JournalofAnthropologicalArchaeology32(2013)774–790 ContentslistsavailableatSciVerseScienceDirect Journal of Anthropological Archaeology journal homepage: www.elsevier.com/locate/jaa Sovereignty and territoriality in the city–state: A case study from the Amuq Valley, Turkey ⇑ James F. Osborne JohnsHopkinsUniversity,DepartmentofNearEasternStudies,GilmanHall117,3400NorthCharlesStreet,Baltimore,MD21218,USA a r t i c l e i n f o a b s t r a c t Articlehistory: Thisarticleinvestigatestherelationshipofstateauthorityandterritoryinthecity–state,usingtheIron Received29June2012 AgeSyro-AnatoliancultureoftheancientNearEastasacase-study.Althoughmoresophisticatedspatial Revisionreceived15May2013 modelingofpoliticalauthorityhasappearedinthepastdecade,archaeologistsarestillpronetoassume Availableonline10June2013 thatterritorialityinancientcity–statesoperatedaccordingtoa‘‘containermodel’’principleinwhich,like themodernstate,politicalpowerisevenlydistributedacrossthelandscapewithinclearboundarydivi- Keywords: sions.ThepresentworkexaminesboththehistoricalrecordfromtheIronAgeontheonehand,andregio- Territoriality nalsettlementpatterndataontheother,toevaluatetheappropriatenessofthisconceptionofterritory Sovereignty andpowerintheSyro-Anatoliancity–stateofPatina,locatedinsouthernTurkey.Textualaccountsand City–state gravity modeling of settlement distributions point toward a pattern of territoriality in which power Settlementpattern IronAge was present inconsistently across the geographical extent of the city–state, and in which borderlines Syro-Anatolia asconventionallydrawndidnotapply.Irefertothisflexiblerelationshipofauthorityandspaceasmal- AncientNearEast leableterritoriality. (cid:2)2013ElsevierInc.Allrightsreserved. Introduction theSyro-Anatoliankingdomsas‘‘city–states’’(Thuesen,2002),de- fined here as independent polities characterized by their small TheSyro-Anatoliancity–statesoftheNearEasternIronAge(ca. scale, by having a single city center that dominated the rest of 1200–700BC) were clustered around the northeast corner of the the settlement pattern economically and politically, and by their MediterraneanSea(Fig.1).Theyaroseoutofthepoliticalturmoil participationinaregionalpoliticalsystemthatinvolvedmultiple thatfollowedthecollapseoftheLateBronzeAgepalaceeconomy neighboringpolitiesofthesamebasiccomposition(Charltonand and the downfall of the Hittite Empire around 1200BC (Bryce, Nichols, 1997: 1; Griffeth and Thomas, 1981; cf. Hansen, 1998;Glatz,2009;WardandJoukowsky,1992).Thesepolitiesthen 2000a,b;NicholsandCharlton,1997;Trigger,2003:94–103),de- existedasindependententitiesintheearlyfirstmillenniumuntil spite the objections of some scholars to this term (e.g., Cowgill, their piecemeal annexation into the imperial apparatus of the 2004: 527; Feinman and Marcus, 1998: 8; Marcus and Sabloff, Neo-AssyrianEmpireinthemid-tolate-8thcenturyBC(Hawkins, 2008:23). 1982; Lipin´ski, 2000). In a process of state formation that is still Embeddedwithintraditionaldefinitionsofthecity–stateisthe only dimly understood, former provinces of the Hittite Empire assumptionthatthispoliticalstructurenecessarilyderivedpolitical and sedentarizing nomadic pastoralists from inland Syria amal- authorityandlegitimacyfromaterritorialstrategythatinvolvedthe gamatedinthe12thand11thcenturiesBCtocreatetheSyro-Ana- ownershipandcontrolofcontinuousstretchesoflanddistinguished toliankingdoms(Bonatz,2000a,b;Bunnens,1995;Gilibert,2011; byclearboundariesandborders(e.g.,CharltonandNichols,1997:1; Giusfredi, 2010; Hawkins, 1982; Lipin´ski, 2000; Malamat, 1973; Hansen,2000b:16;Trigger,2003:94).InthispaperIquestionthis Mazzoni, 1994; Pucci, 2008; Sader, 2000; Schniedewind, 2002; assumptionbyassessingarchaeologicalandhistoricalsourcesfor Thuesen,2002;Ussishkin,1971). territoriality in the Syro-Anatolian city–state of Patina. By using Their roots in the Hittite Empire and subsequent interaction bothmaterialandtextualsourcesasevidence,Itackbetweenobjec- withtheAssyrianEmpirerenderthesestates‘‘secondary’’inneo- tive and subjective frameworks, operating under the assumption evolutionary typologies (Esse, 1989; Fried, 1967: 240–2; Joffe, thatbothsourcesofknowledgecanbeusedtocomplementandsup- 2002; Knauf, 1992; Marcus, 2004; Parkinson and Galaty, 2007; plementtheotherwithjudicioustreatment. Price, 1978). However, it may be more appropriate to consider Afterbrieftheoreticaldiscussionsregardingthecombinationof textsandarchaeologicaldataontheonehand,andsovereigntyand territorialityontheother,thispaperthenevaluateshistoricaland ⇑ Fax:+14105165218. empirical data that shed light on the relationship of political E-mailaddress:[email protected] 0278-4165/$-seefrontmatter(cid:2)2013ElsevierInc.Allrightsreserved. http://dx.doi.org/10.1016/j.jaa.2013.05.004 J.F.Osborne/JournalofAnthropologicalArchaeology32(2013)774–790 775 Fig.1. ConventionalmapoftheSyro-Anatoliancity–states,withregionalkeyinthebottomright.AdaptedbytheauthorfromtheTübingerBibelatlas,BIV14. authorityandterritoryinanearlycomplexsociety.Inbothcasesthe originsofagriculture,urbanism,andtheriseofstate-levelsociety evidenceshowsacomplicatedscenario,oneinwhichterritoryandsov- (Algaze, 2005; Rothman, 2001; Stein, 1999). On the other, the ereigntydonotnecessarilyhavethesamestraightforwardrelationship BronzeandIronAgesoftheancientNearEastfromthethirdmil- astheyhavehad(orareassumedtohavehad)incontemporaryhistory. lennium to the mid-first millennium BC, millennia that cumula- Incontrasttothegenerallyassumedmodelofevenlydistributedterri- tively provide hundreds of thousands of historical documents, torialauthority,powerwasexpressedandexperiencedasapatchyand featureconsiderablylessprominentlyintheliteratureofanthropo- highlyvariegatedphenomenonacrossthelandscapeoftheSyro-Ana- logical archaeology, though significant exceptions do exist (e.g., toliancity–state.Thistypeofterritorialsovereigntymayhavebeena Cooper,2011;Glatz,2009;Parker,2001,2003;StoneandZiman- commonfeatureofsecondaryandcity–statescross-culturally,andIre- sky, 2004; Ur, 2003). This contradictory intellectual history can fertoitasmalleableterritoriality. bereconciledbyanappealtoadialecticalapproachtoarchaeolog- icalproblemsofthehistoricalBronzeandIronAges–notdialecti- calintheMarxian,materialistsenseoftheterm,butinthesense Towardadialecticalapproach advocated by Alison Wylie (1989), who follows Geertz (1979) and others to propose tacking between ‘‘experience-near’’ and ThestudyofcomplexsocietyintheancientNearEasthasanidi- ‘‘experience-distant’’ operating frameworks, an approach that osyncraticdisciplinaryheritageinanthropology.Ontheonehand, combinessubjective/emicandobjective/eticsourcesrespectively. the ancient Near East has featured prominently in the major Thispapersubjectsthedatapertinenttoquestionsofpolitical anthropological debatesofprehistoricarchaeology, especiallythe territoriality in the city–state to two primary modes of analysis. 776 J.F.Osborne/JournalofAnthropologicalArchaeology32(2013)774–790 Thefirstisanexplorationofwhatthetextsandiconographicre- thathasbecomeknownasWestphaliansovereignty:neighboring mainscontributetoourunderstandingofnativeconceptionsand stateswithclearlydemarcatedboundaries,withinwhichstatesov- perceptions of space. Because both the textual and the artistic ereignty is held to be uniformly and evenly distributed across sources are so heavily dominated by the royal figure of the king, space. In crossing boundaries, sovereignty then transfers entirely analysis in this section focuses on how the texts and images in to the neighboring state (Harding and Lim, 1999; Krasner, 1995; question activated people’s imaginations by associating political Ruggie, 1993). Westphalia inaugurated what has been the basic authoritywithaspecificconstellationofsymbols. paradigmforunderstandingstatehoodandterritoryoverthenext Thesecondmodeofanalysisismoreformalinnature.Clifford threeandahalfcenturies. Geertz,theprimaryadvocateofinterpretiveanthropology,recog- JohnAgnew(1994,2009)referstothisgeopoliticalassumption nizedthemethodologicalrealitythatinterpretationofforeigncul- –theconceptualbundlingofsovereigntyandterritory,andwithit tures involves not just ‘‘experience-near’’ concepts, or those theexpectationofpoliticalpowerthatisevenlydistributedacross concepts by whicha culture in questionunderstands its ownac- spacewithinboundedcontainers–asthe‘‘territorialtrap.’’Agnew tionsandbeliefs,butalso‘‘experience-distant’’concepts,thosefor- andotherhumanistgeographers(e.g.,Brenner,1998;Elden,2009, mulationsmade by the researcher thatmake a culture’ssymbols 2010; Paasi, 1996, 2002, 2003; Raffestin, 1984; Sack, 1986) have and patterns of thought and behavior intelligible to others arguedthatthisbundlingisaresultofthegeopoliticalorderthat (1979). Although the goal is always to understand a culture on hasprevailedforthepast350yearssincetheriseoftheterritorial itsownterms,attimesanalysiswillinevitablyinvolvetheapplica- nation-state.Today’sglobalizedpoliticaleconomyillustrateswell tionofmethodsandinterpretationsthatstandoutsideofitsview that the relationship between authority and space is rarely as oftheworld(Sahlins,2004:4).Itshouldbeadded,however,that straightforwardasborderlinessuggest(seealsoAppadurai,1991, the metaphor of near and distant proximity implies the two are 1996). mutuallyexclusive,whentherealityisacontinuumofincreasing Archaeology has been slower to question territory as a neces- ordecreasinginterpretivedistancefromthearchaeologicalsubject. saryroutetopower,althoughthereisagrowingbodyofliterature Ancient inscriptions by contemporary but foreign cultures, for that presents different understandings of the spatiality of state example,occupyanepistemologicalmiddlegroundbetweennear authority. These alternatives include network models (Campbell, anddistant.Wylie(1989)referstothisoscillationbetweenthena- 2009;Keightley,1983;Liverani,1988;Parker,2001;Smith,2005, tive/emic/interpretive/qualitative mode of analysis on the one 2007; Tomaszewski and Smith, 2011) and other reconstructions hand,andforeign/etic/formal/quantitativeontheother,asa‘‘dia- thatmaintainterritorialitytobeanemergentphenomenoncontin- lectical tacking.’’ Archaeologists of the Iron Age Near East are genttohistoricallyspecificcircumstances(Casana,2009;Mantha, well-positionedtoperformsuchtacking,exploringwaysthatthe 2009; Parker, 2003, 2006; Smith, 2003; VanValkenburgh and twoapproachescanbebroughtintoharmony,andruminatingon Osborne,inpress).Thisstudycontributestotheseongoingdiscus- whatitmeansiftheycannot.Theresultisaricheraccountofthe sionsinanthropologicalarchaeologysurroundingtherelationship spatialityofpoliticallifethanwouldbepossibleifeitherwerecon- ofpoliticalauthorityandspaceattheregionalscale.Thearchaeo- ductedalone(cf.Sewell,2005:318–72;Taylor,2008:13;Thurston, logically and textually rich contexts of the ancient Near East, 1997). including the Iron Age Syro-Anatolian city–states discussed here, Inthecasepresentedhere,IexamineIronAgespatialdatafrom areparticularlywellsuitedtoanexplorationofhowearlycomplex theAmuqValleyofsouthernTurkeyasobtainedbythetraditional societies created and experienced politicalterritoriality,and how means of archaeological research, especially regional survey, but this was manifested in the city–state political formation alsoexploretheperceivedandconceivedaspectsofspaceinthat specifically.1 time and region, including data that might be said to fall under Withthismethodologicalandtheoreticalorientationinmind,I theumbrellaofhistory,suchasindigenousinscriptionsandartistic nowturntoterritorialityintheSyro-Anatoliancity–states,andthe programs.Suchadialecticalapproach,thoughnotstrictlyempiri- case-study of Patina in particular. The following analysis begins cal, may be considered an example of what Smith (2011): 168- with historical indications of sovereignty and space, and then 173,followingsociologistRobertMerton(1968)(andcontraBinford interprets the regional settlement pattern data for Patina during (1977)andotherarchaeologists[e.g.,Tschauner,1996;Varienand theIronAgeIIperiod(ca.900–700BC). Ortman,2004]),considersmiddle-rangetheory:thebodyoftheory that exists at the intermediary level between empirical observa- Representationsofterritorialsovereigntyinthetextualand tionsofdataandgrand,generalizingsocialtheory. iconographicrecords PatinawaslocatedintheAmuqValley(Turkish:AmikOvası)in Sovereigntyandpoliticalterritoriality south-centralTurkey(Figs.1and2).Patina’srougharealextentis typicallyderivedfromthenumerousroyalinscriptionswrittenby A recentreviewofsovereigntyinarchaeologyby AdamSmith AssyriankingsaftertheyhadpassedthroughtheAmuqeitheron (2011)summarizesrecenttrendsinarchaeologicalliteraturecon- military campaign or on exploitative economic raids (Grayson, cerningpoliticaldomination,arguingthatpoliticscanbebestde- 1991: 71–6; 1996; Harrison and Osborne, 2012: 125–6; Tadmor scribed as an emergent negotiation between personal will and andYamada,2011).Althoughitssphereofinfluencemayhaveex- sovereignprivilege.Oneaspectofsovereigntythatrarelyreceives tendedeastthroughtheAfrinValleytowardAleppoandwestover criticaltreatment,however,isthetraditionalassumptionofsover- the Amanus Mountains in the earliest decades of its existence, eignty’sterritorialjustification;mostscholarsstillconsidersover- Patina was concentrated in the Amuq Valley proper during the eignty to take place ‘‘within a delimited territory’’ (Smith, 2011: time period that is best understood archaeologically and histori- 416). Theideaofsovereigntylegitimizedbyterritorialholdingshasa longtraditioninWesternthought.Historiansofinternationalrela- 1 A more comprehensive survey of the intellectual history of territoriality in tions typicallytracethe ideato the Peaceof Westphaliain 1648, geography and related disciplines, and the consequences of this history for archaeology,isprovidedbyVanValkenburghandOsborne(inpress).Papersinthe which ended decades-long strife in Europe. In part, peace was accompanyingvolume,TerritorialityinArchaeology(OsborneandVanValkenburgh,in accomplishedbymeansoftheterritorialassignmentsascribedto press), present a series of case studies exploring archaeological manifestations of eachoftheinvolvedparties,leadingtoaprincipleofterritoriality territorialityinearlycomplexsociety. J.F.Osborne/JournalofAnthropologicalArchaeology32(2013)774–790 777 Fig.2. TheAmuqValley,surroundinggeologicalfeatures,andIronIIsettlementsidentifiedbytheauthorfromsurveydatacollectedbytheAmuqValleyRegionalProjects. cally, specifically the 9th and 8th centuries BC (Harrison, 2001: fortifiedsettlementsarealsorepresentedinAssyrianiconography, 120). for example the bronze panels that decorated the gates at the TheAssyrianrecordsprovideacontemporary,thoughnon-na- AssyriansiteofBalawat.ThesegatesportraycitiesinPatinawith tive, perspective on how Syro-Anatolian political authority oper- enceinte walls, fortified gates, and moats (King, 1915: Pls. XXV, atedattheregionallevel.IntheAkkadian-languageannalsofthe XXVII). Finally, there were the a¯la¯ni ša lime¯ti, or ‘‘cities in the Assyrian kings, three levels of settlement hierarchy within the neighborhood.’’ These settlements were never named. Instead, in Syro-Anatolian city–states can be identified from their accounts order to provide a general impression of the extent of inflicted ofsettlementstheyencounteredduringtheirannualmilitarycam- destruction,theywerelistedincountsofthetotalnumberofvil- paigns (Ikeda, 1979; Liverani, 1992). The first is the ‘‘royal city’’ lages sacked by the Assyrian ruler, ranging from a handful to as (Akkadian: a¯l šarru(cid:2)ti), the residence of the ruler and capital city manyas 150 (Liverani,1992: 125).From hisanalysisof Assyria’s (Ikeda,1979:76).Typicalfeaturesofthea¯lšarru(cid:2)tiincludedapal- annualmilitarycampaignaccounts,Liveraniconcludesthatthere ace(e¯kallu),oftenwithatreasury(makku(cid:2)ru),aharem,andasetof were,onaverage,slightlymorethanthreefortifiedcitiesforevery officials(rabûti)includingeunuchs(šare¯ši).Thesecondsettlement royal city, and approximately 20 villages for every fortified city category was the a¯la¯ni dannu(cid:2)ti, or ‘‘fortified cities.’’ These cities (1992:138). were generally named and, like the royal city, were associated Thus, as far as the visiting Assyrians were concerned, the explicitly with a walled fortification system. Syro-Anatolian Syro-Anatolianpolitieswerecharacterizedbyasettlementsystem 778 J.F.Osborne/JournalofAnthropologicalArchaeology32(2013)774–790 Table1 AmuqSurvey(AS)siteswithIronIIoccupation. Amuqsurveynumber Sitesize(ha) No.ofsherds No.ofIronIIsherds IronIIoccupationaccordingtoBraidwood IronIIoccupationdataquality AS6a 0.81 1055 12 No 3 AS10 1.6 28 5 Yes 2 AS15 1.35 87 2 Yes 3 AS16A 1 51 4 No 2 AS17 2.55 175 7 No 3 AS19b 0.18 20 1 Yes 3 AS23 1.44 25 5 No 2 AS33c 0.43 N/A N/A Yes N/A AS36 1.8 168 3 Yes 3 AS44d 1.2 30 6 No 2 AS52 3.22 107 9 Yes 2 AS55e 2.55 75 4 Yes 2 AS74 0.38 9 3 Yes 2 AS75 1.19 166 13 Yes 2 AS84 5.4 304 70 Yes 2 AS89 2.1 512 79 Yes 2 AS91f 3.75 239 7 No 3 AS95 1.44 306 27 Yes 2 AS99 7 442 44 Yes 2 AS106 2.1 200 20 Yes 2 AS119g 2.25 104 1 No 3 AS120 3.6 215 22 Yes 2 AS124 2.7 319 9 Yes 3 AS126 35 183 48 Yes 1 AS127h 0.49 32 1 N/A 1 AS129 4.5 585 49 Yes 2 AS131 3.45 97 16 Yes 2 AS134 0.5 49 5 Yes 2 AS138 1.75 294 5 No 3 AS139 5 218 10 No 4 AS141i 2 N/A N/A Yes N/A AS145i 1.3 N/A N/A Yes N/A AS146i 1.3 N/A N/A Yes N/A AS148i 1.8 N/A N/A Yes N/A AS149i 2.2 N/A N/A Yes N/A AS152 1.44 46 7 Yes 2 AS156 6.25 71 12 No 2 AS158 1 58 4 Yes 2 AS164 1.44 139 8 Yes 2 AS167 10 273 46 Yes 1 AS169 0.375 236 11 Yes 3 AS174 1.125 416 31 Yes 2 AS176 3 233 22 Yes 1 AS182 1 76 5 N/A 2 AS215j 0.33 372 2 N/A 3 AS221 0.16 63 5 N/A 2 AS252k 3.14 36 3 N/A 2 AS253k 4.71 105 3 N/A 3 AS283k 1.72 79 3 N/A 3 AS288 1.276 78 3 N/A 3 a ThesmallproportionofIronIIsherdsofthetotalcollectedmightinclineonetodisregardthissiteashavinghadanIronIIoccupation.However,theunusuallycommon Cypriot-stylepiecesspeakagainstsuchaninterpretation(seePlate1).MostlikelytheIronIIoccupationofthistallmoundisobscuredbysubsequentperiodsofhabitation. b DespiteonlyhavingasingleidentifiableIronIIsherdinitssmallcollection,anIronIIattributionforAS19issupportedbyBraidwood’sinterpretationtothateffect(1937: 22). c Becauseofsignificantdamagetothesiteinrecentyears,theoriginalBraidwoodIronIIattributionistheonlyoneavailable(1937:23). d ConcerningAS44,theAVRPGazetteerreads‘‘Notvisited...Thesitewasnotlocatedinthefield’’(CasanaandWilkinson,2005a:213).Nevertheless,abagofpotteryclearly labeledAS44ispresentintheAVRPcollections,andtheauthorhaspresentedthefindingsfromthissite. e Braidwoodreportsafragmentofa‘‘Hittite’’statuebeingfoundatthissiteandpublishedbyMesserschmidt(Braidwood,1937:25;Messerschmidt,1906). f AlloftheIronIIsherdscomefromthesteepnorthslopeofthetell,anareatargetedbythesurveyorsspecificallytoaccessearlierperiods.Thissuggeststhatsubsequent HellenisticandRomanoccupationsobscuretheIronIImaterialavailableonthesurfaceofthemound,hencethatperiod’srelativelysmallrepresentation. g LikeAS91,thissite’sIronIIlevelisdifficulttoaccessduetoHellenisticandRomanoccupation.ThesingleIronIIsherd,however,isunambiguous. h Althoughnotaceramicvesselperse,alsopresentintheAVRPcollectionsforAS127isabakedclayspoolweight,apparentlyabakedversionoftheAegean-stylespool weightcommoninsiteswithaSeaPeoplesconnectionintheEarlyIronAge(Stager,1995:346–7).ThesehavealsobeenfoundintherenewedexcavationsatTellTayinat (Janeway,2006–2007:138–9)andinIronIIcontextsatÇatalHöyük(Haines,1971:Plate16B). i BecauseofitsproximitytotheborderwithSyria,300mawayorless,thissitewasnotvisitedintherenewedsurvey.Forthisreasonnoassessmentofceramicsfromthe sitecouldbemade.ItsdatingthuscomessolelyfromBraidwood’soriginalassessment,anditsmeasurementsarederivedfromsatelliteimagery. j Likeothersites,weassumethatthelowabsoluteandrelativetotalofIronIIceramicsfromAS215islikelyaresultfromthemound’sheight(29m)andtheobscuringof earlieroccupationsbytheIslamic,Roman,Hellenistic,andAchaemenidlayers. k Althoughitwassurveyed,thedetailsofthissitearenotpublishedintheAVRPreport;itsinformationwaskindlymadeavailabletotheauthorviatheAVRPdatabase.It belongstoagroupofperipheralorhighlandsitesthat‘‘willbepublishedinaseparatevolumededicatedtotheuplandsandmountains’’(CasanaandWilkinson,2005a:203). J.F.Osborne/JournalofAnthropologicalArchaeology32(2013)774–790 779 Table2 Indigenous textual sources from Patina and its neighboring IronAgesitesidentifiedbytheOrontesDeltaSurvey. Syro-Anatolian polities,and thoseof the invading Assyrians, pro- OrontesDeltasitenumber Sitesize(ha) videarelativelydetailedunderstandingoftheexpressionofpolit- ical authority in Patina at the regional scale. The settlement OS11 1.6 OS12 1.2 hierarchy in Patina and other city–states was apparently three- OS16a 1 tiered as suggested by the annals of the Assyrian rulers and by OS32a 1 theterminologicaldistinctionsmadebetweencapitals,fortresses, OS34a 1 andvillagesintheAramaeanandLuwianinscriptionsthemselves. a Siteareaapproximatedfromsatelliteimageryinconjunction Syro-Anatolianrulers,includingthekingsofPatina,createdmonu- withthestatement‘‘All[IronAge]sitescanbedescribedassmall mental inscriptions that extolled not only their ability to build settlements’’(Pamir,2005:72). throughouttheland,butalsotheirabilitytomonitorandkeepit secure.Kingspromulgatedthismessagebyplacingtheseinscrip- comprised of three distinct tiers: a royal city, secondary fortified tions both in the capital and in conspicuous locations in the centers,andsmallruralsettlements.Theconsistentuseofthister- countryside. minologyoverthecourseofseveralAssyrianrulers’reigns(Ikeda, However, this ideology of royal authority being evenly spread 1979: Tables 1 and 2) suggests a durable pattern that was more across the territory of the city–state was not always represented thansimplytheartificeofsingleforeignobservers. soconsistentlyinpoliticaltexts,andaninscriptionfromPatinait- A similar hierarchy may be attested in an indigenous source self complicates this tidy conception of the relationship between from the Syro-Anatolian kingdom of Sam’al, Patina’s neighbor to sovereigntyandterritory.TheOrontesRiverflowsfromthesouth thenorth.Amonumentalstatuefoundinthiscity–statepossesses intotheAmuqPlain,wherePatinawaslocated,beforeturningwest aroyalinscriptionfromthemid-8thcenturythatreadsinpart‘‘In andsouthagaintocontinueitscoursethroughanarrowvalleyun- my days command was given...to establish cities and establish tilitreachestheMediterraneanSea(Fig.2).Thereareasmallnum- towns; and to the inhabitants of the villages my authority ex- berofIronAgesitesintheOrontesDelta(Pamir,2005),including tended’’(Gibson,1975:67),2seeminglyprovidingnativeconfirma- theimportantcoastaltradingsiteofAlMina.Aselaboratedbelow, tionofathreetieredsettlementpattern. scholarsarebeginningtounderstandAlMina,andbyextensionthe Syro-Anatolian royal inscriptions, though briefer and fewer in fewsmallsitessurroundingit,aspoliticallyintegratedwithPatina. numberthantheAssyrianroyalannals,arerepletewithproclama- Thisunderstandingofthedeltawouldappeartosuggestthatthe tionsofroyally-builtcitiesandfortresses(Mazzoni,1994,1995).In entire stretch of Orontes between the Amanus Mountains and astelefromthekingdomofHamath,forexample,theroyalauthor theJebelal-Aqrafromtheriver’sexitfromtheAmuqnearAntakya statesthatafterhavingfinisheddefeatingacoalitionofcity–states toitsdischargeattheseanearAlMinalikewisebelongedwithin that aligned against him, he built ‘‘...all] these strongholds Patina’szoneofinfluence. throughoutmywholeterritory’’(Gibson,1975:11).Likewise,the Yet there is a significant problem with this scenario. An early Phoenician–Luwian bilingual inscription from Karatepe records 8thcenturyBCcuneiforminscriptionknownastheAntakyaStele thatthelocalkingbuiltfortressesthroughouthiskingdom(Röllig, wasfoundinthemidpointofthistributaryvalleyhalfwaybetween 1999:51).Passagessuchasthese,whichrecurrepeatedlythrough- themoderncityofAntakyaandthecoast,andwellwithinthepre- out the Syro-Anatolian realm (see Hawkins, 2000: 231, 315, 402, sumedterritoryofPatina(Fig.2)(Donbaz,1990:5).Thestelerep- 409),pointtoalocalunderstandingofthe‘‘fortifiedcity’’similar resentsaboundarymarker(tahu(cid:2)mu)thatrecordsanAssyrianking tothesettlementtypeidentifiedbytheAssyrians.Nativeinscrip- negotiatingasettlementbetweentheSyro-Anatoliancity–statesof tions collectively indicatethat rulersand elites of Syro-Anatolian Hamath and Bit-Agusi, resulting in Hamath ceding towns to Bit- kingdoms conceived their settlement systems much as described Agusi;theyreportedly‘‘dividedtheOrontesRiverbetweenthem.’’ bytheAssyrians. Aboundarystelebetweentwocity–stateslocatedsquarelywithin Furthermore, the idea of ubiquitous political power was pro- anareathatshould,byallreason,belongtoneitherofthem(see mulgatednotjustbythecontentofroyalproclamations,butalso Fig. 1), is certainly unusual. Some have resolved the problem by by their locations. Inscriptions are found scattered throughout proposing that the stele must have been sent down the Orontes thelandscapeofthesepolities,ofteninplaceswithnoassociated fromanoriginallocationtothesouthandeast,whereaBit-Agu- settlement(Denel,2006). Theinscriptionoverlookingtheancient si/Hamathborderwouldbemoreplausible(Hawkins,1995:96). andmodernroadatKötükaleinthekingdomofMelidisanexam- The motive for the transportation is unclear, and the move is ple,asistheremoterockreliefatKızıldag˘ (Harman(cid:2)sah,2011:63– onlypostulatedtomakeSyro-Anatolianborderspalatabletocon- 4, Fig. 2). In the city–state of Hamath, two Early Iron Age stelae temporaryprinciplesofcartography.Butthestele’slocationisonly werefoundatimportantcrossingsoftheOrontesRiver(Hawkins, problematicifoneassumesaSyro-Anatolianunderstandingofter- 2000:415–19).InPatinaitselfseveralfragmentsofamonumental ritoriality in which borders were discretely marked in space and inscriptionwerefoundreusedinthevillageofJisrelHadid,(now wereunderstoodbyinhabitantstodemarcateterritoriesofevenly DemirKöprü),theAmuqValley’smajorcrossingpointoftheOron- distributed control by political authorities. However, what if the tes. By strategically placing royal inscriptions at prominent river stretch of the Orontes River in the valley between Antakya and crossings, kings maximized their visibility, ensuring that these thecoastwassimplynotundertheauthorityofPatina(seeWeip- monumentsservedasaprominentreminderoftheirroyalauthor- pert,1992:58,n.97)?Thereisnoapriorireasontoconcludethat itytocitizensthroughoutthekingdom. theAntakyaSteledocumentstheaccretionofterritoryonthepart ofBit-AgusiattheexpenseofPatina(Harrison,2001:120).Rather, wemustconsiderthepossibilitythatthisinscriptionmayprovide 2 Unfortunately the pertinent terms are not perfectly understood. K. Lawson a glimpseinto aparticularfeature ofSyro-Anatolian land tenure. Youngeroffersthemorecautious‘‘Inmydaysitwascommandedthroughoutallmy landtoreconstructTYRTandtoreconstructZRRYandtobuildthevillagesofthe Insteadofbeingadiscretelyboundedandintegrateddomain,the dominion’’(Hallo andYounger,2003:156).However,eveninthistranslationthe territory of Patina may have been characterized by a malleable contextimpliesthreedifferenttypesofsettlement.Otherepigraphershaveoffered conceptionofpoliticalspace. etymological and paleographic justifications for understanding TYRT and ZRRY as SimilarinscriptionselsewhereintheSyro-Anatolianrealmsug- citiesandfortressesrespectively(e.g.,DonnerandRöllig,1966:38;Tropper,1993: gestthePatinaexamplewasnotanisolatedcase,suchastheIncirli 68–9).IthankK.LawsonYoungerfordiscussingthispassageanditsproblemswith me. Stele that was found in the city–state of Gurgum, but assigns 780 J.F.Osborne/JournalofAnthropologicalArchaeology32(2013)774–790 territory and settlements from Gurgum and Kummuh to Que, a theavailability ofriverwater,theAmuqValleywasnotirrigated kingdomon the other side of the Amanus Mountains (see Fig. 1) untiltheRomanperiodduetothefactthatit receivesanannual (Kaufman,2007;SwartzDodd,2012:225–9).Likewise,astelewith average of 500–700mm of rain, well above the minimum aLuwian-languageinscriptionbythekingofCarchemish,thelarg- amount necessary for dry farming (Casana and Wilkinson, est city in the Syro-Anatolian realm, was found in the vicinity of 2005b: 28). Aleppo in the city–state of Bit-Agusi (Hawkins, 2000: 143–151). Today the landscape is heavily irrigated, primarily for corn, TheinscriptionrecordsthepurchaseofacitybytheCarchemish- sunflower, and cotton, but this is only a feature of the last few ean king for the price of six hundred mules, demonstrating that decades after several attempts to drain the Lake of Antiochwere Syro-Anatoliansettlementscouldeasilychangeownership(Giusf- finally successful (Çalıs(cid:2)kan, 2008). As Robert Braidwood pre- redi,2010:255–9).Itsproveniencefurthersuggeststhatsuchex- sciently deduced (1937: 8–9), the lake was a late feature. De- changes could take place without consideration for proximity to spite seasonal flooding, a large perennial lake was not yet in theruler’scapitalcity. existence during the lifetime of the city–state of Patina (Wilkin- Foreign Assyrian and indigenous Syro-Anatolian royal annals son,1997:566;2000:176–8),althoughtherewaspossiblyasiz- combine to present a scenario in which settlements were orga- able zone of semi-permanent marshy land created by flooding nized, in broad terms, according to a three-tiered hierarchy and thatwasstillsubstantialenoughtodiscouragesettlementinthat in which political territoriality was promoted by royal decree as area. Pollen cores taken from the adjacent Ghab Valley to the an absolute phenomenon. However, other, more mundane, pas- south show that the area was heavily wooded until roughly sages suggest that in practice sovereignty may have been more 8000BC, then was substantially deforested between 8000 and complicated territorially than the kings presented. This paper 5000BC. By the Bronze and Iron Ages (ca. 3400–600BC), there- nowevaluatesthearchaeologicalrecord,andspecificallytheregio- fore,thelandscapehadbeensignificantlymodifiedtoitsmodern nalsettlementpatterndatafromthekingdomofPatina,toassess agricultural state (Yasuda et al., 2000). If the Ghab Valley pollen the degree to which this data can inform our understanding of cores are tentatively accepted as proxies for the Amuq, then the Syro-Anatoliansovereigntyandspace. Amuqwaslikelyalreadydenudedofitsforestlong beforePatina came into existence. Paleoenvironmental reconstruction thus suggests that the Amuq Plain of the Iron II period was not en- ThekingdomofPatina tirely dissimilar from the Amuq Plain of today, with the excep- tion of varying surface elevations (Casana and Wilkinson, ThenorthOrontesValley 2005b: 30, Table 2.2, and Fig. 2.4). PatinawaslocatedintheAmuqValley(Fig.2),situatedatthe Regionalsettlementpatterns northern extent of the eastern Mediterranean littoral. The valley ispartofthegreatRedSea–EastAfricanRiftValleysystemthatter- The Syrian–Hittite Expedition surveyed the Amuq Valley in minatesslightlyfurthertothenorthattheMaras(cid:2)TripleJunction, the spring of 1936 when Robert Braidwood spent three weeks the meeting point of the African, Arabian, and Anatolian Plates assessing all the visible tell sites in the plain. The results were (Altunel et al., 2009; Sbeinati et al., 2005; Tolun and Pamir, published in Braidwood’s ground-breaking volume Mounds in 1975).TheAmuqisafairlydiscretegeologicalformationthatmea- the Plain of Antioch: An Archaeological Survey (1937). Braidwood sures roughly 535sq.km (330sq.miles) (Yener, 2005a: 2). In accomplished a lot in those three weeks: every visible mound shapeitcanbelikenedtoanequilateraltrianglewithitsapexat was visited, including those up the neighboring Afrin Valley, the northern corner (Braidwood, 1937: 8). On its west lies the and occupational histories were assessed through comparison imposing Amanus Mountain range, extending up to 2250masl. with a ceramic sequence established from the expedition’s exca- TothesouthwestandsoutheastliethefoothillsoftheJebelal-Aqra vations at Tell Judaidah. In total, Braidwood identified 178 sites, andZawiyeDag˘respectively,andtothenortheastisthebasalt-rich alreadyaremarkablyhighnumbergiventherelativelysmallsur- KürtDag˘. vey area. Though his survey methodology was pioneering in its Theplainitselfliesintherangeof80–90masl(CasanaandWil- time, there are limits to Braidwood’s study. There are no pottery kinson,2005b:28).Althoughextraordinarilyflat,theplainhasnot platesprovided, makingitimpossibletoevaluateperiodassigna- hadauniformdepositionalhistory.Geomorphologicalanalysishas tions. Site sizes were not measured formally, either in terms of demonstratedthatinsomeareasthesurfaceoftheplainisessen- areal extent or height. There was no attempt made to subdivide tiallythesametodayasithadbeeninNeolithictimes,whereasin surface collections such that differential occupational zones of a othersthesurfacelevelofearlyperiodsisburiedunderseveralme- single site through time could be estimated. Finally, only tersofaccumulatedsoil.Thisisduetoalluviationfromfloodingof mounded settlements were targeted and there was no attempt the rivers, colluvialdeposits resultingfrom the erosion of moun- to survey the rest of the plain, such that non-tell sites were en- tainslopes,andtheinfluenceoftheLakeofAntioch,allofwhich tirely excluded. This led to a bias favoring the Bronze and Iron affect different parts of the valley (Casana, 2008; Wilkinson, Ages. 2000:169–178). In 1995, research into the settlement history of the Amuq Among the important hydraulic features of the Amuq are was reinitiated by Aslıhan Yener with the Amuq Valley Regional three river systems. The small Kara Su flows from the north Projects(AVRP)(Yener,2005b; Yeneretal.,2000). AVRP’ssurvey and enters the Amuq at its northern corner, while the equally methodology involved taking systematic collections from all the smallAfrinRiverenterstheAmuqviatheconnectedAfrinValley mounds in the plain, including those visited by Braidwood, and to the east. The Orontes River, the only substantial river in the from all non-mounded sites as identified through satellite imag- plain, flows north through the Ghab Valley of western Syria ery and off-site transects (see Casana and Wilkinson, 2005b). In and enters the Amuq from the south, turning westward near addition, the survey was extended into the surrounding high- the sites of Tell Atçana and Tell Tayinat, the two principal lands with the goal of providing a counterpoint to the low- Bronze and Iron Age settlements in the plain. The river then land-exclusive efforts of Braidwood (Casana, 2003: 183–204). skirts the southwestern corner of the plain before heading These efforts have increased the current number of known southwest through the city of Antakya (ancient Antioch) en sites from 178 to 396 (Gerritsen et al., 2008; Swartz Dodd, route to the Orontes Delta and the Mediterranean coast. Despite 2011). J.F.Osborne/JournalofAnthropologicalArchaeology32(2013)774–790 781 IronIIsettlementsintheAmuqPlain ported pottery comes from Cyprus, especially the Bichrome and White Painted components of Cypro-Geometric pottery but also ItisthecollectionsmadebytheAVRPsurveythatformthebasis BlackonRedware(Gjerstad,1948;Schreiber,2003).Tayinatalso of the settlement pattern analyses of the early first millennium producedseveraldozenpiecesofGreekGeometricpottery(Cold- presented here. Identifying which sites in the Amuq might have stream, 1968), including especially the so-called ‘‘pendent semi- belongedtothekingdomofPatinarequiresknowledgeofthelocal circle skyphos’’ which is often taken as a marker of cultural and archaeologyoftheIron Age,and morespecificallytheIron II(ca. economicrelationsbetweentheAegeanandtheNearEastatthis 900–700BC).Fortunately,TellTayinat,theancientcityofKunulua, time(e.g.,Kearsley,1989). ‘‘royalcity’’ofthekingdomofPatina,wasexcavatedbytheOrien- Themorphologicalattributesofmoundedtellsitescompromise talInstituteoftheUniversityofChicagofrom1935until1939un- our abilityto determinea settlement’soccupational history from der the auspices of the Syrian–Hittite Expedition to the Amuq surfacefinds.Thefinalperiodofoccupationisclosesttothesur- Valley(Fig.3).Theexpeditionidentifiedfivephasesofarchitectural face,andthusprovidesthemostcommonceramicmaterialpicked remainswhichweretermedBuildingPeriods(Haines,1971).The up by surveyors, while sites that were occupied for a long time earliestofthesephases,BuildingPeriod1,wasdatedroughlyfrom mightnothavetheirearliestlevelsrepresentedonthesurfaceat 875to825BCandBuildingPeriod2from825untilabout720BC all.AstarkexampleofthisphenomenonisthatofTellSalihiyyah (Haines,1971:66;cf.Harrison,2009a,b).Onceramicandhistorical (AS129),lastoccupiedintheIronAge.Despitebeingattestedhis- grounds,theoriginaldatingofBuildingPeriod2hasbeenlargely torically in the Middle and Late Bronze Ages (ca. 2000–1200BC), confirmed, though modified slightly, from ca. 850 to 738BC. The theseperiodsprovidedlessthan2%oftheAVRPceramiccollection, conquest of the city by the Assyrians in 738BC marks the likely whiletheIronAgeprovided90%.Thisproblemaffectstheearliest endpointofBuildingPeriod2(Osborne,2011). periodsoftelloccupationmostadversely,which,intheAmuq,re- BasedontheTayinatsequenceestablishedbytheSyrian–Hittite fersespeciallytotheNeolithicthroughtheEarlyBronzeAge.How- Expedition a typology of local and imported ceramic wares was ever, a transition from tell-based settlements to farmsteads and madebytheauthor(Fig.4)(Osborne,2011,inpress).Thelocalpot- small villages began in the Seleucid period, roughly the mid-first tery of the 9th and 8th centuries is characterized above all by a millenniumBC(Casana,2007).Oneconsequenceofthistransition largenumberofopenvessels–bowlsandshallowplatterstogether isthattheIronAgelevelofthesetellsisoftentheirlastmajoroccu- constitute 69% of the assemblage examined – and a ware type pational phase, and therefore their uppermost layer (Casana and known as Red Slipped Burnished Ware that is found throughout Wilkinson,2005b:37).ForthosesiteswhoseIronAgeoccupation theLevantatthistime.Regionally,Tayinat’slocalformshaveclose isburiedundermultiplesubsequentoccupations,a relatedprob- parallels with the Levant and northwestern Syria (Cecchini and lemisourinabilitytoaccuratelydeterminesettlementsizeduring Mazzoni,1998;Lebeau,1983;Lehmann,1996,1998);theAmanus theIronAge.Previousresearchershaveattemptedtocompensate MountainrangeonthewestsideoftheAmuqseemstobeadivid- for this problem by looking at other variables pertinent to Iron inglinebetweenLevantineandAnatolianceramictraditions.Im- Agesitetopographysuchasunusualheight,whichisarguedtocor- respondtoIronAgefortifications(Casana2007,2009).Suchcon- siderations can helpfully modify analyses like gravity modeling in instances where the results seem unlikely or counterintuitive (seebelow). Equally challenging is the difficulty of placing these ceramic types into a high-resolution chronological sequence. Although newanddistinctceramicformswereintroducedbythepermanent presenceoftheAssyriansintheIronIIIperiodfollowingtheircon- questofKunuluain738BC,manyoftheformsoftheIronII’scera- micassemblagecontinuewellintotheseventhcentury(Hausleiter andReiche,1999).Isolatingthepre-Assyrianconquestphaseinthe AVRPcollections,therefore,remainsachallengingtask.Thismeans that the results presented here possibly conflate the settlement patternofc.850–738BCwiththatof738–c.625BC. TheresultsofanexaminationoftheAVRPcollections,basedon acomparisonwiththeIronIIceramicassemblagefromTellTayin- at,isprovidedinFig.2andTable1,whichlistsall50sitesfoundto haveIronIIpotteryintheircollections.Inadditiontositesize,ato- talcountofthesherdsinthesecollectionsispresented,aswellas thenumberofdefinitivelyidentifiedIronIIsherds.Thepresenceor absence of an Iron II assignation by Braidwood is also listed for sites AS 1 through AS 178, the last site recorded by Braidwood. ThefinalcolumnevaluatestheconfidencelevelintheIronIIattri- butionaccordingtoahierarchicalscaleadoptedandmodifiedfrom Casana(2009). Confidence scaling helps provide a transparent account of the qualityofthedata,althoughallofthesitesinthetableareincluded intheanalysesthatfollow.Datingconfidenceisrankedintofour categories, from highest quality to lowest: (1) the site possesses an Iron II occupation identified by excavation; (2) the site has a large collection of diagnostic Iron II pottery (>5%); (3) the site hasasmallcollectionofdiagnosticIronIIpottery(65%)atasite Fig. 3. CORONA satellite image of Tell Tayinat (DS1112-2203DA039). Note the with later, and therefore obscuring, occupation; and (4) the site excavationscarfromthe1930sSyrian–HittiteExpedition.Alsonotethelowertown andcitywall,notvisibleontheground. has a small collectionof diagnosticIron II pottery (65%) withno 782 J.F.Osborne/JournalofAnthropologicalArchaeology32(2013)774–790 Fig.4. CeramictypologyoftheIronII(ca.900–700BC)potteryatTellTayinat(Osborne,inpress).Greyshadingindicatesaredslippedandburnishedsurfacetreatment. Typicallocalformsincludeplatterswithringbase(1,2);bowlswithevertedrims(3,4);carinatedbowls(5,7);largebasinswithredslipandburnishontherimonly(6); kraters(8);jugs(14);holemouthcookingpotswithstonetemper(10)andcookingpotswithshelltemperandthickenedrim(11);storagejars(9,12,13);andpithoi(15). ImportedwaresaremostfrequentlyCypro-GeometricpotteryfromCyprus,includingBichromeWare(16)andWhitePaintedWare(17)(Gjerstad,1948);muchlesscommon isGreekGeometricpottery,especiallythepaintedsemi-circleskyphos(Coldstream,1968;Kearsley,1989). identifiablesubsequentoccupation.Thereisonlyasingleinstance (Fig. 2). When it departs the Amuq Valley through its southwest (AS139)ofthislast,leastreliablecategory,suggestingthat,onthe corner,theOrontesentersanarrowzonebetweenthefoothillsof whole, the identification of Iron II sites has been accomplished theAmanusrangetothenorthandwestandthefoothillsoftheJe- withreasonableconfidence.ByfarthemostfrequentIronIIcera- belal-Aqratothesouthandeast.Theentrancetothiszonefrom micformswereround-lippedandsquared-lippedRedSlipandBur- theAmuqwaslateroccupiedbythenextgreaturbancentertofol- nishedWareplatters.Cypriotpotterywascollectedfromanumber lowKunulua,theRomancityofAntioch.Untilrecentlytheonlysite ofsites(e.g.,AS6,AS17,AS84,AS95,AS120,AS138),suggesting alongthevalleyof theOrontes River betweenthe Amuq and the thatthesevesselswerecirculatedwidelythroughouttheplain.The MediterraneanthathasfeaturedindiscussionsofIronAgearchae- rarerGreekGeometricimportsdonothaveasingleoccurrence. ologyhasbeenTellSheikhYusuf,betterknownasAlMina(Wool- ley,1938a,b). AlMinaandtheOrontesDelta BecauseofitsunusuallyhighquantityofGreekGeometricpot- tery,AlMinahasbeenthesinglemostimportantsiteontheLevan- InadditiontotheAmuqValleyproper,thereisanothergeolog- tine littoral for documenting material relations between the icalzonethatmustbeconsidered:theOrontesDelta,asmallstrip ancientNearEastern‘‘East’’andtheclassicalGreek‘‘West,’’espe- of flat, fertile land located at the mouth of the Orontes River cially during the seventh century’s so-called Orientalizing J.F.Osborne/JournalofAnthropologicalArchaeology32(2013)774–790 783 Revolutionoftheclassicalsphere(Burkert,1992).Thetraditional tlementsystemsinavisualandquantitativeway(HodderandOr- viewofAlMinahaslongbeenthatearlyAlMinawaseitherwholly ton, 1976). This technique has its origins in the positivist GreekorhadmanyGreekslivingthere,thattheGreeksatAlMina movementsofgeographyinthe1960s(e.g.,Olsson,1965),andlike hadsettledtherewiththepurposeofimportingNearEasternexo- all quantitative methods, should not be considered in isolation. tic goods to Euboea and the greater Greek world, and that the However,gravitymodelingisahelpfulwayofconvertingnumeri- Greek pottery found at the site was used by these Greek traders calsite-sizedataintoavisualrepresentationofpotentialrelations specifically (Boardman, 1959, 1965, 1990, 1999, 2001, 2002). Re- betweensites,andtodosoinawaythatisconsistentlyappliedto newedresearch,however,hasmadeacompellingcaseforunder- everysite. standing Al Mina as a fundamentally local phenomenon on the Thegravitymodeldependsontwosignificantassumptions.The basis of large quantities of local pottery, the Levantine character firstisthatinteractiondecreaseswithdistancebetweentwosites, of the architectural remains, and more sophisticated notions of due to the rising cost in communicative efficiency that increases ethnicity than simply equating pots with people (Graham, 1986; withdistancebetweensites.Thesecondisthattheintensityofa Lehmann,2005;Luke,2003;Taylor,1959;Waldbaum,1994,1997). site’s interaction with neighboring settlements increases as that ThequestionofwhooccupiedthecoastaltradingsiteofAlMina site increases in size or population; the larger or more populous leadsustoconsiderwhocontrolledthesitepolitically.Geograph- thesite,themoreitinteractspoliticallywithitsneighbors(Alden, icalpassagesinhistoricaltextssuggestthatAlMinawastheIron 1979:170).Inotherwords,interactionbetweentwoplacesiscon- Age town of Ahta, a settlement that was a royal storehouse con- sideredtobedirectlyproportionaltotheirpopulations,butinver- trolled by the kingdom of Patina (Na’aman, 2007: 44; Tadmor selyproportionaltothedistancebetweenthem(Alden,1979:171). and Yamada, 2011: 85–6; Zadok, 1996: 104–5). Analysis of the The precise nature of the proposed interaction – political, eco- materialculturefromtheircontemporarylevels,especiallytheAe- nomic, cultural, or other – is unspecified; here I treat it as an gean and Cypriot ceramics at Tayinat that had to have been im- approximationofpoliticalrelationships,thoughitmightbemore ported from a coastal site and that are of a significantly greater appropriate to speak more vaguely of ‘interaction spheres’. Of quantity at Tayinat than at any other inland Levantine site, also course,theseassumptionswillnotholdineveryinstance.Apolit- indicates close interaction between the two sites (Luke, 2003: icalcenterthatissmallinrelationtomanyofitsothersettlements 12–20;Osborne,2011). maystillbeextremelypowerful.Otherlatentassumptionsinthe ThehistoricalimportanceofAlMinaasanintegralpartofPati- modelmustalsobeaccountedfor,especiallythefactthatitdoes naisnowalsosupportedbysettlementdata.TheplainoftheOron- not take topographic variability into consideration but rather tes Delta was surveyed by Hatice Pamir as a component of the placessettlementsystemsintoanabstractedspace.Infact,thislat- largerAmuqValleyRegionalProjects(Pamir,2005:67–98;Pamir ter assumption is less pernicious to our case than it might be in and Nishiyama, 2002). Pamir’s survey identified five sites in the others: since the Amuq Valley is a flat plain with few significant Orontes Delta that had Iron Age occupation, including the al- topographic barriers, it is perhaps as close a case as one is likely ready-known Al Mina (OS 11) and Sabuniye (OS 12) (Table 2). to get to the imaginary isotropic setting that many quantitative The published drawings include illustrations of several types at- settlementanalysesassume. tested in Tayinat’s assemblages, including Red Slipped and Bur- Gravitymodelingiscomputationallystraightforward.Firstone nished Ware and Cypro-Geometric White Painted and Bichrome measures the distance from every site to every other site. Then Wares(Pamir,2005:Fig.3.11and3.12).Thesesitesbringthetotal everysiteisassignedanindexofthelikelystrengthofitsinterac- numberofsettlementswithintheIronIIcity–stateofPatinato55. tion with every other site based on a simple equation: M =AA/ ij i j (d )2, where A and A are the areas of settlements i and j and d ij i j ij isthedistancebetweenthem(Schacht,1987:1973,n.2).Eachsite Gravitymodeling thushasadifferentinteractionindexforeveryothersite;thehigh- erthenumber,thegreatertheproposedlikelihoodandstrengthof Themapthatresultsfromthesedatashowsahighconcentra- interactionbetweenthem.Eachsite’shighestorderofinteraction tionofsites acrossthe valley,and a smallnumberofsites inthe correspondstoitsnearestinteractionneighbor,anditsnexthigh- deltaoftheOrontesRiver(Fig.2).Thenarrowvalleybetweenthe estordertoitssecondnearestinteractionneighbor.Theresultsof delta and the plain was not surveyed, but erosion from the sur- thesetwodegreesofproposedinteractionarepresentedvisuallyin roundingmountainshascoveredthevalleyfloorwithseveralme- Fig.5. tersofcolluvium(Casana,2008).Togetherwiththesprawlofthe Fig.5isillustrativeinseveralrespects.Firstandforemost,itis modern city of Antakya, this colluvium obscures ancient settle- apparent that AS 126, Tell Tayinat, features prominently; many mentinthisarea. sites both small and large are predicted to have close political TheIronIIsitesintheAmuqValleyarenotdistributedevenly interaction with Tayinat. 18 of the 55 sites have Tayinat as their throughout the plain. Instead, there is a large gap visible in the nearestinteractionneighbor,andanotherninelistTayinatastheir plain’snorthwestquadrant.Thisgapcorrespondscloselywithsea- second,indicatingthatKunulua,thecapitalcityofancientPatina, sonal wetlands predicted by hydrological modeling of the Amuq, waseitherthestrongestorsecondstrongestsiteofpredictedpolit- anareathatisshadedgreyinFig.2.Theaccuracyofthepredicted icalinteractionwithhalfthesettlementsinitshinterland.Nearest hydrologicalfeatureisbestevidencedbysitesAS10,AS15,AS16A, interactionneighborsofthefirstdegreehavebeenlinkedtogether AS 17, AS 19, and AS 23, all of which are close to the wetland’s byblacklinesandthoseoftheseconddegreebylightergreylines. west–northwest edge. These results point towards the existence ItisimmediatelyclearthatTellTayinatformedthemajorpolitical of a semi-permanent wetland area during the Iron II period that center,atleastaccordingtotheassumptionsofthegravitymodel. wasprobablyfloodedannually,possiblyexplainingwhythemajor- Because of its disproportionate size and population, Kunulua is ityofPatina’ssettlementsarelocatedintheeasternandsouthern predicted to have likely had a significant amount of political portionsoftheplain. authority over much of the area, even when distance from other Theissueofsitelocationcanbeexploredfurtherbyassessing sitesisconsidered. therelationshipbetweensitesizeanddistancebetweensitesusing TellTayinatisalsopredictedtohaveheldasignificantamount gravitymodeling.Gravitymodelingisamethodthathasseenonly of political gravity over sites that were far away, including those limitedusageinNearEasternarchaeology(Lupton,1996;Schacht, sitesonthewesternandnorthernedgeoftheAmuqValley.How- 1987),buthaslongbeenrecognizedasahelpfultooltoassessset- ever,thisisatleastpartiallybecausetherearenositesinthearea