaperformance (withthealltooimportanthost(s)inassociation),perhaps appearingonacyclicalbasis,ordrawntothecommunityto participateinasocialgathering/festivalsurroundingsuch importantoccasionsasmarriageceremonies,themeeting offar-flungtradepartners,orthefoundingofalineage’s newhouse.

Furthertracesoftheseinter-communitynetworkscan beviewedthroughreferencetothespecificitiesofobsid- ianconsumption,asforexampleevidencedbythecommon tool-typeofthe‘cornerthinnedblade’.Thesedistinctive implements,definedasbladesthathavedeliberatelyhad theircornersthinnedbyretouchinitiated fromtheend (oftenamedialbreak),andinterpretedassickleelements (Nishiaki,1990),arewellknownfromlaterPPN—Pottery NeolithicsitesofthenorthernLevant/UpperMesopotamia (Nishiaki,2000:205–209,Fig.8.15).TheQdeir1examples (whicharemainlymadeonBingölBblades,followedby afewBingölAandGöllüDa˘gexamples)provideaclear linkbetweenthiscommunityandsuchFinalPPNBcontem- porariesasAbuHureyra,TellSabiAbyad,TellKashkashok II,and Bouqrasinteralia(Fig.9).Whilethesetoolstend to be made of eastern Anatolian raw materials (exclu- sivelyatBouqras),atbothAbuHureyraandQdeir1there

Références

Documents relatifs

Due to their biological role, their ease of measurement (Parrish 2013), and their transferability through diet (Iverson 2009), lipids have been used as biomarkers to study health

These results indicate that for ter- restrial species, micro-habitat resource partitioning and body size discrepancy among species could be the factors influencing dietary patterns

All our samples can be divided into three groups: the sample c-417 where cinnabar was identified as a single cinnabar pigment, samples c-100, c-137, c-155 and c-371 with red

Dans cette section, on présente les faux plans projectifs et on esquisse la preuve du fait que le groupe fondamental d’une telle surface complexe est un sous-groupe arithmétique

Another example is Vinh’s work [32] who used a Szemerédi-Trotter type bound to obtain a different proof of Garaev’s sum-product estimate in finite fields (see Theorem 4 below)..

Moreover Λ is composed of eigenvalues λ associated with finite dimensional vector eigenspaces.. We assume that A is

Puis  on  vient  disposer  le  camion  au  centre  de  la  trame  comme  sur  le  schéma  suivant.  Il  s’agit  maintenant  de  décharger  et  fixer 

The present study investigated the relationship between dietary consumption and alcohol consumption and cognitive disorders among elderly people from rural and urban areas in