ORIENTAL 34
ARS
ORIENTALIS
34
ARS ORIENTALIS VOLUME
34
GuestEditor Alka Patel ManuscriptEditor Ann HofstraGrogg Designer EdnaJamandre ScholarlyPublicationsAssistant CassandraGood EDITORIAL BOARD
Ann C. Gunter,co-editor Martin J.Powers,co-editor DebraDiamond MassumehFarhad Thelma K. Thomas
ISSN 0571-1371
Printedin the United Statesof America C 2007 Smithsonian Institution, Washington,D.C. Cosponsoredby the Departmentof the History of Art, Universityof Michigan,and the FreerGalleryof Art, Smithsonian Institution,ArsOrientalissolicits scholarlymanuscripts on the art and archaeologyof Asia, including the ancient Near Eastand the Islamicworld. Fosteringa broadrangeof themes and approaches,articlesof interestto scholarsin diverse fieldsor disciplines areparticularlysought, as aresuggestions for occasional thematic issues and reviewsof importantbooks in Westernor Asian languages.Briefresearchnotes and responsesto articlesin previousissues of Ars Orientaliswill also be considered.Submissions must be in English,with all non-Englishquotationsnormallyprovidedin translation. Authorsare askedto follow TheChicagoManual of Style,i5th ed. A style sheet is available from the editorialoffice or at the ArsOrientalishome page:www.asia.si.edu/visitor/ arsorientalis.htm. Inquiriesconcerningjournal submissionsand editorialmatters:
[email protected]
EDITORIAL COMMITTEE
SussanBabaie Kevin Carr JosephChang Louise Cort JulianRaby MargaretCool Root JamesT. Ulak J.KeithWilson Ann Yonemura EDITORIAL OFFICES
Ars Orientalis FreerGalleryof Art Smithsonian Institution P.O.Box37012, MRC707 Washington,D.C. 20013-7012 Fordeliveries (DHL, FedEx,UPS, courier): 1050 IndependenceAve.,S.W. Washington,D.C. 20560
Subscription rates: U.S.: Individual $35.00 Shipping and handling $4.00 Institution $45.00
Shippingand handling $4.oo. Foreign:Individual$35.00 Shippingand handling $5.50 Institution $45.00 Shipping and handling $5.50
Subscription-relatedinquires (invoice, payment,and changeof address):
[email protected] Specialsubscriptionratesare currentlyavailableas a membershipoption through the American OrientalSociety.Formore informationwrite the American OrientalSociety, HatcherGraduateLibrary,Universityof Michigan,Ann Arbor,Michigan481og-120r, or accessthe society'shome page at www.umich.edu/-aos. The full text of ArsOrientalisis also availablein the electronicversionsof ArtIndex. Printedon acid-freepaper.
ORIENTALIS
ARS ORIENTALISVOLUME34 (2004)
COMMUNITIES AND COMMODITIES WesternIndia and the Indian Ocean, Eleventh-Fifteenth Centuries This volume is the result of an interdisciplinaryworkshopby the same name, convened at the KelseyMuseum of Archaeologyat The Universityof Michigan,Ann Arbor (November7-10, 2002).
Transliteration The contributions herein have requiredthe transliterationof certain words from Sanskrit,Gujarati,Persian,and Arabicinto Roman characters.Well-known place names, such as Gujaratand Rajasthan, havebeen left without diacriticals,while more specialized terms have been transliteratedfrom their original languages.The system employed for transliterationof Sanskritand Gujaratiterms is the generally acceptedSanskritPronunciationand Diacritic Guide, with slight modifications for Gujarati.The transliterationof Arabicand Persian words has followed the Encyclopaediaof Islamstyle as outlined in the InternationalJournalof MiddleEasternStudies.The only modification to this style is the uniform renderingof the ti marbuitaas -a for both Arabicand Persian.
CONTENTS
7
19
39
62
ESSAY
INTRODUCTORY
ALKA PATEL, GUEST EDITOR
CommunitiesandCommodities:Western Indiaandthe IndianOcean,EleventhFifteenthCenturies
AmericanInstituteof IndianStudies
TOPOGRAPHIES
GRANT PARKER
OF TASTE
IndianTextilesandMediterranean Contexts
DukeUniversity
THE BEGINNINGS
HIMANSHU
TheArtisanandthe Merchantin EarlyGujarat,Sixth-EleventhCenturies
Jawaharlal Nehru University
ARTISANS,
COMMUNITIES,
AND COMMODITIES
MedievalExchangesbetween NorthwesternIndiaandEastAfrica 82
QUSEIR AL-QADIM THE THIRTEENTH
99
134
151
172
IN CENTURY
PRABHA RAY
MARK HORTON
Universityof Bristol
KATHERINE STRANGE BURKE AND DONALD WHITCOMB
A CommunityandItsTextiles
Universityof Chicago
CARVING AND COMMUNITIES
ELIZABETH LAMBOURN
MarbleCarvingforMuslimPatronsat Khambhatandaroundthe IndianOceanRim, LateThirteenth-Mid-FifteenthCenturies
Universityof London
INDIAN TEXTILES FOR ISLAND TASTE
RUTH BARNES
GujaratiClothin EasternIndonesia
OxfordUniversity
LUXURY GOODS AND INTELLECTUAL HISTORY
PHYLLIS GRANOFF
TheCaseof PrintedandWovenMulticolored Textilesin MedievalIndia
YaleUniversity
PATTERNS IN TIME AND SPACE
CAROL BIER
Technologiesof Transferandthe Cultural Transmissionof MathematicalKnowledge acrossthe IndianOcean
TextileMuseum, Washington,D.C.
ALKA PATEL
COMMUNITIES AND COMMODITIES WesternIndiaand theIndianOcean,Eleventh-FifteenthCenturies
THISVOLUMEof ArsOrientalisis the resultof an interdisciplinary workshopby the same name, convened at the Kelsey Museum of Archaeology at the University of Michigan, Ann Arbor, November 7-10, 2002. As is the case with all fruitful intellectual endeavors, the sources of inspiration for the gathering, and consequently for this volume, were multiple. Initially, during my perusals of the Kelsey Museum's various collections for objects that could be useful in teaching (we know that nothing excites students as much as interaction with the actual "subjects" of their studies), I was struck by the museum's textile holdings. Specifically,
the block-printedcotton fragmentsproducedin westernIndiaand excavatedin Egyptseemedto be perfectfor teachingby virtueof theirbeautyand exporthistory.Fortunately,the Indianfragmentshadalreadybeenexhibitedandpublished by RuthBarnes.'As a projectrelatedto teachingbut feasibleas a separateenterprise,a visualizationof the productionand circulationof these importedcloths togetherwith textilesbelongingto the indigenousEgyptianand Levantinetraditions, particularlythe silk, cotton,and linen tirazfragmentsproducedprimarily for elite consumption,2would, I thought,greatlydeepenour knowledgeof nonmodernpatternsof trade,travel,andtransmissionof ideas. The western Indian and tiraz fragments, when considered simultaneously,
seemedto providethe perfectsupportsfordetailedinvestigationof the historical realitiesof the elevenththroughfifteenthcenturies.Amongthe issuesthey urged us to examinewerethe differingconsumerdemandsamongthevarioussocioeconomicstrataof a society.Connectedwith this focuswerethe marketspecificityof import and export commodities and the awareness of these differing consumer demands on the part of merchants trading in these commodities, and perhaps even on the part of the supposedly less mobile communities of craftspeople who
producedthem. These early centuries of the second millennium, however, did not seem to be
as directlyaccessibleas the late fifteenththroughtwentiethcenturies.The latter era saw the ascendancyof navaland colonialpowersover previouslyunexplored regions and unknown societies, which in turn gave rise to many archives such as those of the British, French, Portuguese, and Dutch Foreign Area offices. It is important to note that, for the earlier period treated here, it was the surviving objects themselves that suggested directions for in-depth investigation. This approach diverges fundamentally from that of many studies concentrating on later centuries, in which it is the already archived documentation of commercial travels and transactions that has led to a more intimate investigation of historical patterns. Bearing in mind the lack of surviving archival material dating to the eleventh through early fifteenth centuries,3one method by which the objects' suggestions of historical realities could be investigated further involved collaboration 7
of scholarsfrom variousdisciplines.Thus a groupof historians,archaeologists, art historians,and textual specialistswere invited to participatein the project. Happily,theyacceptedthe invitation. Eventually,the presentationsby the workshopparticipants,and particularly the discussionsthey generated,madeit apparentthata volumecompilinga small numberof thepaperscouldcontributenotonlyto the fieldof IndianOceanstudies but also to otherdisciplines.Certainly,the fieldof IndianOceanhistorywaswell establishedby the groundbreaking workof GeorgeF.Houraniand S. D. Goitein,4 to nameonlytwo prominentfigureswhosepublicationshavebeenespeciallyuseful forthe conceptionand executionof this project.Morerecentcontributionsto the fieldby scholarssuchas K.N. Chaudhuri,SanjaySubrahmanyam, JanetAbuLughod,H. P. Ray,and MarkHorton,5amongothers,havebeen equallyvaluable in both deepeningandbroadeningourknowledgeof the historyof people,things, and ideasmovingthroughoutthe IndianOceanand its borderlands.Thesemore recentworkshave,moreover,also contributedto the very scholarlymethodologiesbywhichwe identifyandapproachhistoricalprocesses. Withoutthese importantstudies, the presentvolume would not have been possible.Indeed,it can be said that, in additionto the KelseyMuseum'stextile collections,the broaderintellectualinspirationforthe workshopandthis collection of paperswas the groundworklaid by previousscholarship.Therefore,this volume'sdevelopmentof the frameworksanddatapresentedbyprecedingstudies is, we hope,the most fittingtributeto them. Theprimarycontributionthe presentvolumeseeksto maketo IndianOcean studiesand otherfieldsis its emphasison a methodologicalframeworkanchored in regionalspecificity.As is true of most disciplines,the initial stagesof developmentarecharacterizedby approachesthat areoverarchingin theirgeographical, and sometimesalso chronological,scopes.This qualityis demonstrated,for example,by Chaudhuri'sinfluentialpublications.In Tradeand Civilisationin the Indian Ocean (1985),and Asia beforeEurope(1ggo), Chaudhuri first applied to the
studyof the IndianOceanthe usableelementsfrom the Braudelianframework, whichwasitselfdevelopedas a methodologicaltool forinvestigationof the Mediterraneanandits borderingsocieties.6 Chaudhuri'sanalysisof the conceptof Asiaconcludedat the outsetthatit was "essentiallyWestern,"since "thereis no equivalentwordin any Asian language ...thoughexpressionssuchas the 'Seaof China'or the 'Seaof Hind' heldcertain analogousmeaningsin Arabicandsome of the Indianlanguages."7 Nevertheless, exploringthe applicabilityof the Braudelianframeworkto the IndianOceanwith the additionof religiousideologiesas fundamentalin the shapingof social and materialrealities8-required a less generalbut still somewhatoverarching 8
ALKA PATEL
treatmentof culturallydisparateregions.In placeof Asia as an identifiableand intellectuallyviablewhole,Chaudhuri'sworkproposedthe identificationof "four [distinguishable]IndianOceancivilisations,"namelythose of Islam,Sanskritic India,SoutheastAsia,andChina.9 The distinct identificationsof Islam and SanskriticIndia are particularly noteworthyfor our purposes.Not only in Indian Ocean studies, but also in most other scholarlyfields, the eleventh through fifteenth centuries of the history of South Asia have traditionallybeen analyzedthrough the two ends of what has effectivelybecome an intellectualpolarity between "Islam"and "India."It shouldbe remembered,however,that one componentof this polarity is a spatiallyspecificdesignation,while the other is a socioreligioussystem with a regionalorigin, but without a fixed physicallocation over time. Thus the intersectionof these two cultural processes-for both "India"as well as "Islam"are,it must be confessed,alwaysin the making-is surelyuniquein its regional manifestation.This work,then, proposesto keep in sight the historical specificityof the spaceof westernIndia and its culturalsurroundings,and therebyexplorethe uniquityof the negotiationsand changesinitiatedby Muslim and other communitiesin the region. Sincethis workhopesto contributeto the studyof IndianOceansocieties,it cannot remainmethodologicallyintroverted,concentratingonly on the events taking place within the region under investigation.Thus severalof the papers herein explore commercialand other exchangesbetween the western Indian coast of Gujaratand other geographicallyremotebut long-connectednodes of the Indian Oceanworld.As Hortonpoints out in his essayin this volume, the analysisof a specificregion,definedas historicallyand culturallyunique,and its interactionwith otherIndianOceansocietiespresentsan importantcontribution to a methodologybecomingincreasinglyprominentin Indian Ocean studies.1o Moreover,the twofoldapproachof exploringthe uniquityof a region,alongwith the connectionsit maintainedwith other,equallydistinctplaces,breaksthe hegemony of overarchingframeworks.Generalapproachesand ideas had certainly been necessaryduringthe initial stagesof investigation.Thanksto the advances they broughtabout,we can now beneficiallyconcentrateon smallerparcelsof study,whichwill in the endreciprocallyaidin nuancingthe largerframeworks. One final note must be madepriorto an introductionof the articlesin this volume.In lightof the manyexcellentcontributionsto the workshop,selectionof the papersto be includedwas especiallydifficult.Sincethe workshopwas underpinnedby the KelseyMuseum'sholdingsof westernIndianand tiniz fragments, a sizablenumberof the contributionstreatedtextileproductionand commercial exchange.In orderto providereaderswith a widerstudyof IndianOceanmate9
COMMUNITIES
AND COMMODITIES
rialhistoryduringthe elevenththroughfifteenthcenturies,othercommoditiesof exchangewerealso emphasized.This aim of wide representation, in conjunction withlimitedspace,necessitatedthe selectionof paperstreatingthe circulationof a varietyof commoditiesandideasoversomepaperson textilehistory. The articlesby GrantParkerand HimanshuPrabhaRayelucidateimportant precursorsto the patternsof commerce,travel,andtransmissionof ideasin place during the elevenththrough fifteenthcenturies.Parker'sarticlemakes several noteworthycontributionstowardour understandingof Indo-Romancommercialandculturalcontactsspanningthe firstcenturyB.C. throughthe fifthcentury A.D. Archaeological datarecoveredfromRomansitesin the Italianpeninsulaindicatethatcommercialtrafficwasnot unidirectional,flowingonly fromthe Mediterraneanto the Indiansubcontinenton voyagesof goodsacquisition.Remnants at severalRomanports of Indiantextilesand teak,the lattermost likelycoming from seagoingvessels,remindus that shipsof Indianoriginalso madetheirway far westward.Thesedata,togetherwith finds of importeditems such as pepper and ivory at numeroussites, suggestthat consumptionof luxury commodities took placein variousambitswell outsidethe urbanitas,or cosmopolitanismand sophisticationof Rome.Along with imperialismand Christianity,then, Parker proposesa thirduniversalism,namelya universalismof the exotic.Justas India representedthe extentof the knownworldand the limits of empireand religious mission, its commoditiesby synedocherepresentedthe extent of the unknown world,the ultimateother.This representationmadethe acquisitionand possession of thesecommoditiesall the moredesirable. Ray'scontributionturns our attentionto westernIndia, and specificallyto Gujarat.Herinvestigationof the movementsof merchantandartisangroupsduring the firstmillenniumof the CommonEra,andof the roleof statestructuresin this mobility,providesan importantchronologicalbackgroundto laterinvestigations.First,she points out that inscriptionsdatingto the secondcenturyA.D., issued duringthe reignof the KsatrapaRudradamana,alreadyidentifyGujarat as a regionalentity numberingamongthe territoriesof this ruler.Althoughthe precisebasesfordefiningthis regionalidentityin the mindsof contemporaneous merchants,artisans,andothercommunitiesremainunclear,the region'sidentificationas suchis neverthelessmethodologicallyimportant.It identifiesthe beginningsof a recognitionof the area'sculturaluniquity,whichseemsonly to solidify andbecomemorepreciselyarticulatedthroughoutthe followingcenturies. Supportedby Parker'swork,Rayalso notes that despitethe paucityof references to long-distancetraveland trade in surviving Sanskritcanonicaltexts, archaeologicalfindsandinscriptionsprovethatthe communitiesof westernIndia were clearlyimportantparticipantsin maritimetravelfor commerceas well as 10
ALKA PATEL
pilgrimage during the early centuries of the Common Era. Moreover, far-trading merchants, who most often had access to the surplus capital required in long-distance commerce, were not the only groups that traveled. Artisan communities were mobile both inter- and intraregionally, at times relocating from, say, Gujarat to Malaya in order to avoid onerous state taxation on services and goods and other state-enforced demands. While the state did not, apparently, foment trade and production in an active way on an inter- or intraregional level, it did attempt to draw revenue from commercial and artisanal activities. Indeed, this mobility of artisan communities motivated by the need to avoid taxation was most likely an essential mechanism in what scholars have long termed the pan-Indic nature of religious ideas, iconography, and artistic styles during the early centuries of the Common Era."l Perhaps most important, Ray'swork proposes a fundamental methodological shift. Rather than relying on paradigms of historical rupture, or the dichotomy of rural and urban ambits for investigating processes of historical change, she suggests that we examine these processes by focusing on another crucial factor: communities of people, who identified themselves on various bases such as religion or occupation, were certainly affected by-and in many cases even precipitatedpalpable changes in patterns of taxation, trade routes, and religious endowments. From the evidence presented in her work, it would seem that communitieswhether merchant, artisan, or other-more than state structures or even regional identities, were integral to both the maintenance as well as the alteration of historical patterns of commerce and travel. Ray'sproposed shift of the scholar's gaze to communities underpins Horton's presentation of the material evidence of connections between northwestern India and East Africa during the eleventh through fifteenth centuries. First, this article insightfully points out that there was a wide availability of certain commodities throughout the various regions bordering on the Indian Ocean. By virtue of their 11
COMMUNITIES
AND COMMODITIES
ports and inland networks,these regionsessentiallyconstitutedthe activecatalystsin commerce.Thusthe find of Chineseporcelainsin EastAfrica,for example, does not necessarilyindicatedirectlinks betweenthe latterand the Chinese mainland.Bywayof explanation,Horton(togetherwith Parker)bringsourattention to the importanceof cabotage,or tramp-trading,in nonmoderncommercial networks.Ratherthan callingat only one emporiumper trip, merchantgroups and individualsmost often tradedin severalports of call in orderto increasethe profitandefficiencyof theirlong andperilousoceanicjourneys.Thispracticeled to a widespread,surprisinglyconsistentvaluationof certainitemsthroughoutthe IndianOceanregions.Thisconsistencyis an importantindexofthe interconnectednessamongfar-flungregionson the basisof tradedcommodities,an interconnectednessthathasheretoforebeenlittle acknowledgedin studiesof the eleventh throughfifteenthcenturies. Onceagain,archaeologicaldataindicatethatmerchantsandmiddlemenwere not the only groupstravelingthe seawaysand overlandroutes.Communitiesof craftspeoplealso traveledlong distancesand in fact sometimessettled at their points of disembarkation.Horton proposespreciselythis scenarioas an explanation for the rapidrise of cloth productionin EastAfricaduringthe eleventh throughfourteenthcenturies.He suggeststhat the movementof artisansfrom northwesternIndia-perhaps Gujaratitself- andtheirsettlementalongportsof the EastAfricancoastbroughtconcomitanttransfersof technology,namelycotton cultivationandthe weavingandprintingof textiles.Weknowthesepractices to havebeen part of a long-establishedand well-developedtraditionin Gujarat and surroundingareasof northwesternIndia.Hortonclaimsthis to be the plausibleexplanationforthe acceleratedrise (in only threehundredyears)of a textile industryin EastAfrica,an industrythat had not been evidencedin the region priorto ca.A.D. 1000. Fromthis example,Hortonfurtherproposesthat artisanmobilitywas probablymorecommonthanpreviouslythoughtin IndianOceanscholarship.Rather than making severalround-tripvoyages-the common practiceamong merchants and their tradingguilds-migrating artisan communitiesmore likely settled in their port of call or nearbyareas.Since their settlementoften lasted overseveralgenerations,theseartisancommunitiesin all probabilityassimilated themselveswithin theirnew socioculturalhomelands,overtime relinquishingat leastsome of theirdistinctidentitymarkers.Therefore,the indicesin the historical recordof their migrationfrom greatdistances,includingoral transmission of origin storiesacrossgenerations,frequentlybecameless fixed and overtime diminishedin historicity.Nevertheless,these origin storiesshouldbe borne in mind by both anthropologistsand archaeologists,as they canbe possibleindica12
ALKA PATEL
tors of the migrationand culturalassimilationof artisancommunitiesthat had originatedacrossthe ocean. Next, the articleby KatherineStrangeBurkeandDonaldWhitcombprovides, among other things, a unique historicalperspective.It highlightsthe effectsof changingIndianOceantradepatternsduringthe mid- to latethirteenthcentury on the quotidianand luxuryitems consumedand tradedby a singlemercantile familyat the importantport of Quseiral-Qadim,Egypt.Basedon archaeological data,the workposits a crediblenarrativethat is singularlyeffectivein depicting the use andsignificanceof importeditems.Weseethatcottonblock-printedcloth from westernIndia, most likely Gujarat,was a ubiquitousitem in the archaeologicalsiteknownasthe Sheikh'sHouse,servingforgarmentsaswellas furniture coverings.Theseprintedtextilesweredecoratedlargelywith aniconicmotifsand Arabicpseudocalligraphy, providinga strongindicationthattheywereproduced specificallyfor Islamicmarketsin the PersianGulfand the RedSea.The market specificityof these cloths, evidentby virtue of their motifs,becomeseven more apparentwhen comparedwith the textile exports from northwesternIndia to SoutheastAsia,presentedlaterin this volumeby Barnes. Quotidianitems are not the only commoditiestreatedby Burkeand Whitcomb. The excavationsof the Sheikh'sHouse at Quseiral-QadimWest and of anotherresidenceafterthe settlementshiftedeastwardwithin the sameportarea revealthatseveralluxuryitemsnumberedamongthe families'possessions.These include Islamic glazed ceramicsand Chinese celadons.Stratigraphicanalysis indicatethat these items were retainedamong the households'belongingsover generations.Consistentwith Horton'sproposalsforthe EastAfricancoast,Burke andWhitcomb'sfindsalso suggestthat certainluxuryobjectswerenot for trade but ratherwereprizedpossessionsfor tradingfamilies,probablyitems gifted to them by prosperousmerchantsor otherparties.Moreover,the valuationof suitable gift items changedin seemingsimultaneitythroughoutIndianOceannetworks.Bythe secondhalfof the thirteenthcentury,Islamicglazedceramics,likely originatingin the regionsborderingthe PersianGulf(seeHortonin this volume), lost theirstatusof covetedgifts as the popularityand demandforChineseporcelainswason the rise. ElizabethLambourn'scontributionto this volume bringstogethermany of the issues addressedin the previousworksfrom differentperspectives.In her investigationof marblecenotaphsproducedat the importantGujaratiport of Khambhat,she proposesthat these creationsweredistinctiveand highlyprized commoditiesboth in theirareaof originand at othernodesof the IndianOcean networkextendingfromthe coastof EastAfricato SoutheastAsia.Thusthis study refutesthe misconceptionthat small,lightweight,portableobjectssuchastextiles 13
COMMUNITIES
AND COMMODITIES
werebyandlargethe onlycommoditiesto be shippedvastdistanceson the Indian Ocean'swaterways.Heavystoneproductsalsotraveledfarandwide,fulfillingnot only the demandsof consumersuponarrivalat theirdestinationbut also serving as balastforshipsduringtheirjourneys. Lambourn'sanalysesof the gravestones'iconographicprogramsalso provide insight into the possiblecollaborationof variousartisancommunitiesresiding andworkingin a singleregion.Hercomparisonsof the gravestones'iconographic motifswith thoseof othernorthwesternIndianproductions,suchas architecture, printedtextiles,and manuscriptillustrations,stronglysuggestthat motifs were sharedby craftspeopleacrossspecializations.Moreover,thesemotifswereshared by the variousreligiousambitsof northwesternIndia,encompassingJainismand BrahmanicalHinduismas well as Islam.Thesegravestones,then, alongwith the buildings,textiles,andmanuscriptsusedas comparanda, leadus to reconsiderthe rigidityof ourdefinitionsof objectsas "Islamic," "Hindu,"or "Jaina." Theseconsiderationsof iconographyalso havebearingon the "international" marketingof the gravestones.Lambournmakesthe extremelyimportantpoint that printed textiles from Gujaratwere alreadyin circulationthroughoutthe regionswheredemandforthe gravestoneswasalsosubstantial.Itis possible,then, thatthe motifswhichwerecommonplaceon these ubiquitousprintedtextilesin effectstimulatedthe demandforthe stonegravememorials.Clearly,the predilection forblock-printedGujaraticlothcontinuedlargelyunchangedin theseregions from the elevenththroughnineteenthcenturies.The carvingin stone of motifs thatwerefamiliarandpleasingfromtextilescouldhavefomented-and perhaps actuallyintroduced-a tasteforthe gravestonesaswell. In additionto the insights regardingthe local manufacturingprocessesof these stone gravememorials,Lambourn'sworkalso casts light on the mobility of theirproducers.The gravestonesfound at sites in regionsother than western Indiastronglyindicatethatartisanstrainedin Gujaratistoneworkingoftentraveled overseaswith their commissions.At the time of their assembly,these intricate memorialsrequiredthe highly specializedskills of craftspeopletrainedin the stone-carvingtraditionsof Gujarat,so that their presenceat the sites of the gravestonesis most stronglyimplied.Moreover,in consonancewith the scenario proposedby Horton,it is likelythat some of these artisangroupssettledat these new placesand probablyintegratedthemselvesinto the indigenouscommunities. Some gravestonesfound at sites in Javaand Sumatra,but clearlybelonging to the Khambhatcorpusbasedon the executionof theircarving,evincea confluence of westernIndianandlocalmotifs.Lambournproposesthatone of the ways this confluencecouldhavebeen achievedwas throughthe long-termpresenceof Gujaratistoneworkersin theseareas,permittingthe amalgamationof indigenous 14
ALKA PATEL
motifs and stylistic elements with an execution recognizable as belonging to the Khambhat stoneworking tradition. As noted in the discussion of Burke and Whitcomb's article in this volume, there is compelling evidence of the market specificity of certain commodities traded throughout Indian Ocean networks during the eleventh through fifteenth centuries. The groundbreaking work by Barnes on the export of western Indian textiles to Southeast Asia confirms and adduces to this evidence. While the cotton textiles traded to Egypt and the Red Sea ports were printed primarily with aniconic and pseudocalligraphic motifs, those traded to areas such as eastern Java and the island of Sulawesi tended to be replete with the colors and figural motifs that were culturally significant to Southeast Asian communities. These included lavish court and entertainment scenes set against large swaths of deep and vibrant red backgrounds. This specificity of motifs indicates that the communities of Southeast Asia were by no means passive consumers of the textiles shipped to them from the ports of Gujaratand northwestern India. They most likely communicated their preferences either indirectly through the repeated success of some cloths and not others or perhaps directly by "placingorders"with merchants traveling between the points of manufacture and consumption. Barnes also makes note of the very different roles western Indian textiles played in Southeast Asian communities when compared to their Near Eastern counterparts. Unlike the everyday functions they fulfilled in the ports along the Persian Gulf and the Red Sea, covering both human limbs and furnishings, Gujarati cotton cloths, as well as the more expensive and highly valued patolis (doubleikat silk), were markers of familial and dynastic wealth and status in Southeast Asian societies. As such, they were often carefully preserved and retained among a household's possessions over generations. Indeed, it is noteworthy that patoltiswere exported only to Southeast Asia and not westward to the Persian Gulf and the Red Sea. This trend is explicable, at least in part, by the generally higher status of western Indian cloth in Southeast Asia. The labor-intensive and therefore high-pricedpatolas would not have been salable commodities in the western reaches of the Indian Ocean world, where Gujarati cloth was used exclusively in daily garments and considered rugged enough even for the covering of floors, walls, and furniture. A forceful and consistent effortakin to a powerful "advertisingcampaign"-would surely have been required to convince Egyptian consumers that a labor-intensive, low-quantity, and expensive cloth was produced by the same region that, to them, was known primarily for informal, everyday clothing and upholstery. In contrast to the preceding analyses of textiles as commodities for transport and trade across vast distances and throughout vastly different societies, 15
COMMUNITIES
AND COMMODITIES
the articlesby PhyllisGranoffand CarolBier once again shift our gaze-this time to textiles as conveyersof multiplemeanings.Granoff'sworktreatsa previously unexploredtype of consumptionof textiles, and close to home within the geographicalregion of their production.Ratherthan commoditiesfor sale or exchange,the studypoints out how cloth,particularlythe varietywovenfrom threadsof severalcolors,was the metaphoricalcornerstoneof reflectionfor the or IndianRealists,a highlyinfluentialschoolof medievalIndian Nyaya-Vaisesika philosophers.It wasthe enigmaof the multicolored(citra) cloth, creatinga color of its own, thatsparkedreflectionon perceptionand ontology,leadingto debates with theirarch-rivalsthe Buddhists.The samemulticoloredcloth that had been carriedlongdistancesandservedvaryingfunctionsrangingfromthe mundaneto the ceremonialwas,in a verydifferentsphereof consumption,madeto servephilosophicalfunctions.In fact,citraclothwasinstrumentalin the Nyaya-Vaisesika's abandonmentof somelong-heldphilosophicaltenets. Bier'sworkfurtherdevelopsthe ontologicaldistinctionbetweenthe textileas objectand commodityand the abstractmathematicalprinciplesinherentin the weavingandpatterningof cloth.Thisarticleinnovativelypositsthatweaversand printersof clothhada practicalawarenessandunderstandingof keymathematical principles,with which membersof the intelligentsiaof the medievalIslamic worldengagedon a purelytheoreticallevel. In effect,textiles could haveserved as the conveyersof mathematicalideasfromthe Indiansubcontinentto the Near Eastandpossiblyviceversa.Not onlydid thesehighlyportablecommoditiesserve as links betweenepistemologicallydifferentcultures,but they also couldbe seen asthe conduitsof interactionamongvariousstrataof the samesociety.Thanksto the practicalengagementwith mathematicalprinciplesrequiredby their vocations, artisanscould participatein the highly esotericdiscussionsof these very principlesthatweretakingplaceamongthe intellectualelitesof theirtime. In essence,this volumecompilesarticlesthat relyprimarilyon the surviving, "unarchived" materialremainsas the basesfor investigationinto historicalpatternsof IndianOceantradeduringthe elevenththroughfifteenthcenturies.It is hopedthatthisworkwill collectivelyleadto futurecontributionsin IndianOcean historythattreatthe manylivesof commoditiesin nonmodernworlds-ranging from the commercialto the symbolicand philosophical-and also of the communitiesthatproducedthem.
16
ALKA PATEL
NOTES
1.
Ruth Barnes,IndianBlock-PrintedCotton Fragmentsin theKelseyMuseum(Ann Arbor:Universityof Michigan Press,
see SelectedBibliography. 6. See esp. Chaudhuri,Asia beforeEurope,
1993). Barnes's publication of the
7. Chaudhuri,Asia beforeEurope,22. 8. See Chaudhuri,Asia beforeEurope,24,36. 9. Chaudhuri,Asia beforeEurope,49ff. lo. Many recentworkssharein this methodological shift, including AngelaSchottenhammer,ed., TheEmporiumofthe World:MaritimeQuanzhou,1000-1400 (Leiden:Brill, 2001); Momin Mohiuddin, MuslimCommunitiesin MedievalKonkan (610 -1900 A.D.) (New Delhi: Sundeep Prakashan,2002). See also Selected Bibliography. ii. See esp. V. S. Agrawala,IndianArt (Varanasi:PrithviPrakashan,1965), 2; PramodChandra,On theStudyof Indian Art (New York:Asia Society,1983), 39-80: Chandra,TheSculptureofIndia (Washington, D.C.:National Galleryof Art,
NewberryCollection of Indian textiles at the Ashmolean Museum, Oxford University,forms a valuablecompanion to the Kelseypublication:IndianBlockPrintedTextilesin Egypt:TheNewberry Collectionin theAshmoleanMuseum, Oxford,2 vols. (Oxford:ClarendonPress, 1997). 2.
While these fragmentshavebeen treated piecemeal in some earlypublications,the most comprehensiveanalysisremains FlorenceE. Day,"DatedTirazin the Collection of the Universityof Michigan," Ars Islamica 4 (1937):421-46.
3. The immediate exception that comes to mind is the workby S. D. Goitein with the Geniza documents. Forour purposes, his "Indiabook" is of course particularly noteworthy,which was at least partially compiled as LettersofMedievalJewish Traders:TranslatedfromtheArabicwith IntroductionsandNotes (Princeton: Princeton University Press, 1973). For his
other articlesconcentratingon the Geniza documents' information on the India trade,see SelectedBibliography. 4. GeorgeF.Hourani,ArabSeafaringin the Indian OceaninAncientandEarly Medieval Times (1951; rev. and ed. John
Carswell,Princeton:Princeton University Press, 1995); Goitein, Letters of
MedievalIJewish Traders. 5. Some foundationalworksinclude K. N. Chaudhuri,Tradeand Civilisationin the Indian Ocean:An EconomicHistoryfrom theRiseofIslam to 175o(1985;reprint, Cambridge:CambridgeUniversityPress, 2001); Chaudhuri,Asia beforeEurope: Economyand Civilisationof theIndian Oceanfrom theRiseof Islam to i75o(gg0o; reprint,Cambridge:Cambridge UniversityPress,2000). Forother works, 17
COMMUNITIES
AND COMMODITIES
24-25.
1985), 20.
SELECTED BIBLIOGRAPHY
Abu-Lughod,Janet.BeforeEuropean Hegemony:TheWorldSystem,A.D. 12501350.New York and Oxford: Oxford
Ray,Himanshu Prabha,and J.-F.Salles,eds. TraditionandArchaeology.Delhi: ManoharPublishers,1996.
University Press, 1989.
Subrahmanyam,Sanjay.ThePortuguese Bloom, JonathanM. PaperbeforePrint:The Historyand Impactof Paperin theIslamic World.New Haven and London:Yale UniversityPress,2001. Chakravarti,Ranabir,ed. Tradein Early India. New Delhi: OxfordUniversity Press,2001.
Empire in Asia, 1500-i7oo: 1993.
Wink, Andre. TheSlaveKingsand theIslamic Conquest,lzth-13thCenturies.Vol. 2,AIHind: TheMakingof theIndo-Islamic World.Oxford:OxfordUniversityPress, 1999.
Goitein, S. D. "Fromthe Mediterraneanto India:Documents on the Tradeto India, South Arabiaand EastAfrica from the Eleventhand TwelfthCenturies." Speculum 29 (1954): 181-97.
. "NewLighton the Beginnings of the KarimMerchants."Journalof the Economicand SocialHistoryof the Orient 1.2
(1958):175-84.
. "FromAden to India."Journalof the Economicand SocialHistoryof theOrient 23.1-2
(1980): 43-66.
Horton, Mark,and JohnMiddleton. The Swahili:TheSocialLandscapeof a MercantileSociety.Oxford:Blackwell, 2000.
Ibn Battuita. TheRhela(India, Maldive Islandsand Ceylon),trans. Mahdi Husain. Gaekwad'sOrientalSeries122. Baroda: Oriental Institute, 1953.
Jain,V. K. Tradeand Tradersin Western India. Delhi: MunshiramManoharlal, 1990.
Ray,Himanshu Prabha,ed. TheArchaeology of Seafaring.Delhi: PragatiPublications, 1999.
18
ALKA PATEL
A Political and
EconomicHistory.London:Longman,
GRANT PARKER
TOPOGRAPHIES OF TASTE Indian TextilesandMediterraneanContexts
Abstract SouthAsiantextilesweretradedwithin multivalentnetworkslinkingthe Roman Mediterraneanandthe IndianOcean,bylandandby sea.Thisarticlebeginswith a brief surveyof the availableevidencebeforeexaminingthe attitudesevinced in Romansourcesabout such commodities.The emphasisis on consumption. Citiesplayeda prominentrole in consumption,but evidentlynot to the degree that ancienttexts suggest.The articleexaminesthe complexwaysin whichsuch commoditieswerelinkedwith Romanworldviews,andindeeddid muchto shape them.It emphasizesboth the particularismof specifiedoriginsandthe universalist aspectsof metropolitanconsumption.
FORREASONSBOTHSPATIAL AND TEMPORAL, the prehistoryof our common theme might profitablyembracethe Roman empire.The Mediterranean in the period of roughlythe first centuryB.C. to the fifth centuryA.D. exhibits some of the same patternsof exchangethat receivemore detailed discussion with regardto other areasand later periods. Indeed,though fartherwest, the ancient Mediterraneanbrings to light some of the social processesin which South Asian textiles, along with other objects,became involved.It is here that we can observethe social constructionof the exotic, and indeed of a pointedly generalizedconceptionof "the East."This articleundertakesthe treatmentof the Mediterraneanlife of SouthAsiantextiles.In the pagesthat followI hope to showthat this line of inquiryis productive,especiallyin view of the richvariety of availablesources. The articlewill begin with a brief overviewof South Asian textiles in the ancientMediterraneanas well as a considerationof their routesof circulation before proceedingto some specific questions concerningtheir consumption. These are as follows:To what extent was the taste for such exotic goods an urbanphenomenon?How weretextiles and other objects"mapped"in Roman minds?Here I proposethat this mappingdeservesconsiderationboth in terms of their supposedprovenanceand of the routesby whichthey weretransported. Finally,I shall consider the social meaning of this South Asian "exotic"in broaderterms, relativeto Romancosmologies.If production,distribution,and consumption are the three paradigmaticphases of economic activity,'then this article will concentrateon the third of these. Yet, as I shall show with regardto mentalmaps, this third stagerelatesto the earlierones in some surprisingways.As one aspect of the life of commodities,the phenomenologyof consumptionconcernsthe waysin which"India"existedin the minds of Roman consumers.2 19
TextilesPlus Atthe outsetit is necessaryto identifythetextilesthemselvesandthe otherobjects with which they havebeen found. Forthe primeevidencewe must look outside of the Mediterraneanitself,to the Red Sea coast of Egypt.Since the late 1970s, archaeologicalworkhas focusedon a numberof sites,notablyQuseiral-QadIm3 and, morerecently,Berenike.4In fact,the site of Berenikecontinuesto put forth importantmaterialthathasonlybegunto changethe overallassessmentsof longdistancetradealongthe monsoonroute. Cottonhas been amongthe variedfinds at Berenike,beginningon a modest scale in 1994.5Sincethen, finds haverun to hundredsof fragmentseach season. Thoseof the S/Sgrouphavebeen identifiedas Egyptianor Nubian;6on the other hand,the Z/Zcottonswerelikelyimportedfromthe Indiansubcontinent.7 Onthis scoretherewas some doubtin the earlieryearsof the dig, but circumstantialevidencenow points clearlyin this direction,not leaston the groundsthatthe subcontinentwas a majorproducerof cotton duringthe periodunderdiscussion.8 Whilethe designson some of the textilefragments(e.g.,rosettesand lotusbuds) still requiredetailedstudy,the generalconclusionis thattheseareof Indianorigin. Someof the textilefragmentsappearto havebeenpartof shipbuildingandrepair (sail and rope),particularlyin light of timber remains(Indianteak) that most likelycome fromseacraft.9Yetthe majorityprobablyrepresenttradegoods or,in smallerquantities,the personaleffectsof thoseinvolvedin the monsoontrade. Thesecotton finds area valuableadditionto the archaeologicalrecordof the searoutebetweenEgyptandIndia,especiallygiventhe scarcityof cottonremains in climateslessdrythanthatof the EasternDesert.Cottonis mentionedin one of the earliestGreco-Romanreferencesto India,namelyin the Historiesof Herodotus (ca.485-ca. 425B.C.). In this "mosteasterlypartof the inhabitedworld"there are, accordingto him, "wildtreeswhich producea kind of wool which is more attractiveandof betterqualitythanthatof sheep,andwhichis usedbythe Indians for clothing" (3.106).1'
The desirabilityand noveltyvalueof this productare immediatelyapparent. Thiscottonor "treewool"also featuredamongthe accountsof the historiansand scholarsaccompanyingAlexanderon his campaignto the eastin 327-325B.C. For example,the navalcommanderNearchusis quotedin Strabo'sGeography(15.1.20 C6g3)on the use of cottonin garments;Strabomentionssilk in the samebreath. Finally,the Periplusof theErythraean Sea,a ship captain'smanualfromthe midfirst century A.D. written in Greek,makes several referencesto the transport of cotton on the monsoon route. Both cloth (chapters48, 49, 5i) and garments (chapters 48, 51, 59) occur among goods brought into Egypt, whereas exports from Egypt to Arabia, India, and the East African coast include various kinds of garments (e.g., 20
GRANT PARKER
Bronze statuette of Indian origin, found at the site of Khor Rori, southern Oman; probablysecond century A.D. H. 8.6, w. 4.5 cm. American Foundation for the Study of Man, FallsChurch,Virginia, 3103.
chapters 6, 24, 56).11At a port called "Ganges"in the Ganga delta it was possible to acquire high-quality cotton, in the form of garments: "On [the Ganges] there is a port of trade [emporion] sharing the same name as the river, Ganges, through which malabathron, Gangetic nard, pearls, and cotton garments of the very finest quality, the so-called Gangetic, are transported" (chapter 63). It is typical of the Periplusthat various objects are linked in the context of a particular port. The designation of quality, diaphorotatai, has connotations of distinctiveness as well as value. Silk is another textile frequently mentioned by the Periplus as being transported from the subcontinent to the Red Sea. This is one of only two products, along with nard, to originate from all four of India's major exporting regions mentioned there. The southwestern ports of Muziris and Nelkunda offered "silk cloth" (othonia serika, chapter 56). Barbarikon offered in the Indus delta "Chinese pelts, cloth and yarn" (sirika dermata kai othonion kai nema sirikon, chapter 39);12 and the Ganges delta likewise, in addition to silk floss (erion serikon, chapter 64). The very word for "silk" provided the Greek and Latin name for China and its inhabitants-the "silk people" (Seres). This connection, in which the word for the object chronologically precedes that for place and people, is an interesting case of a cognitive geography related to commodities, as I shall discuss further below. Indeed, silk presents special problems concerning provenance, especially as it is unclear whether it was produced in the subcontinent at this time.13 Before we can investigate the social context of the textiles, it is important to consider other objects that formed part of the same exchange networks. There is plenty of evidence of the importation of spices to the ancient Mediterranean. Less clear is their exact provenance, when the same ships would combine goods from the Arabian peninsula with those from South Asia (see below). Archaeological evidence found at Berenike, in the form of peppercorns, resonates with the Periplus and other sources to illustrate that pepper was the most significant of these. White and black pepper came from the piper nigrum trees in Kerala on the southwest of the subcontinent, whereas long pepper (piper longum) originated in the north. What is striking is the heightened demand for pepper, as for other spices, that emerges from literary sources. Whereas pepper had been used pharmacologically in the Mediterranean since the late fourth century B.C., during the period of Rome's ascendancy it took on major culinary proportions. In the nearly five hundred recipes transmitted under the name of Apicius, who lived at the time of Augustus and Tiberius, pepper features almost without exception, even for wine and sweets. Apart from pepper, the Apician corpusmentions eight spices that must have come from South or Southeast Asia: ginger, putchuk, nard-leaf, cinnamon, spikenard, asafoetida, sesame seed, and turmeric. A much longer list of less 21
TOPOGRAPHIES
OF TASTE
exoticseasoningscouldbe madefromApicius,i.e., of ingredientsthatcamefrom within the boundsof empire.'4 Preciousstonesconstituteanotherclassof commodityimportedfrom India: e.g., diamonds,emeralds,sardonyx,turquoise,and onyx. All of these are mentionedin the Digest,a syntheticworkof Romanlawcodifiedin A.D. 533underthe emperorJustinian;these, like the other forty-ninekinds of commoditiesmentionedin thatlist, weresubjectto dutyuponentryinto Alexandria(39.4.16.7).Literarytexts often mentionthese in a social-criticalvein, especiallyas linkedwith conspicuousconsumptionon the partof Romanwomen (e.g.,Propertius,Elegies 2.22.10).
The above-mentionedpassagefrom the Digest mentions "Indianeunuchs" amongthe taxableitems, and thus points to humantrafficas well.'5Conversely, the Periplus mentionsthat the rulerof Barygazawould readilypurchaseconcubines and "singingboys"broughtfromEgypt(chapter49). Morefrequentlyseen in texts is a Mediterraneaninterestin, and thus demandfor,wild animals and birds,especiallyparrots. On the other hand, craftgoods constitutea surprisinglysmall and idiosyncraticgroup of objects.Among them is a statuetteof a young Indian woman engagedin her toilette,attendedby two small servants.This finelycraftedivory objectstandstwenty-threecentimeterstall and was madein such a wayas to be viewedfromthreesides,perhapsas the leg of a smalltable.'6Itwasfoundin Pompeii in the Via dell'Abbondanza,andit is quitepossiblethatthis wasat the home of a long-distancetrader.TheIndianprovenanceof this objectis beyonddoubtin view of comparandafromthe subcontinent,particularlythe Begramivories.Furthermore,it is clearthat it madeits wayto Pompeiibeforethe volcaniceruption of A.D. 79.A secondclassof craftobjectsis harderto accountfor,in partbecauseit lacksanycontext:a numberof marblebustsnowin the GalleriaBorghesein Rome arediscerniblyof Severandate(earlythirdcenturyA.D.). Theirunusualfeatureis that they have"Indian"hairstyles:cirrusknots that can be paralleledin Indian sculptureratherthanRomanportraiture.'7 Thestorybehindtheseheadsis a completemystery. ByLandandby Sea The period around3300-2200 B.C. sawthe vigorousexchangeof goods between the MesopotamianandIndusvalleycivilizations.Sincethe 1930S evidenceforthis exchangehas come from Mohenjo-Daroand Harappaas well as from Bahrain, designatedas Dilmun in Sumeriansources.Closerto the period underdiscussion, thereis furtherevidenceof networksof exchangewithin the PersianGulf: tradeactivitiesof the Seleucidswerea continuationof mucholderpatterns.'8 It is 22
GRANT PARKER
worthwhile to bear this long-range background in mind while considering routes during the Roman period. Two major routes linking continental Asia with the Mediterranean come immediately to the fore: the network of overland routes between East and West Asia, collectively called the Silk Road, and the monsoon route, extending from the Red Sea to the west coast of India. Each of these has been the subject of considerable scholarly attention, quite apart from wide popular interest. For one thing, there are serious questions as to whether this trade network touched on peninsular India, or indeed as far as the Mediterranean.19 Even if some of the more grandiose talk about a Silk Road in our period is exaggerated, and even if the term itself is misleading, there is clear evidence of overland networks stretching from Central and South Asia to the eastern Mediterranean. For example, the Parthian Stations of Isidore of Charax charts an itinerary from Antioch to Afghanistan, crossing the Euphrates at Zeugma.20While this work of about A.D. 25 is fragmentary and jejune, it does suggest a route of overland trade. Needless to say,this in no way precludes the possible existence of other land routes. The fact that this period saw hostilities between the Roman empire and its Parthian neighbors renders such evidence for trade all the more significant. When considered alongside the abundant archaeological and epigraphic evidence from Palmyra, it presents further proof of overland trade across western Asia, some of it in the same period as the monsoon trade.21 What we might call the monsoon route is described by the elder Pliny, stageby-stage from the mouth of the Nile to peninsular India (Natural History 6.10o6). As part of his conspectus of world geography, he lists the settlements along this route as well as the distances between them. It stretches from the delta up the Nile, then across the Eastern Desert for 30g Roman miles or 12 days. Once having reached the city of Berenike, Pliny's implied travelersails out of the Red Sea to the southern point of the Arabian peninsula, and then for forty days with the monsoon at his back, to the west coast of India. Pliny mentions a number of ports on the west coast of India, along with comments as to their viability and their access to inland commerce. Though Pliny'slist of ports differs substantially from that of the Periplus,the passage strikes much the same tone, especially in view of its focus on the logistics of sea travel. Severalconflicting accounts of the "discovery"of the monsoon exist in ancient Greekand Latin texts. Even allowing for the implicit ethnocentrism in the concept of discovery, such texts do make it clear that the monsoon and the route it made possible were a source of fascination. It is thus no accident that Pliny's account has a moralizing edge: he prefaces the description by saying that in any one year it absorbs so million sesterces of the Roman empire's wealth (i.ioi). While Pliny in 23
TOPOGRAPHIES
OF TASTE
the compendiousNaturalHistorydoes havea tendencytowardmoralizingcomment,it is strikingthata similarsentimentis repeatedin a differentcontext.Thus, while discussingspicessystematicallyaccordingto theirorigin,he has this to say aboutthe southernpartof the Arabianpeninsula,"HappyArabia"(ArabiaFelix): Butthe title "happy"belongsstill moreto theArabianSea,forfromit come the pearlswhichthatcountrysendsus. Andby the lowestreckoningIndia, Chinaandthe Arabianpeninsulatakefromour empireloo million sesterces everyyear-that is the sumwhichourluxuriesandourwomencost us. Forwhatfractionof theseimports,I askyou, nowgoesto the godsor to the powersof the lowerworld?(12.84)22 Hereit is importantto noticethatPlinyconflatesthreepartsof Asia,in a generalizingmanner,as sourcesof luxurygoods.23 But this conflationis not merelya matterof mentalmaps:in fact, it closely reflectsone aspectof ancienttrade.Cabotageor tramp-tradingfeatureslargein a majornewaccountof nonmodernMediterranean history.24 Thistradinginvolved the exchangeof goods en route-i.e., an ongoingprocessof purchase,sale, and exchangeat anygivenpoint on a commercialseavoyage,as opposedto long-distanceshippingthatinvolvesa singleoriginanda singledestination.As regardsthe monsoon route,this type of tradeis in factimpliedby the Periplus:its recurrent patternis one of selectivepurchaseand saleof specificgoods,at particularpoints en route. To take anotherexample,the monsoon route, as describedin the Periplus, extendsfromthe RedSeato Indiain onlyone of itsvectors:by another,it stretches down the east coast of Africa,perhapsas far as Zanzibar.Again,we are dealing herewith commoditiesthat areso distinctivelynon-Mediterranean thatthe differencesbetween the routes are easily minimized. From the first centuryA.D. Axum (ancientEthiopia)playeda majorrolein the monsoontrade,and fromthe fourthand fifth centuries,at the heightof Axumitepower,its traderseclipsedthe roleof thosefromRomanEgypt.25 The abovediscussionshould serveto underlinethe coexistenceof land and sea routes:thus, to judge from their materialand documentaryevidence,the port of Ostia tells us mostly about maritimetrafficwithin the Mediterranean, andPalmyraaboutoverlandtrade.As differentexcavationsvie forattention,there is a dangerthat one is recognizedat the expenseof the other.Evenif any "balance sheet"betweentradenetworksis impossibleto obtain,it is at least importantto grasptheirsimultaneousexistence.Certainlythe Berenikeexcavationhas receivedconsiderableattentionin both the popularand the academicpressesin 24
GRANT PARKER
recentyears.Compellingas its evidencemightbe, it shouldnot shift the discussion entirelytowardsearoutes. UrbanismandUrbanity If the foregoingdiscussionhas focusedon commoditiesand the routeswhereby they reachedthe Mediterranean,we can proceedto examinethe social patterns within which they became entwinedonce havingtraveled.How can the Mediterraneanconsumptionof SouthAsiangoodsbe mapped?In particular,to what extent arewe to regardconsumptionan urbanphenomenon?26 It will be necessaryto considerthesequestionsboth in the narrowersenserelatingspecificallyto SouthAsiangoodsandmorebroadlyrelatingto exoticgoods. First,in the narrowersense:to takeone of the commoditiesdiscussed,silkdoes not fitthe patternof urbanconsumption,judgingfromthe archaeologicalrecord. Certainlythereareliteraryreferencesindicatingits rolein the livesof metropolitan Romansand especiallyelitewomenof the city.Yetfindshaveclusteredmuch moreextensivelytowardthe fringesof the Romanempire,andby no meansonly to majorsettlements.TheseincludeBavariaon its Rhine-Danubefrontier,Kentin Britain,and Sarmatiantombsin Kerchand on the Volga.Someof thesesilk finds haveoccurredon points situatedon traderoutes,e.g., Arsinoeand Panopolisin Egyptas well as Palmyraand DuraEuropos.Thesetend to be fromthe thirdand fourthcenturies,andthusfromthe laterRomanempire.27 Onlysomeofthe settlementslistedaboveinvolvecities. Silkis, anyway,somethingof a specialcase,giventhatIndiawasnot its sole (or evenmain)point of origin.Beyondthat,the mainarchaeologicalevidenceforthe presenceof SouthAsiangoodsin the Mediterranean worldcomesfromBerenike. Certainlythis is true of cotton,now thatthereis clarityon Indianoriginsamong the findsdiscussedabove.If,on the basisof literarytexts,we mighthaveexpected the city of Rometo be the main consumer,this assumptionpresentsa problem. Romewas,afterall,the ancientMediterranean world's"consumercity"parexcellence;by the firstcenturyA.D. its populationmight havebeen arounda million people,makingit a majormegalopolisevenby modernstandards.28 Or does it presenta problemafter all? Self-evidently,many of the objects broughtfromSouthAsiawill havebeen moresubjectto decayin the moistersoil andairof Italythanthatofthe RedSeacoast.Fromthis pointof viewit is not surprisingthatRomeofferslittle evidencedirectlyrelevantto the presentdiscussion. Nonetheless,thereare some suggestionsaboutthe kind of evidencethat is now lost to us. Thecity of Romehadits own pepperhouses,the horreapiperatoriabuilt in A.D. 92 by the emperorDomitianon the northernslopeof the Palatinehill, this beingsomethirteenyearsafterthe elderPliny'sdeathin A.D. 79.29 Thereis no doubt 25
TOPOGRAPHIES
OF TASTE
thatsuchbuildingswereof considerableproportions.Further,the continuingstatus of pepperas a luxuryitem well into lateantiquityis underlinedby a remarkablestatistic:when in A.D. 408 the city of Romewas blockadedby the Visigoths underAlaric,the RomanSenateofferedhim, togetherwith 5,000 poundsof gold, 30,000 pounds of silver and other gifts, and no less than 3,000 pounds of pepper
subjectto his withdrawal,as we learnfrom the fifth-centuryhistorianZosimus (Historia nova 5.35-42). Such a huge quantity is hard to square with what would
otherwiseseemto be decliningproportionsin peppertradein the laterperiod.At the sametime, the enormityof this statisticis in itselfgroundsforsuspicion.Yet, even if Zosimusexaggeratesthe quantity,such a storyat least remindsus of the highprestigeattachedto pepper. Evidenceof this natureis fragmentaryand unreliablyanecdotal.But it does not detractfroma moregeneralizedimageof Romeas a consumercity.In classical studiesthe currencyof this term owesmuch to its use by MosesFinley,30 followingWernerSombartandMaxWeber:thatis, the citythatderivesits maintenance fromtaxesand rentsratherthan its own production,in whichrespectthe medievalcityprovidesa contrast.31 Indeed,the concept of a consumercity in antiquityhas been the sourceof muchcontroversyfollowingtheworkof Finley.32 Now it is perhapsrenderedmoot, given,for example,thatthe city of Romehas been shownto interactin manifold and complexwayswith the economiesof its hinterland.33 Still,we areleft with a veryconsiderablediscourseabouturbanlife,bywhichancientpeoplemademuch of Rome'sconsuminghabits.Itwasunderthe ruleof Augustusthatthe citybegan to earnthe statusof worldcapital,partlythroughthatemperor'sambitiousbuilding program.34 In the firstcenturyA.D. therewas a fine line betweenthe cultured eleganceof villas and gardens,on the one hand,and the conspicuousconsumption of a nouveauricheformerslavesuchasthe fictionalTrimalchioin Petronius's Satyrica.A moraltraditionin Romanthought,seen in socialcommentatorssuch as the youngerSeneca(4 B.C.-A.D. 65),laygreatstoreon sucha distinction,which parallelsthat betweenleisureand slothfulness.35 It is in this sense that I referto urbanity(Latinurbanitas)as a senseof sophistication,expressedin the mostobvious sensein relationto city life. It is in the firstcenturyA.D., followingthe end of civilwar,thatthe conceptof urbanitasis most clearlyvisiblein the Romanworld, the word now carryingwider connotations of sophisticationbeyond merely denotingthe city.36Urbanitasshouldbe seenas a discourseaboutthe citythatlays muchemphasison its consuminghabits. Constantinoplecan be consideredin the samelight,providingperhapsmore evidenceof a demandfor specificallySouth Asian goods.37When the emperor Constantinefoundedthis cityin A.D. 324 he intendedthatit shouldbecomea new 26
GRANT PARKER
Rome -a strategic and symbolic alternative to the older metropolis. This is a role it did in fact assume, particularly while the empire in the West suffered major political and military setbacks in the fifth century A.D.38 If the consumer city model persists, through a combination of ancient discourse and modern scholarship,39then there are two modest points to be made here, specific to the current topic. First, it should not blind us to practices of exotic consumption outside of urban settlements. Finds of peppercorns and silk in less urbanized, more remote parts of the empire are already a suggestion to this effect. More important, the Bay of Naples shows significant evidence of this life of luxury in, but not limited to, the cities of Pompeii and Herculaneum.40 The life of cultured leisure celebrated, for example, by the poet Statius (ca. A.D. 45-96) in Silvae 2.2, is situated far away from the hustle and bustle of urban life, at a villa in Sorrento.4'Such a place presented an ideal setting for Epicurean thought, which centered on the capacity of philosophy to bring freedom from anxiety. These lavish villa and garden settings are a reminder that consumption was by no means limited to cities. Second, the evidence for specifically exotic commodities points to the phenomenon that we might, for current purposes, more securely call transit cities. Points along the routes discussed above show so much evidence of consumption that they reveal a close connection between the practices of distribution and consumption. Implicit in this insight is a warning to avoid falling into hoary notions of center and periphery in assessing such networks. This observation is particularly applicable, both in general and specifically, to Berenike and Palmyra. Little was known about Berenike before the excavations of the past decade;in a relatively short time, a picture has emerged of a settlement that functioned not merely as an entrepot but as a settlement that harbored wealth and made the life of consumption possible. It would be wrong, therefore, to consider Berenike a mere conduit. The larger archaeological context now emerging makes it clear that some of the objects found there would have been used on-site ratherthan automatically transported to the Mediterranean proper.42And furthermore, though excavations at Ostia Antica have thus far failed to turn up anything distinctively South Asian, the site does offer many different indications of wealth in the private and public spheres, the kind of wealth associated with luxury consumption.43After Rome's conquest of the Carthaginians in 146 B.C., its maritime activity increased, and the city grew, reaching its height under the emperor Trajan. In sum, it is certainly interesting and significant that silk shows widespread diffusion, suggesting a many-centered "Roman Mediterranean"with several points of consumption. There can be no denying the power of Rome, and later Constantinople, as centers of consumption, but the comments above offer two modifica27
TOPOGRAPHIES
OF TASTE
tions to the model.First,the Bayof Naplesgoesbeyondanynarrowdefinitionof urbanism:it is herethatwehavemuchevidenceof the life of highconsumptionpartsof a lifestylethat mightbe havebeen consideredleisure(otium)by its participantsor decadence(luxuria) byits critics.Second,Berenikenowappearsfrom excavationsto havebeen not merelya placefor the transmissionof commodities butalso a placefortheirconsumption,evenif ata modestlevel. Foreshorteningthe East Wehaveconsideredabovesomeof the spatialdimensionsof Mediterraneanconsumption,as practicedby Romans.Butit is anothermatterentirelyto adumbrate the mentalmapsengenderedby the objectsthemselveswith regardto their supposedorigins.Wheredidtheseexoticgoodsseem to comefrom,andhowdidtheir supposedplacesof originrelateto Romanworldviews? Letus begin at a transitcity closeto Rome.44One of the most remarkablesectorsatOstiaAnticais a largeforumof the tradeguilds,the PiazzaledelleCorporazioni. It is in the form of a doublecolonnade,locateddirectlybehind a theater. Bothwerebuiltin the time of Augustus;bothwerepublicspacesin the heartof the city.Leadingoff the colonnadeweresixty-onesmallrooms.In frontof mostwere mosaicsindicatingthe occupationor originof their incumbents,eithervisually or verballyor both. The fascinatingaspectof the Piazzale,for currentpurposes, is the visual representationof the tradingguilds on the largemosaic floor that survivestoday.45Thereis no Iberianelementin evidence,butthis maybe because only half of the mosaicssurvivetoday.Includedin the survivingmosaicsarerepresentationsof CarthageandAlexandria,andof ArelateandNarboin Gaul.Thus diversepartsof the Mediterraneanare represented-in whatwe can even call a kind of map,giventhatthe floorrepresentsgeographicalspace.Presumably, merchantsusinga particularroomwereengagedin tradewiththe Mediterranean port indicatedon the mosaic,orperhapsoriginatedfromthatplace. Seenin this light, the mosaicsare emblemsof Mediterranean"connectivity," to use a key term from the majornew studyby PeregrineHordenand Nicholas Purcell.46 It wouldbe reasonableto think thatit tells us moreaboutthe distribution than the originsof commodities,eventhoughit mightstill addto a senseof the exotic. Significantly,this map of distributionis articulatedat a placethat is closeto a majorcenterof consumption,namelythe cityof Rome. Wewouldbe hardpressedto find an equivalentmap of the originsof luxury goods in the ancientMediterranean.It is not clearthat any such representation exists. Whatdoes survivein profusionis geographicalwritingfrom,likewise,a high point of Mediterraneanconnectivity,namelythepax Romana thatfollowed Octavian/Augustus's conquestof hislastremainingrivalsin 31B.C. Mostobviously, 28
GRANT PARKER
thereis the lengthysurveyof the "inhabitedworld"(oikoumene)by a geographer and historianat the time of Augustus,Strabo(64 B.c.-after A.D. 24). Within his systematiccoverageof differentpartsof the world,theirtopographiesand indigenous customs,Strabohas little to say aboutcommoditiesassociatedwith particularareas.Certainlyhis commentsaboutIndiancommoditiesare extremely limited comparedwith the nearlyexclusivefocus on them in the Periplus. 47Yet there are indicationsthat merchantsplying the monsoon trade,not unlike the authorof the Periplus,providedmuchof Strabo'sinformationforhis description of India,particularlythe eastcoastof India.48 was the equallyextensiveNatuSeveraldecadeslaterthan Strabo'sGeography ralHistoryof the elderPliny(brieflydiscussedabove),with its compendiousview of nature(Natura).ThoughPliny'sdebtto earlierwriters,especiallythose of the Hellenisticperiod,is obviousandsignificant,his textis verymucha productof its times, and its subtextof Roman"worldempire"impossibleto overlook.49 It is in this laterworkthatwefindbothsystematicgeographyanda consciousnessof commodities.Sucha combinationis not surprising,giventhe encyclopedicscope of Pliny'swork,withinwhichdifferenttopicsoccurin differentbooks.Thephenomenon of encodingcommoditiesspatially,e.g.,"Gangeticnard"in Peripluschapter 63,aswe havenotedabove,is by no meanslimitedto Plinyor evento geographical writers.In the caseof exoticgoods,it is profuselyevidencedin the Deipnosophists of Athenaeusof Naucratis,who flourishedaroundA.D. 200. Thisis a worksupposedly representingdinnerpartyconversation;it centersmostly on literarytopics butinvolvesmuchdiscussionaboutthe wherewithalof ancientdinneranddrinking parties-e.g., drinkingvessels.Farbeyondits model of the symposionfrom Plato'sclassicalAthens,it is thusa celebrationof exoticconsumption.50 In the Greco-Romandiscourseaboutluxurygoods,variouspartsof western Asiarecurwith frequency,both as the sourceof luxuriesand as locationsof consumption.Certainlythis discourseis an importantpartof classicalageAthenian representationsof the Achaemenidworld.5'Suchan ethnographicprofileof the Iranianworld continues into later periods of the Greco-Romanworld.52 Even if this observationseems at firstlike a literarycommonplace,there is a serious point here. Partof SouthAsia that was known as "India,"i.e., the Indusvalley, was a satrapy(province)of the Achaemenidempire,as reorganizedby DariusI in 522-521 B.c. (HerodotusHistories 3.96-106). The courts of the Achaemenid kingsthemselvescollectedluxurycommoditiesand also providedcentersforthe exchangeof geographicalinformation.Thus,if the spatialencodingof goodswas a commonphenomenonin ancientMediterranean societies,thenit is no accident that Greeksand Romansassociatedboth Persiaand India with luxury:that is, Indiaas an ultimatesourcebut Persia,i.e., the Achaemenidcourt,as a centerof 29
TOPOGRAPHIES
OF TASTE
luxuriousconsumption.53 By this reckoning,the centralizingand hoardingtendenciesof the Achaemenidcourtprovidedone factorin Rome'sgeneralizingview of luxuriesfromthe East. Buttherearealso otherfactorsthat madewesternAsiaand the northwestern IndianOceanlittoralsubjectto this processof foreshorteningin Greco-Roman thought. Most obviously,severalWest Asian lands offeredthe Mediterranean eitherthe sameor similarcommodities.Thusthe aromaticsassociatedwith Arabiaseemednot radicallydifferentfromSouthAsianspices.Forexample,Plinythe elderfrequentlymentionsthe two groupstogetherin his discussionof spicesand aromaticsin his Natural History 12.26-50.54 In the case of spices,we might well askwhetherRomanswouldlink theseexoticproductswith anyparticularlandor with the tropicsin general. Further,these same or similargoods are being transportedalong the same routes.If we allow,with the Periplusas well as Hordenand Purcell,that tramptradingwas the predominantmeans of exchange,then we should all the more expectsomegeneralizationin Romanminds.Specifically,theArabianandIndian peninsulasboth formedpart of the monsoon route,and both producedexotic commoditiesforconsumptionmainlyin the Mediterranean. Finally,we see direct comparisonbetweenIndia and Egypt (or Ethiopia)at severalpointsin Strabo'sextensivedescriptionof India,forexample,in his Geography15.1.13C69othe deltasof the Indusandthe Nile,the animalslivingthere,and the physicalappearanceof the inhabitants.In this passageStrabocomparesEthiopia with Indiaas well, i.e., both the upperand lowerNile valleywith the Indus valley.55He is herepresentingthe lessfamiliarwith referenceto the morefamiliar. This is a case of translationin the wider,culturallylocatedsense of that term.56 Thereis no doubtthatStrabo'sreaderswouldhavebeenmorefamiliarwith Egypt, directlyadjoiningthe Mediterraneanand the breadbasketof the Romanempire. Butatthe sametime this comparisonreflectsthe routesthatbroughtIndiancommoditiesto the Mediterraneanandespeciallyto the city of Rome.Thelastpartof the monsoonroutefromIndiato Romewould havecoincidedwith the shipping routesforRome'scornsupply.57 In sum, Romangeographicthinkingevidencesan overlapbetweenIndiaand otherlandsof the northwesternIndianOcean.Variousreasonscan be advanced for this overlap,includingthe common (tropical)climatethat differsfrom the Mediterranean climate.Thusit is thatspicesareby definitionnon-Mediterranean commodities,comingas they do fromthe tropics;herbs,on the otherhand,were well known in the Mediterraneanworldfrom an earlierstage.Here it is worth emphasizingthatthe monsoonroute(evenmorethan the land route)unitesdifferentlands.Fromthe pointof the Romanconsumer,it matteredlittlefromwhere 30
GRANT PARKER
exactly a particular object came; what did matter is that it came from afar;what is more, it is western Asia that was the major source of luxury goods. We have here what initially seems like a paradox: there was a strong tendency to mark commodities spatially in the Roman world,58 and, at the same time, a foreshortening of the East whereby different parts of the monsoon route were conflated from the perspectives of Mediterranean peoples. In the absence of a distinctive notion of East and Southeast Asia in Roman topography,59 we might well describe India as Rome's FarEast. However, this is not really a paradox at all, if we bear in mind that we are speaking less about formal, scientific geographies (notably the Geographyof Claudius Ptolemy, second century A.D., with its extensive list of coordinates) than about popular ideas concerning the extent of the world. And it is in this latter, everyday sense that commodities had such power to connote.
TheUniversal,the Particularandthe Exotic Two kinds of universalism emerge in JacquesAndre and Jean Filliozat'sL'Indevue de Rome, the major book synthesizing Roman visions of South Asia.60 The first of these visions involves Alexander, whose campaign, which took his troops to modern Pakistan in 326 B.C., stretched as far east as the Indus valley. Though the campaign did not stretch far west of the Aegean, it was presented to contemporaries as "world conquest." That it might have been-within particular worldviews that celebratedAlexander as a hero. The image owes much to Alexander'sown manipulation of a public image: he presented himself, by turns, as the god Dionysus and as the hero Heracles, both of these being linked with the edges of the earth (and with the East especially, in the case of Dionysus).6' It is this imperial brand of universalism that underlies much of the political discourse of the Roman empire. Though Alexander himself was to remain a controversialfigure, a conception of world empire outlived him and was in many cases associated with him.62 At the level of philosophy, some of the major branches of Hellenistic (i.e., post-Classical) thought are discernibly the products of an Aegean world very different from that of the classical city-state (polis). Both Stoicism and Epicureanism have been described as "refugeephilosophies" in their political disengagement, particularly in their resignation at the curtailment of citizen participation in public life.63A particular kind of universalism was central to Stoicism, and it is the Stoic thought of Posidonius (ca. 135 B.c.-ca. 50 B.C.) that informed Strabo'sconception of the world.64Stoicism was also a major influence on the elder Pliny.65 The second universalism identified by Andre and Filliozat is in the realm of Christianity. The idea that salvation is open to all souls, at least potentially, was (and in some parts of the world still is) a motor-force behind Christian mission31
TOPOGRAPHIES
OF TASTE
aryactivity.66 Severalrelatedelementsmaybe identifiedhere:first,God'spoweras lordof all, a conceptgoingbackto Judaism.67 In ByzantineChristianitythe term Pantokrat6r (Lordof All) wasone of the mostimportantepithetsof God.Further, therewas Jesus'sinjunctionto the disciplesto bringthe good newsto all people throughoutthe world.68 Underlyingthesetermsandtextsis the ideathatsalvation is open to all who believe.It is these relatedsensesof universalismthat spurred Paul'smissionarytravelsin variouspartsof the Mediterraneanand underliesthe theologyof his letters,preservedin the Christianscriptures. Withinthe capacioustopicof the perceptionand interpretationof SouthAsia by Mediterraneansocieties,such an analysisis undoubtedlyapt. India,being on the "edgesof the earth"was a benchmarkof imperialconquest,and thus important to the self-presentationof emperorsbeginning with Augustus.69Within Christianitythe storyof missionconcernsthe apostleThomas.Accordingto the apocryphalActsof Thomas(aSyriactext datingto the latesecondcenturyA.D.), it is Thomaswho wassentto bringthe good newsto India,a journeyhe undertakes onlywith reluctanceandone thatled to a martyr'sdeath.Thememoryof Thomas continuesto playan importantrolein the Christianityof southernIndiato this day.70In both RomanimperialismandChristianity,Indiathuspresentedthe ultimatetest of success:if Romanpoweror Christianityprevailedevenin India,then it couldclaimto be a genuinelyworldwidephenomenon. ButAndreand Filliozat'saccountof universalismsis not necessarilycomprehensive.The discussionin this articlewouldsuggesta thirduniversalism:thatof commodities.It mightinitiallyseemsurprisingto viewcommoditiesin this light, but I believethereis a caseto be madefora "global"senseof commoditiesin the ancientMediterranean.As we have seen, Athenaeus,more so even than Strabo and Pliny,presentsone objectafteranotherwith referenceto its supposedpoint of origin.Thetrafficin long-distancecommoditiesto RomeandotherMediterraneancenterscreateda certainkindof geographyfortheirconsumers,a universalism realizedatthesepointsof consumption. It mightbe most apt,then, to describethis universalismas a universalismof the exotic:the sensethatexoticobjects,whatevertheiractualorigins,expressthe extent of the world. By "exotic"here I mean radicalotherness,particularlyan othernessthatis materiallyexpressedandtherebyopen to exchange.In a Roman worldthat witnesseda greatertrafficin luxuryitems than everbefore,this phenomenon reachednew levels.Paradoxically,this trafficconstitutesa universalism of particulars,based on particularpoints of origin, as they are imagined. But it does deservethe designationof a universalismbecauseit involvespoints of articulation,namelythe sitesof consumptiondiscussedabove.This is another wayof sayingthatcommoditiesrepresenteda wayforRomansto think aboutthe 32
GRANT PARKER
extent of the "inhabitedworld"(orbis terrarum),largelyremovedfrom the scientific geography,a kind of populargeographythat continuedto be made and remadein ongoingpatternsof socialpractice.I havesuggestedthatthis phenomenonappliedespeciallyto the northwestIndianOceanzone,with Indiabeingpart of this zone and sometimeseven representingit by synecdoche,the principleof part-for-whole. The exoticaspectof Mediterraneanconsumptionmight seem on the surface to be a highlyparticularisticphenomenon,especiallygiventhe frequentstipulations of origin in texts such as Athenaeusor Pliny or the Periplus. But if it is in factthe othernessof objects,morethan anythingelse,that confersexoticstatus, then a broaderview of provenanceis appropriate.Forit is an aggregateof these particularismsthatalloweda totalizingview of the "inhabitedworld"(oikoumene or orbisterrarum)in a text suchas Pliny'sNatural History.Indiahadan important placewithinthis compositeexotic.Aswe haveseenmostclearlyin relationto silk, therewasa certaindegreeof fudgingbetweenvariouszonesof originin the larger Indian Ocean basin. But it is the Piazzaledelle Corporazionithat points most clearlyto a universalismof the exoticduringRome'simperialage,suggesting,as it does, thatvaluederivedas muchfromthe verydynamicsof Mediterraneanconnectivityas fromanyspecificprovenance.
33
TOPOGRAPHIES
OF TASTE
NOTES
I should like to thank Alka Patel,Emily Mackil,the anonymous readersof ArsOrientalis and, not least, the conferenceparticipants for their comments and suggestions. 1. JohnMoreland,"Productionand
Exchangein HistoricalArchaeology,"in CompanionEncyclopediaofArchaeology, ed. GraemeBarker,2 vols. (London: Routledge,1999), 637. 2. It is in this special sense that I havemade bold to use the term "RomanIndia"in the largerresearchprojectfrom which this articlederives:this term constitutes an oxymoron and provocation,given that the Romanempireneverdid gain political or military control over South Asia, howevervividly "India"might have featuredin Romanimperialimaginations. The currentessaycoverssome of the same ground as GrantParker,"Ex OrienteLuxuria:Indian Commodities and RomanExperience,"Journalof the Economicand SocialHistoryof the Orient 45.1(2002): 40-95, but with significantly differentemphases.Of course there is an inherent ethnocentrism in such a history of representations.Fora different approachto some of the same material, see RomilaThapar,"IndianViews of Europe:Representationof the Yavanasin EarlyIndian History,"in CulturalPasts (Delhi: OxfordUniversityPress,2002), 536-55. 3. Donald S. Whitcomb and JanetH. Johnson,ed., Quseiral-Qadim 1978: PreliminaryReport(Cairo:American ResearchCenterin Egypt, 1979); Whitcomb and Johnson,QuseiralQadim1980:PreliminaryReport, American ResearchCenterin Egypt Report7 (Malibu:Undena, 1982); Gillian Vogelsang-Eastwood,ResistDyedTextiles from Quseiral-Qadim,Egypt(Paris: Association pour l'Etudeet la Documentation des Textilesd'Asie,1ggo); Carol 34
GRANT PARKER
Meyer,GlassfromQuseiral-Qadim and theIndian OceanTrade(Chicago: OrientalInstitute, 1992). 4. StevenE. Sidebothamand Willemina Z. Wendrich,ed., Berenike1994:Preliminary Reportof the1994Excavationsat Berenike (EgyptianRedSea Coast)and theSurvey of theEasternDesert(Leiden:Research School CNWS,1995);Berenike1995 (1996); Berenike1996(1998);Berenike1997 (1999);Berenike1998 (2000), and continuing. Fora recentsynthesis, see W. Z. Wendrich,R. S. Tomber,S. E. Sidebotham,J.A. Harrell,R. T. J. Cappers,and R. S. Bagnall,"Berenike Crossroads:The Integrationof Information,"Journalof theEconomicand Social Historyof theOrient46.1(2003):46-87. 5. A. M. I. VanWaverenand W. Z. Wendrich, "Textiles,"in Berenike1994,63- 68. 6. J.P.Wild and F.C. Wild, "TheTextiles," in Berenike1996, 236.S/S and Z/Z threads arenamed for the direction of weaving: the names match the slope at the centerof each letter.See further E. J.W. Barber, PrehistoricTextiles:TheDevelopmentof Clothin theNeolithicand BronzeAges with SpecialReferenceto theAegean (Princeton:PrincetonUniversityPress, 1ggo),66. 7. J.P.Wild and F.C. Wild, in Berenike1997, 272,with considerablecaution in inferringprovenience.A. J.Veldmeijer, "TheCordage,"in Berenike1997, 267, relatessuch evidence to the question of when cotton was firstcultivatedin Egypt. 8. E. J.W. Barber,PrehistoricTextiles,32-33. 9. F. C. Wild and J.P.Wild, "Sailsfrom the Roman Port at Berenike,Egypt," Internationaljournalof Nautical Archaeology30.2(2001): 211-20. lo. "India,"i.e., the Indusvalley now mostly in Pakistan,arises in Herodotus's account of the PersianWarsbecause it was a satrapy(province) of the Achaemenid empireset up by Darius and thus
forms part of the backgroundto his main narrative.Herodotus'emphasison the naturalwealth of "India,"and thus of Darius'sempire, is no coincidence. ii. LionelCasson, ed., ThePeriplusMaris Erythraei(Princeton:Princeton UniversityPress,1989).Foran attemptto readthe workin light of the objectsand routesinvolved, see GrantParker, "PorousConnections:The Mediterranean and the Red Sea,"ThesisEleven67 (2001): 59-79. 12. Since the word dermatadenotes the skins of animals, the adjectivesirikacannot here be translatedas "silk." 13. See, in general,Barber,Prehistoric Textiles,30-32, withfurtherreferences. The single best overviewof spices in the 14. Roman empireis J.Innis Miller, The RomanSpiceTrade(Oxford:Clarendon, 1969).However,the certaintywith which the authoridentifies particularspices from ancient literaryreferencesadds a note of doubt about its reliability.Miller's controversialpresentationof the
i8. Jean-FrancoisSalles,"TheArab-Persian Gulf under the Seleucids,"in Hellenismin theEast,ed. Amelie Kuhrtand Susan Sherwin-White(Berkeley:Universityof California Press, 1987), 75-108. For the
19.
"cinnamonroute"(153-72) nowreceives support from AndrewDalby,Dangerous Tastes:TheStoryof Spices(Berkeley: Universityof CaliforniaPress,2000), 3641.Cinnamon and cassia (sometimes identified as a separatespice) had found their way to the easternMediterranean
126-28.
22.
23.
24.
1998),119-37. For a trenchant analysis of
the supposed Silk Roadas a "modern fabrication"-after all, the term was coined only in the mid-nineteenth century- see WarwickBall,Romein the East:The Transformation ofan Empire (London:Routledge,2000),138: "Any
bytheseventhcenturyB.C.; they originateboth in SriLankaand in the northern parts of SoutheastAsia. 15. Foranother example, compareHorace, Satires2.8.14-15. 16. BhagwantSahai,IconographyofMinor Hindu and BuddhistDeities(New Delhi: AbhinavPublications,1975)157-79.The statuette is in the collections of the Museo ArcheologicoNazionale in Naples. 17. R. M. Cimino, AncientRomeand India (Delhi: MunshiramManoharlal,1994),
long view, see now especiallyHimanshu PrabhaRay,TheArchaeologyofSeafaring in AncientSouthAsia (Cambridge: CambridgeUniversityPress,2003). Indeed, the question of the Silk Roadmay seem a nonstarter,given the degreeof anachronismand even of romantic fantasyinvolved in positing a single transcontinentalroad. In its extreme form this idea of a single roadcannot be maintained. However,it is equally importantto insist that there was overlandtransportof silk, one of the productsunder discussion, and for that reasondeservesconsiderationhere. For one, FergusMillar sounds a skeptical note: "CaravanCities:The RomanNear Eastand Long-DistanceTradeby Land," in ModusOperandi:Essaysin Honourof GeoffreyRickman,ed. MichelAustin, Jill Harries,and ChristopherSmith (London:Institute of ClassicalStudies,
25.
1969), 175-271. 26.
ancient 'route' ... was at best simply a
20
broadchannel of communications across a region for the movement of people, goods and ideas. It rarelyfollowed any fixed pathway.It was nevertranscontinental." Modern Kavunlu,formerlyBelkis. R. J. A. Talbert,ed., TheBarringtonAtlas of the Greekand RomanWorld(Princeton: Princeton UniversityPress,2000), map 67, F2.
21.
35
Fora particularlyimportant document TOPOGRAPHIES
OF TASTE
137,showing how the city'swealth was relatedto tradeactivity passingthrough it, see J.F.Matthews,"TheTaxLawof Palmyra:Evidencefor Economic History in a City of the RomanEast,"Journalof RomanStudies74 (1984): 157-80. On the link between Palmyraand South Asia, see the telltale aside in Appian'saccount of the civil wars,while narratingMarcus Antonius'sunsuccessful attackon that city: "for,being traders,they bring Indian and Arabianobjectsfrom Persiaand get rid of them in Roman territory"(Roman History5.9). Pliny,NaturalHistory,vol. 4, bks. 12-16, trans. H. Rackam,rev.ed. (Cambridge, Mass.:HarvardUniversityPress,1986). BelowI shall delve more deeplyinto the worldviewimplied here. It should be noted that the Latinadjectivefelixis most obviously translatedas "happy"but also carriesstrong connotations of wealth, as in the Pliny passagequoted. Comparen. 47below for the Greekequivalent,with its similar semantic field. PeregrineHordenand Nicholas Purcell, TheCorruptingSea:A StudyofMediterraneanHistory(Oxford:Blackwell,2000), e.g., 140, 142. N. V. Pigulewskaja,Byzanzaufden Wegen nachIndien (Berlin:Akademie-Verlag, AD
27
Certainly,to considera majortheorist of a millennium later,the consumption of luxury goods is centralto Ibn Khaldun's account of what constitutes cities and makes them attractiveto nomadic people:e.g., 1.224 and 1.276 in Franz Rosenthal,trans., IbnKhaldun,The Muqaddimah:An Introductionto History, 3 vols. (New York:Bollingen, 1958), 1:25253,308-9, respectively. ManfredG. Raschke,"NewStudiesin Roman Commercewith the East," AufstiegundNiedergangderromischen Welt2.9.2 (1978): 625, 727-28.
28. PeterGarnseyand RichardSaller,The RomanEmpire:Economy,Societyand Culture(Cambridge:Cambridge
Press, 1996-2000),10:782-811;
35. J.P.Toner,LeisureandAncientRome
University Press, 1987), 6. 29.
and
"Rome and Italy," in ibid., 11:405-43. (Cambridge: Polity, 1995).
LawrenceJ.RichardsonJr.,A New Topographical DictionaryofAncientRome (Baltimore:JohnsHopkins University
36. Edwin S. Ramage,Urbanitas:Ancient Sophisticationand Refinement(Norman,
Press, 1992), 194-95.
37. Pigulewskaja,Byzanz. 38. Certainlythis is the romanticimage of the city, as expressedin W. B. Yeats's
30. Moses Finley,TheAncientEconomy,2nd ed. (Berkeley:Universityof California
Okla. 1973), 112.
poem, "Sailing to Byzantium" (1908).
Press, 1976).
31. Garnseyand Saller,RomanEmpire, 48-49. 32. Fora useful reviewof approaches, ultimatelyreaffirmingthe consumer model as the best available,see C. R. Whittaker,"Do Theoriesof the Ancient City Matter?"in UrbanSocietyin Roman Italy,ed. T. J.Cornell and KathrynLomas (London: Routledge, 1995), 9-26.
For
attemptsto find alternatives,see Helen M. Parkins,ed., RomanUrbanism: Beyondthe ConsumerCity (London: Routledge, 1997). At stake, among other
things, is a minimalist approachto ancient economic practice.Thus, accordingto the consumer model, "the cities of the ancientworld dependedon the exploitation of their hinterlands through rents and taxes for their sustenance and ... aimed for a large
measureof regionalself-sufficiency.The corollaryof this ... was that therewas little need for,or actual developmentof, urbanmanufacturingand inter-regional trade."David J.Mattingly,"Beyond Belief?Drawinga Linebeneath the Consumer City,"in Parkins,Roman Urbanism,210. 33. Hordenand Purcell, CorruptingSea, 105-8.
34. Nicholas Purcell, "Romeand Its Development underAugustus and His Successors,"in CambridgeAncient History,ed. A. K. Bowman et al., 2nd ed. (Cambridge:CambridgeUniversity
39. In tracinglinks between ancient and modern, RayLaurence,"Writingthe RomanMetropolis,"in Parkins,Roman Urbanism,1-20, interestinglycompares the dystopianimage of Rome in the Latin satiristJuvenal(Satires3) with that in LewisMumford, TheCityin History:Its Origins,Its Transformations, andIts Prospects(New York:Harcourt,Brace and World, 1961), e.g.,
233.
40. JohnH. D'Arms,Romanson theBay of Naples:ASocialand CulturalStudyof the Villasand theirOwnersfromi5oBC toAD 400 (Cambridge,Mass.:Harvard University Press, 1970); Paul Zanker,
Pompeii:Publicand PrivateLife,trans. D. L. Schneider(Cambridge,Mass.: Harvard University Press, 1998).
41. R. G. M. Nisbet, "Felicitasat Surrentum (StatiusSilvae11.2),"JournalofRoman Studies 68 (1978): 1-11.
Thus, for example,Wild and Wild, in "TheTextiles"in Berenike1996,discussing the use of textiles on-site: "Intheir homes Spartanfurnitureand rough finished walls would havebeen disguised under an astonishing rangeof high-class coverlets,hangings and curtains. Berenikians,even in the fifth centuryAD, had high expectations"(236). 43. RussellMeiggs,RomanOstia,2nd ed. (Oxford:Clarendon,1973). 44. Fora recentreassessmentof Ostia Antica, see esp. Nicholas Purcell,"ThePortsof Rome:Evolutionof a FacadeMaritime," 42.
36
GRANT PARKER
in "RomanOstia"Revisited,ed. Anna GallinaZevi and Amanda Claridge (London:BritishSchool at Rome, 1996), 267-79. He arguesagainstdescribingit as "theport of Rome"and appliesinstead the idea of a "faqademaritime,"involving an "interlockingand complexwhole" (279).As the volume title suggests, Meiggs'sRomanOstiaremains fundamental; see 283-88, 449-51 on the
Piazzaledelle Corporazioni. 45. Approximatelyhalf of the mosaics survive. See further,KatherineM. D. Dunbabin,Mosaicsof the Greekand RomanWorld(Cambridge:Cambridge University Press, 1999), 60-65.
46. Hordenand Purcell,CorruptingSea, 123-71. 47. Note Strabo Geography 15.1.13C69o, 15.1.18C692, and 15.1.30C700, which I
regardas exceptions to this trend. Especiallyworthy of note are 15.1.18C692, a generalcomment on its fertility (eudaimonia,literally"wealth"),and 15.1.69 C718 on rival displays of wealth.
48. "Asregardsthe merchantswho currently sail from Egyptby the Nile as faras India, only a small number havesailed right up to the Ganges,and even these aremerely privatecitizens and not useful as regards the account [in Greek,historian]of the placesthey haveseen' (StraboGeography 15.1.4C686).This striking passagereflects Greekand Romanprejudiceagainst traders,especiallyas Strabogoes on to describeembassiesreceivedby Augustus from India that are,by implied contrast, acceptablesources of topographical knowledgebecausethey are relatedto the emperor.In the more mathematicaland less discursiveGeographyof Claudius Ptolemy (ca. AD 1oo-ca. 178), written
about a century after Strabo,it is clear that commercialtravelershavegenerated much of the information, not least for the Bayof Bengal.
49. A. Wallace-Hadrill,"TheElderPliny and Man'sUnnaturalHistory,"Greeceand Rome37 (1990): 80-96.
50. David Braundand JohnWilkes, ed., Athenaeusand His World:ReadingGreek Culturein theRomanEmpire(Exeter: University of Exeter Press,
62. Diana Spencer,TheRomanAlexander: Readinga CulturalMyth (Exeter: Universityof ExeterPress,2002). 63. MarthaC. Nussbaum, The Therapyof Desire:TheoryandPracticein Hellenistic Ethics(Princeton:Princeton University Press, 1994).
2000).
51. Edith Hall, InventingtheBarbarian: GreekSelf-DefinitionthroughTragedy (Oxford:Clarendon,1989);Christopher Tuplin,AchaemenidStudies(Stuttgart:
64. KatherineClarke,BetweenGeography and History:HellenisticConstructionsof theRomanWorld(Oxford:Clarendon,
Steiner, 1996), 132-77. 52. Horace Odes 1.38.1.In this regard, R. G.
65. Wallace-Hadrill,"ElderPliny";Mary Beagon,RomanNature:The Thoughtof Pliny theElder(Oxford:Clarendon,1992). 66. Denise KimberBuell, "Raceand Universalismin earlyChristianity," JournalofEarlyChristianStudies10.4
M. Nisbet and MargaretHubbard compareother examplesof eastern (and especiallyPersian)luxury.A Commentaryon Horace:OdesBook1(Oxford: ClarendonPress,1970), 423-24. 53. Amelie Kuhrt, TheAncientNear East,c. 3000-330 BC, 2 voIs. (London:Routledge, 1995), 648.
54. In the case of many spices Pliny confuses source and entrepot. Miller,RomanSpice Trade,20-21. 55. See also StraboGeography15.1.16C691, invoking Herodotus'sdescriptionof Egypt alongsidethat of Alexander's general Nearchus on India, and 15.1.18-19
1999).
(2002): 429-68.
67. Psalms 24:1-2: "Theearth is the Lord's and all that is in it; for he has founded it upon the seas, and establishedit on the rivers"(New RevisedStandardVersion). 68. Matthew29:19-20: "Gothereforeand make disciples of all nations, baptizing them in the name of the Fatherand of the Son and of the Holy Spirit,and teaching them to obey everythingthat I have commanded you."CompareMark16:15; Luke 24:47. And see Acts i:8: "you will be
C692-693.
56. JamesClifford,Routes:Traveland Translationin theLateTwentiethCentury (Cambridge,Mass.:HarvardUniversity Press, 1997).
57. G. E. Rickman, TheCornSupplyof AncientRome(Oxford:Clarendon,1980). 58. On this phenomenon, see further Parker, "Exorienteluxuria,"78-87. 59. JohnFerguson,"Chinaand Rome," AufstiegundNiedergangderromischen
my witnesses in Jerusalem,in all Judaea and Samaria,and to all the ends of the earth" (all New RevisedStandard Version). 69. ResGestae34. See also the emperorTrajan in Dio Cassius Roman History 68.29.1.
70. See, e.g., PaulThomas, Christiansand Christianityin India and Pakistan (London: Allen and Unwin, 1954).
Welt 2.9.2 (1978): 581-603.
60. JacquesAndre and JeanFilliozat,L'Inde vuedeRome(Paris:BellesLettres,1986),11. 6i. Andrew F.Stewart,Facesof Power: Alexander'sImageand HellenisticPolitics (Berkeley: University of California Press, 1993).
37
TOPOGRAPHIES
OF TASTE
HIMANSHU
PRABHA
RAY
THE BEGINNINGS TheArtisanand theMerchantin EarlyGujarat,Sixth-EleventhCenturies
Abstract This articleaims at locatingtradingactivitiesin Gujaratprimarilythroughan analysisof archaeologicaldata,inscriptions,and textual sources.The continuity between the early and medievalperiods is essentialto understandingthe underlyingstructuresin terms of the hierarchyof routes, utilization of local resources,and organizationof watertransportin the region.The first section of the articlehighlightsdiversetrendsin the study of northwesternIndia,particularlyin the context of Indian Ocean sailing. The primaryfocus in the latter part of the articleis the period from the sixth to eleventhcenturies,which marksan expansionof contactswith the PersianGulf and other parts of the IndianOceanregion.In traditionalhistoricalwritingthe emphasishas been on agrarianexpansion and the spreadof Sanskriticculture through movements of brahmanasat the behest of the emergingregionalstates and the consolidation of a feudal order.In contrast,this article shifts the emphasisfrom the centralityof the state and addressesissues relatingto the organizationof craft activities,interactionbetweenthe artisanand tradinggroups,and diverseconsumptionpatterns.The articlealso arguesfor autonomyof religioustraditions, apartfrom the state,and highlightsthe complexreligiouslandscapeof Gujarat, thereby contestingthe somewhatmisplacedemphasison Brahmanizationor Sanskritizationof the region.
IT IS GENERALLY BELIEVED that from the seventhto the fifteenthcenturies, the Indian Ocean had become a "Muslimlake,"Buddhismhaving declinedin the Indian subcontinentand Hinduism having become inwardlooking.'This view persistsdespitethe fact that earlynarrativeliteraturein Sanskrit,Prakrit, Pali,Tamil,and othervernacularscontainsglowingaccountsof maritimetravel by merchants,craftspeople,musicians,and others.In addition,the diversityof evidenceprovidedby a rangeof archaeologicaland architecturalsourceswhen collatedfurnishesfascinatingvignettesforan understandingof the maritimeorientationof communitiessettledin northwesternIndia. This articleattemptsto drawout evidencefrom archaeologicalfinds, architecturalremains,and earlyinscriptionsto presentthe earlyhistory of trading communitiesin northwesternIndia.As will be discussedin greaterdetailbelow, archaeologicalfinds have generallybeen used to supportpreconceivedtheories presentedby historians,while religiousstructureshave been studied either in termsof styleand chronologyor as objectsof royalmunificenceprovidinglegitimizationto the politicalelite.In contrast,this articleexaminesreligiousshrines within theirsocialcontextas indicatorsof communityidentity,includingthat of 39
tradinggroups.2The picturethat emergesis that of a diversereligiouslandscape patronizedbyvariedcommunities. Traditionally, threebroadtrendshaveinfluencedhistoricalstudieson ancient Indiain the postindependenceperiod.Oneis the Marxistschool,whichsupports the Indianfeudalismtheory.3The paradigmopposingthe theoryof Indianfeudalism emphasizesthe centralityof the political structurein initiatingchange and suggeststhree discontinuousphases of urbanizationaroundwhich other economicactivitiescoalesced.Thesephaseshavebeen definedas the Harappan (third-second millenniumB.C.), earlyhistoric (600 B.C. to third-fourth centuriesA.D.), and earlymedieval(sixth-seventh-twelfth-thirteenth centuries)periods.4Threemajorhistoricalprocesseshavebeenpostulatedforthe earlymedieval periodwithin this framework:expansionof statesociety;assimilationandacculturationof tribalpeoples;andintegrationof localreligiouscultsandpractices. A thirdmodel for understandingthe developmentof tradeis evidentin writings by AndreWink. He arguesagainstthe declineof Romantrade,since GreekByzantinetradersagainbecameactivein the Indiatradefromthe fourthto sixth centuries.5In the fifth-sixth centuries,Persiancommercesynchronizedwith the ascendancyof the Sasanianempire,and with the coming of Islamtherewas an increasein tradingnetworksin the IndianOcean.Wink suggeststhat controlof tradewasa motivatingfactorin the Arabconquestof Makran,Sindh,Kathiawar, and Kachh.The period until the tenth centurywas characterizedby Araband Persiansettlementsalongthe IndianOceanlittoralandthe emergenceof an integratedMuslimtradingempire.Furthermore,Buddhismlargelydisappearedfrom Indiain the earlymedievalperiod,thoughit continuedto flourishin Gujaratuntil the ninth century.Finally,due to restrictionsstipulatedin the Dharmasdstras (LawBooks)on maritimetravel,the Hindupopulationturnedto "agrarianpursuitsandproduction,awayfromtradeandmaritimetransport."6 The positionadoptedin this articleis differentfrom the abovethreemodels in severalways.In contrastto the rural/urbandichotomyandearlyhistoric/early medievaldisjunctionsupportedby both R. S. SharmaandB. D. Chattopadhyaya, this articlehighlightsthe centralityof the communityin the studyof the pastand continuityin the settlementpatternas evidentfrom the archaeologicalrecord. Therewere,no doubt,shifts,gradations,andgrowingcomplexityin the varietyof settlementtypes overtime, but few or no ruptures.Forexample,the lowerShetrunjivalleyandthe areaaroundPadriandHathabin Bhavnagardistrictemerged as the coreregionof the earlyMaitrkarulersin the fifth centuryA.D., as evident from inscriptions.This settlementwas by no means a "newbeginning,"since archaeologicalexplorationin the region has providedevidencefor twenty-two earlyhistoricsiteslocatedin a linearpatternalongthe riveranda multitiersettle40
HIMANSHU
PRABHA
RAY
0
X0
'!5 ',
,
!.'!
;f
< i V5gi !,.7 'XRadhanur
,=g
X
_,,
0
,
>
KACHH
~~~~~,~~
'fiaMtt
~~~~~~Himatnagar|
ggd><_ :0
*~~~Adhol g
2XXSaGUJARAT
;
*Bhuj
-: -:
=
-
g
0
2NORTH ';"7'i!i',
;
-
/
-
---
PLAIN
Surendranagar I 111lI
7
1|
Dhandhukaf
*JA=ng
X
g
8
~O'.Ikot *:
o th eSAURAcTRA
HILLSB
- tesi
-sorbarsd~~~~
t
-
i;
- -
*~~~~~o th sure Vers~~~Veval
=~~~~~
~utdhr
tu
eorirs
NarmadaRiver
arlon
Bharuch s d
~~~~~~~~~~~~~~~Gogh
ted onT a
allisn
are tion th Thbi arbtrf ile
continuou ocuato
a
SO b
e
te
Th isRiveer. Amel anotheriprtant arcaeogia
lmen h,r,:h.
aJo
*
Siiarly,_he
r
the earliest se Ma of Guart
W_
to
T UT
|
a
site lctedusra
_teStrni
an excaain
of th sit fro
con-
tefrtcnuyBC
RIve tthfis-eodcnuisADThseteetgaulyxPLAnddINa
ment hisracmple |~~~~~~~~~ncitos
ofGujarat. Map __
~~ ~ ~ ~ ~ ~ totetreoddvso-ftw,mrkt
hl
sretle h
is eodcnuyAD
fu importahesorted ncito
~rfr
by
f Karp n
reto .D' the
thrcee
division
tand of
inscritions.r Whielded tscond-uuscnuryaDncition
ua
pc
frua spce
noratio
urdmn
from
ngr-iaa
(nagraigama-
ofKtesiefrmthefrapa Renudradmana
fanapada) Maixth thAuD. varity ofesterrito493-76)aentoionl exadveriose cepigrynaphs S i BeG I p drlanga bhutwee riearliesubntse(palentlocadraka, gasmall bheami,patraso thera Ghelo
4ThE
firslttleevidencenotatie-SponsorTed Thvertohe setrade,nthoghraualers werpnedonstanl deiinga mateansflorgtax ingadcnrolnehrlcaiebuies.raeivle
a complexhierarchyof transactions,such as gifts to those in authorityand, at the local and regionallevel,barterand monetaryexchanges.Only commodities requiredby powerfulgroupswerecontrolled.It is significantthatwhile archaeological data and distributionnetworksof ceramicsindicatelocal, regional,and oceanic interaction,starting from the third-second millennium B.C. onward, inscriptionsfromGujaratarelargelysilentabouttradingactivitiesuntil well into the fifth-sixth centuriesA.D. From this period onwardinscriptionsreflectan increasingcomplexityin commoditiestradedand in the natureof transactions conducted.,,Thepost-tenthcenturymarksa furtherchange,with shopsandmarketsin the vicinityof templescontributingmaterialsto the performanceof rituals and festivities.Forexample,a thirteenth-centuryrecord,inscribedon a longslab of polishedblackstone fromone of the templesat Somanatha,refersto the purchaseof shopsbythe benefactorandtheirdonationto the temple.12 WhileWink'semphasison continuityof tradingactivityin thewesternIndian Oceanis welltaken,studieson trademustinvolvea discussionof transportation, shifts in coastalcenters,and emergenceof new settlements.For example,Bharuch,Sopara,and Kalyanwereimportantoutletsfor tradein the earlycenturies of the CommonErabutgavewayto Valabhiaroundthe middleof the firstmillenniumA.D., whenit becameprominentunderMaitrkarule.Arabattackson Valabhi aroundthelatterhalfof the eighthcenturyledto its abandonment.13Therewasthe consequentrise of Stambhatirtha,or Khambhat,mentionedin the Kavigrantof the RastrakutarulerGovindaIII (A.D. 827).Anotherimportantcoastalsettlement (modernPrabhasPatan),whichcontinuedto play wasthatof Somanathapattana an activerolein maritimetrade. Maritimetravelin the ancientperiodinvolveda varietyof watercraft.Cargo carriersformedthe foundationof tradingventures.Merchantsand tradersin some cases certainlyowned ships and watercraft,but they neithermannednor sailedthem.'4Moreoften, goodsand cargoeswereentrustedto the captainof the vessel,who was then responsiblefortheir saleand profit.Thusmaritimeactivity involveddiversegroupsfromthe ownersof watercraftto thosewho commanded themandstill otherswho sailedthem.Hencetherecanbe no simplecasteattributions of the communitiesinvolvedin tradingactivity.Moreover,the normative ruleslaid down in the Dharmasastras need to be balancedagainstthe narrative literaturein Sanskrit,which providesextensiveaccountsof maritimetravelby merchants,craftspeople,andmanyothers. Similarlypertinentto this discussionarecommoditiesandcrafts,whichin the archaeologyof Gujaratreferto ceramics,bronzes,beads,andotheritemsof ornamentation.Thesehavegenerallybeen interpretedas indicatorsof the ethnicityof tradinggroups,with little discussionof eithertheir contextor changesin their 42
HIMANSHU
PRABHA
RAY
ha
S4X
0 KM
300
150
I Jetalsar
FOPRTIFICATION- \JUNAGARH
l I .
I
II
Junagarh
\\ Khapra< KodiaCavies
MS
AdichadiVavLake
-
RailwayL1 Station
BoIv&4
1
sW
I te
{f11asj1:
iR
Mosquej / 4-Verival_\
\
LJ
J /1
11
il
I IAS
V
*~~ \IR
>
C
Edict ~
4)Navghan Vat'LakeUflO
fromAkt,dag
to
A.
N
wz1
-
B
5-1
1L
1L
ar
~~~~~~~~s
aM
ORTIFICATION
Ir
l obetsfunEnRuarthaebenc
use ovrtimeSever
teasvincfo
Archaeologicalmonuments in and around Junagarh.
msentofeieDresserl2-
typectamporaeiGjaa have beefon,tencty-ied
Gujra arevidnc
and thirteen of these are clustered around Jun&garh. The area around Junagarh provides a fertile stretch, and, as I will discuss later, it formed a core region in the early period. This area was also the location for royal inscriptions and religious shrines. Other find-spots of amphorae sherds as emereSomanathapattana, CleGuarly, include gious structres,such coastal centerstyeamples,e such as Dwarka, ienteffth-sixt wenturfives and Valabhi, Nagara, arDe withind theaparamtrhfpterso.iti cnstextaied others. tof bhseae "impohrtsneed as Valabhi Sites such into centers the midamong developed political by dle of the first millenniUM A.D. In contrast, others, like Somanathapattana and buton hie.Ohrfidsosomhresed eiiu of great Dwark&, incoitnsumptin. roandwere pilgrimage centers sanctity. At Soman&thapattanaarchaeological evidence of historical settlement dates to the fourth century B.C., but relimoblit wieSoaashptt One of the suhtsheaers, fistrateiesnu A.dope byconrafspepl,
asiameants"netoimroin
43
THE BEGINNINGS
conomicualiveithood
Moblit alamtesof paidedninoresistring
the state-sponsoredefforts to control labor. The best-knownexampleof this strategyis the Sanskritinscriptionof the chief of a guild of silk weaversfrom a now-losttemple at Mandasorin CentralIndia. The guild moved from Lata, identifiedas the region between the Narmadaand Tapti rivers,to Dasapura (modernMandasor)"bringingtheir childrenand kinsfolk."In spite of having adopteda varietyof occupations,the silk weaversretainedtheir collectiveidentity and met togetherto make an endowmentto a templeof the sun god "with hoardedwealth."'18 The importanceof gifts to the templein the social life of the communityis thus evident. The religiouslandscapeinfluencedthe tradingsystemin severalways,from molding cultural preferencesand choices to active participation.Religious shrineswereboth consumersof a varietyof commoditiesused in ritualas well as importantlocalesfor tradingactivity,as indicatedby shopsand marketswithin or in the vicinityof templepremises.Thebuildingof newtemplesstimulatedeconomic growth,therebytransformingboth the geographicand social landscapes of the region."9 At the same time thereare severalinstancesof a differentialtax on commoditiesrequiredfor religiouspurposes.Moreimportant,of course,severalcommoditiessuchas textileswereimbuedwith multiplemeaningsandwere both items of common consumptionas well as productsfor elite and religious requirements.This complexityof interactionbetweenpatronsand otherinterest groupsand religiousstructurescan only emergewhentradingactivityis studied not merelyin commercialterms,butissuesof identityand statusmarkersarealso broughtinto the discussion. TheMaritimeNetwork Archaeologicalinvestigationin northwesternIndia has providedinformation on severallong-lastingcoastalsettlements.Forexample,the site of Mandvion the estuaryof the RukmavatiRiverat the entranceto the Gulfof Kachhhasbeen known as a port town with links to both Omanand the EastAfricancoastsince archaeologicalexplorationconductedin the mid-nineteenthcentury.In addition to a rangeof ceramics,importantfindsincludelocaland regionalcoins,aswell as a Byzantiumsolidusof Heracliusdatedto A.D. 638andan Arab'Umayyaddinarof A.D. 716.
Dwarka,locatedon the north coast of Saurastra,is anothercontemporaneous sitewith a long periodof settlementfromthe firstcenturyB.C. to almostthe present.Itsimportancestemsboth fromits coastallocationaswellasthe religious significanceof itstemplesandsacredassociationas a centerof pilgrimage.21 Other continuouslysettledsitesalongthe GujaratcoastincludePorbander,Somanatha, and Valabhiat the headof the Gulfof Khambhat,as I will discussin a latersec44
HIMANSHU
PRABHA
RAY
tion. While sites such as Valabhidevelopedinto politicalcentersby the middle of the firstmillenniumA.D., others,like Somanathaand Dwarka,werepilgrimage centersof greatsanctity.At Somanathaarchaeologicalevidenceof historical settlementdatesto the fourthcenturyB.C., butreligiousstructuressuchastemples emergeonlyin the fifth-sixth centuriesA.D. Farthersouth, Chaul,referredto as Campavatior Revatiksetrain the Epics, is mentionedin inscriptionsand literarysourcesfrom the earlycenturiesof the CommonErato the seventeenthcentury.In a copperplaterecordof 1094,three majorcoastalcentersareidentifiedand include,in additionto Chaul,Thaneand Soparaalso in northwesternIndia. While no systematicarchaeologicalwork has been undertakenin the area,villagersdiggingwells in coconutplantations unearthedcrucialinformationregardingsettlementfrom 250 B.C. to A.D. i6oo, thoughfora briefperiodfromA.D. 250 to 500, the siteseemsto havebeenunoccupied.22 Epigraphsdatingfrom the tenth-eleventhcenturiesreferto these coastal settlementsas linked to eachotherthroughland routesas well as to those in the interior.23
The Sasaniansruled Iran and adjacentcountriesfrom 224 to 651 and were activeparticipantsin the tradeof the IndianOcean.The Nestorianannalsrefer to trade with India and Sri Lanka during the reign of Yazdigird1 (399-421), while
Tabarireportsthathis successor,BahramV (421-38),marriedan Indianprincess and receivedas a dowrythe port of Daibalin the Indusdelta,togetherwith the adjacentparts of Sindh and Makran.24
Archaeologicalexcavationsconductedat Siraf in the PersianGulf indicate its beginningsin the Sassanianperiodwhen a fortwas locatedat the site. David Whitehousecontendsthat in additionto militaryfunctions,SassanianSirafwas alsoa commercialport,asevidentfromimportedobjectsatthe site.25Aftera break in settlement,Sirafreemergedas a majorcenteraroundA.D. 700. Sohar,situated close to the mouth of the PersianGulf on the coast of Oman,reachedits greatest size in the tenthcenturyA.D., occupyingseventy-threehectareswith a defense wall protectingit on the seawardside.Thesetwo sitesin the PersianGulfparticipatedin a regionaltradenetworkin the ninth to eleventhcenturies.Thisnetwork includedsettlementsin the Indusdelta,suchas Banbhoreon the Makrancoast,as wellas severalcentersalongthe westcoastof IndiaandMantaion the northcoast of SriLanka. ArchaeologicalexcavationsatBanbhorehaverevealeda fortifiedcityenclosing 167hectaresin the earlyIslamicperiod,with an innerresidentialandcommercial areaand an industrialsectoroutsidethe defenses.The nucleusof the port was a mosque-one of the earliestin the region-with two Kuficinscriptionsdatedto A.D. 727
45
and 907.
THE BEGINNINGS
Chineserecordsreferto Persian,or "Possu,"merchantsfrom the fourth to twelfthcenturies,but it is only fromthe ninth centuryonwardthat the Persians initiatedregularsailingto SoutheastAsiaand China.Sirafformeda majorport for this purpose.The earliestreferencesto Muslimmerchantsengagedin longdistancetradeoccur in an Omani text by Al-Rahildatedto the ninth century. Thesereferto Al-Qasim,a learnedIbadishaikhinvolvedin the aloeswood trade with China.The second merchant,Maymuin,lived in Basraand was relatedto two Ibadiimams. Thoughin the literarysourcesthe focusis on luxurygoods,the archaeologicaldataindicatetradein subsistenceitemsandbulkcommodities.In additiontherearereferencesto a largenumberof artisansat Siraf,includingshipwrights,weavers,metalworkers, jewelers,andpotters. Thereare severalreferencesto small merchantswho traveledvast distances and tradedin a varietyof goods. Thirteenth-centuryaccountsof a merchantat Qais describehim as wishing "to carryPersiansaffronto ChinawhereI understandthatit hasa highprice,andthentakethe dishesfromChinato Greece,Greek brocadeto India,Indiansteelto Aleppo,glassof Aleppoto Yemen,andthe striped material of Yemen to Persia."26
The somewhatlaterGenizadocumentsindicateinterlinkagesbetweenmerchantgroupsfrom differentregionsand distinctionsin the scope of operations and organizationof tradingactivity.Forexample,the distinctionbetweenretailersandwholesalerswaswelldelineated.Manyof the retailersweretradingartisans who marketedtheirown products.Thetermtajir wasusedforbigmerchantswho tradedin awidevarietyof goods.Theyweredividedintothreecategories:onewho storedlargequantitiesto sell when the pricewas high;anotherwho transported goods fromone countryto another;and a thirdwho sentshipmentsabroad.One suchmerchantwasNahraybin-Nissim,a wholesalemerchantmentionedin 250 of the Genizadocuments.27 Oneof the characteristicfeaturesof maritimetradefromthe ninth-tenth centuriesonwardwasthe locationof marketsin fortifiedsettlementsalongthe Indian Oceanlittoralandfartherinland.Rulesgoverningthe paymentof taxesandregulatingthe functioningof the marketswereoften inscribedon copperplatesand provideuseful insightsinto the organizationof the tradenetwork.The Quilon copperplatesof SthanuRavifromthe Malabarcoastaresignificantin connection t theChristianchurch.A marketwaslocatedwithin with tradingrightsgrantedt the precinctsof the fortifiedsettlementatthe portof Quilon,whilethe churchwas situatedoutsidethe fortificationwall.28 Alongthe Konkancoast,therearereferencesto the fortifiedmarketcenterof Balipattana.The Kharepatanplatesof Rattarajadatedto A.D. 10o8 list gifts to the templeofAvvesvarabuiltbyRat.t.araja's fatherandsituatedinsidethefortifications. 46
HIMANSHU
PRABHA
RAY
Theseincludeda measure(gadyana)of goldfromeveryvesselcomingfromforeign landsanda coin (dharana)of goldfromthosecomingfromthe coast,exceptthose from Chauland Candrapura.Also locatedwithin the fortificationsweresettlementsof femaleattendants,oilmen,gardeners,potters,andwashermen.29 PilgrimtrafficacrosstheBayof Bengalwasindeedmultidirectional,thoughthe movementof worshipersfromthe Indiansubcontinentto the Indonesianarchipelagois often underplayed.The most significantfor this studyarethe Kelurak andPlaosaninscriptions,asthe formermentionsa gurufromGaudidvipaandthe settingup of an imageof Manjusri.Thelatteris a fragmentarystoneinscriptionin Sanskritof the ninthcentury.Itis nowin the JakartaMuseum(no.D 82), thoughit is knownto havecomefromCandiPlaosanin CentralJava.It refersto theworship of the Buddhatemple(jinamandira)by groupsof peoplebowedby the burdenof devotionwho continuouslyarrivedfrom Gurujaradesa, i.e., northwesternIndia (satata-gurujaradesa-sa
magatais ... sugatabhakti- bhra-pranatai).30
As
the portionof the inscriptionbetweenthe two adjectivesis damaged,the identity of the personsfromGurujaradesa is difficultto establish. TheEmergenceof TerritorialIdentity The earliestreferenceto territorialdelimitationin northwesternIndiaoccursin the Arthasastra,a treatisefor governance,that has often been used as a base for writingeconomichistory,supplementedbytheinscriptionsofthe Ksatrapas dated to the earlycenturiesof the CommonEra.31The archaeologicaldatawould,however,indicatethat in the fourth-third centuriesB.C., a muchsmallerunit within Saurastra-specifically, the area around Junagarh-formed the core area of economicandpoliticalactivityin the region.At leastfivecoin hoardscontaining more than 4,800 silvercoins havebeen recoveredfrom Junagarhdistrict.These areessentiallysmallcoinsof aboutone gram,eachproducedfroma singledie and somerestampedcoinsof the Magadhajanapada.32 Thisdistinctiveterritorialidentityby no meanstranslatedinto isolationfrom otherpartsof the subcontinent,asis evidentfromthe epigraphicdatapreservedon the hill atGirnar,about1.5kilometersto the eastof Junagarh.Inscribedon the hill arerecordsoftwo of themajordynastiesofthenorth-the Mauryas(317to 186B.C.) and the Guptas(datedin the reignof king SkandaguptaA.D. 455-67) -while the third,engravedon the westernside,belongsto the KsatraparulerRudradamana (secondcenturyA.D.).Thesethreerecordsalsopresentinformativecontrasts. Rudradamana's dynastywas one of six familiestermedthe westernKsatrapas, or satraps,who ruled in Saurastraand Malayain the earlycenturiesof the CommonEra.Thesewereallied to otherrulerswho used the title ksatrapa(literally,viceroy)on their coins and in inscriptions.33 In contrastto many of their 47
THE BEGINNINGS
contemporariesin Centraland peninsularIndia,the Ksatrapasused Greeklegends on theircoinage.Prominentamongthe variousissuesarethe silvercoins of Nahapana,with the bust of the rulerand a legendin Greekscripton the obverse andsymbolssuchasthe thunderboltandarrowandinscriptionsin Kharosthiand Brahmion the reverse.The Greeklegendon the obverseis a transliterationof the inscriptionsin Brahmior Kharosthion the reverse.34 Thedistributionof Ksatrapa coins is an indicatorof tradenetworksthat extendedwell beyondthe political boundariesof the rulers.In additionto centersin northwesternIndia,thesehave been foundat severallocationsin CentralIndia,andtwo coin moldsoccuras far eastas the Andhracoast.35 Rudradamana's recordis of crucialimportancein establishinga time frame for the perceptionof the regionas distinctive.This inscriptionfor the firsttime providesterritorialidentityto northwesternIndia.UnlikeAsoka'sGirnarinscription, copiesof whichoccurwidelyin the subcontinent,Rudradamana's recordis regionallyspecificboth in the natureof its messageand in the territorialdemarcationof his kingdom.,6The primaryobjectiveof the epigraphwas to recordthe restorationof LakeSudarsanaby Mahaksatrapa in whosereignit Rudradamana, hadbeendestroyedby a storm.37 Nearlythirtystoneinscriptionsof the westernKsatrapashavecometo light,of whichnine havebeenfoundatthe villageof Andhauandone eachfromKhavada, Mevasa,and Vandhavillagesin Kachh.38Theseinscriptionshavebeen used primarilyforworkingouta chronologyof thewesternKsatrapasandtheirgenealogical connections,but they aresignificantin manyotherwaysas well.Whilesome of the epigraphsfrom the westernDeccan,Sarnath(nearVaranasi),and Taxila (now in Pakistan)recorddonationsmadeto Buddhistmonasticestablishments, thosefromKachhtendto be memorialstones(yastior lasti)or referto the digging of wells.39 TheKsatraparecordsfoundatJunnar,Karle,andNasikprovideinsights into the conflictbetweenthe Ksatrapasandthe Satavahanas of the Deccan.These politicaland militaryconflictsseemto be underscoredby the overstrikingof the KsatrapaNahapana's silvercoinsbytheSatavahana rulerGautamiputra Satakarni, e.g.,thosefoundin the JogalthembihoardatNasik.40 Anothervexedquestionrelatesto the ethnicidentityof theseKsatraparulers. Perhapstheonlyindicatoris theNasikrecordof Usavadata,Nahapana'sson-in-law who refersto himselfas a ?aka.n4Thetermsakais generallytranslatedas a "Scythian,"referringto movementsof communitiesfromCentralAsia.Noris Usavadata the onlysakadonorat the Buddhistestablishmentof Nasik.ThesakaDamachika Vudhika,a writerfromDasapura,also donateda caveandtwo cisterns.42Despite these donationsto Buddhistreligiouscenters,the Ksatrapasretainedthe use of the Greeklanguage,as evidentin the legendson theircoinsdiscussedabove. 48
HIMANSHU
PRABHA
RAY
Nahapanais perhapsthe only rulerof the subcontinentwho findsmentionas Manbanosin the PeriplusMarisErythraei(Periplusof the ErythraeanSea)(chapter 41),a first-centuryA.D. Greektext. The Periplus(chapter47) furtherstates, that "beyondBarygazatherearemanyinlandpeoples... and abovethese [to the north]aretheverywarlikeBactrians,whohavetheirownkingdom."43 ThePeriplus refersto the continuedcirculationof coinsknownas drachmsandengravedwith inscriptionsin Greekin the markettown of Barygazaor Bharuch.These coins wereissuesof kingssuchas Apollodotusand Menanderwho ruledin northwesternIndiaafterAlexander'scampaignin India. Thesereferencesled W. W. Tarnto suggestthat tradebetweenIndiaand the West had gone on since the Greekconquestof the subcontinent44 and that the maritimeconnectionbetweenSaurastraand the Westexpandedunderthe reign of Menander(150-135B.C.). Menander'sempireis thoughtto haveextendedfrom Mathurain the eastto Bharuchin the west,andthe regionfromKabulto Mathura hasalsoproduceda largenumberof Greekcoinsfromthe reignof Menander.45 MortimerWheeler,on the otherhand,arguedfortradeinitiatedbythe Roman empireafterhis excavationsatArikameduin 1946,46and sincethen this view has been the prevailingorthodoxy.As mentionedearlier,severalobjectsfound in Gujarathavebeencitedas evidenceforRomantrade.Whileanalysisof the bronze handle from Akota shows analogieswith specimensfrom Pompeii,thereis little doubtthat red polishedware(RPW)is locallyproducedpottery.Madefrom fine well-levigatedclay,RPWoccursin stratadatedfromthe firstcenturyB.C. to the fifth centuryA.D. and has been recordedat more than four hundredsites in Gujarat,with up to 160 stylisticvariations.47 Someof the excavatedRPWsitesare Dwarka,Somanatha,Amreli,Devnimori,Samlaji,Nagara,andTimbarva,among others,manyof whicharealso known fortheirearlytemples.Thusthe distribution of RPWis both alongthe coast as well as in centersfartherinland.It is significant,however,thatRPWsherdsdatedto the firstfivecenturiesA.D. havebeen foundat archaeologicalsitesaroundthe PersianGulf. ForeignersandTrade Themobilityof merchantsenabledthem to migrateand shift theirareasof operation to new regionsand ports. Forexample,the merchantsof Sirafand Sohar shiftedto the portsof Qaisandotherregionsafterthe declineoftradein the upper PersianGulf. This mobilityalso meant that merchantswere often stationedon foreignshores,and there are severalreferencesin literatureand inscriptionsto groupscategorizedas outsiders.On accountof the complexityof economictransactionsin the ancientperiod,determiningethnicidentitiesis a taskfraughtwith difficulties. 49
THE BEGINNINGS
Fromthe firstcenturyB.C. to the secondcenturyA.D., whilemanyof the Arabs of the easternMediterraneanregionswere Romansubjectsor Romancitizens, others,suchas Nabataeans,Palmyrenes,and Sabaeanslivedbeyondthe frontiers of the empire.In the Indiancontext,referencesto yavanasareubiquitousin SanskritandTamilsourcesandreferto a rangeof nonlocalgroupsfromWestAsia.48 Atthe sametimetherearereferencesto foreignersin chargeof provincesin the subcontinent,especiallyin Gujaratandthe Konkancoast,suchas in the Junagarh inscriptionof Rudradamana, recountingthe historyof LakeSudarsana.Manyof the earlyreferencesto the yavanasindicatethatGreekssettledin the northwest,49 though this is not the patternelsewherein the subcontinent.Anotherproblem is the inabilityof the literarysourcesthemselvesto distinguishamongdifferent ethnicidentities,as in the caseof allusionsto Romans,Arabs,Indians,andEthiopiansin Greekaccounts. This debatenotwithstanding,there are severalinstancesof foreignsettlers beingtaxedby local politiesin the earlymedievalperiod.The evidencefromthe subcontinentindicatesthe presenceof non-Indianmerchantssuchas the Tajikas andtheTuruskasin differentareas.Turuskadanda is a termthatoccurscommonly in the Gahadavalainscriptionsof the Gangabasinand has been takenby several writersas a tax on Turkishsettlers.50 A yona (yaivana)-raja of Sanjayataor Sanjanlocated on the north Konkan coast is mentionedin an inscriptionfrom Nagarjunakondadatedto the fourth century A.D. 51 The Chinchani copper plates mention that the entire district (mandala)of Samyana(Sanjan)was made over by the RastrakitarulersKrsna II (reigned 878-915) and Indra III (reigned 915-927) to Sugatipa Madhumati
(Muhammad)of the Tajikacommunity.The latterconqueredthe chiefs of all the harbors(velakuila) of the neighborhoodon behalf of his masterand placed his own officialsin them.52He establishedfree ferryserviceat two streamsand also a feeding house at Samyana.In addition, he createdan endowmentfor repairsand also for offeringworship(naivedya)to the goddessDasami.A later grant of the eleventh century refersto the region as being under the rule of Camun.daraja, manyof whoseofficialsbearnamesof Arabicorigin.A little later, termsbetweenthe rulingelite and the yavanasseem to havesoured.The Kharepatanplatesof AnantadevaI datedto A.D. 1095referto the rulerdrivingout the vile yavaniaswho had devastatedthe Konkanregion.53 TheEmergenceof Valabhi:Capital,PortTown,and ReligiousCenter The history of hostilitiesand invasionsin northwesternIndia goes back to an earlierperiod,as discussedabove.In the fifth centuryA.D., the Junagarhinscription refersto hostile invasionsduringthe earlyyearsof Skandagupta's reign,but 50
HIMANSHU
PRABHA
RAY
thesewerefinallyrepulsed.5At the sametime,Valabhi,locatedon the Bhavnagar creekat the head of the Gulf of Khambhat,emergednot merelyas an outlet for maritimetradebut also as the capitalof the rulingdynastyof the Maitrkas(A.D. 493-776)and the core areafor religiousconsolidation.5ContemporaneousSanskritliteraturesuchas the Dasakumaracaritam (Storyof TenPrinces)by Dan.din Referencesto its wealthyresidescribesValabhias a prosperoustradingcenter.56 dentandtravelingcommunitiesof traders(vaniggrama), arefoundin the copper platedonationsfromToramana's reigndatingto the latefifth-early sixth centuriesA.D. 5 self-assuranceof its tradingcommunityis evidentfromthe charter of Visnusenaof A.D. 592,whichreiteratesthe customarypracticesfollowedby the groupand the acceptanceof them by the ruler.This consolidationof economic activityalongwith expansionof politicalauthorityandincreasinginvolvementof brihmanasand Buddhistviharas(monasteries)in agrarianintensificationmarks a departurefromthe tradingpatternsof earlierperiods. Underthe earlyMaitrkarulers,the capitalwas shiftedfromJunagarhor Girinagarato Valabhi,andthe recordsof the earlyMaitrkarulerssuchas Dronasimha and DhruvasenaI were issued from Valabhi,e.g., the BhamodraMohotaplate of Dronasimha(year183or A.D. 502).58 The list of administrativepersonnelof the Maitrkasis fairlylarge,59 andan importantsourceof staterevenuewassulka(tolls) andthe personin chargeof collectionis referredto assulkika.So Three copperplatesdated in the reign of the Hun.a Toraman.a(fifth-sixth centuriesA.D.) recordgifts madeby the tradingcommunityof Vadrapallito the orNarayan.la templeof JayaswamI belongingto the queenmother.MaharajaBhuita andMaharajaMatrdasalso madedonationsof certainvillagesto the temple.The main commoditiesreferredto in the platesaremolasses,salt, cotton, and grain, anda detailedlist of revenuestransferredis also enumerated.Theserevenueswere calculatedon the basisof vesselload, donkeyload, and cartload of the produce. Vadrapalliwas probablylocatedeightkilometerswest of Sanjeli,and signatories to the donationincludedtradersfrom Ujjain,Kannauj,Mathura,and perhaps A goldsmithconstructeda lakenearthe temple.61 Man.dasor. Ofinterestto this discussionis the charterofVisnusenadatedto A.D. 592, issued fromLohatain the Kathiawarregion.Thischarteris addressedto a list of officials, as is usual in Maitrkarecords.The charterassuresprotectionto the community of merchants(vaniiggraima) establishedin the regionandendorsestheircontinued functioning.62It providesa detailedlist of seventy-twotraderegulationsor customarylawsto be followed.Someof the regulationsareof greatinterestto this discussion.Forexample,it is specifiedthatmerchantsstayingawayforayearwerenot requiredto payan entrancefee on theirreturn.Otherclausesspecifydutiesthat wereto be paid.A boat full of containers(bhandia-bhrta-vahitrasya)was charged 51
THE BEGINNINGS
twelvesilvercoins, but if the containerswere for religiouspurposes,they were chargedonlyone anda quartersilvercoin. In the caseof aboatcarryingpaddy,the chargewashalfthis amount.Otheritemsfrequentlytransportedbyboatincluded driedgingersticks,bamboo,wine,leather,buffaloes,camels,andbulls.63 Indigoor nila is mentionedas one of the items exportedfrom areassuch as Gujarat,andthe charterof Visnusenamentionsa tax on the pressingof the indigo dye.As mentionedearlier,indigo figuresas one of the items of tradefrom India westwardto Egypthandledbythe businesshouseof Ibn'Awkalin the periodfrom A.D.
980 to 1030.64One of the three varieties mentioned includes Sindani indigo,
Sindan/Sandanbeingidentifiedwith the Konkancoastof the subcontinent. The varietyof taxableobjectsmentionedin the inscriptionis an indication of the diversenatureof tradein the region.These includedoil mills, sugarcane fields, wine, cumin seed, black mustard,and coriander.The inscriptionalso refersto a tax on dyersof cloth,weavers,shoemakers,andretailershawkinggoods on foot. Otherssuch as blacksmiths,carpenters,barbers,potters,etc., could be Clearly,the Maitrkas recruitedforforcedlaborunderthe supervisionof officers.65 wereattemptingto widentheirresourcebase.A parallelprocessthatis evidentat this time relatesto changesin the religioustopographyof the region-a religious topographythatcontinuedto be variedanddiverse,withlittle evidenceof assimilationor acculturationwithin the ideologicalframeworkof a statesystem.On the contrary,the rulersprovidedpatronageforseverallocalcults. ReligiousTransformationand RitualSpace The settingup of the SudarsanaLakeat Junagarhis often quotedas an example of centralizedMauryancontrol,but what is seldom discussedis the religious landscapearound the site. Though no structurescontemporaneouswith the Asokan edicts at Girnarhave so far been identified,the earliestremains are of a spaciousmonastic complex at Intwa on a hill three kilometersnorth of Girnar.An inscribedsealing found in a cell dates the monasteryto the reign of RudrasenaI (A.D. 199-222).66 The religiousmonumentsat Junagarhinclude three sets of rock-cutcaves:the KhapraKodiacaves,the BabaPyaracaves,and the cavesat Uperkot.In additionto these, a Visnutemplewas establishedat the site in the Guptaperiod. Anotherextensivesitewaslocatedthreekilometerssouthofthe foot of Girnar at Boriaamid thickjungle,while a somewhatlaterfourth-to sixth-centuryA.D. site was excavatedfartherinlandat Devnimoriin the Meshvovalley,120 kilomeAt a distanceof aboutsix hundredmetersto the east of ters east of Ah.madabad. the sitewasa smallmoundcontainingbricksanda smallSivalingasuggestingthat it wasa Saivatemple.67 52
HIMANSHU
PRABHA
RAY
1
4th-century A.D. image of lajja-gaurt
from Dhank in Junagarhdistrict. 210.44 courtesy American Institute of Indian Studies, Gurgaon.
Perhapsone of the earliestshrinesin Gujaratwasexcavatedat the site of Padri in the TalajatalisIlof Bhavnagardistrictof Gujarat,hardlytwo kilometersfrom
theGulfof Khambhat, andwasdedicatedto a fertilitydeitytermedlajjd-gaur7.i8 Two terracottaplaquesof the goddesswere found on the floor of the structure, while a squareslateplaquewith the imageof lajja-gauniwas foundon the surface. Other imagesincludedsandstonefiguresof Gan1esa and Visnu.69Sixth-seventh centurylajfii-gaurifigurineshavebeenreportedfroma numberof sitesin Gujarat, andthe worshipof the goddesscontinuedinto a laterperiod(fig.1).7? Evidenceforthe worshipof anothernon-Sanskriticlocal goddesscomesfrom the recordof one of the earlyMaitrkarulers,Droniasimha.He donateda village for providingsandalwood,incense,lamps,oil, and garlandsto the shrine of the goddessPand4uraja, otherwisenot knownfromtextualsources.Dron.asimhaalso builta templeto Kottammahikadeviwithin the boundaryof Trisangamakaidentifiedwith Tarsamia,aboutfivekilometersfromBhavnagar,and the grantto the
goddesswasresumedbyDhruvasena II(reigned A.D.
639 -4o).7'
The sectarianaffiliationof the Maitrkasis evidentin the sixth century,when DhruvasenaI (reigned519-49) refersto his predecessorsas worshipers of Siva (parama-mahesvara) and himself as worshiper of Vityu (parama-bhdgavata). Theseepithetsappearedin a grantof land to Buddhistvihdras.72Thus even here thereis no evidenceof unambiguoussectarianaffiliationofthe rulersor of grants madeto templeswith the objectiveof legitimizingtheirrule.In most casesdonations aremadeforrepairandworshipof existingtemples,e.g.,the copperplateof SiladityaI (reigned599-614). The rulermadea grantof two piecesof land along worship with a tank (Yamala vopf) forrepairof a templeof ivaaswell as for of1the deity.Worshipof the deityinvolvedbathing(snapana)of the image,applicationof poundedsandalwood (gandha), incense (dhapa), flowers(pu.spa), garland (malya), (gtta) anddance(nrtya).73 lightingof lamps(dipa), andperformanceof music The earliesttemple contemporaneouswith the Maitrkasin Sauramara is the kilometers from and templeof Gop locatedaroundsixty-five the coast datedto the latesixth-earlyseventhcenturiesA.D. (fig.2). In the nexthundredyearsa large numberoftempleswereconstructed,mainlyalongthe coastline,butnoneofthese bear any inscriptionsand cannotbe associatedwith royalty.A distributionmap 53
THE BEGINNINGS
5 , '^l !-
-
3
2
2
6th-century A.D. temple at Gop, Jamnagardistrict. A 34.5courtesy American Institute of Indian Studies, Gurgaon.
of the templesitesindicatesa clusteringof templesatKhimesvaralocatedbetween the Vartuand Bhadarrivers.Thereis thus physicaldisjunctionbetweenthe religiouscenterslocatedalongthe coastandthe politicalcenteratValabhiatthe head of the Gulfof Khambhat. A distinctiveaspectof the templesof westernand CentralIndia at this time was theirdedicationto Suryaworship.The Bhadreniyakagrantof SiladityaI (ca. 610-i1) recordsthe donationof two hundredpadavarttasof land in the villageof Bhadreniyakafor worship(pujd) of the sun god (Adityadeva)establishedin that village.Worshipinvolvedbathingof the image,applicationof sandalpaste,flowers,lightingof oil lamps,and performanceof vocal and instrumentalmusicand and maintenanceof serdances,cost of sacrificesand offerings(bali-caru-satra), vantsof the gods.74 The Mandasorstoneinscriptionof the time of KumaraguptaI and the Indore copperplategrantof Skandaguptashow that sun worshipin CentralIndia was patronizedby guilds,but similarepigraphicinformationis not availablefortemreferto rest houses attachedto ples in Gujarat.The copperplatesof Toramaiina templeswheremedicinaldiet and therapywas providedto mendicants,men and womenin the serviceof the deity,anddevoteesanddisciples.Thesearein addition to the lakereferredto aboveand othergrantsmadeto the temple.Archaeological explorationaroundSanjeliled to the discoveryof a largenumberof bricks,but no Thus,in additionto beinga religiouscenspecificstructurescouldbe identified.75 ter,the templewaslinkedto servicesprovidedto the communityatlarge. Though there is no mention of Jainismin the nearlyone hundred copper plate grantsof the Maitrkaperiod, JainaliteraturedescribesValabhias a great pilgrimagecenter (tirtha),and archaeologicaldata provide evidence of earlier shrines (fig. 3). Jainametal and stone images stylisticallydated to ca. 500-600 and from Khedbrahmain northern A.D. were found at Akota near Vadlodara Gujarat.Around the same time Jainarock-cutcaves were excavatedat Dhank 54
H IMAN SH U PRAB HA RAY
4~~~~~~~,~
3
4
3 Bronze image of Jina from Vala. 132.73
courtesy American Institute of Indian Studies, Gurgaon. 4 4th-century A.D. seated Buddhaimage from Dhank, Junagarh.210.39 American Institute of Indian Studies, Gurgaon.
in Saurastra. Somewhat later in the eighth century, a Jaina monastic settlement of Digambara affiliation is known to have existed at Nagasarika or Navasari in southern Gujarat.76 As compared to that of the Jainas, there is abundant Buddhist presence both archaeologically and in inscriptions, though no structures contemporary to the Asokan edicts have so far been identified. But perhaps the most prolific references to Buddhist monastic establishments occur in the Maitrka inscriptions, which locate these around Valabhi, though there are no surface indications of any structures at the site.77Clay seals with the Buddhist formula Ye Dharma Hetu Prabhava were, however, found in archaeological excavations at Valabhi.78 The inscriptions record, among others, the names of two merchants, Kakka and Ajjita, who were responsible for the building of two viharas, which were named after them.79Reference to fields belonging to Buddhist monks occurs in Dhruvasena II's grant.80 The Vala plates of Dhruvasena I record the grant of the village of Pippalarunkhari and the right to collect its produce to a Buddhist vihara at Valabhi built by his niece Dudda.8'Dudda's vihara also received the donation of a village by Guhasena.82Dhruvasena I also donated the village Vataprajakato a vihara at Valabhi founded by Acarya Bhadanta Buddhadasa.83Three copper plates of Dantivarman of the Gujaratbranch of the Rastrakuitafamily dated saka samvat 789 (latter half of ninth century) refer to the donation of forty-two villages, after the king had bathed in the river Puravi identified with Purna in Surat district. These villages were granted to the Buddhist vihdra at the sacred pilgrimage center of Kampilya in Farrukhabad district in the northwest and were to be enjoyed by a succession of pupils of the Aryasangha. The money was to be used for defraying the worship expenses of perfumes, flowers, frankincense, lamps, and ointments and for the repair of the temple.84Thus the wealth, ritual, and festivities involved in Buddhist worship at this time is abundantly clear (fig. 4). 55
THE BEGINNINGS
In contrastto the earlierphasein northwesternIndia,when merchantsfunctionedeitherin an individualcapacityoras a collective,the articulationof lineage identitybecameimportantfor tradinggroupsfrom the eighth centuryonward. The inscriptionsrecordingtheirbenefactionscontainlong eulogies,orpra?'astis, butverylittleinformationon theiroperationsoractivities.Anotherchangeoccurs in the natureof religiousaffiliation.While the earliergrantsinvariablyrecord gifts madeto Hindutemples,duringthe eleventh-twelfthcenturiespatronageto jainashrinespredominatesin northwesternIndia.Thereis a concomitantrecording of detailedfamilyhistoriesof the merchantdonorsin inscriptionsas well as 85 The objectiveof the dharmakathMs or religiousnarthe writingof biographies. rativeswas the propagationof jaina religiousprecepts,but the main characters of HaribhadraSuniand invariablywerethe merchants,e.g.,the SamaraicchakdhM of UddyotanaSuni,the latterwrittenin the eighthcentury the KuvalayamilakahM A.D. Anothergenreof literaturethatexpandsat this time is legalliterature,a good an anonymousworkcontainingmodel forms examplebeingthe Lekhapaddhati, of legaland other documentswith datesrangingfrom 755to 1476. This shift in narrativewritingis matchedby increasedlineagestratificationamongthe jaina preceptorsor teacherswho exhortedthem to a path of righteousness.Several instancesof a lineagemerginginto anotherand the mobilityof the communities furtheramplifythis increasingsocialcomplexity.816 Conclusion In thefinalanalysis,ratherthanrupturesanddiscontinuities,occasionallyrevived by externalstimulus,a complextradingnetworkinvolvinga varietyof groups includingartisans,craftspeople,and transportersmarksthis period of the history of Gujarat.The historyof these communitiesis evidentfromthe donations thattheymadeto religiousestablishments.Thustradeandtradingactivitycannot be studiedin isolationfromthe diversereligiouslandscapethatdevelopedin the region.Theattemptin this articlehasbeento highlightthe locationandarchaeology of religiousstructures,therebyplacingthe templein its widersocialbase.It also needs to be appreciatedthat the templewas by no meansunique;insteadit waspartof a diversesacredgeography.Fora comprehensive appraisalof craftsand communitiesin northwesternIndia,the intertwinedstrandsof religiousarchitecture,economicactivity,aqndpoliticalinterventionneedito be examinedand
NOTES
1. M. N. Pearson,TheIndian Ocean
(London and New York:Routledge, 2003),95,75.
JohnE. Cort, "Communities,Temples, Identities:Art Histories and Social Histories in WesternIndia,"in Ethnography and Personhood:NotesfromtheField, ed. MichaelW. Meister (Jaipurand New Delhi: RawatPublications,2000),101-28. 3. This theory creditsthe introduction of iron around1000B.C. with expansion of agricultureand production of a "surplus" in the middle Gangavalley.This surplus was then invested in tradingventures, which emergedaroundthe middle of the firstmillennium B.C. Tradewith the Romanempireled to the developmentof urban centersin the earlycenturies of the Common Era,but with the decline of the Romanempireand consequentlyits trade with India, these towns and cities were abandoned.A period of agrarian expansion and self-sufficientvillage economy followed, lasting until A.D. 1000, when foreigntraderevivedunder the Arabs.R. S. Sharma,IndianFeudalism,c. A.D. 300-1200 (New Delhi: Macmillan, 2.
1965).
4. B. D. Chattopadhyaya,TheMakingof EarlyMedievalIndia (New Delhi: Oxford University Press, 1994),167. The period
between the sixth-seventh and twelfththirteenth centuries showed developments vastly differentfrom the society of the earlierperiod. "Stateformationwas a crucialagentof changein this respect,in the sense that it broughta measureof cohesion among local elements of culture by providingthem a focus" (35).While urban centerswerecharacterizedas centersof political power,surroundedby largeagriculturalhinterlandsand located along traderoutesin the earlyhistoric period, the characterof these urban centerschangedin the earlymedieval period into "nodalpoints in local 57
THE BEGINNINGS
exchangenetworks"(181). 5. AndreWink,Al-Hind: TheMakingof an Indo-IslamicWorld(New Delhi: Oxford University Press, 1990), 45- 64.
6. Wink, Al-Hind,72. 7. The site of Hathabwas the largest,located close to the sea with an areaof more than forty hectares,while the largestnumber of sites-i.e., eleven-fall in the category of three to twelvehectares.It is significant that while thirteen sites werelocated in the blackcotton soil zone, six were situatedclose to the coast. The site of Padri,situatedthree-four kilometers inland, was known for extractionof salt. Ashit BoranPaul, "TheEarlyHistoric Settlementand SubsistencePatternin the ShetrunjiRiverBasin,Bhavnagar District, Gujarat," Puratattva 30 (19992000): 99-105.
8. S. R. Rao,ExcavationsatAmreli(Baroda: Museum and Picture Gallery,1966). 9. R. N. Mehta,"Valabhiof the Maitrakas," Journalof theOrientalInstitute(Baroda) 13 (1963-64): 240-51.
lo. The office of the drangikawas peculiar to the Maitrkasand includeda numberof villages in its jurisdiction;e.g., three grantsreferto villages belonging to the Mandali-dranga.The termpetha has been found in one grant,which refersto Vatagramain Dipanakapettah in Bilvakhatasthali.The conclusion is that a pethawas largerthan a gramaor village but smallerthan a sthali.The occurrence of the termpetha in two copperplate grantsfrom CentralIndia and datedto A.D. 529 and 533indicates the Central Indian connections of Gujaratat this time. K. J.Virji,AncientHistoryof Saurashtra(Bombay:KonkanInstitute of Arts and Sciences, 1952), 234-35. i. Both literaryand inscriptionalsourcesof
the tenth and post-tenth centuries indicate a rangeof commodities traded such as grains,pulses, salt, oil, ghee,
cittena,translatedas "withvaried stripes of differentcolours"A. L. Basham,"The MandasorInscriptionof the Silk Weavers,"in Essayson GuptaCulture,ed. BardwellL. Smith (New Delhi: Motilal
textiles, betel leaves,arecanuts, coconuts, spices, incense, ivory,coral, gold, etc. V. K. Jain,Tradeand Tradersin WesternIndia (New Delhi: Munshiram Manoharlal, 1990), 56. 12.
GeorgeBuhler,"TheCintraPrasastiof the Reignof Sarangadeva,"Epigraphia
Banarsidass, 1983), 99. In their new place
of residence,the silk weaversadopted severalnew occupations. "Somebecame
Indica 1 (1892): 271-87, verses 49-57:
pastmasters in archery ... others became knowers [tellers] of varied tales ... while
He [Tripurantaka] ... presented ...
eight drammas[forthe purchase]of sandalwood.... The guild of the gardenersshall furnish daily 200 white roses and 2000 fragrant oleander blossoms. He assigned ...
six drammasfor the monthly expenditure in order to procure ...
a pure vedika.Two manakasof husked rice, one manakaof Phaseolusmungo,four karsasof clarifiedbutter and as much oil for the lamps. Fivebetel nuts of good quality ... two manas of fragrant
gum [guggulu]for the time of offering incense ... so leaves of betel
pepperfor the preparationof bitaka. 13. Jain, Trade and Traders,73.
14. The Anavadainscription (1291) refersto
gifts to a temple by the nau-vittakasor ship owners.IndianAntiquary,41:21. i5 These include sites such as Valabhi, Nagara,Dwarka,Somanatha,etc. i6. Dilip Rajgorrefersto finds of hoardsof Roman coins from Ahmadabadand Kera,but none of these are availablefor study.D. Rajgor,"RomanCurrencyin Gujarat,"in New DimensionsofIndology, ed. R. N. Mehta (New Delhi: Bharatiya
others, unassuminglyvirtuous, fully devoted to discoursingon righteousness [dharma] ... some remained excellent in
their (ancestral)craft, and others, knowersof the self (atma) mastered astrology"(97-98). 19. CynthiaTalbot,PrecolonialIndia in Practice:Society,Regionand Identityin MedievalAndhra(New Delhi: Oxford UniversityPress,2001), 87:"Moreover, temples offeredarenasfor the formation of new sociopolitical identities, and the patronageof specifictempleswas often motivatedby such factors.The patronage of templescan hence be viewed as both a symptom and a cause of the dynamism of medievalAndhrasociety: it both reflectedongoing processesand helped shapefurtherevolution." 20. P.N. Vasa,"Findof a LateByzantine Solidus and ArabUmayyadDinara from Kutch," Numismatic Digest 14 (1990):14, 30-33. 21.
22.
Vidya Prakashan, 1997),197-203. 17. R. D. De Puma, "TheRoman Bronzes
from Kolhapur,"in Romeand India:The AncientSea Trade,ed. V. Begleyand R. D. De Puma (New Delhi: OxfordUniversity Press,
1992),
101-2.
i8. The inscription describesdesigns woven by them on silks as varnantara-vibhaga-
Z. D. Ansari and M. S. Mate,Excavation atDwarka (Poona:Deccan College ResearchInstitute, 1966). V. D. Gogte, "Discoveryof the Ancient Port of Chaul,"Man and Environment 28.1 (2003):
23.
67-74.
Jain, Trade and Traders,1o9-16.
24. David Whitehouse, "SasanianMaritime Activity,"in TheIndian Oceanin Antiquity,ed. J.Reade(London:British Museum, 1996), 344. 25. Whitehouse, "SasanianMaritime Activity,"342. 58
HIMANSHU
PRABHA
RAY
26. Quoted in Whitehouse, "Sasanian
MaritimeActivity,"342. 27. S. D. Goitein,A MediterraneanSociety: TheJewishCommunitiesof theArab Worldas Portrayedin theDocumentsof the CairoGeniza,vol. 1,EconomicFoundations (Berkeley:Universityof California Press,1967),149-50,155. 28. MeeraAbraham,TwoMedievalMerchant Guildsof SouthIndia (New Delhi: Manohar,1988),1lo. 29. V. V. Mirashi, "TheInscriptionsof the Silaharas,"CorpusInscriptionum Indicarum,vol. 6 (New Delhi: 1977),i85. 30. J.G. De Casparis,SelectedInscriptions from theseventhto the ninthcenturyA.D. (Bandung:MasaBaru,1956),188-89. 31. While defining two types of sangha,or oligarchyorganization,the Arthasastra (11.1.4) places Saurastrain the first categorytogetherwith Kambojaas living by tradeand vocation of arms (vartas'astropajivinah). The authorof the text Kautiliyaidentifiesmembersof this type of sanghaas engagedin cattle rearingor tradesin peacetime, but taking to armswhen necessary.Trans.R. P. Kangle,KautiliyaArthasastra(Delhi: Motilal Banarsidass,1988),pt. 3,124. 32. Dilip Rajgor,Punch-MarkedCoinsof EarlyHistoricIndia (California:Reesha Books International,2001), 63-70. 33. The Ksatrapasareknown to haveruledin regionsas farapartas the northwest frontiersalongAfghanistanas well as in the northernplains atMathura.The rule of the WesternKsatrapasis generally datedbetween the second half of the first centuryA.D.and the beginning of the fifth centuryA.D.During this time they issued coins in silver,copper,and leadaswell as copperalloyedwith leadand arsenic. 34. Thus while the legend on the coins of Apollodotus reads:"BasileusSuteros Appolodotoy,"the coins of Nahapan.a carries"RannioSaharatacNahapanac."
AmiteshwarJhaand Dilip Rajgor,Studies in the Coinageof the WesternKsatrapas (Nashik:Indian Institute of Researchin Numismatic Studies, 1994), 25.
35. Jhaand Rajgor,Studiesin the Coinageof the WesternKsatrapas,73. 36. The epigraphis in Sanskritprose, which also characterizedsubsequentMaitrka epigraphs.The inscription refersto Rudradamana'sselection as rulerby all groups (varnas),his close alliance with Satakarniwho ruled over the Deccan, and his personalachievementsas a regionalruler.After defeatingthe Yaudheyas,he extended his reign over eastern and westernAkaravanti,Anupa, Anarta,Saurastra,Svabhra,Maru, Kachh,Sindhu-Sauvira,Kukura, Aparanta,and Nisada. EpigraphiaIndica
39. Three of their inscriptions from Andhau near Khavdamention the erection of a yasti or funeralmonument in memory of Satrunsasa,a monk (srdmanera). Epigraphia Indica 35 (1963-64): 190-92.
Another recordfrom Wandh refersto the erection of funeral monument in memory of the servantof the mother, mentioningAtimutagotraand the name of Kachhas Kasadesa.EpigraphiaIndica 37 (1961):109. See also H. D. Sankalia,
of Prehistoricand HistoricArchaeology Gujarat(New Delhi: Munshiram Manoharlal, 1987), 84.
40. H. P.Ray,Monasteryand Guild:CommerceundertheSatavahanas(New Delhi: OxfordUniversityPress,1986), 53. 41. Emile Senart,"TheInscriptionsin the Cavesat Nasik,"EpigraphiaIndica8
8 (1905-6): 36-49, no. 6.
(1905-6): 85-86.
37. Suvisakha,the son of Kulaipa,a Pahlava who was appointedin chargeof the province, carriedout the actual repairsof LakeSudarsana.CandraguptaMaurya had originallybuilt the lake in the fourth century B.C. and Tusasphaadded conduits to it in the reignof the Mauryan rulerAsoka in the third century B.C. 38. Fourof the Andhau inscriptions are on long, narrowslabsof stone and in several cases are inscribedlengthwise. The purpose of all four is to recordthe erection of funerarymonuments by relativesof the deceased. R. D. Banerji, "TheAndhau Inscriptionsof the Time of Rudradaman,"EpigraphiaIndica16 (1921-22): 19-25. Eightothers areknown from the Buddhistmonastic sites of Nasik, Karle,and Junnarin Maharashtra. R. Jamindar,"SomeObservationsfrom the KsatrapaEpigraphsfrom Kaccha,"in Art, Cultureand NaturalHistoryofKutch, ed. S. K. Bhowmik (Vadodara:Department of Museums, 1976-77),
EpigraphiaIndica16(1921-22): no. 17.
92.
See also
238-41,
Senart," Inscriptionsin the Cavesat Nasik,"95, no. 26. 43. LionelCasson, ThePeriplusMaris Erythraei(Princeton:Princeton UniversityPress,1989), 81. W.W. 44. Tarn,TheGreeksin Bactriaand India (1938;reprintChicago:Ares Publishers,1984), 361-62. 45. Tarn,Greeksin Bactriaand India, 227. 46. R. E. M. Wheeler,A. Ghosh and Krishna Deva, "Arikamedu:An Indo-Roman TradingStationon the EastCoast of 42.
India," Ancient India
2 (1946): 17-124.
47. Nancy Pinto Orton, "RedPolishedWare in Gujarat:A Catalogueof TwelveSites," in Romeand India:TheAncientSea Trade, ed. Vimala Begleyand R. D. De Puma (New Delhi: OxfordUniversityPress, 1992), 46.
48. H. P.Ray,"TheYavanaPresencein Ancient India,"Journalof theEconomic and SocialHistoryof the Orient31(1988): 311-25.
49. The grammarian,Panini, a residentof Salaturain the vicinity of Taxilastates,in his Astddhyayi(grammaticaltreatise) 59
THE BEGINNINGS
that the feminine form of yavanais
yavnani(4.1.49). This form, accordingto Katyayana,denoted Greekwriting (Varttika3 on Panini 4.1.49) and is furtherparaphrasedby Patanijaliin the middle of the second century B.C.Pan.ini also associatedthe ydvanaswith the kambojasand describedthem as having shavenheads (Panini Ganapatha178 on 2.1.72: yavanamuaand kambojamunda). Ydvanaswereregardedin the law books as degenerateksdtriyas (Manu10.43-44;
Mahabharata santiparva 12.200). Milindapaniha(1) atthi Yonakandm ndnaputabhedanamSagalanndma nagaram.... In the Girgi Samhitd,a work on astrology,a section named the Yuga Purana,contains some historical information, including the story of the attackon Saketaand Pataliputrain which the yavan.astook part. The association with the northwestcontinues in the VisnuPuranadatedto the latterhalf of the firstmillennium A.D., which refersto the yavanasas living in the west of the country.The VisnuPurana:A Systemof HinduMythologyand Tradition,trans. H. H. Wilson (reprint,Calcutta,1972) 11,3, 141-44. 50. R. S. Avasthyand A. K. Ghosh, "References to Muhammadansin Sanskrit Inscriptionsin Northern India, A.D. 730 to 1320,"Journalof IndianHistory15 (1938):161-84;16:24-26,171. 51. EpigraphiaIndica34(1960-61): 197-203. 52. EpigraphiaIndica32(1957-58):45-60. 53.Mirashi,"Inscriptionsof the Silaharas," 115-20.
54. J.F.Fleet,"JunagadhRockInscriptionof Skandagupta,"CorpusInscriptionum Indicarum3(1886):no. 14. 5 Bhattarakais mentioned as the founder of the Maitrkadynasty in inscriptions, though none of his own recordsare known. He is creditedwith the acquisition of royalpower (rajyasri)through the
princes (raja-putra),palaceofficers (rajastrength of his hereditaryservants, sthaniya),ministers, city officers friends, and guilds (maula-bhrta-mitra(drangika),headmen (mahattara), sreni).The phrase used in the inscripirregularsoldiers,spies, regularsoldiers, tions is borrowedfrom the Arthasastra riderson elephantsand horses, etc. "E. (maula-bhrta-sreni-mitrdmitra-atavibala Hultzsch, PalitanaPlatesof Simhaditya: 9.2.1) and is an indication of armed The Year 255," Epigraphia Indica 11 (1911soldiersmaintained by tradinggroups.E. The same list is repeatedin 12): 16-20. Hultzsch, "GaneshgadPlatesof DruvaseHultzsch, "PalitanaPlatesof Dharasena na I,"EpigraphiaIndica3 (1894-95):318II:GuptaSamvat252," EpigraphiaIndica 23; Sten Konow,"PalitanaPlatesof 11 (1911-12): 80-85. Palitanaplatesof DhruvasenaI,"EpigraphiaIndica11 II (year252 or A.D. 571) record Dharasena "A N. G. Majumdar, (1911-12): 105-15; in the village of Ekalika. of land grants the of Passagein the Inscription History of Saurashtra,245. Virji, Ancient 6o. Maitrakasof Valabhi,"IndianAntiquary and M.S. University Thakkar, 61. Mehta 48 (1919): 207-11. CopperPlates. 56. Dandin, Dasakumaracaritaof Dandin, 62. D. C. Sircar,"Charterof Visnusena ed. M. R. Kale(Delhi: Motilal BanarsiSamvat649," EpigraphiaIndica30 (1953dass, 1986),164,332. One of the stories in the text is set in the city (nagari)of 54):163-81, no. 30. 63. Sircar,"Charterof Visnusena." Valabhiin Saurastra,wherethe chief of 64. Norman A. Stillman, "TheEleventh sea traders(navika-pati)lived. He CenturyMerchantHouse of Ibn 'Awkal possessed immense wealth, like Kubera. (A Geniza Study),"Journalof the 57. R. N. Mehtaand A. M. Thakkar,eds., M. Economicand SocialHistoryof the Orient S. UniversityCopperPlatesof the Timeof 16 (1973): 15-88. Toramana(Vadodara:MS University, 65. Sircar,"Charterof Visnusena,"178. 1978). 66. The legend reads:maharaja-rudrasena58. L. D. Barnett,"BhamodraMohota Plate D. Mitra, vihare-bhiksusanghasya. of Dronasimha,"EpigraphiaIndica16 of Sahitya plates (Calcutta: 17-19. The Navalakhi BuddhistMonuments (1921-22): Sansad,1980), 141. B. Ch. Chhabra, SiladityaI (year286 or A.D. 6o5) found in a village about eight-ten miles from Findingsin IndianArchaeology(Delhi: Jundgarhrecordsthe gift of a village to SundeepPrakashan),181-83; J.Ph. Vogel, "Sealsof BuddhistMonasteriesin forty-fourbrahmanaswho had moved from Sangapuri(identifiedwith the Ancient India,"Journalof the Ceylon presentShahapurnear Junagarh).H. M. Branchof theRoyalAsiaticSociety, Bhadkamkar,"NavalakhiPlatesof centenaryvolume, n.s., 1 (1945): 27-32. 67. R. N. Mehtaand S. N. Chowdhary, SiladityaI, EpigraphiaIndica,"1i (19l12):174-80. The names of the brahmanas ExcavationatDevnimori (Baroda:MS are curious. Some look like family names, Universityof Baroda,1966). while others are names of individuals. 68. The 7.14 hectaresite with a 3.2 meter thick The name Boppasvamimay show that he habitationaldeposit has provideddata for was very likely a Tailangbrahmana(175). three culturalperiods. PeriodI is early 59. The Palitanaplatesof Sinhaditya(year B.C.), Period II is Harappan(3000-2600 matureHarappan(2500-1900 B.C.) and 255or A.D. 574)referto severalpeople who were addressedregardingthe transferof the third period is earlyhistoric (first land to a brdhmana.These included century B.C. to firstcentury A.D.). The 6o
HIMANSHU
PRABHA
RAY
third period has yielded a rangeof ceramicssuch as redware,red slipped wareand graywaretogetherwith bangles,beads, and terracottaobjects. Ashit BoranPaul, "AStudyof the MaterialRemainsof the EarlyHistoric PeriodDiscoveredat Padriof Bhavnagar District, Gujaratin India,"Pratnatattva (Journalof the Departmentof Archaeology, JahangirnagarUniversity)6 (June 2000):53-66. 69. V. S. Shinde, "TheEarliestTempleof The RecentEvidencefrom Lajjagauri? Padriin Gujarat,"Eastand West(IsMEO) 44.2-4 (1994):481-85. 70. These include Kanadand Variavin Surat,
Kaviin Bharuch,PavtiJetpurin Baroda, Tarsangand Bavakain Panchmahal, Valabhipurand Gagdhaliin Bhavnagar, Amreli in Amreli district, Dhanakin Rajkot,and Sathalin Ahmadabad.V. H. Sonawane,"SomeRemarkableSculptures of Lajjagaurifrom Gujarat,"Lalit Kala:Journalof OrientalArt,23(197778):27-35. 71. "TwoNew ValabhiCopperPlates,"
Journalof theBombayBranchof theRoyal AsiaticSociety,20 (1897):1-1o. These copperplatesalso referto a temple at the site. 72.Virji, AncientHistoryof Saurashtra,33. 73. G. Buhler,"AValabhiGrantof SiladityaI, Dated 290," IndianAntiquary9 (1880): 237-39. 74. R. D. Banerji,"TheBhadreniyakaGrant of SiladityaI, GE292,"EpigraphiaIndica 21 (1931-32):116-1g, no. i8. 75. Mehtaand Thakkar,M.S. University CopperPlates,23. 76. M. A. Dhaky and U. S. Moorti, The Templesin Kumbhariya(New Delhi: AmericanInstitute of Indian Studies, 2001).
77. Buddhistvihiras in Gujaratas recorded in Maitrakainscriptions: 1. Dudda vihdraor vihdramandala
nearValabhi(copperplate grantof DhruvasenaI (A.D. 535).It included Bhatarakavihara,Gohakavihara, Abhayantrikavihara,Kakkavihara, Buddhadasavihdra,Vimalagupta vihara,and Sthiramativihara. 2. YaksaSuiraviharabuilt by Yaksa Sara,which formed the head of the monasteriesfor nuns. It included Parn.a-Bhattavihdraand Ajjita vihara. 3. BappapadiyaviharanearValabhi built by AcaryaBhiksu Sthiramati for monks from foreigncountries: nanddesapranatvagatarthyabhiksus -anghasya. 4. Vamsakataviharabuilt by SiladityaI. 5. Yodhavakavihdrawith waterreservoir. 78. Diskalker,"TenFragmentsof Stone Inscriptionsand a ClaySeal from Vala," Annalsof theBhandarkarOriental Research Institute20.1 (1938-39):i-8. 79. A. S. Gadre,"FiveValaCopperPlate Grants,"Journalof the Universityof Bombay3(1934):79. 8o. Sanghaksetra,sthavirabrahmadeya khetra,lines 36-39, Journalofthe Bombay Branchof theRoyalAsiaticSociety,n.s.i (1841):50-53; Bavyasthavirakaksetra (line 49 in "Grantof DharasenaIV," IndianAntiquaryi (1872), 14-18; and satka ... Koutumbaksetra(line 59 in "Grantof SiladityaIII,"EpigraphiaIndica 4 (1896-97):74-81. 81. "AGrantof KingDhruvasenaI of Valabhi,"IndianAntiquary4 (i875): 104-7. 82. "AGrantof King Guhasenaof Valabhi," IndianAntiquary7 (1878):66-68; Indian Antiquary4, (1875):174-76;Epigraphia Indica13(1915-16):239. 83. "AnUnpublishedGrantof King Dhruvasena I,"Journalof theRoyalAsiaticSociety,n.s., 1 (1895):379-84. 84. D. R. Bhandarkar,"Platesof Dantivarman of Gujarat,"EpigraphiaIndica6
(1900-1901): no. 28.
85. Jain,Tradeand Traders,chap.1. Another noncanonical Jainatext significantfor this paperis the Vasudevahimdiin two parts;the firstpart dated to the sixth centuryA.D. and the second a few centuries later.A. P. Jamkhedkar, Vasudevahimdi:ACulturalStudy(Delhi: Agam Kala Prakashan, 1984), 8.
86. Anita Sharma,"TradingCommunities, Social Mobility under Religious Patronage:Aspects of EarlyMedieval WesternIndia from VIII-XVth Century,"(MPhil diss., JawaharlalNehru University, New Delhi, 1996).
6i
THE BEGINNINGS
1
MARK HORTON
ARTISANS, COMMUNITIES, AND COMMODITIES MedievalExchangesbetweenNorthwesternIndiaand EastAfrica
I
Decoratedplasterniches in a house in Lamu,eighteenth or nineteenth century,that were used to display imported pottery and glass. Photographby the author.
Abstract Studiesof the medievalIndian Ocean have long been directedtowardnarratives of its regions,with an implicitnotion that they maybe connectedthrough trade.But there is little real understandingof the complexand changinginterrelationshipsacrossthe ocean.This articleexploresthe links betweentwo areas, EastAfricaand northwesternIndia,andbringstogetherthe archaeological,oral, and documentaryevidenceto suggesta complexinteractionbetween the two areasfromthe eleventhuntil the fourteenthcenturies.In particular,this article arguesthrougharchaeologicaldatathatmerchantsand middlemenwerenot the only groupstravelingthe seawaysand overlandroutes.Communitiesof artisans also traveledlong distances,and in fact sometimessettledat theirpoints of disembarkation.
STUDENTSOF THE ARCHAEOLOGY of the East African coast recognize that long-distancetradingconnectionswith westernAsiawerea centralfeature behind the developmentof the Swahiliculture.,The towns of the Africancoast wereessentially"portsof trade"into whichwerecollectedthe rawmaterialsand commoditiesof the Africanmainland(ivory,slaves,skins,aromaticwoods,timber,resins,minerals,and gems),whichwereexchangedfor manufactureditems carriedby shipsfollowingthe monsoonwinds fromthe portsof the PersianGulf region.Thisrelativelysimplemodelof long-distancetradeis supportedbyhistorical accountsin the Arabicgeographies2 and by finds of importedpottery,beads, and glassfromexcavatedstructuresand rubbishpits;indeedthe proportionand rangeof importsfoundin theseAfricansitessuggestthatimportedmaterialculture was a ubiquitousfeatureof the lifewaysof these communities.The bulk of these findscamefromthe PersianGulfportsor theirhinterlands.This "bilateral trade"with the gulf andits peripheriesoriginatedin pre-Islamictradingsystems, and has continuedinto the contemporaryperiod,with a dhow tradein timber mangrovepolesthatfinallyceasedin the 198os. However,this model of bilateraltradebetweenEastAfricaand the ports of westernAsia maybe misleadingfor a numberof reasons.A reassessmentof the mercantileorganizationof the westernIndianOceanis underwayincorporating recentarchaeologicaldiscoveriesfromvariousEastAfricanports.3Thesearchaeological discoverieshave prompteda realizationthat the documentarysources introducea significantbiasastheyoriginatelargelyin the PersianGulfregionand naturallydwellon tradefromthere.Otherareasof the westernIndianOceanare less well covered,includingwesternIndiaas well as the southernroutethat may havelinkedSouthandSoutheastAsiawith the Africancoast. 63
It is very difficultto reconstructcomplexnetworksfrom the archaeological evidencealone,as the presenceor absenceof artifactsdoes not providedefinitive indicationof traderoutes.Forexample,Africancommoditiesmayhavereached westernIndiavia two legs, one along a sea routebetweenthe PersianGulf and EastAfrica,andthen pickedup by shipssailingbetweenthe RedSeaandwestern Indiapassingvia Aden,Hormuiz,or Kish;4it is known that therewereconsiderablequantitiesof Africanrawmaterialsin thesetowns. So a commoditysuchas ivorycould reachIndiafrom EastAfrica,but the items that reachedEastAfrica in exchangemay not be specificallyIndian.Indeed,they could havecome from anywherewithin the tradingregion.ThusIndiandiagnosticartifactsin the East Africanassemblagesmight not be expectedeven though tradewith India was extensive. It is also difficultto gaugethe scaleof the trade.When the quantitiesof viable imports to East Africa (mostly glass and pottery,from aroundthe Indian Ocean)arecomputedfromarchaeologicalexcavations,the actualnumberof vessels involvedis comparativelysmall.The proportionof importedglazedpottery foundat EastAfricansitesbetweenthe eighthandfourteenthcenturiesis around 5 percentof the total ceramicassemblage,with some variationovertime, place, and context.5However,whenthis proportionwas calculatedat the excavatedsite of Shangain termsof actualnumberof vesselslost in an entiretown,only140vessels werelost per year duringthe fourteenthcentury,fewerthan one vesselper householdperyear.In the eleventhcentury,onlytwentyvesselswerebeinglostper year,and in the ninth centuryonly fivevessels.6As verylittle importedpotteryis foundoutsidethesetownsites,thesedatesgivea realisticindicationof the import rate-hardly enoughto payforthe hugequantitiesof ivoryand othercommodities thatwereapparentlyleavingthe Africanports. The notion of an exchangein the formalistsense,7as envisagedby some economic anthropologists,maybe misleading.The importedpotteryand glassthat is foundmayhavebeena gift exchangeto cementrelationsbetweentraders;these were socially embeddedcommoditiesthat have symbolic meaningsabout the prestigeandstatusof the tradingclass.Theactualexchangemayhaveusedlessvisibleitems,suchas cloth (whichnormallydoesnot survivein wettropicalregions), foodstuffs,or evencoinage,whichtendsto be conservedandrecycled.Supportfor the ideathatthe importedpotteryandglassweregiftswith socialmeaningcomes from the coastalarchitecture,whichincorporatesnichesspecificallydesignedto displaythese importeditems in merchants'houses (fig. 4)8 Glazedpotterywas also built into walls to decoratetombs and the vaultsof mosquesand mihrabs. Importedpotteryandglasswereseenas prestigiousgifts (andIslamicglazedpottery was laterreplacedby Chineseporcelainsand by the nineteenthcenturyby 64
MARK HORTON
WesternIndian Ocean between the eleventh and fourteenth centuries.
.
Fust
pX4iraf
al-Qaili;
0l L
ofthese gifts cod >
X
[ he
fR
ce
_ an
>
r w d I I|
1000
500
M
l inbho~~~re-Daybul hambhet
st t
I
-
B
~~~~~~INDIA
AFRICAI|I|III_
MalmInII|I|I-_
M osab
I
m
b
_.
Potr .
thawhr tegoswraculymd
an gls.
r
o drcnia
zUFALA
Sofaa , 65
Dtch
> _
ARTISANS,
an
Sctts
G
= COMNTE,
_ AN
l|
k k U^ ~~~~~~~Unguja
t
COMODIIE
udrlaepitdwrs,nttaetmsere.Teoos
Dtchandwer Sotihe Poteyem goodsergae actallye wae)mnttade.
gas
pers.Te noietidonors
tors of trade but denote membership in a common Indian Ocean culture.9 Perhapsthe most fundamental problem for both historians and archaeologists is distinguishing between commodity exchange and the movement of artisans. If the Indian Ocean is viewed as a "Braudeliansea,"then its cultures were interlinked through both trade and settlement; communities of artisans were likely to be as 65
ARTISANS,
COMMUNITIES,
AND COMMODITIES
2
2
Bronze lion figurine discovered in an eleventh-century pit at Shanga, 59mm high. Photographby the author.
mobile as the commodities they were making.10By concentrating simply on commodities, the existence of artisan communities working at great distances from their homeland, is rarely considered. Yet when, with European contact, detailed historical descriptions of the ports around the Indian Ocean rim became available, the port towns were already complex multicultural entities. The origin of particular artifacts maybe ambiguous, and possiblywas intended to be so. An example is an eleventh-century bronze lion figurine, found at Shanga, and apparently modeled on the African rather than Asiatic lion but stylistically very close to a range of brass lion figurines from Hindu contexts in western India (fig. 2). It is likely that Indian artisans, resident in East Africa, made the lion figurine; they may also have introduced lost-wax casting."1 The Shanga lion may be part of a wider pattern of artisans from around the Indian Ocean working in African communities and presumably over time, becoming fully assimilated into the local culture. Such artisans might have included stonemasons carving the twelfthcentury Kufic inscriptions found at Kizimkazi and Tumbatu or the bead makers producing carnelian and rock crystal beads.
EastAfricaand NorthwesternIndia:DocumentaryEvidence Documentary evidence for contacts between northwestern India and East Africa is very slight indeed. Al-Mas'uidi,whose account is reliable as he visited East Africa in 916, noted that ivory was one of the main exports; the tusks went first to Oman and "from there are sent to China and India."12In India ivory was much sought after. Al-Mas'uidigave no indication, however, as to the Indian ports at which the ivory arrived. A little later, al-Biriani (ca. 1030) gave the first specific indication of an East Africa-western India connection by commenting that the port of Somanatha became famous "because it lay between Zanj and China.'' This comment may again have referredto an ivory trade, which was the major African export of this time and one of the main items that the Persian Gulf merchants used to acquire Chinese goods in their famous "China trade." Al-Idrisi (iioo-ii66), who never visited East Africa and whose descriptions can be confusing, suggested there was a trade in African iron to India, "where it carries a good price, for it is the object of great commerce and consumption in the 66
MARK HORTON
Indies,andalthoughit existsin the islesandminesof this country[India],it is not as good as the iron of Sufala,eitheron accountof its abundance,goodness,malleability.''14 Al-Idrisiwasnot specificregardingthe areaof Indiato whichthe iron wasbeingtransported,buthe does referto ironminesin Sufala(whichrepresents the southernpartof the EastAfricancoast)as well as farthernorthnearMalindi and Mombasa.Recentworkon EastAfricaniron suggeststhat indigenousfurnaceswerecapableof producingsteelof a high quality,in a traditionthatextends backto the earlyIronAge.Fromexcavatedsites,thereis muchevidenceforsmithing andsmeltingbetweenthe tenthandtwelfthcenturies,althoughsmeltingmay havealsotakenplacein the interior.'5 Fromthe twelfth century,thereis evidenceof Indiancommunitiesliving in the port cities of the IndianOcean.Indianmerchantswererecordedat Kish (at the entranceto the PersianGulf) in 1170,a port thathada particularlycloserelationshipwith EastAfricaat this time, and seemsto havebeen the transhipment pointforAfricanslaves,ivory,crystal,andambergris.'6 Twelfth-century Kishwas a very diverseport, with Arabs,Jews,Indians,and Persiansall prosperingfrom this Africantrade.Indeed,so importantwas this Africantradeto the merchants of Kishthat the rulerof the town led a navalattackon its rival,Aden,to retain controlof it.17 It is only at the beginningof the sixteenthcenturythat we havespecificdocumentaryevidencefor residentIndiancommunitiesin EastAfrica,specifically at the ports of Malindi,Mombasa,Kilwa,and Pate.'8Vascoda Gama'sexpedition of 1498located Indians residentin Malindi and Mombasa,but wrongly identifiedthem as Christiansratherthan Hindus.He had no difficultyin finding a pilot who could guide his small fleet to westernIndia from Malindi to A little laterTomePires (ca. 1512-15)describedthe main exportscarCalicut.19 ried in ships from Khambhatand Aden to Kilwa,Malindi,and Mogadishuas rice, wheat,soap, indigo, butter (and lard),oils, carnelians,and coarsepottery ("likethat from Seville"),and all kinds of cloth.20 DuarteBarbosa(ca. 1517-18) noted the great profits made in East Africa by merchantsfrom Khambhat,21 dealing in ivory, gold, and ambergris.By the sixteenth century,East Africa reliedon cloth from Khambhat;as one Portuguesewriterput it, "Allthis coast dresses in these cloths and has no others."22 The scale of the cloth trade was
clearlysubstantialon the eve of the Portuguesearrivalin the Indian Ocean. The documentarysourcesare silentas to whetherthis was a tradethat grewup in the fifteenthcenturyor was part of an ancientpattern. Thereare difficultiesin fillingthe periodspanningthe twelfththroughearly sixteenthcenturies.Ibn Batttita,who visited EastAfricain 1331,left a long and detaileddescriptionof the regionbut did not makea singlereferenceto tradeor 67
ARTISANS,
COMMUNITIES,
AND COMMODITIES
contactwith India.The elaborateclothesthatthe rulerand courtiersof Mogadishuworeweremadefromsilk,Jerusalemstuff,andEgyptianlinen;therewasalso a flourishinglocal cloth industry.23TheArabgeographerAbulal-Fida'(1273-1331) referredagainto the iron mines of Malindibut did not mentionthe tradein iron to India.24Takingall this materialtogether,most scholarshave concludedthat therewaslittlecontactbetweenEastAfricaandIndia,exceptperhapsa short-lived tradein iron,untilthe developmentof the clothtradein the fifteenthcentury. OralTraditionsand Chronicles Perhapsmorerevealingaboutthe connectionsbetweenEastAfricaand Indiaare the oral traditionsand chroniclesof the Swahilitowns. Most prominentarethe storiesaboutthe Debuliorsometimesthe Dibapeople.TheSwahilihavelittleidea who these peoplewere,when they arrived,or wherethey came from,but these peopleplayan importantpartin the originsof a numberof coastalsettlements, rangingfromZanzibar,Pemba,Kilwa,to the Lamuarchipelago,rememberedas A casecan be madeon phoneticgroundsthatSwahilinames theirfirstlandfall.25 of Debuli/Dibain the oral traditionsrepresentthe port city of Daybul/Dibalin northwesternIndia. Storiesabout the Debuli referto earlytimes in Swahilihistorybut are then confusedby the arrivalof latergroups.In one version,for example,the Debuli werewreckedoff KiwayuIsland,just to the north of the Lamuarchipelagoand They sailed down the coast in mitepe intermarriedwith the local population.26 (a local type of sewn boat, often associatedwith the early settlementof the coast) and built stone houseson Pemba,Zanzibar,Mafia,and Kilwa.On Pemba they are rememberedas a violent people, who forcedthe populationto labor carryingheavy stones for their new buildings.On Zanzibarthey were "cruelly At Kilwa,traditionsplace the Debuli as "rulusing men as beastsof burden."27 ing"beforethe Shirazidynasty,which may dateto the late eleventhand twelfth centuries.228
While it is very likely that Debuli/Dibarepresentsthe port city of Daybul/ Dibal,locatedat the mouth of the Indus,other suggestionshavebeen madefor alternativelocations.29Controversyalso surroundsthe preciselocationof Daybul, although one candidateis Banbhore,where excavationshave unearthed Banbhorewas probablydesertedby the the remainsof a substantialmosque.30 twelfthcentury,althoughthe placenameDaybulseemsto havebeentransferred to Thatta,whichis alsoin Sindhandis sometimesknownas Debal-Thatta.Given that the main prosperityof Daybullies beforethe twelfthcentury,theseAfrican traditionsmaywellbe accurateanddo indeedreferto a pre-Shiraziperiodof Swahili history.It is unlikely,however,thatlargegroupsof Daybulicitizensactually 68
MARK HORTON
"settled"in EastAfrica.It is moreprobablethatDaybulwasa placewellknownon the coast,presumablythroughtrade,and may havebeen employedas a placeof mythic origin.Africanorigintraditionsare complexand generallynot literalin character.3' Similartraditionsaboutthe "origin"of anotherAfricangroupcoming fromShirazcan alsobe explainedin thisway.32 ArchaeologicalEvidence While these oraltraditionslackprecisionand havethe potentialfor severalconflictinginterpretations,there remainsome interestingarchaeologicallyattested links amongEastAfrica,Banbhore,andothersitesin the Indusdelta. Ceramics Fora numberof years,findsfromexcavationshavesuggestedthatcontactbetween EastAfricaand northwesternIndiawas more extensivethan historicalsources suggest;pottery,beads,cloth and coins providespecificindicationsof theseconnections. The most ubiquitousevidenceis ceramic.Indianpotteryis found at numeroussites in EastAfrica,datingfrom the sevenththroughfifteenthcenturies. EvenearlierIndianceramicshavebeen foundat the exceptionalsite of Ras Hafunin northernSomalia,whichhasIndianpotterydatingto the second-third centuriesA.D.33The main publishedaccountsof IndianpotteryarefromManda, Kilwa,andShanga.34 Becausethe sherdsare stratified,they can providea reliablechronologyfor contact between India and Africa. Scientificallyexcavatedand dated parallels from Indiaare extremelyillusive,however,and some of the materialmay come from the Deccan or even farthersouth. Attributionto India is largelythrough fabricand form,whichis verydistinctiveto both EastAfricanpottery,as well as unglazedwaresfromArabiaand the PersianGulf.Fourgroupsof Indianpottery haveso farbeenidentified: Grass-temperedgray ware
Grass-tempered graywaresareessentiallywaterjars,knownas chatties,with narrow-neckedmouthsand elbowrims often with exaggeratedprofiles.Thepasteis gray-strawbuff,verysoft, with manyairholes,causedby the burningout of chaff and vegetation.Some of the watervesselscan be massivein size. The jarswere probablymadeon the slowwheel.Thepotteryis foundin stratifiedcontextsfrom the eighthcenturyonward,with examplesfromMandaand UngujaUkuu,but is commonbetweenthe eleventhandfourteenthcenturieson numerousothersites. Theshapeandformsofthesewaresareextremelylonglived,andsimilarwaterjars arestill madein Gujarattoday. 69
ARTISANS,
COMMUNITIES,
AND COMMODITIES
7
A
B
C
D
~~~~~E
0 MM
100
200
3
3
Decorated Indian red warepottery found at Shanga,dating from mideleventh to fourteenth centuries.
Grog-temperedmarronware Grog-tempered marron wares have a distinct purple paste with a gray core and some airholes along with grog inclusions. They are found not only in East Africa but also along the Persian Gulf and the Red Sea, where they are termed "purple wares."35The forms include large jars and carinated bowls all dating from the twelfth-fourteenth centuries. The rims have a simpler form than the gray wares, and the vessels are better made. Red-slippedorangeware Red-slipped orange wares have a very distinctive orange-red to buff pink fabric, with flaky fracture and striated airholes and very thin vessel walls. They are coated in a red haematite slip, which is often worn away.They are fine jars, with narrow necks and often exaggerated rims. These wares may have their origin in the Indian red-polished wares but occur in east Africa only from the tenth to fourteenth centuries. They were found in particularly large quantities in excavations on the Zanzibari island of Tumbatu, where they represented 20 percent of the imported assemblage in the thirteenth century. Decoratedredware Decorated red wares are the only decorated Indian ceramics and have been found exclusively at Shanga, where in certain phases they constitute up to 30 percent of the imported assemblage. The fabric is a red to orange-red paste, with white flecks as a temper, and the surface is often burnished smooth to a red or reddish brown color. Decoration includes incised lines and punctuates in bands and panels. The forms are largely bowls and jars, and the rims often have complex forms. These can be quite large vessels, with rim diameters of between 15 and 20 millimeters. They date from the mid-eleventh through the fourteenth centuries (fig. 3). 70
MARK HORTON
Thepresenceof theseclassesof Indianpotteryatboth EastAfricanaswell as Persian Gulf sites is unexplained.At certainsites, in certainphases,it is relatively common,but at othersit runs at only about3 percentof the total importassemblageand abouto.i percentof the total potteryfound.36 While thereis an early presenceof Indian ceramicsbetweenthe seventhand tenth centuries,the bulk of importationspansthe eleventh-fourteenthcenturies.Surprisingly,it becomes rareby the end of the fifteenthcentury,when the Indian cloth tradewith East Africawasat its height. TheIndianwaresfoundin EastAfricaareallveryutilitarianandcontrastwith the fineglazedwaresthathavebeenfoundandarediscussedaboveaspossiblegift exchangesbetweenmerchants.Mostarewaterjars,and it is possiblethey arrived by boat, carryingwaterfor the voyageacrossthe ocean. Some came ashoreand may have been replacedby other,possiblylocally madejars for the homeward journey.This explanationwould indicatea directmaritimeconnectionbetween northwesternIndiaandEastAfricaforwhichthereis otherwiselittle evidence. However,as the rangeof Indianpotterytypesalsoincludesbowlsandcooking vessels,it is possiblethatthe potterywasownedbyIndiancommunitieslivingand workingin EastAfrica.Thesecommunitiesmayhavebroughtpotterywith them or soughtto haveit carriedfromIndia;TomePiresnoteda tradein coarsepottery in the earlysixteenthcentury,37 and culturalreasonsmaybe the best explanation as to why this very inferiorpotterymovedseveralthousandsof miles acrossthe IndianOcean.It may providea signaturefor an Indianpresencein EastAfrica, whichbecomesmorelikelywherethe proportionsriseto aboveio percentof the importedassemblage,as occursin certainlevelsatShangaandTumbatu. Beads An Indianconnectionwith EastAfricais best demonstratedin the manufacture and tradein beads.Beadsare a relativelycommon find at EastAfricansites and seem to havebeen one of the significanttradingitemswith the Africaninterior. Beforethe tenthcentury,coastalsitesmadetheirown beadsfromshelland semipreciousstones;glassandstonebeadswerea relativelyrareimport.Thelocalshellbead industryused the marinegastropodAnadara,and largenumbersof bead grindershavebeen found in levelsup to aroundgoo. Thesebead grinderscomprise discardedsherdsof pottery (both locallymade and imported)and sometimes beach pebbleswith numerousgroovesworn into their surface.The shell beadsthemselvesarefairlyrare,suggestingthatthe indigenousinhabitantsof the coastmadethemmostlyfortradepurposeswith the interior.38 Glassbeadsdo occurin verysmallnumbersin the samelevelsasthe shellbeads and beadgrindersandwereprobablyimportsfromwesternIndia.Theyaregen71
ARTISANS,
COMMUNITIES,
AND COMMODITIES
erallywoundoblatebeads,usingyellowor,morerarely,a cobaltblueglass.39 Less commonarefoldedbeads,with two colorsof glass,anddrawnbeads.Thesebeads tend to be largerand muchbettermadethan glassbeadsfromafterthe eleventh century. Around1050 therewas a transformationof the beadindustry,and glassbeads become very common. From stratifiedand sieved excavationsat Shanga,40 46 beadswerefoundin phase12 (ca.1050), 71 in phase13 (ca.1075),peakingat 146 in phase15(ca.1200), anddecliningto only13beadsin phase20 (ca.1375).Thesebeads arethe mass-produced"trade-wind" beads,whichareeitherwoundor drawn(at Shangathesewerein roughlyequalproportions).Thewoundbeadsarelenticular, spheroid,bicone,andoblatein shape,andthe colorsareyellow,green,blue,black, turquoise,red,andwhite;the proportionsof shapeandcolorchangedovertime. The site of Mkokotonion Zanzibarseemsto havebeen a centerfor eitherthe importor manufactureof glassbeadsduringthe thirteenthand fourteenthcenturies.41Largenumbersof canebeads,alongwith someglasswaste,arestill found in the sand along some five hundredmetersof beach,wherethey seem to have erodedfrombeachsideoccupationdeposits.42 Twotest pits, dug in 1989 through these deposits,suggestthat thereare severalmillion glassbeadsstill presenton this site -a concentrationfarlargerthanwouldbe expectedon a settlementsite. The moltenglassfoundwith the beadsmayindicatethattheywere"finished"on theAfricancoast,havingbeenimportedasrodsorevenglassrawmaterial.43 Mkokotonilies closeto Tumbatu,whichhada highproportionof Indianpotteryin its thirteenth-fourteenthcenturylevels;from Mkokotoniitself, coins from South Indiahavealsobeendiscoveredalongwith two hoardsof Chinesecoppercoins.It seemsthatthe sitewasa hugedepotforthe finishingoff, sorting,anddistribution of Indiantrade-windbeads. The stonebeadsfound in EastAfricamay be importsor locallymade,using IndiantechnologiesorevenbyresidentIndianartisans.Thecarnelianbeadindustry of Khambhatis well known;Piresnotedthem as tradeitemsto EastAfricain the earlysixteenthcentury.44 Carnelianbeadsoccurredthroughthe sequenceat Shangain small quantities,with little statisticalvariationin importrate.45One group,however,werewasters,discardedafterthe holesthroughthe beadhadbeen incorrectlybored.While these could havebeen dud beadsimportedfrom India, it is possiblethat theywerewastersfrom carnelianbead-makingactivityin East Africa,as theyarecruderough-outsthatwereneverfinishedandpolishedbutdiscardedafterthe hole boringhadfailed.If this is the case,thesebeadssuggestthe presenceof artisanson the coastwho had accessto the complexdiamonddrilling and polishingtechnologyneededto make the stone beads (fig. 4). The raw material,a quartzchalcedony,is readilyavailablein EastAfrica.Workin other 72
MARK HORTON
AI~~~~~~
.
X3w {;,
'A-4.r'^'
VII
;<.:
A
A
,,4
*~~~)
'
)
> 'at
<
''
N
,ffi,3~~~~~~~~~~~~~~~a
4
The remains of small workshops or kiosks at Shanga,found in front of a largestone building dating to around looo. Photographby the author.
quartzstoneis also evidencedalongthe coast.In particular,rockcrystalseemsto havebeen an importantexport,supplyingthe famousFatimidworkshops.Waste lumpsof crystalhavebeen found,alongwith wasterbeads,wherethe quartzhad shatteredor the holes had been incorrectlybored,as with the carnelians.These crystalbeads,whichweremorecommonthancarnelian,datemostlyto the period between950 and1250. Stonebeadsrepresentthe problemof recognizingcommoditiesor artisansin the archaeologicalrecord.While it was certainlycheaperto mass-producebeads in Indiaforexport,the presenceof rawmaterialsmayhaveprovideda contextfor a localbead-makingindustry.Thisworkmayhavebeen undertakenby localAfricans, but, at leastinitially,they would havehad to acquirea complextechnology involvingdiamonddrill bits. The easiestwayfor this technologytransferto take placewas the physicalsettlementof Indianartisans,who mayhavebeen encouragedto livein EastAfricato helpin the complexworkingof rockcrystalforexport to FatimidEgyptin the tenth-eleventhcenturies. Cloth The only pieceof medievalcloth, datingto the eleventhcentury,to havesurvived from EastAfricais a small squarethat was found sandwichedbetweentwo gold coins in a hoardfoundfromMtambweMkuu.Theclothis madefromz-spuncotton andwasdyedbluewith indigo.46It is probablyof Indianorigin. The manufactureof cloth is more visible in the archaeologicalrecord.The main evidencecomesin the formof spindlewhorls.Theseweremadefroma vari73
ARTISANS,
COMMUNITIES,
AND COMMODITIES
---
5
5
A stone tank at Shangathat may have been used in cloth making, divided into two parts, dating to the early fourteenth century. Photographby the author.
etyof sources- discardedceramic,chloriteschist,ivory,andterracotta.Themost common wererubbedsherdswith a hole boredthroughthe center.Theycan be dated,both throughthe warethat they used and from theirstratifiedpositionin the archaeologicalsequence.The best evidencefor the chronologyof cloth-makalthoughit is a patternfound elsewhere.Here,it was ing comes from Shanga,47 discoveredthat cloth-makingdatesfrom looo, and the numberof whorlspeaked in 1075.Theythen fall off gradually,theirdepositionendingaround1300. Whorls weremade from importedsgraffiatopotteryand very rarelyfrom latertypes of imported pottery.The purpose-madedecoratedterracottawhorls date to the twelfthcenturyandwerecommonatShangaandKilwa.48It is alsopossibleto suggest the locationof the actualcloth-makers.At Shanga,wherea high concentration of spindlewhorlsaswellas Indianpotteryhavebeenfound,therearedetached timberworkshopsthatarelaterincorporatedinto the houses,with adjacentstonelinedtanksthatmayhavebeenusedforthe fullinganddyingof cloth (fig.5). This evidence would seem to suggest that, before iooo, cloth was being importedto EastAfricaready-made,butthen,rapidly,a localclothindustrydeveloped. It is possiblethat this industrymay havebeen developedby Indiantextile workerssettledthere.Theindustrycouldhaveusedflax,but cotton is morelikely. IbnBattita noted a local industryof clothweavingin Mogadishu,the productof whichtook its namefromthe town andprovidedmuchwealthforthe merchants; Chinesesourcesdatingto the thirteenthcentury it wasexportedas faras Egypt.49 suggestthat raw"red"cotton was importedinto Zengba(Zanzibar),presumably to turn into cloth.50Cotton can be successfullygrownin EastAfrica,and it may havebeen introducedto some of the mainlandplantations. By the laterfourteenthcentury,however,the local textile industryseems to havecollapsedin the faceof stiff competition,as finishedtextileswereimported fromwesternIndia.The demandforclothwas substantialin EastAfrica,not just in the coastaltowns but also as a tradeitem for the Africaninterior.Cloth was replacingalternativessuchas skinsandbarkclothamongthe IronAgecommunities and seemsto havebeenthe main tradeitem for exchangingof ivoryandgold. 74
MARK HORTON
41'
S)A
V
.
%,e
t
t
,
6S
6
Partof the Mtambwehoard of over silver coins, found in 1984,on Pembaisland. Photographby the author.
2,000
The prestige of the Indian cloth to the African communities was related to its colorful use of dye that was not locally available. Duarte Barbosa noted in 1517that the merchants of Sufala, located on the southeast coast, still grew their own cotton to make white cloth but unraveled imported dyed cloth from Khambhat, thread by thread, and incorporated it in their own textiles: "With this thread and their own white they make much coloured cloth, and from it gain much gold."'5 The rapidgrowth of the EastAfrican cloth industry from iooo is another example of "technology transfer,"as there is no evidence for local cloth-making before this date. While only the whorls survive, the manufacture of cloth required supplies of raw materials, fibers as well as dyestuff and fuller's earth, and the technology to both spin and weave. By 1300, the products of this industry could compete with Egyptian cloth, an advance that may have been achieved by the settlement of clothiers from overseas who were able to introduce the entire "package"at once. Given the long history and technological sophistication of the North Indian textile industry, this is the likely origin, although it is difficult to prove. By the fifteenth century, however, the local textile industry collapsed in the face of Indian competition, which was now able to export cloth that was both cheap and superior in decoration and quality.
Coinage The East African towns used locally minted coins from the late eighth century.52 The earliest coins come from stratified deposits from Shanga and comprise minuscule silver coins, weighing between o.1 and 0.3 grams, with a name and a short statement such as "Muhammad / billah yathiq."The coins were made in coin flans, a method described by Hamdani in the Yemen, examples of which have actually been found in excavations at Banbhore, the possible site for Daybul in Sindh.53 These eighth-ninth century coins from the Lamu archipelago develop into the silver and copper coinage of the Swahili towns. This series is best known from the Mtambwe hoard, excavated in 1984on Pemba island; it had been buried after 1066, the date of the latest Fatimid dinar found with the hoard.54The locally made coins 75
ARTISANS,
COMMUNITIES,
AND COMMODITIES
wereof silverandcontainthe namesof localrulerswith rhymingcoupletson their reverse,a practicefound on both Fatimidcoins and coins of the Amirsof Sindh (fig. 6). Coppercoins werealso made in EastAfrica.They are more commonly foundwith similarcoupletsandcontinuedto be madeuntil the fifteenthcentury. Theevidencesuggeststhatwhilethe EastAfricancoinagewasmadeforlocal rulers, it was influencedby coining practicesaroundthe Indian Ocean,including thoseof Sindh.Itmaynotbe coincidentalthatthe traditional"first"landfallof the "Debuli"in the Lamuarchipelagois alsowherethe firstevidencecanbe foundfor numismaticconnectionswith the Indusdelta. Conclusions If we wereto relysolelyupon documentaryevidence,then much of the tradeof the IndianOceanmaritimeworldwouldbe invisible.Ratherthana smallnumber of "superhighways" linkingthe keyregionswith high-profilecommodities,there was a complexcrisscrossingof traders,linking togetherdifferentareas.Some commoditieswerenot recordedat all, and most are not visible in the archaeologicalrecord,either.Theycan be inferredthroughsecondaryevidence-trade in liquids,for example,mayleavepotteryvesselsbehind,a cloth trademayleave spindlewhorls,a beadtrademayleaveglasswaste.The tradein particularitems may be surprising:raw cotton may be tradedratherthan the finishedtextiles, or wroughtiron ingotsratherthan completedobjects.The relationshipbetween EastAfricaandnorthwesternIndiais an excellentexampleof this hiddentradein a varietyof commodities,suchas clothandbeads,thatwerevitalto the prosperity of the wholesystem,butwhichareseldomrecordedin the texts. Thereare tantalizingglimpses of multipleconnectionsbetween regionsto suggestthatsomevoyagesmayhaveconnectedmorethantwo regions,a situation suggestedby Piresin the earlysixteenthcentury,citing India,EastAfrica,and It makesconsiderablenavigationalsenseto sail fromnorthwesternIndia Aden.55 to EastAfricaandto returnviaArabia,as this canbe donein two monsoons,or to travelto EastAfricavia Arabia,and returndirectlyto India.Suchmultiplevoyagescouldresultin muchhigherprofits,as eachleg contributesto the overallgain of the voyage,in verymuchthe samewaya triangulartradedevelopedin the eighteenth-centuryAtlantic.Thusa possiblescenariois that Indiancloth and beads wereexchangedin EastAfricaforivory,timber,andslaves.Theshipsthentraveled to the PersianGulf,wherethey unloadedtheirEastAfricancargosand obtained manufacturedgoodsto takebackto northwesternIndia.In someoftheselegs,the boatswouldneedballast,andthis explainswhybricksthatoriginatedin the gulf, or morelikely from the Indusdelta,are found in EastAfrica.56 In some cases,it seemsstonearchitecturewasbeingcarriedacrossthe oceans.57 76
MARK HORTON
7
7 Twelfth-centuryKuficinscription from Tumbatuisland, forming part of a highly ornate mihrab. The inscription uses local Porites coral, although the style of the lettering is similar to inscriptions from Sindh, as well as Sirafin the PersianGulf. Photographby the author.
Communities of Artisans While the idea that foreign merchants played important roles in the foundation of port towns in East Africa has now been rejected, the possibility of substantial foreign artisan communities is rarelyconsidered. It is difficult to explain, for example, the transfer of technologies (such as bead- and cloth-making and ironworking) unless these technologies arrived with a community of artisans. Such communities are very difficult to locate in the archaeological record, as within a few generations they may have been assimilated. In some cases it is possible to demonstrate their presence, as in the case of the Shanga lion, which was made in East Africa but modeled on a small Hindu figurine, possibly from the Deccan.,8The ceramic and craft-working evidence does suggest that the period from about iooo was important in the exchange of technologies and the movement of artisans. While Indian artisans may have moved to East Africa, Africans may have moved, as well, around the Indian Ocean, not as slaves but as genuine artisanal communities. The exciting discovery of an African community living at Sharma on the southern Arabian coast around looo, evidenced by substantial proportions of African pottery, is very similar to the situation in which Indian pottery is found in East Africa at the same date.59At Sharma, around 30 percent of the pottery is of African origin, quantities that cannot be explained through a trade in ceramic containers (most of the pottery is cooking pots) but must be an ethnic indicator of a resident African community. If Africans were resident in southern Arabia, then other communities may well be located in other areaswith strong trade links, which could well include northwestern India. There are communities in Gujarat who claim African descent and are known as the Sidis.60They may not have arrived as slaves, as is generally held, but could have been artisans and part of a diaspora that goes back to the tenth century. The similarity in Kufic styles among Zanzibar, Siraf, and the Indus delta during the late eleventh and early twelfth centuries points to such a craft connection among the three areas (fig. 7).61 While there is currently insufficient evidence to reconstruct in detail the relationship between northwestern India and East Africa from the eleventh until the fourteenth centuries, it is clear that a complex interaction existed between the two regions involving the movement of commodities and artisans in both directions. Hopefully archaeological discoveries will shed light on these contacts in the future. This volume has opened up a new approach to the study of the Indian Ocean, and while little at present can be proven, questions can now be asked that can produce new methodologies in the future. 77
ARTISANS,
COMMUNITIES,
AND COMMODITIES
NOTES
1. H. N. Chittick, "TheEastCoast,
Madagascarand the Indian Ocean,"in TheCambridgeHistoryofAfrica,vol. 3, ed. RolandOliver(Cambridge:CambridgeUniversityPress,1977), 183-231, esp. 218; MarkHorton and JohnMiddleton, TheSwahili:TheSocialLandscapeof a MercantileSociety(Oxford:Blackwell, 2000), 80-82,204-5.
Thomas M. Ricks, "PersianGulf Seafaringand EastAfrica,"African HistoricalStudies3.2 (1970):339-57. 3. MarkHorton, Shanga:TheArchaeologyof a Muslim TradingCommunityon the CoastofEastAfrica(London:British Institute in EastAfrica,1996); Neville Chittick,Manda:Excavationsat an Island Porton theKenyaCoast(London:British Institute in EastAfrica,1984); Neville Chittick,Kilwa:An IslamicTradingCity on theEastAfricanCoast,2 vols. (Nairobi: BritishInstitute in EastAfrica,1974); Felix Chami, TheTanzanianCoastin the FirstMillenniumAD (Uppsala:Societas ArchaeologiicaUpsaliensis, 1994). 4. Ivory,crystal,and ambergris,all African products,are listed as being tradedby Jewishmerchantsin Aden and Kishin the thirteenth century.S. D. Goitein, Studies in IslamicHistoryand Institutions (Leiden:Brill, 1966), 339. 5. Horton, Shanga, 244, wherethe mean proportionis 5.6 percent,with variations per phase rangingfrom 2.3 percentto 7.4 percent.Chittick,Manda, 65, claims proportions up to 28 percent,but this was from beach deposits wherebroken pottery had been unloaded from the ocean voyages.His figuresfor occupation deposits aresuspect, as potterywas not systematicallyrecoveredthrough sieving. T. H. Wilson and A. L. Omar,"Archaeological investigationsat Pate,"Azania 32 2.
(1997):31-76, give only 1 percent as
imports duringthe comparableperiod. 6. MarkHorton, "TheEarlySettlementof 78
MARK HORTON
the SwahiliCoast"(PhD diss., University of Cambridge, 1984), 310. These calcula-
tions arebased on the density of pottery in each phase, the length of each phase, and the total areaof the site occupied at this time. 7. GeorgeDalton, "EconomicTheoryand PrimitiveSociety,"AmericanAnthropologist 63 (1961):
1-25.
8. Horton and Middleton, Swahili,112; Mark Horton, "PortCities and Their Merchants on the EastAfricanCoast,"in Citiesin the World1500-2000, ed. Adrian Greenand RogerLeech (Leeds:Maney Publishing, 2006), 15-30. Very similar
niches havebeen found in fourteenthcentury houses in EastAfrica,which may havebeen used to displayforeignluxury goods, obtained through gift exchanges. 9. DerekKennet,"AnArchaeologicalStudy of the Sasanianand IslamicPeriodsin Northern Ras-al-Khaimah(U.A.E.)" (PhD diss., Universityof London,2001), 120 -39, shows a remarkablesynchronization in the proportionsof both glass and glazed pottery between Shangaand Kush, suggestingthat both sites werepart of a common culture of the westernIndian Ocean ratherthan direct tradepartners. 1o. K. N. Chaudhuri,Asia beforeEurope: Economyand Civilsationof theIndian OceanfromtheRiseof Islamto 1750 (Cambridge:CambridgeUniversity Press, 1ggo), is a most ambitious study of
the pre-PortugueseIndian Ocean,yet the idea of an almost invisibleworld of artisans crossingcultures is barely considered.See Horton and Middleton, Swahili,209, for criticism of the world systems approach,as it has been applied to EastAfrica. i. M. C. Horton and T. R. Blurton,"'Indian' Metalworkin EastAfrica:The Bronze Lion Statuet-tefrom Shanga,"Antiquity 62.234 (1998): 11-23; Horton, Shanga, 358.
The bronze lion figurine in figure3 is
based upon contemporaryIndian figurines,possibly from the Deccan region,but may havebeen made in East Africa, as the metal composition is quite differentfrom the Indian examples, while the lion itself seems to havebeen modeled on the African species. It providescircumstantialevidence for Indian metalworkersresidentin East Africa at this date. 12.Al-Mas'fidi, "Mujujal-Dhahab,"in The EastAfricanCoast:SelectDocumentsfrom theFirstto theEarlierNineteenthCentury, G. S. P. Freeman-Grenville,comp. (London:ClarendonPress,1962), 15-16. 13. Quoted in EdwardSachau,Alberuni's India, 2 vols. (London:Trubnerand Sons,
on an Expeditionof the King of Kish (Qais) againstAden,"Bulletinof the Schoolof OrientalandAfricanStudies16 (1954): 247-48.
Hakluyt Society, 1962) 2:377-79.
24. Abual-Fida,"Geographied'Aboulfeda," in EastAfricanCoast,Freeman-Grenville,
18. E. A. Alpers,"Gujaratand the Tradeof East Africa, C. 1500-1800," International
23. 25.
JournalofAfricanHistoricalStudies9.1 (1976): 23. N. Chittick, "Indian Relations
with EastAfricaBeforethe Arrivalof the Portuguese,"Journalof theRoyalAsiatic Society 2 (1980): 117-127.
25-26.
Diary ofHis TravelsthroughAfrican
Paideuma 28 (1982): 85-100.
Waters, 1497-1499, ed. and trans. Eric
Axelson (SomersetWest:Stephan Phillips, 1998), 47:"Herewe found four ships of Christiansfrom India.... These Indiansareswarthymen. Theywearfew clothes. Theyweargreatbeardsand the hairsofthe headareverylong and plaited. They do not eatbovine meataccordingto whatthey say.And theirlanguageis differentfrom that of the Moors;some of them know a little Arabicfrom continuous communication they havewith them."See also MarkHorton, "The PortugueseEncounterwith the Swahili Townsof the EastAfricanCoast,"in CulturesoftheIndianOcean,ed. Jessica Hallett and ConceicaoAmaral(Lisbon: National Commission forthe Commemorationof the PortugueseDiscoveries,
1910), 2:104.
1998), 382. 20.
Tome Pires, "SumaOriental,"in East AfricanCoast,Freeman-Grenville,12526.
Duarte Barbosa,"TheBook of Duarte Barbosa,"in EastAfricanCoast,Freeman-Grenville,132. 22. MichaelN. Pearson,PortCitiesand Intruders:TheSwahiliCoast-India and Portugalin theEarlyModernEra (Baltimore:JohnsHopkins University Press,1998),48, quoting an account of Almeida'svoyagein 1506. 23. Ibn Battuita,The TravelsofIbnBattuta, 21.
79
CRAFTSPEOPLE,
COMMUNITIES,
SirJohnGray,"TheWadebuliand the Wadiba,"TanganyikaNotesand Records 36 (1954): 22-42; Gray,Historyof ZanzibarfromtheMiddleAgesto 1856 (London:OxfordUniversityPress,1962),
26. A. H. J.Prins, "TheMtepeof Lamu, Mombasaand the ZanzibarSea,"
19. Vascoda Gama,Vascoda Gama:The
14. Quoted in GabrielFerrand,Relationsde
voyageset textesgeographiquesarabes, persanset turksrelatifd l'Extreme-Orient de 8e au i8e siecles(Paris:E. Leroux,1913), 177 (my translationfrom the French). i5. Horton, Shanga,363.C.-M. Kusimba,The Archaeologyand EthnographyofIron Metallurgyon theKenyaCoast(PhD diss., BrynMawrCollege, 1993).C. M. Kusimba,D. J.Killich and R. G. Cresswell, "Indigenousand ImportedMetals at Swahili Siteson the Coast of Kenya," MaseaResearchPapersin Scienceand Archaeologyii (1994),supplement,63-77, identifiedlow carbonbloomery steel among artifactsfrom the Swahilisites of Ungwana and Galuthat arelikely to have been locally made. They also recorded cruciblesteel from the same sites that may havebeen imported from India. 16. Ricks,"PersianGulf Seafaring,"353-54; DavidWhitehouse, "Kish,"Iran14(1976): 146-50; AndrewWilliamson, "Hurmuz and the Tradeof the Gulf in the 14thand 15th Centuries,"Proceedingsof the SeminarforArabianStudies4 (1973): 52-68. 17. S. D. Goitein, "TwoEye-WitnessReports
1325-1354, trans. H. A. R. Gibb (London:
27. Gray,HistoryofZanzibar,24. 28. Neville Chittick,"The'Shirazi'Colonization of EastAfrica,"JournalofAfrican History 6.3 (1965): 289-go.
Gray,HistoryofZanzibar,26, considered that they came from Dabhol, a port a hundredmiles south of Bombay,and that they lived duringthe fifteenth century. The Diba are sometimes linked with the Maldives ("DibDib" islands),but both these interpretationsarealmost certainly wrong. The alternativeplace name of Daybulas Dibal indicatesthat the African traditionsarereferringto the same location in the Indus delta. 30. Monique Kervan,"MultiplePorts at the Mouth of the RiverIndus:Barbarike, Deb, Daybul,LahoriBandar,Diul Sinde," inArchaeologyofSeafaring,ed. H. P.Ray 29.
(Delhi: Pragati Publications, 1999), 70-
154.The limited excavationsat Banbhore arepoorly reported,but see "Excavations at Banbhore,"Pakistan3 (1966):49-50; "Banbhore,"PakistanArchaeology5 (1968):176-85; S. M. Macaque, "The
GrandMosque at Banbhore,"Pakistan Archaeology6 (1969): 182-209; F.A. Khan, Banbhore:A PreliminaryReporton the Excavations RecentArchaeological (Karachi:Departmentof Archaeology and Museums, 1969).
31. In the fifteenth century,however,there
AND COMMODITIES
was a merchantliving at Kilwawith the nisbaor surname of"al-Daybuli." Chittick, "'Shirazi'Colonization,"290. 32. Horton and Middleton, Swahili,52-61. 33. MatthewSmith and Henry Wright,"The Ceramicsfrom RasHafun in Somalia," Azania
23
(1988):115-41. The Hafun
pottery may,however,point to the wellattestedclassicaltradebetween the Red Seaand India, which lies beyond the scope of this article. 34. Chittick,Kilwa,ii, 329 -30; Chittick, Manda, 101; Horton, Shanga, 300-303. A
study of the Indian pottery from Zanzibarand Pembawill appearin Mark Horton, Zanzibarand Pemba:ArchaeologicalInvestigationsofan Indian Ocean Archipelago(London:BritishInstitute in EastAfrica,forthcoming). 35. Donald S. Whitcomb and JanetH. Johnson, Quseiral- Qadim1980:PreliminaryReport,American ResearchCenter in Egypt Report7 (Malibu:Undena Publications,1982), pl. 45. Kennet, "ArchaeologicalStudy,"115-117,292, describes"pink"wareswhich are probablythe same type. 36. Horton, Shanga,272-73. 37. Pires, "SumaOriental,"in FreemanGrenville,EastAfricanCoast,125. 38. Horton, Shanga,323. 39. Horton, Shanga,329; Chittick,Manda, 181-83. 40. Horton, Shanga, 329 -32.
41. Horton, Zanzibarand Pemba. 42. So common arethe glass beads, the local
inhabitantsareableto pick them out from the sand and string them into necklacesto sell to tourists. 43. It is normally held that all the glass beads were imported into EastAfrica ready made. At Mogadishuglass-bead-making equipmentwas found in the nineteenth century,including crucibles,colored pastes, sticks of glass, and colored beads. JustusStrandes,ThePortuguesePeriodin
EastAfrica,Englished. (Nairobi:East AfricaLiteratureBureau,1961), 78. 44. Pires, "SumaOriental,"in FreemanGrenville,EastAfricanCoast,126. 45. Horton, Shanga,332. 46. Horton, Zanzibarand Pemba. 47. Horton Shanga,337-41. 48. Chittick,Kilwa,428-29. 49. Gibb,IbnBattuta,374. 50. FriedrichHirth and W. W. Rockhill, ChauJu-Kua:His Workon the Chinese andArabTradein the Twelfthand ThirteenthCentures,EntitledChu-yanchi (SaintPetersburg:Printing Office of
58. Horton and Blurton,"IndianMetalwork." 59. Fromthe site of Sharma.I am very gratefulto Axelle Rougeullefor showing me the materialfrom her excavations there,whereAfricanpottery constitutes in some areasup to 30 percentof the total assemblage,and virtually all the cooking vessels areof African origin. Axelle Rougeulle,"CoastalSettlementsin
the Imperial Academy of Sciences, 1911),
6o. S. S. P.Freeman-Grenville,"TheSidi and
SeminarforArabian Studies 29 (1999): 134.
the Swahili,"Bulletinof theBritish
126. 51. Duarte Barbosa,"Book,"in Freeman-
Grenville,EastAfricanCoast,128. 52. Helen W. Brown,"Coinsof EastAfrica:
An IntroductorySurvey,"Yarmouk
Association of Orientalists 6.1 (1971):3-18;
J.de V.Allen, "Habash,Habashi,Sidi, Seyyid,"in Africaand theSea, ed. J.C. Stone (Aberdeen:AberdeenUniversity
Numismatics 5 (1993):9-16; Brown,
Press, 1984). Timothy Insoll, The
"EarlyMuslim Coinage in EastAfrica: The Evidencefrom Shanga,"Numismatic Chronicle152 (1992): 83-87; Horton,
Archaeologyof Islamin Sub-Saharan Africa(Cambridge:CambridgeUniver-
Shanga, 368-77.
53. Khan,"Banbhore,"45;ChristopherToll, 'Kitabal-jauharatainal-'atiqatainalma'i' atayn min al-safra'wa' l-baida'by al-Hamddn7,Studiasemitica Upsaliensia 1 (Uppsala:Almqvist and Wiksell, 1968), chap.46; J.C. Jetzer,"TerracottaCoin Moulds,"NumismaticInternational Bulletin15 (1981): 348-49. 54. M. C. Horton, H. W. Brownand W.A. Oddy,"TheMtambweHoard"Azania 21 (1986): 115-23; Horton, Zanzibarand Pemba. 55. Pires, "SumaOriental,"in FreemanGrenville,EastAfricanCoast,125-26. 56. Chittick,Manda, 15-16. 57. ElizabethLambourn,"Carvingand Communities:MarbleCarvingfor the Muslim Patronsat Khambhatand around the Indian Ocean Rim, Late Thirteenth-Mid-Fifteenth Centuries," in this volume. 8o
Southern Yemen: The 1996-1997 Survey
Expeditionson the Hadramawtand MahraCoasts,"Proceedingsof the
MARK HORTON
sity Press, 2004),151, suggests that the
Sidis at BabaShormay havebeen involvedin the mining and production of carnelian,but at presentdirect evidence is lacking. 61. David Whitehouse, "Siraf:An Islamic City and Its Rolein IslamicArt,"Storia Della Citta 7 (1978): 54-58; Kervan,
"MultiplePorts,"137,includes a floriate Kuficinscription from a site in the Indus deltathat is stylisticallyidenticalto an inscription from Tumbatuin EastAfrica.
.... . ....
Ot:
7r
.. 1W
tk;
--ow
41
KATHERINE
STRANGE
BURKE AND DONALD
WHITCOMB
QUSEIR AL-QADIM IN THE THIRTEENTH CENTURY A CommunityandIts Textiles
I
Quseir al-Qadim, the Sheikh's House from the north.
Abstract During the 1982excavationsat Quseir al-Qadimon the Egyptiancoast of the RedSea,a corpusof well-preservedtextileswas discoveredin an areacalled"the Sheikh'sHouse."Thearchaeologicalnatureof this corpuspresentstwo assetsover a museum collection:the possibilityof close datingand contextualassociation with otherartifacts.Thisarticlediscussesthe collectionof textilesin the Sheikh's House from these perspectives.The archaeologicalcontexts allow an association with otherartifactswithin architecturalsettings.A moreimportantfeature of these contextsis the presenceof hundredsof fragmentsof letterswritten in Arabicon paper;this textualdocumentationallowsa reconstructionof the individualswho used the textilesand, morebroadly,an extremelydetailedexamination of their tradingcommunity.On a largerscale,the archaeologicalnarrative maypresentthis communityin relationshipto the peopleof the CairoGeniza,to the archaeologyof Fustat(Cairo),and its mercantilerolein the largerpictureof IndianOceanhistory.
SHEIKHABIJMUFARRIJ returnedfromthe mosquewherehis son wasa reader, walkingsouthwarddownthe slope(fig.1).He enteredthepassagewaythatdivided his businessfromthe new walls of his son'shouse. SheikhIbrThimibn Mufarrij hadrecentlytakenovermanyof the dailyaffairsof the business,andthe oldhouse hadbeendividedinto two households,one forthe old sheikhand a secondforhis son'sfamily,the lattermakinga suiteof roomsabovethe harbor.Eachhousehad a largefamilyroomandtwo smallerroomsforsleepingandstorage.Twowinding staircasesled to rooftopenclosuresmadeof reedsforsleepingin the hot months. On the seasideof the passagewaywerea seriesof doorsto the shuna,or storehouse. The sheikhstill keptwoodenkeysto the doorsundera loose stone of the threshold;he had written "miftahal-HajjBaraka"on the one belongingto his trustedclient to distinguishit from his own key.The largerrooms of the shuna werereallyopenyardsfilledwith sacksof grainandcontainersof flour;therewere many containers,each labeledand correspondingto an inventory.This operation hadbecomeso complexthathis friend,SheikhNajib,wasbecominga trusted clerkfor storageand dispatchof individualconsignments.Hereand therewas a sackof datesor a bundleof flax. The smallerroomswerefilledwith a varietyof foodstuffs,perfumesand spices, metalworkand garments,individualtrading speculations,andspecialordersforgoodclients. SheikhAbuiMufarrijremovedhis robewith its fancyembroideryand dressed in an everydaylinen thaub,perhapsone with a tasteful checkpattern.His son IbrThimhad recentlyfurnishedhis housewith pillowsin the new styleof Indian 83
ew~~~~~r
Aw-
'sE*
i4-
-
?..,;
2
X,.t^,9S
:bs
2
Quseir al-Qadim, the Sheikh's House looking east towardthe sea.
prints,those with a red ground.Forhimself,the old sheikhpreferredthe classic Indianbluepatterns,his favoritebeingthe "treeof life"alternatingwith a saddled f1l, or elephant.These coveringswere resewnby his wife as tearsappearedand wouldsoon haveto be replacedwith the simpleblockpatternsbeingbroughtwith today'sshipments.Prosperityandmorecompetitionamongthe merchantsmeant the qualitygoodsof earlierdayswerein declineandreplacedby massproduction. Gonewerethe dayswhenthe headof the Karim!merchants,the Ra'isal-Tujjjar, wouldcalluponthe sheikhforassistance,makinghim almosta governmentagent. Indeedthe largestand most importantexports,the grain and flour to feed the Haramain,the holy citiesof Makkaand Madina,seemedto be past.The sheikh's frustrationsand mood were not improvedby a messagefrom his youngersons (Muhammad,'Umar,and Hasan).Theyhad gottenthemselvesstrandedby winter winds in the Yemenfort and werepleadingfor heavycoatsand some of their mother'scakes(whichwererathergood). Livesfrom Textsand Artifacts The above narrativeis constructedfrom the excavationsat Quseir al-Qadim on the Egyptiancoast of the Red Sea (fig. 2). The setting is the areacalled "the Sheikh'sHouse,"revealedduringthe 1982 season. This areawas a knoll on the southernedgeof the site,locatedabovethe sabkhaor mudflatsof the silted-inbay of the Romanperiod (fig.3). Upon this knoll was a single structuralcomplexof stoneand mudbrickwalls,within whichwasa stratifieddepositionof remarkable preservation;this archaeologicaldepositionfeaturednumerouslettersdiscarded with othertrashandrefuse,pottery,glass,seeds,andespeciallytextiles. This corpusof textilesis closelysimilarto the collectionof block-printedtextiles in the KelseyMuseumat the Universityof Michigan,extensivelyresearched by RuthBarnes.The textilesfrom Quseiral-Qadimhavealreadybeen published Thearchaeologicalnatureof this corpushastwo by GillianVogelsang-Eastwood.' assets over a museum collection:the possibilityof close dating and contextual 84
KATHERINE
STRANGE
BURKE AND DONALD
WHITCOMB
4 c=
048 I3
62
4 T'A
36c-
~s_ -
JlO ~~~~Jgb 33
~~~~~
1
I
22
/n 8
1
5-
17
48131
--~~~~~(9)
45
:
5
:-\.
3~~~12_D_B
20
3
3
Plan of the Sheikh'sHouse excavation.
association with other artifacts. This article will discuss a group of textiles from these perspectives. As Barnes perceptively notes in her study: "The excavations' published reports offer important new material concerning the circumstances of the settlement in the past. Yet it takes much imagination to sense the reality of life behind these fragments.ou2 Fortunately, a set of these textiles was discovered in contexts that allow a reconstruction of their association with other artifacts, in architectural settings, and textual documentation of the individuals who used them. This combination of artifacts and documentation will allow an extremely detailed examination of this trading community, its relationship with the Geniza and archaeology of Fustai, and its role in the larger picture of Indian Ocean history.3The textiles were found with hundreds of fragments of letters written on paper. These Arabic letters do not belong to an archive; they are part of the rubbish accumulated in the Sheikh's House.4 They do record precise details of the mercantile enterprise of Sheikh AbuiMufarrij, his family and associates. These letters bear dates
that clusterwithin the firstfour decadesof the thirteenthcentury(1200-1240); likewise, the coins found in this area were also minted during the late Ayyuibid 85
QUSEIR AL-QAD1M
IN THE THIRTEENTH
CENTURY
, ..
:4,.e
.
%.
4a.
.I
5
!
_
6~~~~~~
4
ThreeAyyuibidcoins, RN 82-695 (locus Kgb-67),RN 696 (locus Kgd-i), RN 698 (locus Jloc-17). 5
Two wooden keys, RN 82-524, RN 82-560 (locus Jgd-9). 6 Arabicletters, RN 82-1059, RN 82-1061(locus Kioa-li).
period,to 1249 (fig. 4),5 with only a few residual Fatimid issues and no Mamlik coins. The occupationof this house seemsto havelastedhalf a centuryand presents an unusualcontextualizationfor the abundantrangeof artifactsand their architecturalsetting. The Sheikh'sHouseas Context Eight of the resist-dyedtextiles from Quseir al-Qadim published by Gillian Vogelsang-Eastwood were found in the Sheikh'sHouse. This complexof mudbrick-walledstructuresis interpretedas two adjoininghouses, associatedstorerooms,6and a passagewaybetweenthe houses and storerooms.The two houses areon the westandsouthsidesof the knoll,andthe storeroomsline up northwest to southeastalongthe east side.Eachhouse consistsprimarilyof one largeroom (approximately5.5x 4 meterseach) and two smallerrooms alongsidethe large room,usually3 x 2.5 meters.Eachpreserveda stairwayto the secondflooror roof, with some extantwoodentreads.The storeroomswerelarger(ca.4 x 2-5 meters) and undifferentiated,each enteredoff the corridor.7Twowooden keys, secreted beneaththe thresholdof one entry,seemto confirmthis interpretation.Bothkeys are large, and one is inscribed (fig. 5).8
Due to the aridenvironment,awiderangeof artifacttypeswaspreservedatthe site,includingleather,fiber,cloth,andpaper.Thousandsoffragmentsoflettersand documentswrittenin Arabicon paperwerepreserved,includingover 1,500from the Sheikh'sHousealone (fig.6). LiGuohasbegunthe publicationof theseletters with a selectionof almost one hundredwell-preservedexamples.9These documents,consistingmostlyof businessletters,shippingnotes,and accountrecords, 86
KATHERINE
STRANGE
BURKE AND DONALD
WHITCOMB
7
4i31i.i,I~~~~~~%J)
8
7 Resist-dyedtextile, RN 82-922 (locus Kgb-63). 8 Resist-dyedtextile, RN 82-945 (locus Kgb-63). 9
Resist-dyedtextile from locus Ei8d-4.
9
detail the businesstransactionsthat were undertakenby SheikhAbu Mufarrij and his son SheikhAbi IshaqIbrThimibn AbuiMufarrijas they participatedin the IndianOcean-Red Sea-Nile valleytradeduringthe lateAyyfibidperiod (ca. Thestratificationofthe letterswithinthe roomsindicatesthatIbrahim 1200-50).10 wasrunninghis own partof thebusinessfromthe earliestoccupationofthe house; that most of the lettersareaddressedto his fatheris probablyindicativethatAbui Mufarrijownedthe storeroomsout of whichhis son also operated." Most of the extantlettersare missingthe names of the senderand recipient, but of those that retainthis informationaroundthirty areaddressedto Ibrahim, thirty to AbuiMufarrij, and another thirty to the Shuna of Aba Mufarrij.12 Con-
centrationsof lettersaddressedto Ibrahimarefound in both main living rooms, the largestroomsof eachhouse, althougha few morearescatteredaboutalmost Almost all of the lettersaddressedto all areasof both houses and storerooms.'3 AbuiMufarrijare in the largeliving room of the North House (Roomc in fig.3), with only one in a storeroomand two in the SouthHouse.This distributionmay be due to the superiorpreservationof the North House,whichwasbuilt afterthe SouthHouse.Nevertheless,the predominanceof lettersaddressedto Ibrahimin RoomC also seemsto indicatethatthis washis room,althoughhis fathermayalso haveusedthe spaceas an office. Theimaginedreconstructionofthehouseanditsoccupantsthatbeganthisarticle relieson the informationin the Arabicletters,organizedin termsof the resistdyedtextilesandotherartifacts.Halfof the eightresist-dyedtextilespublishedby fromthe Sheikh'sHousewerefoundin the storerooms,and Vogelsang-Eastwood halfin the housesor the main corridor.Twofragmentswerefoundin AbuiMufarrij'sliving room (Room c of the North House).14 Both are cotton tabby weaves with
ablock-printedtreeoflife motifsetin a squareframeof naturalcoloron adarkblue ground.Number52 has a stylizedrosettealternatingwith the tree of life (fig.7), while number53alternatesthis motif with a saddledelephant(fig.8).15The tree of life motif on both pieces has ratherdetailedflankingbirds and quadrupeds; estimatesthis styleto be earlyin the life of this motif,while Vogelsang-Eastwood latertextilesbeara muchsimplifiedverticalline flankedby crescents(fig.9)V6 87
QUSEIR AL-QADIM
IN THE THIRTEENTH
CENTURY
10
11
14
12
13
15
10
Drawing on paper, RN 82-1096
(locus Kgb-48). 11
Small ostracon, RN 82-135
(locus Kgb-48). 12
Basket, RN 82-659 (locus Jloc-17). 13 Matting, RN 82-661 (locus Jgd-4). 14
Rope coils, RN 82-635(locus Jgd-4). 15 Leather shoe, RN 82-574
(locus L8c-2).
Thesetwo above-mentionedtextilesoccurin a well-stratifiedcontext,beneath the lowerof two floorsexcavatedin Roomc, in one of the earlieststratafromthe use of the room (Kgb-63).They occur within a rich artifactassemblageof 380 paperfragments(fig.io), two of whichareidentifiedby Guoaslettersto Ibrahim,"7 one silvercoin, an ostracon (fig. ii), over twenty textile fragments(more than half dyedblueor red),severalglassfragments,fragmentsof iron nails andbronze implements,manypiecesof basketry(fig.12), matting(fig.13),and rope (fig.14), two fragmentsof leathershoes (fig. 15),bones, shells, seeds (includingabout130 date pits) and about400 sherdsof pottery,5 percentof which is glazed (fig. i6c, j, and 1).Fiveof the glazedsherdsare "mustardware,"likely importedfrom the Yemen(fig.16a,b),'8andone is a celadonimportedfromChina(fig.16d).Thereare also sherdsfromat leasttwo unglazedorangejarsbearinga standingcobramotif alternatingwith an elongatedcrescent,in brownor redpaint (fig.16h,i). Thejars bearsomeresemblanceto Nubianpaintedpottery. In Room d, a small room behind the stairwayin the southeastcornerof the South House, a small (approximately 3 x 5 centimeters)fragmentof cotton with wovenblue and naturalstripeswas found,with a red resist-dyedscrapof cotton sewn to it. "Thepatternconsistsof a single row of dots (fivein total), tracesof a curvedline andpossiblypartof a largedot in naturalon a redground.Thepattern is clearon both sidesof the cloth"(fig.17).19 As in Roomc of the North House,this fragment was found in the lowest stratum of this room (Kioa-20), indicating its
use duringthe earlyoccupationof the house.In the samecontext151sherdsof pot88
KATHERINE
STRANGE
BURKE AND DONALD
WHITCOMB
16
Ceramics from Room c, North House. A
B
Jgd-4,RN 82-339 red ware, greenish yellow glaze on interior and exterior rim, brown paint, moderate medium sand Jgd-4,RN 82-330 red ware, greenish yellow glaze on interior and exterior rim, brown paint, moderate medium sand
c Kgb-48, RN 82-340
D
E
F
G
H
I
cream ware, bluish white glaze on interior and exterior,clear glaze, on lower exterior, moderate medium sand Jgd-4,RN 82-330 grayware, darkolive gray (celadon), comb incised, stoneware Jgd-4,RN 82-44 red ware, cream surfaces, comb incised, moderate medium sand Jgd-4,RN 82-44 tan ware, red surfaceson interior and exterior,blackenedexterior,common medium sand and pebble Jgd-4,RN 82-44 darkred ware, traces of cream surface (slip?), incised, common medium sand and chaff J9d-4/K9b-36,RN 82-95 red-brownware,yellow-orange slip on exterior and interior neck, darkbrown paint, common medium and moderate coarse sand
~~~~~B
H
I
Jgd-2/4, RN 82-97
red-brownware,yellow-orange slip on exterior and interior neck, darkbrown paint, common medium and moderate coarse sand i Jgd-4,RN 82-337 cream ware, turquoise glaze on exterior and interior shoulder,impressed decoration on exterior;joint seam on interior,moderate medium sand
O0CM
O7L
5
16
L
Kgb-59, RN 82-307
white ware, darkblue glaze on interior and exterior,incised, frit M
K Jgd-4, RN 82-233
white ware,white glaze with black crackleon interior,greenish clear glaze on exterior,not base, frit
'J
N
89
Kgb-59, RN 82-307
white ware, clear glaze on interior and exterior,not base, porcelain Kgb-63,RN 82-48 light gray-orangeware, buff-cream surfaces,
QUSEIR AL-QADIM
IN THE THIRTEENTH
red-brownpaint on exterior,moderate medium sand o Kgb-59,RN 82-307 red ware,buff slip on exterior,brown paint, incised lines, common medium sand p Kgb-49,RN 82-49 orange-tan ware, cream surface on exterior, red-brownpaint, moderatemedium sand
CENTURY
19
17
21
20
17
Resist-dyedand striped textiles, RN 82-939 (locus Kloa-20).
i8 Resist-dyedtextiles, RN 82-937,938 (locus Kloa-15). 19
Resist-dyedtextile, RN 82-927 (locus Kioa-ii) 20
Rope sandal, RN 82-621 (locus Kloa-li).
21
Resist-dyedtextile, RN 82-931 (locus Kloa-8).
tery were found, 3 percent of which were glazed. These included fourteen sherds of green- or yellow-painted mustard wares and one imported Chinese celadon. Other finds were nine fragments of Arabic letters, seeds (including seventy-five date pits), fragments of bone, glass, metal, matting, rope, wood, a scrap of sewn cotton, and a dozen other textile fragments, both dyed and undyed. In the South House, in Area f, a room or part of a corridor near Room d, two fragments of resist-dyed textile were found, both dyed red. One, Eastwood catalogue number 56, exhibits a faded diamond-pattern, with four-lobed rosettes between them on a light red ground. Eastwood catalogue number 57 is coarsely woven with traces of a geometric and possibly foliate design in red on a natural ground.20 These fragments (fig. 18) were found in one of the medial strata (Kioa15),not the earliest, together with 291 sherds of pottery (1.5 percent of it glazed), some glass, the usual collection of metal, bone, shell, seeds (including 1,025date pits), matting, rope and wood fragments, a dozen textile fragments, of which half were dyed red or blue, and eighteen letter fragments. Passageway d, the corridor outside and to the east of the North House, runs between it and the row of storerooms along the east side of the compound, ending at the South House. This passageway was full of artifacts from the final use and abandonment of the house. One of the resist-dyed fabrics was found in the uppermost stratum (Kioa-11).It is coarse cotton with a "crude"blue design consisting of mainly interconnecting circles, on a natural ground (fig. 19).21 The stratum in which it was found contained 157sherds (.os percent of which was glazed), both 90
KATHERINE
STRANGE
BURKE AND DONALD
WHITCOMB
3..
..
~~~~~~~~~~~~~23
22
22
Glassbeads from locus Kgb-41 (RN 82-392)and locus Kgb-49 (RN 404) in Room c of the North House. 23
String, RN 82-455 (locus Jloa-2), RN 82-565(locus Jgd-1i), RN 82-576 (locus Jloa-ii).
local ceramics and Far Eastern imports such as celadons and porcelains. Also in the same context were a woven fiber fan, fragments of rope sandals (fig. 20), fragments of leather shoes, a ceramic net weight or bobbin, fragments of glass, wooden and metal artifacts, matting and rope fragments, various types of organic debris such as bone, shell, and seeds (including 1,206 date pits), sixty-eight fragments of letters written in Arabic on paper,22an "ostrichicon" (a fragment of ostrich egg shell with Arabic writing, see below) and sixty-one textile fragments including two pieces of silk with Arabic or pseudocalligraphic inscriptions. In the storeroom immediately adjacent to Area f (referredto as Shuna f) of the South House three tiny scraps of resist-dyed cloth were found. All are faded red on a natural ground, and two of them have geometrical designs. The third appears to be decorated in a band of calligraphy or pseudocalligraphy (fig. 21).23 They are in the uppermost stratum (Kioa-8) and thus from the last use of the storeroom. In the same stratum were found a glass bangle, several glass fragments, some 149 sherds of local pottery (again only .05 percent of which was glazed), metal scraps, a bronze coin, bone, shell, seeds (including only thirty-one date pits), matting, rope, wooden implements, seventy-six fragments of Arabic letters, including one with Greek and Coptic writing on the back, and a few textile fragments including a fragment of a face veil.24 The final piece of resist-dyed cloth comes from the uppermost stratum (Jloa2) in the far northeastern storeroom, known as Shuna b. Shuna b appears to have been an unroofed courtyard, bounded by walls on all but the eastern, unexcavated side. The textile from this last use of the courtyard is about 20 X 12 centimeters of fine cotton dyed dark blue. At one end of the cloth a large brown square encloses a dark blue rosette-in-square pattern, next to which is an unidentifiable form in natural color.25This fragment was found in context with 297 sherds of pottery, including two sherds of imported celadon, six sherds of Yemeni mustard ware, glass, a glass bead (for examples of glass beads, see fig. 22), a few scraps of bronze, fragments of bone, seeds (including 234 date pits), wood, many fragments of matting and rope, a bobbin or ball of string wound around a piece of wood (fig. 23), 151textile fragments, thirty-eight letter fragments, and one whole letter complete with string binding. 91
QUSEIR AL-QADiM
IN THE THIRTEENTH
CENTURY
24
24
Drawing of saqiyya(waterwheel) incised on a ceramic vessel, RN 82-346 (locus Kgb-69, 70,71).
In all of the above-mentioned loci the pottery corpus, consisting mostly of Nile valley ceramics with a few Nubian and even Chinese imports, is not unusual for a ceramic corpus of this period (fig. 24).26 Celadons (green-glazed stoneware) and white-glazed porcelains were widely traded in Egypt and the Near East, and ceramics of Nubian provenance, or even Nubian influence, are common in Egypt. Excavations at Fustat, for example, have yielded very large quantities of Chinese ceramics, dating from the ninth through the fifteenth centuries, with the largest quantities dating from the tenth through fourteenth centuries.27Analysis of these ceramics shows that in the late Ayyfibid period, correlating with the Southern Song dynasty in China (1127-1279),Egypt was importing celadons produced at Yue, Longquan, and Fujian Province white wares and yingqing (a white "sugary" ware usually glazed light blue) wares from Jingdezhen city and Fujian Province, and brown wares from Fujian and Guangdong Provinces.28The celadons from the Sheikh's House have parallels in the Fustat materials. For example, the style and execution of the incised and combed peony petal motif on sherd RN 330 (see fig. i6d), as well as its olive glaze and dark gray body, are seen in a small group of sherds identified as Yaozhou ware, dating to the Northern Song period (96o-1127).29 Similar sherds were identified as Northern Celadon (Yaozhou ware) at Athar, another Islamic port on the Yemeni coast of the Red Sea in use from the early Islamic period until the late eleventh century.30That this and other similar celadon sherds are traditionally dated earlier than the main occupation of the Sheikh's House perhaps indicates the value of fine Chinese imports as "heritage wares," kept in a family for generations, or possibly that Yaozhou-type wares at Quseir were imitations made in southern kilns, especially in Guangdong Province, during the Southern Song period. This contextualization of the block-printed Indian textiles presumes that 92
KATHERINE
STRANGE
BURKE AND DONALD
WHITCOMB
-~~~
4
25
Drawing of ship, black ink on paper, RN 82-1097 (locus Kgb-27).
they weresomewhatrareand valued.Their depositionin the Sheikh'sHouse is a kaleidoscopicvision of the physicalelementsof dailyuse in the earlythirteenth century.The port of Quseir producedlittle on its own; foodstuffs and objects wereimportsof variablerarityandvaluation,whereinprofitand livelihoodwere derived.The sheikhand his communitywerebound to take a broadview of the commerceof the IndianOcean,as well as participationin the life of the Nile valley,of Egyptunderthe Ayyuibidrule (fig.25). As moreof his lettersareread,these dualaspectsmaybe elicitedto reflectwhatmaybe lesspreciselydeterminedfrom artifactsalreadystudied. Afterthe Sheikh Sometime after1250, Ibrahim,his now widowedmother,and the restof the family abandonedthe house and businessand movedaway.The port in the late thirteenth and fourteenthcenturieschangedits character.Settlementshifted to the seashore,and dwellingsbecameless permanent,being madeof reedsand thatch (barasti) construction.Quseircame to resemblemany coastalplacesin Sudan, Arabia,and elsewherein the regionof the IndianOcean.This laterport presents the paradoxof "rich"artifactualcontentsin a "poor"architecturalsetting.This very differentdepositionalhistoryrepeatsthe contextualizationof architecture with potsherds,textiles,seeds,coins, andletters.Whenthe lettersfromthis Eastern Areahavebeen studied,otherpersonalitiesand their communitymay allow forthe constructionof othernarratives. This latercommunitytestifiesto the greatprosperityof Egyptand the Indian Oceanduringthe Bahr!Mamlik period,with tradedartifactsfromIndia,China, Syria,and even Tekrur(WestAfrica).Numismaticsand anotherlargecorpusof lettersfound in the excavationsconfirmthis conclusion;indeed, almost all the coins of the EasternAreawereMamlik, with almosthalf issuedin the late fourteenth century.Tradeitems continued,includingChineseporcelainsand more resist-dyedtextiles. In addition,preliminaryanalysisof the plant remainssuggests a dramaticshift from Mediterraneanproductstowardan Indian Ocean "spicetrade.`3'This picturewould conformto a majorreorientationof external politicalandcommercialrelationships,of whatJean-ClaudeGarcinhascalledthe "livingspace"of medievalEgypt.32 LastFragmentsandAnotherCommunity Perhapsthe most difficulttask for an archaeologistis to determinefinality,the endof settlementoccupation,evenwhentheremaybeevidenceof violentdestruction and abandonment.At Quseir,a high ridgeabovethe beachrevealedsignsof a previousexcavation,debristhrownout of the cemeteryincludinghumanbones 93
QUSEIR AL-QADIM
IN THE THIRTEENTH
CENTURY
and ostrichicons.The exampleof the latteralreadymentionedfrom Passageway d (Kioa-ii,82-197)is quiteunusual.All remainingfragmentsof these ostrichegg shellswith Arabicwritingcome fromthis cemetery.Twoof thesehavebeen published and bear dates, 1465-75 and 1485-95, and texts that reinforce a funerary
utilization.33 This excavationalso produceda bronzecoin (A22d-2, 78-459)issued under Qa'it Bay (reigned 1468-96), falling in the same date range (1468-96). It is
not surprising,then, that fragmentsof resist-dyedtextiles from this same locus findcomparisonsdatedto the fifteenthcentury.34 Finally,one maynotethatassociatedceramicsincludeblue-and-whitedecoratedporcelainsandpaddle-stamped handmadepotterynot foundelsewhereon the site.35 The consistencyof evidence in this cemeteryis in markedcontrastto the excavationsin the EasternArea,where coins of Barqiuq(reigned1382-99) and associatedceramicsdate from the fourteenthcentury.This latestmanifestation at Quseiral-Qadimis recentlycontextualizedin researchat Quseirfort in the moderntown.36The evidencesuggestsgrowthin the fifteenthcentury,as noted this is the settlement,underthe moderntown,thatattractedthe ranby Garcin;37 dombombardmentbythe Portuguesein 1540.38 Soonafterthe latterevent,Joaode Castroclearlydistinguishesthe "mostmiserabletown"fromthe ruinshe associatedwith the classicalLeukosLimen(AlbusPortus).39 Quseiral-Qadimbecame an occasionalcemetery,as ruins often do, and an attractionfor curiousvisitors from a Napoleonicsoldier (who lost a uniformbutton)40to GustaveFlaubert,41 littlerealizingits historywith the sheikh,his community,andtheirtextiles.
94
KATHERINE
STRANGE
BURKE AND DONALD
WHITCOMB
NOTES
1. Gillian M. Vogelsang-Eastwood,Resist
Dyed TextilesfromQuseiral-Qadim, Egypt (Paris: AEDTA, 1ggo). 2.
Ruth Barnes,IndianBlock-PrintedCotton Fragmentsin theKelseyMuseum,the UniversityofMichigan,KelseyMuseum Studies8 (Ann Arbor:Universityof Michigan Press, 1993), 7.
3 The Geniza refersto the documents found storedin the Ben Ezrasynagogue in Fustat/Cairo.Manyof these letters describethe community and its commercial relationships,as describedmost thoroughly in S. D. Goitein,A Mediterranean Society:TheJewishCommunitiesof the ArabWorldas Portrayedin theDocuments of the CairoGeniza,vol. 1,Economic Foundations(Berkeley:Universityof CaliforniaPress,1967). 4. Thatthis is a nonarchivalrandom collection cannot be overemphasized.In this aspect, chronology,and paleographic details, there is a strongsimilarity to the Cairo Geniza documents. Li Guo, "Arabic Documents from the RedSea Port of Quseir in the Seventh/Thirteenth Century,"pt.1,"BusinessLetters,"Journal of Near Eastern Studies 58 (1999): 163.
5. This period comprisesthe reignsof the Ayyuibidsultans al-Malikal-'Adiland his sonal-Malikal-Kamil.Precisedatesonthe lettersreadrangefrom1215to 1235,though this is only a small fraction of the total inventory.Guo,"BusinessLetters,"162. 6. SheikhAbuiMufarrij'sshunaitself is mentioned in the Quseir documents as the destination for deliveries.See, for example,text 1 (RN 1026a;Guo, "Business Letters")and text 5 (RN 988;Li Guo, "ArabicDocuments from the RedSea Port of Quseir in the Seventh/Thirteenth Century,"pt. 2, "ShippingNotes and Account Records,"JournalofNear EasternStudies6o (2001): 99 -101. 7. This descriptionof a shunaaccordswell with that offeredby M. Amin, Layla'Ali 95
QUSEIR AL-QADIM
IN THE THIRTEENTH
Ibrahim,and M. Muhammad,ArchitecturalTermsin MamlukDocuments:648923 H/125o-1517(Cairo:AUCPress,1ggo), 71-72. 8. The keys are describedin FredrikT. Hiebert, "CommercialOrganizationof the EgyptianPort of Quseir al-Qadim: Evidencefrom the Analysisof the Wooden Objects,"ArcheologieIslamique 2 (1991),157,fig. 1o(82-524). 9. Guo, "BusinessLetters,"161-go; Guo, "ShippingNotes,"81-ii6; Li Guo, Commerce,Culture,and Communityin a RedSeaPortin the ThirteenthCentury: TheArabicDocumentsfromQuseir, IslamicHistory and Civilization Studies and Texts (Leiden:Brill, 2004). lo. Guo discusses this semiofficialrole that the sheikh apparentlyplayedin the grain trade ("BusinessLetters,"176,178,179,n. 7). The primaryinteractionwas with the cities of the Hijazas discussed in Guo, "ShippingNotes,"89,93;and Guo, Commerce,Culture,and Community,45. A letter from anotherhouse (P8b-i8;78592)describesthis commerce in grain. See GladysFrantz-Murphy,"TheRedSea Port of Qusayr:ArabicDocuments and NarrativeSources,"in Donald S. Whitcomb and JanetH. Johnson,Quseir al-Qadim 1980:PreliminaryReport, AmericanCenterin Egypt Report7 (Malibu:Udena Publications,1982), 273-77. 11. Foran explanationof the workingsof the sheikh'sbrokeragebusiness accordingto the Quseir documents and letters, see Guo, "ShippingNotes,"81-85;and Commerce,Culture,and Community, 9-69. 12. Guo, Commerce,Culture,and Community,2-3. One learnsthe names of the sheikh'sother sons in various letters (Guo, "BusinessLetters,"165);in the abovenarrative,their identitywith those strandedin QasrYamani(Guo, "Business CENTURY
Letters,"181-85)is the presentauthors' conjecture. 13. Li Guo, personalcommunication to Donald Whitcomb, March20, 2003. 14. Vogelsang-Eastwood,ResistDyed Textiles,cat. nos. 52, 53. 15. Vogelsang-Eastwood,ResistDyedTextiles, 52,112,fromKgb-63 (82-922,945). 16. Vogelsang-Eastwood,ResistDyed
Veil from Egypt," Costume 17 (1983):
Textiles, 52, 111, cat. no. 51. 26. Some ceramiccorporawith similar
materialsare: ClaireHardy-Guilbert, "Julfar,cite portuairedu Golfe arabopersique'ala periode islamique,"
Textiles, 19,113, cat. no. 55, with a from Kloa-ii (82-927). 17. See text 4 in Guo, "ShippingNotes,"98;
Archeologie Islamique
and text 55in Guo, Commerce,Culture, 18. Whitcomb and Johnson, QuseiralQadim 1980,149; EdwardJ.Keall,"The
Arche'ologieIslamique
(1991):5-46.
GeorgeT. Scanlon, "FustatExpedition:
The Surveyof a Town on the Tihamah Plain of North Yemen,"WorldArchaeology14 (1983):379 -91; Christopher Ciuk and EdwardJ.Keall,ZabidProjectPottery Manual, 1995:Pre-Islamicand Islamic CeramicsfromtheZabidArea,North Yemen,BARInternationalSeries655
Preliminary Report, 1978,"Journal of the
Southampton,
2002).
19. Vogelsang-Eastwood,ResistDyed Textiles, 55, 113, cat. no. 58.
Vogelsang-Eastwood,ResistDyed Textiles, 54, 114, from Kioa-15 (82-937). Vogelsang-Eastwood,ResistDyedTextiles,
(1984):1-38.
dell'Istitutouniversitarioorientaledi Napoli 49 (1989):246, 259. 31. An earlierdiscussion of this locational
distinction in seeds and other finds appearedin Donald Whitcomb, "Quseir al-Qadim, Egypt:Textand Context in the Indian Ocean SpiceTrade,"al-'UsuralWusta:TheBulletinofMiddleEast
Christina Tonghini, Qal'at
Ja'barPottery:A Studyof a Syrian FortifiedSiteof theLate lith-14th Centuries(Oxford:OxfordUniversity,
Medievalists 32.
1998). A fuller discussion of the ceramics
at the Sheikh'sHouse and at Quseir alQadim can be found in a 2007 University of ChicagoPh.D. dissertationby KatherineStrangeBurke:"Archaeological Textsand Contexts on the RedSea: The Sheikh'sHouse at QuseirAl-Qadim." 27. Bo Gyllensvard,"RecentFindsof Chinese Ceramicsat Fostat,"pt. 1,Bulletinof the MuseumofFarEasternAntiquities, Stockholm 45 (1973):92. Also see Gyllens-
(82-905); check patterns on linen were
much more common; a particularly interestingpatternwas from Kgb-36
(1975):93-117; Robert Lockhart Hobson,
"ChinesePorcelainfrom Fostat," BurlingtonMagazine(September1932):
(82-913).
Includingone addressedto Ibrahimb. AbuiMufarrij.Li Guo, personalcommunication to Donald Whitcomb, Novem-
1o9 -13; Oscar Raphael, "Fragment from
Fustat,"Transactionsof the Oriental Ceramic Society
(1923 -24):
17-25;
Tatsuo
Sasaki,"ChineseCeramicsExcavatedat Fustat, 1985,"TradeCeramicsStudies6
2002.
23. Vogelsang-Eastwood,ResistDyed 96
KATHERINE
STRANGE
7.2 (1995): 25-27.
Jean-ClaudeGarcin,"Pourun recoursa l'histoire de l'espaceve,u dans l'etude de l'Egyptearabe,"Annales:Economie,
Societes, Civilisations35 (1980): 436-51. 33. Frantz-Murphy,"Arabicdocuments,"
vard, "RecentFindsof Chinese Ceramics at Fostat,"Pt. 2, Bulletinof theMuseumof FarEasternAntiquities,Stockholm47
deredfragmentswerefound in Kgb-38
ber 25,
Aththar, 217-108, Saudi Arabia," Annali
AmericanResearchCenterin Egypt21
53,113, cat. no. 55. The sheikh's embroi-
22.
2
Dynamics of Zabid and Its Hinterland:
RebeccaBridgman,Report2001, The IslamicPottery:SomePreliminary Findings(Southampton:Universityof
21.
2 (1991): 161-203.
Cassirer, 1971),81-95, 271-91.
28. TsugioMikami, "Chinaand Egypt: Fustat,"Transactionsof the Oriental CeramicSociety45 (18o-81): 69. 29. Gyllensvard,"RecentFinds"pt. 2,97, pl. 3:1, 2, 4. 30. JurisZarins, "ArabSouthernRedSea Portsand the EarlyChinese Porcelain Tradeas ReflectedPrincipallyfrom
Axelle Rougeulle,"Lesimportations de ceramiqueschinoises dans le Golfe arabo-persique(VIIIe-XIe siecles),"
and Community, 251-52
20.
Vogelsang-Eastwoodpublisheda wellpreservedveil from this site:Gillian M. Vogelsang-Eastwood,"AMedievalFace-
33-38. 25. Vogelsang-Eastwood,ResistDyed
discussion of the derivationof motifs,
(Oxford: Tempus Reparatum, 1996), 108;
(1986): 99-103; George T. Scanlon, "Egypt and China:Tradeand Imitation,"in Islamand the TradeofAsia:A Colloquium, ed. Donald S. Richards(Oxford:Bruno
Textiles, 53,113, cat. no. 54. 24.
BURKE AND DONALD
267-83; the "ostrichicons" are figs. 18,19.
A similar funerary"ostrichicon"from Quseir is publishedby AshrafSalamafor the Misr InternationalUniversity (website:archnet.org).Another reportis from the Universityof Southhampton excavations:D. A. Agius, "'LeaveYour Homeland in Searchof Prosperity':The OstrichEggin a BurialSiteat Quseir alQadim in the Mamluk Period,"in Egypt and Syriain theFatimid,Ayyubidand MamlukErasIV,ed. U. Vermeulenand J. Van Steenbergen,OrientaliaLovaniensia Analecta 140 (Leuven: Peeters, 2005), 355-80. The term "ostrichicon" was
suggestedby the late Prof.Robert Braidwoodand is retainedhere in his memory.Another referenceto ostrich egg WHITCOMB
shells maybe noted in the observationof Ibn Battuta,when he arrivedon the African coast south of Quseir:"We landed there and found on the shore a hut of reeds, shapedlike a mosque, inside were ostrich egg-shells filled with water...".H. A. R. Gibb, The TravelsofIbn Battuta,1325-1354(London:Hakluyt Society,1962), vol. 2,107. 34. Vogelsang-Eastwood,ResistDyed Textiles,34, cat. no. i. 35. Donald S. Whitcomb, "IslamicPottery," in Whitcomb and JanetH. Johnson, Quseiral- Qadim1978:PreliminaryReport (Cairo:American ResearchCenterin Egypt, 1979),1o8. 36. CharlesLeQuesne, "QuseirFortVisitor Centre:ArchaeologicalReport," American ResearchCenterin Egypt, Cairo, 1999. 37. Jean-ClaudeGarcin,"K. usayr,"Encyclopedia of Islam,2nd ed. (Leiden:Brill, 1982), 519. 38. Albert Kammerer,LaMerRouge: l'Abyssinieet l'Arabiedepuisl'antiquite (Cairo:Societe royalede geographie d'Egypte, 1929), 1:8i. 39. This seems to be the earliestcitation of this erroneousidentification.See Donald Whitcomb, "Quseiral-Qadim and the Locationof Myos Hormos,"Topoi6 (1996):747-72. 40. Excavatedas a coin, B4a-5,78-446. 41. GustaveFlaubert,Voyageen Orient (1849-1851) (Paris:Librairiede France, 1925), 121.
97
QU$EIR AL-QAD1M
IN THE THIRTEENTH
CENTURY
ELIZABETH
LAMBOURN
CARVING AND COMMUNITIES MarbleCarvingforMuslimPatronsat Khambhatand around theIndianOceanRim,LateThirteenth-Mid-Fifteenth Centuries
Abstract Foraperiodof somesixcenturies-from thelatetenthto thesixteenthcenturyA.D. -Khambhat was the premierport of westernIndia.Positionedat the interface betweennorthernIndiaand the IndianOceantradingnetwork,Khambhatwas not onlya transitpointforcommoditiesfrombothareasbutalsoa productionsite in its own rightand a majorcenterof Muslimpatronageof architectureand fine marblecarving.Thisarticleexaminesthe productionof marblecarvingsforMuslim patronsatthe portof Khambhatandits seaborneexportto patronsaroundthe IndianOceanlittoral,fromEastAfricato easternJava.Takingthe approachthat everystonecarvingis a documentthat can tell the storyof its own manufacture, and couplingthis approachwith stylistic,iconographicand epigraphicanalyses, the articleexaminesthe processesof the commodity'sproductionat Khambhat, andpatternsof consumptionby patronsin Gujaratandabroad.
TODAY,STONEAND EVENSTONECARVERS travelinternationally. Inthepreindustrialperiod,however,the high densityof stone,translatedas a high weight, andthe difficultiesof transportassociatedwiththis guaranteedthatstone,andthe stonecarverswhoworkedit, generallystayedcloseto the areain whichit wasquarried.Stonecarvingin the premodernperiodhasbeencharacterizedprincipallyas a localcraft,basedaroundlocal sourcesof stoneandusinglocalterminologyand techniques.,It is not surprising,then,thatstoneis nevermentionedas a commodityin the literatureon IndianOceantradeduringthe premodernperiod. Nevertheless,there were exceptionsto this general"rule."The transportof marblearoundthe Mediterraneanduringthe Romanperiod,both as a rawmaterialand in the formof finishedor partlycarvedpieces,is perhapsthe best documentedandmostwidelystudiedexampleof nonlocal,sea-basedcirculation.2 But the Mediterraneanmaynot be the only geographicalareawherestoneandindeed stone carverscirculatedin this manner.While little attempthas been made to translatethese models to Asia, there is growingevidenceboth in the field and acrossexistingpublicationsthatcarvedstonewas sometimescommissionedand shippedacrossvast distances.Whilestill in its earlystages,this researchsuggests the existenceof a numberof discretenetworksof stone distributionaroundthe IndianOceanandSouthChinaSeaduringthe pre-industrialperiod,particularly, but not exclusively,involvingIslamictombstones.Networkscan be tracedout of FujianProvincein easternChinato Brunei,the Philippines,and even as far as northSumatrafromthe thirteenthcenturyonward.3Thereis strongevidencefor an intensetradein Islamictombstonesaroundthe Indonesianarchipelagofrom the earlyfifteenthcenturyonward.4A numberof Vaisnavaand Buddhiststatues 99
1
Marblegravememorial of'Umar al-Kazeruini(d. 734 A.H./A.D.1333), Kazertinitomb complex, Khambhat.FromJamesBurgess, On theMohammadanArchitecture ofBharoch,Cambay,Dholka, Champanir,and Mahmudabadin Gujarat,ArchaeologicalSurveyof Western India 6 (London:W. M. Griggsand Sons, 1896),pI. 23.
found on the coasts of northeastern and western Sumatra may have been imported from southern India sometime between the eleventh and fourteenth centuries.5 My own research on the marble carving produced for Muslim patrons at the port of Khambhat in Gujarat, India, quite unexpectedly led to the discovery of a significant number of pieces bought or commissioned by Muslim patrons around the Indian Ocean rim, from sites in East Africa to Java,between the late thirteenth and mid-fifteenth centuries.6 Other networks may well remain to be identified since the possibility of a sea-based spread of stone and carved objects has hardly been looked for in these regions. It is still far too early to be able to define and analyze the larger patterns of this phenomenon in Asia. For the moment, a great deal of empirical research remains to be carried out to establish the extent of these networks and to document and analyze individual groups of stone carvings. As the following article will show, however, this empirical starting point by no means precludes the development ofanalytical or interpretive approaches to individual groups. I will discuss the westernmost example of this phenomenon identified so far: the production and export of marble carvings from the Gujaratiport of Khambhat during the period spanning the late thirteenth to mid-fifteenth centuries. While I have had occasion to publish several groups of exported Khambhat material in detail,7 this article is an opportunity to look at the full spread of export production and, most important, to set it in the context of production for the home market. The article focuses particularly on the communities involved in the production of marble carving for Muslim patrons at Khambhat and on the various Muslim communities that purchased or commissioned this work. What was the background of those involved in this carving, and how were workshops organized, particularly around work for overseas patrons? Who purchased Khambhat marble carving 100
ELIZABETH
LAMBOURN
7 t
~~~~~~~~
...~~S
I~
''
*
...
3
.
.
3
2
2
Marblecarving of the Jaina tlrthankaraVasfipujya,dated v.s. 1215. Privatecollection, 1271/A.D. Khambhat.Photographcourtesy of the American Institute of Indian Studies, Centre for Art and Archaeology,Gurgaon. 3 Marblecarving of the yaksi Ambika, thirteenth century. Private collection, Khambhat.Photograph courtesy of the American Institute of Indian Studies, Centre for Art and Archaeology,Gurgaon.
and why, and did the profile of these consumers differ substantially between local and export markets? Though we have no carvers' "signatures"and no contemporaneous accounts of this production, this article is grounded on the idea that "any object worked in stone is a document that, correctly understood, describes its own manufacture."8The material provides a rich body of technical, stylistic, epigraphic, and even iconographic data illuminating aspects of stonecarving and its trade.
The Corpusof KhambhatMarbleCarving The marble carving produced for Muslim patrons at Khambhat has come to be embodied in the minds of most scholars by the grave memorial of the merchant 'Umar al-Kazeriini (d. 734 A.H./A.D. 1333),situated at the center of his magnificent tomb complex attached to the congregational mosque at Khambhat (fig. 1). His memorial was the first Khambhat grave to be published with illustrations, and the volume it appeared in, James Burgess's On the Muhammadan Architecture of Bharoch, Cambay, Dholka, Champanir, and Mahmudabad in Gujarat (1896), remains a classic to this day.9However, al-KazerLini'sgrave is only the most cited and reproduced example of a large corpus of carved marble, including foundation inscriptions and architectural carvings, produced for Muslim patrons at Khambhat from the early thirteenth century right up to the end of the nineteenth century.'0As we will see, this production grew out of the strong preexisiting marble-carving activity at the port for Jaina and Hindu patrons. A few surviving fragments, such as a relief of the tirthankaraVasupujya, dated v.s. 1271/A.D. 1215 (fig. 2) and a thirteenth-century figure of the yakst Ambika (fig. 3), testify to the quality of this production."lAlthough evidence can be found for the production of fine marble carving for Muslim patrons at Khambhat through the 101
CARVING
AND COMMUNITIES
late nineteenth century, there is only one period for which we have evidence of a stable and sustained production, and that is the period between the late thirteenth and mid-fifteenth centuries. The large numbers of carvings that survive for this period and the uniformity of their style suggest that one or several marble workshops were established at the port and working specifically for Muslim patrons. Significantly, it is also from these centuries that there is evidence of the export of Khambhat marble carving.
4 Inked rubbing of the marble headstone of Hasan al-JIraq (d. 699 A.H./A.D.
1299), Somandtha-pattana
FromAnnual Reportson Indian Epigraphy (1954-55).
Marble Carving for Muslim Patrons at Khambhat The corpus of Khambhat carving for the period between the late thirteenth and mid-fifteenth centuries (excluding, for the moment, the export production) includes some forty-three cenotaphs and/or headstones,12nineteen foundation inscriptions,13one land grant,14and eight architectural fragments, including mihrab elements.J These numbers suggest that grave memorials constituted the mainstay of production, followed by foundation inscriptions and architectural elements. This breakdown of types fits with what we know of building traditions and materials at Khambhat. In the alluvial plain of central Gujarat,where Khambhat and the other major cities of Gujaratwere located, the only locally available building material was brick, and the majority of construction is likely to have been in brick supplemented by timber for roofs, internal supports, and timber frames.'6 All stone would have had to be brought in from other areas-northern Gujaratfor marble, Saurastra for lower qualities of stone such as sandstone -or else extracted from existing structures."7Either way, stone constituted an expensive material and in the context of Muslim religious architecture was in all likelihood reserved for focal areas such as foundation inscriptions, mihrabs,and doorways. This corpus probably represents only a fraction of the marble actually produced for Muslim patrons during this period, since the later decline of the port and difficulties in obtaining marble seem to have led to an increase in the recutting of earliergravememorials. Before further discussing Khambhat marble carving, it is useful to give a short overview of the style of material produced and the development of the corpus over this period. While there is a considerable variety of headstone design during the 1280s and 1290S, some key features emerge quite clearly. Inscriptions constitute the main type of decoration, but most examples also carry a lamp motif, something of a fashion on grave memorials throughout the Islamic world at the time.18 Though there were variations in design, the majority render the lamp motif in a rich, rounded relief, placing it at the center of an arched niche and flanking it with split plantains. The headstone of Hasan al-'Iraqi (d. 699 A.H./A.D. 1299) (fig. 4) from Somanatha-pattana on the coast of Saurastra is a good example of this type and also provides the earliest dated rendering of the Kufic bismillah design that was to 102
ELIZABETH
LAMBOURN
4*
4,
*v.
0i'a
SE! i f
i
1XE1ir f <
|7?P
15S
-.
as~'' 1
-
_
*
5
Marbleheadstone of Muhammad al-Astarabadi(d. 683A.H./A.D. 1284). Dar al-Atharal-Islamiyya, Kuwait. Photographby the author. 6
Marbleheadstone of IkhtiyaralDawlawa al-DYn(d. 716A.H./A.D. 1316). Khambhat. Photograph by
the author. 7 Larmihrab,marble, first half of the fourteenth century. Museum of Historical Stone Carving, Shiraz, Iran. Photographcourtesy of Sima Arshadi,Curator,Museum of Historical Stone Carving, Shiraz.
mark Khambhat carving for the following century and a half. The headstone of Muhammad al-Astarabadi (d. 683A.H./A.D 1284)provides an example of an undecorated, purely epigraphic headstone produced around the same period (fig. 5). By the second decade of the fourteenth century the basic design had been refined to include an ornate parikara(a frame of miniature niches containing lamps) and a cusped arch head as, for example, in the headstone of Ikhtiyar al-Dawla wa al-Din (d. 716A.H./A.D. 1316) (fig. 6). By the early fourteenth century there is far less variation in design than during the later thirteenth century, suggestive of a large-scale production relying, in part, on precarved stock. However, as al-Kazergini'sheadstone demonstrates (see fig. i), this was not the only model available. Tombstones also varied in size, presumably to accommodate different budgets. Headstones of both types -lamp with split plantains and bismillahdesign- continued into the 1340S.The high-relief lamp with or without split plantains also provided the main decorative motif for cenotaph sides (see fig. 1) and mihrabs,as exemplified by the so-called Lar mihrab(fig. 7). There appears to have been a lull in fine grave memorial production during the second half of the fourteenth century.19However, fine marble memorials, this time without substantial architectural carvings, reappearat the end of the century and continue until the mid- or even late fifteenth century. The fifteenth-century headstone type was heavily influenced by earlier models such as al-Kazeruini'sheadstone, and the focus of decorative elaboration was on the Kufic bismillahacross the shoulder of the stone. The headstone of AbuiBakr al-Quraishi (d. 818A.H./A.D. 1415) at Khambhat exemplifies this type (fig. 8). Some of the most stunning examples of carving from the fifteenth century are the large arched footstones, carved in flatcut relief with tree and plant motifs, that have survived on many Khambhat graves in northern Sumatra, for example, the footstone of grave 6 in the Teungku Sareh burial group, possibly recording a death in 834 A.H./A.D. 1430-31 (fig. 9). Produc103
CARVING
AND COMMUNITIES
8
8
Marbleheadstone of AbfuBakr al-Quraishi (d. 8i8 A.H./A.D. 1415). Khambhat.Photographby the author.
tion forboth the homeandexportmarketsappearsto tail off towardthe middleof the fifteenthcentury,possiblyas a resultof increaseddemandfor marblecarvers forprojectsat the newcapitalof Ah.madabad. Epitaphsand other inscriptionsconfirmthe high statusof workedmarbleat Khambhat.Whileit is usuallyimpossibleto builddetailedbiographiesof the individualsmentionedin inscriptions,theirtitles and nisbasappearto point to three majorgroups:merchants,religiousdignitaries,and noblesand officialspostedin Gujarat.20 Epitaphsincludeseveralkhwajas,the greatmerchantsof Iranandneighboring countries,and a number of merchantsholding the title malikal-tujjdr, including 'Umar al-KazerfinL.2A number of nisbasand 'urfs(aliases) point to specifictrades,forexample,two 'alamgdrs(lance or bannermakers),a lakar7ya(associatedwith or relatedto wood), and two fapfl7yas (fatol is an adjectivemeaning "soft"andis appliedto betelnuts).22Otherheadstonesrecordthe demiseof a qadi,
two imams(one of whichwas also a muft!),and threesadrs.Amongthe noblesor officialsmentionedone can singleout Ikhtiyaral-Dawlawa al-Din, the bahrbek (commanderof the sea)at Khambhat(seefig. 6),23Shamsal-Din Mahmud(d. 732 A.H./A.D. 1331),al-dabir (secretary),24and Fakhral-Dawlawa al-Din Muhammad al-BiitiharI,a Tughluqofficialandarmycommander,the patronof thejami'mas104
ELIZABETH
LAMBOURN
9
9
Marblefootstone, grave6, Teungku Sareh,Samudera-Pasai(Sumatra), possibly 834 A.H./A.D.
1430-31.
LeidenUniversityLibrary,Legatum Warnerianum,The Netherlands, Or.23.48i, photograph 386.
jid.25Many of the names recorded on Khambhat inscriptions have geographical nisbas referring to places outside India, for example, in Iran, the Arabian peninsula, and the Near East. However, these nisbas do not necessarily indicate that they were recent arrivals, and many families may have been resident in western India for generations.26 Khambhat Carving around the Indian Ocean Rim27 Undoubtedly the most surprising aspect of Khambhat marble carving is the fact that this production found an international appeal well beyond its origins. To date twenty-eight grave memorials, complete and fragmentary, and four groups of architectural elements have been found at fifteen sites around the Indian Ocean rim (fig. io). These include a cenotaph side panel in two parts from Kilwa in present-day Tanzania;28a mihrab,a portal facing, and other carved elements produced for the Fakhr al-Din mosque in Mogadishu;29one tombstone from Aden;30three tombstones from Dhofar in southern Oman;31a mihrab found at Lar in southern tombstones at Somanatha-pattana on the Saurastra coast (see Iran (see fig. a fig. 4),33 Rander in southern Gujarat,34 Goa,35and Madayi in northern Kerala;36 two tombstones from on the coast of Rajpuri Maharashtra;37 funerary inscription from Trincomalee in eastern Sri Lanka;38a group of twelve cenotaphs and/or tombstones from Pasai in northern Sumatra (of which see fig. 9)39 and finally three cenotaphs and/or tombstones and a carved column from Gresik in eastern Java.40 All these pieces are carved in the creamy white marble characteristic of western India. Moreover, many pieces carry traces of temple spolia in a style typical of medieval western India, irrefutable proof that the marble itself came from this region.41 The design and decoration of the pieces, their carving technique, and the style of their calligraphy all correspond to known Khambhat prototypes. Surprisingly, although many of these carvings have been published, only the three stones 7);32
105
CARVING
AND COMMUNITIES
from Dhofarin Oman and one gravefrom Pasaiwere ever paralleleddirectly with Khambhatmaterialand sourcedto the port.42Gatheredtogether,theypaint a wholly unexpectedpictureof exportactivity;and theremay still be room for growthin the numbers. Eightwhite marblecolumnsin the 887A.H./A.D. 1482Manstirlyyamadrasaat The Jubanin the Yemenshouldprobablyalsobe addedto the Khambhatcorpus.43 columnsare carvedwith typicallyIndianmotifs such as a suspendedchain and floralbands.A recentstylisticanalysiscomparingthem againstvariousregional traditionsof Islamicarchitecturein Indiafoundthe strongestparallelsto be with Whilethis analysisdid not point to Khambhatas GujaratiIslamicarchitecture.44 a possiblesite of production,it seems reasonableto suggestthat these eight columns mayhavebeenmanufacturedatKhambhatsinceit is currentlythe onlysite in westernIndiawith a provenoverseasexportproduction.45 Reportsof further Khambhattombstonesin the Maldivesand Aden,if substantiated,may increase Forthe moment,howthe scale and spreadof this exportproductionfurther.46 ever,these new additionsyield a total corpusof Khambhatcarvingfor Muslim brokendownin the followingmanner: patronsof iio pieces,47 Corpusof KhambhatCarvingfor MuslimPatrons,LateThirteenthLateFifteenthCentury TABLE1
TOTAL Khambhat IndianOceanRim andEnvirons (excludingKhambhat) (bytype) Headstonesand/or CenotaphsandFunerary Inscriptions
43
FoundationInscriptions (includinglandgrants)
20
Architectural Carvings TOTAL (by location)
28
71
0
20
8
11
19
71
39
110
Thequalityof KhambhatmarblecarvingobviouslyraisesthePpossibilitythatthese pieces left Khambhat,not betweenthe thirteenthand fifteenthcenturiesat the time of theirmanufacture,butduringlatercenturies,throughspoliationafterthe declineof the port.In this regard,it is importantto be ableto establishclearconnectionsbetweenthe individualmentionedin the epitaphand the sitewherethe graveis or waslocated.Whilethis connectioncannotbe madein all cases,enough 106
ELIZABETH
LAMBOURN
Somanath-pattana: i tombstone
m
8 columnsKhmht Juin:it
10
to
10
10
Diffusion of Khambhatmarble carving around the Indian Ocean rim, late thirteenth to mid-fifteenth centuries.
links can be establishedto supportthe view that this patternof dissemination was generatedby contemporaneousexport productionratherthan laterspoliation. Oneof the bestknownexamplesis thatof ShaikhAbuiBakral-Damiri(d.714 A.H./A.D. 1315), whose headstonewas found at Dhofarin southernOman. In this instance,we are fortunatein havingthe near contemporaneousaccountof Ibn Battuita,who speaksof al-Damirias one of the religiousdignitariesat the court of the Rasuilidgovernorof Dhofar.48Similarly,certain of the gravememorials foundat Pasaiin northernSumatrarecordthe epitaphsof membersof the family of the sultansof Samudera-Pasai, for example,the gravememorialof a daughter of SultanZainal-'Abidinof Pasai(d.831A.H./A.D. 1428),whosenamesandtitlescan be corroboratedbylocalhistoriesandotherlocallycarvedgravestones.In the case of the headstoneof Hasanibn Muhammadibn 'Allal-'Iraqi(d. 699 A.H./A.D. 1299) at Somanatha-pattana on the Saurastracoast,an olderinscriptionfromthe same town informsus thathis familyoriginallycameto SomanathafromHormuizand hadbeentheresinceatleastthe mid-thirteenthcentury.49 OverseasPatronsand the Tastefor KhambhatMarble For the moment, there is no evidence to suggest that Khambhatmarblewas exportedforthe gravesof Khambhatnativeswho had died abroador weresettled outsidethe port.50The only individualwho can be clearlylinked to Khambhatis 'Ala'al-Din 'All,the son of MalikMufarrah. Sultani(d. 768A.H./A.D. 1366),whose His father funeraryinscriptionwas foundat Rajpurion the Maharashtracoast.51 was a well-known noble and patron of the Tughluqperiod who was initially appointedto the post of nd'ib(delegate)of the 'iqta'(administrativegrantof land fromthe centralstate)of Khambhat.52 On the whole,then, the datacarriedin the 107
CARVING
AND COMMUNITIES
epitaphsof Khambhatheadstonesfound outsidethe port underlinethe ideaof a luxuryproductionthatappealedto localelitesacrossethnicor regionalidentities. The nisbaof ShaikhAbuiBakral-Damiri(d. 714 A.H./A.D. 1315) suggeststhathe was of Egyptianorigin,comingfromone of the two smalltownsnamedDamirain the Nile delta.The Rasuilidshad close contactswith Mamlik Egyptduringthe fourteenthcentury,andtheymayhaveplayedapartin thearrivalof anEgyptianshaikh at the Rasuilidcourt.Otherheadstonesmarkthe gravesof merchantsand nobles. As we have seen, Uasan al-'Iraqi,who died in 699 A.H./A.D. 1299 at Somanathapattana,belongedto a merchantfamilyoriginallyfrom Hormuizin the Persian or GreatLeaderand Noble, Gulf;however,his titles sadral-suduirwa al-akcibir, suggestthathe also playeda religiousand administrativeroleat the port (seefig. 4). One of the inscriptionsfrom Trincomaleerecordsthe demisein 729 A.H./A.D. 1329of the daughterof a Syriannoble,AmIrBadral-DinHusainibn'Alial-Halabi. in northernSumatrais especiallyrich Thegroupof epitaphsfromSamudera-Pasai in informationaboutthe purchasersof Khambhatmarblecarving,andthe graves seemto belongto the familyand courtiersof the sultansof Samudera-Pasai. Two of the earliestimportedKhambhatgravesin northernSumatracommemoratethe gravesof a descendantof the penultimate'Abbasidcaliphandhis wife, or perhaps a femalerelative.'Abdallahibn Muhammadibn 'Abdal-Qadiribn 'Abdal-'Aziz ibnal-MansirAbuiJa'faral-'Abbas! al-MuntasirbillahAmiral-Mu'minlnKhalifa Rabbal-'Alamindied in 809 A.H./A.D. 1406, and his wife, or femalerelation,Sitti Jahanbint al-Malikal-Mu'azzamwa al-Amiral-Mukarram... Khandied in 816 A.H./A.D. 1414.Anotherpairof Khambhat tombstonescommemoratesthe graveof a certainNa'inaHusamal-DinibnNa'inaAminwho diedin 823 A.H./1420 C.E. The title nd'indseemsto be the Arabicrenderingof a Tamilcasteor communityname and suggeststhat Husamal-Din was a TamilMuslim,probablyone of the many Tamilmerchantsknownto havebeenbasedatSamudera-Pasai.53 It is not difficult to understandthe success of Khambhatmarblecarving beyondits local market,since it ranksamongthe finestmarblecarvingin India of the periodandholdsits own with the highestqualitystonecarvingfromacross the Islamicworld.The color,texture,and translucencyof the primarymaterial, coupledwith the refinementof its carving and design, must have elicited the admirationof all who saw it. Khambhat'sstatusas the premierport of western Indiacertainlyguaranteeda maximumaudienceamongthe manyforeigntravelers and merchantswho visitedthe port,prayingin its mosquesand perhapsvisiting its shrinesand cemeteries.It is even possiblethat news of this production reached Syria by the early 1320S, since a passagein the Taqwirnal-bulda~n (Table of Countries), a geographical treatise by Abti al-Filda',the Ayyuibid prince of Hama in Syria, finished in 721 A.H./A.D. 1321, describes Khambhat as "a beautiful 108
ELIZABETH
LAMBOURN
city,biggerthan Ma'arrain Syria.Itsbuildingsarebuilt in brickand thereis also white marble."54 Ab5 al-Fida"stext was laterincorporatedinto Shihabal-Din alQalqashandi'sgreatMamlfikencyclopedia,the Subhal-a'shafisina 'atal-insha', finishedin 814 A.H./A.D. 1412,55 andthe successof this text in the Islamicworldmay marblecarvingworkshops. havecontributedto awarenessof Khambhat's One may also wonderwhetherthe attractionof Khambhatmarblecarving beyonditsproductionsitewasmotivatedin partbythe closeparallelsbetweencertain decorativemotifson the gravememorialsand contemporaneousexporttextiles producedin Gujarat.TherearestrikingparallelsbetweencertainKhambhat footstones,with rowsof mangoand bananatreesor with a single,largetree (see fig.9), andresist-dyedcottonsmanufacturedin westernIndiaforexportthroughout the Near East,the Arabianpeninsula,and SoutheastAsia duringthe same period.56 Although these cottons were relativelycommon commodities, their motifs might reflectsimilarmotifs used on now lost high-statustextilessuch as velvetsand silks.57 It also seemslikelythat overseasclients'interestin Khambhat carvingrosein proportionto the rarityof fine stonesin theirown regionandthe level of sophisticationof local stone-carvingtechniques.Evenin westernIndia, no othercoastalsitehasso farshownevidencefora similarlydevelopedandwellestablishedmarble-carvingtradition,althoughcentersdid exist inland. Beyond westernIndia,the workshopsat Khambhathad a quasimonopolyon whitemarble and its carving.Geologicaldataon EastAfrica,Oman,SouthIndia,northern Sumatra,andeasternJavaindicatethatnoneof the stonesourcesor stone-carving traditionsin theseregionscouldcompetewith Khambhat.58 Thesinglefactorthatmadesucha widespreadexportphysicallypossible,however,was Khambhat'sport. With marbleaveragedat twenty-sevenquintals,or threeEnglishtonspercubicmeter,largecenotaphgraveswouldhaveweighedseveraltons.59 Watertransport,whichin Gujaratcan be equatedwith sea transport, overcamethis problem.Once loaded,Khambhatmarblecould travelenormous distances.Indeed,carvedmarbleconstituteda valuableballastto help load and balanceships. Export Patterns
Gravememorialswerebyfarthemostpopularexportcommodity(seetablei). The relativelylargenumberof survivinggravememorialsfound outsideKhambhat (twenty-eight)in comparisonto survivalsin andaroundtheport(forty-three)initiallysuggeststhatexportproductionwas extremelyimportantto the Khambhat workshops,constitutingover half of gravestoneproduction.The numbersfor Khambhatshould probablybe adjusted,however,to take accountof the port's declinein latercenturiesandthe widespreadrecarvingandrecyclingof its carved 109
CARVING
AND COMMUNITIES
marble.Evenif the twenty-eightgravememorialsfound outsideKhambhatrepresentonly half the originalexportproduction,it is clearthat tombstonesnever becamea widelytradedcommodity.Thesetombstonesarethe fruitof individual purchasesor commissionsdirectlyfrom the Khambhatworkshops;the private characterof these purchasesexplainswhy Khambhatmarblecarvingneverfeaturesas a tradeproductin the sourcesof the period. One strikingaspect of the spreadof Khambhatmarblecarvingis the contrast between the late thirteenth and fourteenthcenturies,on the one hand, and the fifteenth century, on the other. During the first period, Khambhat marbleappearsto havebeen exportedexclusivelyin the westernIndianOcean, from East Africa to Sri Lanka,and to numeroussites. By contrast,in the fifteenth century,export focusedalmost exclusivelyon the easternIndianOcean and on just three sites;the single exceptionis the group of pillarsat Jubanin the Yemen.The shift is difficultto explain since it does not correlatewith any known change in the patternof trade aroundthe Indian Ocean during these three centuries.As far as we know,tradeand contactsbetweenKhambhatand Samudera-Pasai wereas activein the fourteenthas in the fifteenthcentury,and Khambhatgravestoneswere availablefor export as earlyas the late thirteenth century.Their appearancein Samudera-Pasaiwould seem to be due to local changesin taste and perhapsburialpracticesin the earlyfifteenthcentury.By contrast,the cessation of exports to Samudera-Pasaiin the mid- to later fifteenth centuryappearsto correspondboth with a decline in marblecarvingat Khambhatitself and with a period of intense political upheavalin SamuderaPasai.Thereis no evidencefor a regularproductionof tombstonesat Khambhat duringthe secondhalf of the fifteenthcentury,and it is duringthis periodthat earlierheadstonesbegin to be recarvedand reassembledat Khambhatfor local usage (as opposed to recarvingfor export,which was practicedfrom the early fifteenth century).Although there are no suggestionsof any particulareconomic downturnat Khambhatduring this period, it is possiblethat its marble carversfound largerand more regularcommissionsfrom the sultan and his court at Ahmadabadthan their traditionalbase of merchantpatronageat Khambhat.At the other end of the chain, these trendscoincidedwith a period of intensepoliticalinstabilityin Samudera-Pasai that in all likelihoodhindered tradeand made complexoverseascommissionsunworkable. Communitiesof Carversat Khambhat Thecorpusof marblecarvingfromKhambhatwasmanufacturedin a cosmopolitan environment,reachingan internationalmarket,andfindsits placewithin the vastnetworksof long-distancetradethatcrossthe IndianOceanandSouthChina 110
ELIZABETH
LAMBOURN
WI.I 11
Detail of marble mihrdbniche commissioned by Khalis 'atiq 'Umar al-Kdzertini,726 A.H./A.D. 1326. Khambhat.Photographby the author.
Sea. However, explored at a micro, ratherthan macro, level, it also reveals valuable information about the organization of marble-carving workshops at Khambhat and the technical development of this craft in response to Muslim patronage. Sculptors and Marble Carvers Though we have no signatures or mason's marks on Khambhat carving, using Peter Rockwell's approach to carved objects as documents that tell the story of their own manufacture, a considerable body of information can be built up about the training and methods of the marble carvers who worked for Muslim patrons at the port. Since carving specialties tend to be organized around particular stones, it is almost certain that they were marble specialists and trained in the local "western Indian" tradition of carving, principally applied to temple sculpture and decoration. Since Khambhat is not situated in an area with natural stone reserves, it is unlikely that stone carvers, let alone specialized marble carvers, were permanently based at Khambhat except during large building projects or when there was a large-scale demand for certain types of marble carving. To this day it is the Gujarat-Rajasthanborder area that sustains resident bodies of marble carvers. It follows that these carvers originally came to Khambhat from other areas of Gujarat, probably northern Gujarat and southern Rajasthan, in response to patronage at the port. Given a sustained demand, however, it is not impossible that groups then settled there on a more permanent basis. The background of these carvers is worth exploring in light of the strength of influence from local temple sculpture on Khambhat carving for Muslim motifs at Khambhat display a profoundly patrons.60The mihirabsand mihra~b sculptural approach. Most mihrabsin India, and indeed in the Islamic world, are 111
CARVING
AND COMMUNITIES
w
-
12
. .
13
12
PafichamukhiVav,Dabhoi, midthirteenth century A.D. Photograph courtesy of the American Institute for Indian Studies, Centre for Art and Archaeology,Gurgaon. 13 Detail of fig. 12. Photograph courtesy of the American Institute of Indian Studies, Centrefor Art and Archaeology,Gurgaon.
either made up of flat slabs, carved with a niche motif, or are built niches, lined with decorated stone. By contrast, Khambhat mihrabssuch as that from Lar (see fig. 7) and the mihrab niche commissioned by al-Kazertini's freed-slave Khalis in 726 A.H./A.D. 1326 (fig. 1i) are sculpted from single blocks of marble, with no joins or additions. The niche of Khalis' mihrdbis carved to a depth of more than fourteen centimeters, with the lamp "hanging" at its center almost in the round, attached to the back of the niche by only three thin bridges of marble. The carving out of the mihrabniche and the treatment of the lamp both suggest an approach founded in traditions of sculpture in the round rather than low-relief carving. While the transfer of sculptural approaches to mihrab design seems perfectly logical in the Indian context, the Khambhat mihrabsappear to be the only stone mihrabsin the subcontinent to have allowed, or perhaps even encouraged, this transfer. A second aspect of the Khambhat mihrabs that ties them specifically to contemporaneous sculpture is their clear trompe l'oeil intent and technical display. The central niches of Khambhat mihrabs(see figs. 7, 11) have a life-size lamp hanging from minutely rendered chains. The shape of the lamp corresponds to the general shape and size of contemporaneous Islamic glass lamps, of which numerous examples survive from MamlulkEgypt and Syria. The sole inaccuracy is that the 112
ELIZABETH
LAMBOURN
J,
~
~
4
15
16
14
14
Detail of lamp motif with split plantains, undated marble headstone from A.D. 1310s-1320S. Khambhat.Photographby the author. 15
Wedding procession of the Jaina tirthankaraNeminatha, palm-leaf manuscript of the Subahukatha, dated v.s. 1345/A.D. 1288,Sanghavina PadaJnianaBhandar,Patan,Gujarat. FromMori Chandra,JainMiniature PaintingsFromWesternIndia (S. M. Nawab:Ahmedabad, 1949). 16 Marriageof the Jainattrthankara Pargvanatha,palm-leaf manuscript, thirteenth century A.D. Jfiana Bhandar,Jaisalmer,Rajasthan. FromSarabhaiManilal Nawab, The OldestRajasthaniPaintingsFrom JainBhandars,JainArt Publication SeriesNo 5 (S. M. Nawab: Ahmedabad, 1959).
Khambhat lamps hang from chains attached to the rims of the lamp rather than from chains attached to the handles on the body of the lamp, as in real examples. Nevertheless, the intention is clearly illusionistic, and from a distance the marble lamp hanging at the heart of the niche might have been mistaken for a real glass lamp suspended by a metal chain. Moreover, it is possible that the whole was originally painted, thereby furthering the illusion.6' This interest in trompe l'oeil appears to connect to similar notions in slightly earlier Indic architecture in Western India, for example at the fort of Dabhoi, dating to the mid-thirteenth century.62At a pavilion overlooking a tank, the so-called Pafichamukhi Vav (figs. 12, 13), sculptors introduced a high-relief frieze of figures rendered in three-quarter size, positioned as if they were seated within the pavilion overlooking the water. The "inhabited" balconies are a visual joke, evoking the real balconies within the structure and the living people who once moved inside and around them. There is also evidence of interaction with contemporaneous Jaina iconography. While lamp motifs were something of a fashion on mihrdbsand tombstones across the Islamic world during this period,61none include the split plantains seen on Khambhat carving (fig. 14).The motif appears to derive from representations of temporary structures, such as cauris, or wedding pavilions, as seen in Jain manuscript illustration. A miniature in a manuscript of the Subdhukatha,dated v.s. 1345/ A.D. 1288, depicts the wedding procession of the Jain tirthankaraNeminatha (fig. 15).The illustrator has picked the critical moment as the procession approachesthe bride in her marriage pavilion and Neminatha decides to renounce the world, and with it his bride. The pavilion has a distinct cusped roof; a lamp is suspended from the center of the structure and is flanked by two stacks of water pots and two split plantains. An even closer parallel comes from a scene in an undated thirteenthcentury manuscript showing the marriage of the tirthankaraParsvanatha (fig. 16). Here the mono-scenic mode of narration provides a visual unit even closer to that at the head of Khambhat gravestones and mihrdbs.It is important to stress that the niches with split plantains shown on Khambhat headstones and mihrabsare not marriage pavilions: the stacked pots, not the plantains, signal the marriage func113
CARVING
AND COMMUNITIES
^
~~~~~~~~~w
,
mod
tof
17
17
Mosque foundation inscription dated 674A.H./A.D. 1276,red sandstone. Sultanpurnear Gurgaon, Haryana. Ghulan Yazdani,"The Inscriptions of the TurkSultans of Delhi,"EpigraphiaIndicaArabic and Persian Supplement (1913-14), pl. loa.
tion of these structures. It is my suggestion, however, that they were added as the result of a visual habit, a natural association between the Islamic motif of the lamp hanging within a cusped arch requested by Muslim patrons and the cauri depicted in Jaina narrative cycles. Although pavilions constructed from split plantains were used for a variety of functions in India, in Jaina manuscript illustration and narrative carving the form appears to be employed almost exclusively for wedding scenes. It is thus very specifically from this one iconographic element that the motif passed into the decorative repertory of carving for Muslim patrons at Khambhat. Although the parallels come from Jainamanuscript painting, there is abundant evidence for a close correlation between narrative cycles in manuscripts and those found in architectural carving. It would not be unreasonable to suggest that identical scenes were carved in stone during the thirteenth century, probably by the same craftspeople who later carved Muslim headstones at Khambhat. Writing and Carving Islamic Inscriptions at Khambhat Muslim patronage must also have introduced new demands that radically altered the composition, working methods, and even the technology of marble workshops. The majority of carving on Islamic tombstones and foundation texts consists of inscriptions (the text or epitaph and associated Qur'anic verses), written in Arabic or Persian and carved in flat-cut relief. This method was a revolution for local stone-carving practice, where inscriptions were not a primary form of decoration and were unsually incised (compare fig. 2). Flat-cut relief inscriptions demanded several new and important contributions. First, they required presumably literate individuals capable of composing texts in Arabic or Persian, selecting programs of Qur'anic verses, and penning the texts for transfer to stone.64 Second, they required someone capable of carving them. The prominence of Islamic inscriptions on Khambhat marble carving means that one or several skilled scribes must have been closely associated with the workshop. Certainly by the early fourteenth century, when carvings were being produced in large numbers, one or several professional scribes must have been involved at the very heart of production. Without "signatures"it is difficult to establish their backgrounds. The style of script found on Khambhat headstones, described as a however, and on inscription slabs from as early as the 1280S-best 114
ELIZABETH
LAMBOURN
18
i8 Marbleheadstone of Shaikh
Arjfun al-Akhsi (d.633A.H./A.D. but produced in the 128os or 1290S. Petlad. Inked rubbing from G. Yazdani,"Inscriptions in the Tomb of BabaArjun Shah, Petlad,"EpigraphiaIndo-Moslemica (1915-16)pl. 14a. 1236)
spidery naskhwith riqd'flourishes -relates closely to styles of script on silver and gold coinage,farmans, and foundation inscriptions from the mid-thirteenth century, under the Delhi sultans.65A foundation inscription dated 674 A.H./A.D. 1276 from Sultanpur near Gurgaon in Haryana (fig. 17), is one example, while other inscriptions in a similar style are known from Uttar Pradesh, the Panjab, Delhi, and even Bengal. These parallels raise the question of whether professional scribes and calligraphers, who had previously worked in the ambits of the Delhi courts, settled or worked at Khambhat even before the Muslim conquest of 704 A.H./A.D. 1304-05. As one would expect, research into contacts of this type is almost nonexistent. Professional scribes were not necessarily the only ones involved in penning inscriptions, however. In less active times, more informal models of association may have existed. Paradoxically, we appear to have more information about the latter case than the main production. The only "signature"on Khambhat carving comes from a headstone produced for the grave of a local holy man named Shaikh Arjan al-Akhsi who died in 633A.H./A.D. 1236and is buried at Petlad. On stylistic grounds, the tombstone appears to be later than the date of death given on the epitaph and was probably carved in the 1280S or 1290S, thus before the Muslim conquest of Gujarat (fig. 18).66The final line, underneath the epitaph proper, carries the phrase "katibuhu AbuiBakr ibn Mahimud ibn Isma'il al-Jauhari"(its 115
CARVING
AND COMMUNITIES
19
19
Headstone of Amin al-Din ibn Ardashiral-Ahwi (d. 630 A.H./ A.D. 1232), marble. Khambhat. Photographby the author.
thejewelerorjewelseller).67 writeris AbuBakribn Mahmidibn Isma'ilal-JauharI, Muslimnamesarerichin data,and his nametells us thatAbfiBakrwas at leasta third-generationMuslim,while his nisbaal-JauharIsuggeststhat he, or possibly an ancestor,was eithera jeweleror a gem trader.As Khambhatwas an important centerof tradein semipreciousand preciousstones, the nisbafits well with the known Khambhatcontext;however,it is difficultto extrapolatemore aboutthe preciserelationshipbetweenAbfiBakr,the "writer"of the inscriptions,and the Khambhatworkshops.AbuiBakrmayhavebeena professionalscribe,butthenthe nisbaal-Khattat,the scribe,would haveseemedmore appropriate.On the other hand,AbuiBakrmayhavebeen a localjeweleror gem traderwith a particulartalent forcalligraphywho offeredto composeandwriteout the programbecausethe tombstonewasforthe graveof a localsaint.The barakaorblessingsderivingfrom this piousactperhapsexplainshis urgeto recordhis nameandhis role. The ideathata locallyavailableliterateMuslimhelpedpen fine inscriptionsis not as strangeas it mayseem, particularlyfor "one-off"inscriptionsratherthan regularproductions.A studyof "signed"Musliminscriptionsin Gujaratbetween the fourteenthandsixteenthcenturiesshowsthat,especiallyoutsideAhmadabad, local and obviouslyliteratedignitarieswereoften associatedwith the penningof foundationinscriptions.68 Onlyone signedinscriptionidentifiesthe composerasa othernamesfailto giveanyclearindicationof the composer'soccupation scribe,69 (thoughall are Muslim),while four specificallymention religiousoccupations. Thus a mosquefoundationinscriptionfrom Mangroldated784A.H./A.D. 1382-83 the Persianportionof a bilinwaswrittenbya shaikh(Shaikh'UmaribnAdam),7o gual inscriptionfromthe Mangrolfortdated797A.H./A.D. 1395wascomposedby a qadi(QadiBadr[-i] Zahir),7' a 860 A.H./A.D. 1456mosque inscription from Patan,72 and a 920 A.H./A.D. 1514 foundationinscriptionfor a jami'masjidfrom Junagarh This recordwouldsupport wereboth writtenby the imamsof the saidmosques.73 the idea that professionalscribesprobablyhad little interestin recordingtheir contributionsand thus remainedanonymous,whereas"enlightenedamateurs"~ 116
ELIZABETH
LAMBOURN
Ip~~~~C 1
SP.~~~~~~~~~~~~~~I
----w*~~~~~~~~
z
4.
20~
~n-'
-'U
20
_
-,
_
_
_
~~~~~~~~~~~~~~Mml
_
-
k
I
I
-.4
_
j
r
20
cenotaph, Side of al-KazeruinV's Kazertinitomb complex, fig. i. Photographcourtesy of the American Institute of Indian Studies, Centrefor Art and Archaeology,Gurgaon.
asked to compose and pen a text for an inscription were far more eager to record their contributions for posterity. The convention of carving Islamic inscriptions in flat-cut relief also raises the question of how this carving technique was introduced and disseminated at Khambhat and whether it subsequently became a subspecialization within the craft of marble carving. The slightly rustic, hesitant qualities of the earliest marble headstones from Khambht- for example, the headstone ofAmin al-Din al-Ahwi in 630 A.H./A.D. 1232 (fig. 19) -seem to support the idea that this was a new type of carving that demanded new skills and probably the development of new carving tools. It is certainly not a carving genre that appears at Khambhat full-fledged and at the peak of its ability. Once again, the physical evidence is too sparse to interpret comfortably. On the one hand, the gradual improvements in the carving of these headstones over the thirteenth century supports the idea that a new skill base was being built up at Khambhat, finally culminating in the headstones of the 1280s. On the other hand, it is possible to envisage contacts with the more developed traditions of Islamic inscription carving in the Delhi sultanate to the north that might have led to a sudden advance in inscription carving in the 128os. Again, research is sorely lacking, although a passage in a Jaina text, the Prabandhacintamani of Merutunga, refers to the transport of a statue from the Mammana (Makraina?) quarry,apparently situated in the lands of the Delhi sultan, to a temple near Satrunjayaduring the ministership ofVastupdla and supports the idea of direct contacts between these two regions.74Whatever the route of development, by the 128os (in any case well before the Muslim conquest of Gujarat) Khambhat marble carvers were capableofproducing Islamic inscriptions as fine as anything produced within the Delhi sultanate or in the main cultural centers of the wider Islamic world. There is some evidence that inscription carving developed as a separate specialization at Khambhat and possibly in Gujarat more broadly. The sides of alKazeriini's cenotaph show an odd break in the rhythm of the hanging lamps along its sides (fig. 20). The frieze of lamps is interupted by a plantain and then starts up 117
CARVING
AND COMMUNITIES
21
21
Mihrdbwith inset, incised foundation inscription plaque dated 733A.H./A.D. 1333.Mosque of Hilal Maliki, Dholka, Gujarat. Photographcourtesy of the American Institute of Indian Studies, Centre for Art and Archaeology,Gurgaon.
againon a new slab,to be completedagainby a secondplantain.By contrast,the epigraphicprogram,the full eighty-threeversesof suiraYa'S7n,runscontinuously acrossboth slabs.In this case what appearsto havehappenedis that a standard length cenotaphsidewas firstcarvedwith a friezeof lampsbut was then discoveredto be too shortto accommodatethe plannedepigraphicprogram.The sides hadto be extendedby an additionalslab,andthe Qur'anicinscriptionswerethen carvedacrossbothpanels.Themistakedemonstratesthatdecorativereliefcarving wasexecutedbeforeinscriptioncarving.Thisexamplealso suggestsan absenceof coordinationandplanningin the executionof the two phases,evenforwhatmust havebeenone of the mostprestigiouscommissionsof the decade.Sincea mistake of this sort is unlikelyto havecome aboutif a singlemarblecarverwas executing the wholepanel,it suggeststhatthe decorativereliefandinscriptionwereexecuted by two separatecarvers.Furthersupportforthis ideacomes froma uniqueseries of inscriptionsin the 920 A.H./A.D. 1514 jami'masjidin Junagarh,Saurasra (now known as the Borvadmosque).The mosquepreservesa fine carvedmihiraband a foundationtext inscribedon a separateslab of marble.The foundationtext is silentas to the carversand scribesinvolved;however,two crudePersianinscriptions incisedon the uppercornersof the mihirabprovideexceptionaldetailsabout who executedwhatpartofthe commissionandevenhowtheywerepaid.As edited NarbadTahya byZ. A. Desai,the firstinscriptionreads:"stonecarver[sangtarash] the executorof the mihirabof the jami' mosque."The second, longerand more detailedinscriptionstates:"Thistabletbearingthe datewas executedby the stone carver[sangtarash], Radh,son of . .. the land of the village of Umaralahas been 118
ELIZABETH
LAMBOURN
_
4
,
0
-~~~~ ~~~,~
-
'
4'I%*svsseJl
_
~~~~~~~~~~~~~~~. ~~~~~~~ ...
.
......
~j-,,2 ~ ~ 1?1~
22
22
Cenotaph of Ahmad Shah I (d. 846 A.H./A.D. 1442). Ahmadabad. Photographcourtesy of the American Institute of Indian Studies, Centre for Art and Archaeology,Gurgaon.
granted to him by the great Khan Khawwas Khan. The in'ami land [is allotted to] the writer of the text, Sayyid Miran Nuir,the imam of the jami' mosque."75 Thus the mihrab and the foundation tablet were carved by two different carvers, and the text of the inscription tablet was written by Sayyid Miran Nuir,the imam of the mosque. Although the Borvad mosque inscriptions are so far unique in the surviving evidence, they do hint at a separation of skills that was most likely in effect across western India. If these ideas are correct, the tombstones from Khambhat represent the fruit of an exceptional amalgam of specialized skills-those of the composer and writer of the text, and of the sculptor and the inscription carver. The difficulties of gathering all these skills together in a single time and place helps explain the unique character of Khambhat marble carving among the Islamic stone carving of western India. Thus, although we find many ornately carved mihrabs in Gujarat, for example in the 733A.H./A.D. 1333mosque of Hilal Malik! at Dholka (fig. 21), they are rarely inscribed on the body of the mihrab;instead, separate inscription slabs are accommodated within the mihrabstructure. In funerary epigraphy,too, cenotaphs such as that of Ahmad Shah I (d. 846 A.H./A.D. 1442) at Ah.madabadare finely carved but do not carry inscriptions (fig. 22).
WorkshopOrganization It is difficult to be precise about the structure within which Khambhat carvings were produced. The consistency of script, inscription programs, and decoration from the 128os onward points to a stable group of craftspeople working in what might be called a Khambhat style. There was a great deal of variation within these parameters, however, and, until the 1310S, the surviving numbers are simply too few to argue for the existence of a separateworkshop serving only Muslim patrons. Thus we can probablyassume these "new"products were manufactured within an existing atelier, one that worked primarily for Hindu and Jainapatrons. Although the main elements of marble carving for Muslim patrons at Khambhat had been formulated prior to the actual Muslim conquest, this event appears to have had a major effect on production, as Muslim communities now became the principal patrons. This development appears to have led to a substan119
CARVING
AND COMMUNITIES
*
.
b
>
;i;
_
A
~~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. --
e
..
..,..-...
'
V..
. _.......'..,
24
23
23
Partlycarved cenotaph lid, grave6, TeungkuSareh.LeidenUniversity Library,LegatumWarnerianum,the Netherlands, Or.23.481,photograph 384. 24
Marblecenotaph of Fatimaal-'Alaw! (d. 738A.H./A.D. 1337),Nagara. Photographby the author.
tial increasein the productionof gravememorialsand an expansioninto other areassuch as architecturaldecoration.By the second decadeof the fourteenth centurywe haveevidenceof large-scaleproductionof gravememorialsalongwith foundationinscriptionsand architecturalelements.The stylistic consistencyof this materialand its numbersindicatethat a, or perhapsseveral,workshopsnow specializedin carvingmarbleforthe Muslimmarket.Thismaterialprovidesgood dataon the processesor stagesof productionand the divisionof laborwithin the workshopor workshops. On the whole, productionappearsto havebeen heavilyhierarchicwith limited spacefor inputfromthe purchaseror patron.The basicunit of marblecarving for Muslimpatronswas the slab,used for head and footstones,for cenotaph sides and lids, for foundationinscriptions,and for certainparts of mihrabsand lintels.Onlymihrdbnichesrequiredblocksratherthanslabsof marble.Thedominance of inscriptionsin Khambhatcarvingand the low-reliefstyle of most noninscriptionaldecorationmeantthatthese slabshad to be cut, possiblysawn,then smoothed and even partiallypolishedbeforethe next stage of carving,namely the transferof inscriptionsand designsto the marblesurfacefor carving.Since this processwasrelativelylengthyandrepetitive,the workshopor workshopsvery likelyheld a stock of pre-preparedslabsreadyfor the next commission.No such slabshavesurvivedat Khambhat,but a numberfrom Pasaiand Gresikconfirm that marbleslabswerepreparedto a smooth,semipolishedfinish,with the major planesof the design,suchas the tripartitedivisionof the slabfor cenotaphsides, ready-carved(see fig. 9). Cenotaphlids were probablyalso precarved,as in the partlyfinishedlid of grave6 from the TeungkuSarehburialgroupat SamuderaPasaiin Sumatra,whichrecordsa deathin 834A.H./A.D. 1430-31 (fig.23).Slabsseem to havebeen preparedwith a singledecorativeschemein mind, regardlessof the finaldesignselectedby the patronor purchaser.Thusthe cenotaphof Fatimaal'AlawI(d.738A.H./A.D. 1337)atNagara,justnorthof Khambhat,eschewsthe central friezeof hanginglampsgenerallycarvedon Khambhatcenotaphsidesduringthe fourteenthcenturyand is insteadcarvedwith the full eighty-threeversesof sura Ya'Sin. Despitethis variation,the cenotaphsides were clearlydivided into the threeplanesrequiredfor the lamp designand are thus likelyto havecome from precarved stock (fig. 24).
The same rigidseparationof carvingstagesis seen further"up"the chain of production.Thesmallmuseumattachedto the ShriB. D. RaoCollege,Khambhat Campus,preservesa partlycarvedcenotaphside panelwith the centralband of 120
ELIZABETH
LAMBOURN
:=:~~~~T V..
*:
~.
.7
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~*~
25
25
Unfinished cenotaph panel, fourteenth century, marble. Museum of Shri B. D. Rao College, KhambhatCampus, Khambhat. Photographby the author.
hanging lamps completed but the upper and lower calligraphic bands untouched (fig. 25). As mentioned earlier, al-Kazertini's cenotaph sides appear to have been prepared in the same sequence. The logic behind the stages here is technical, since the relief frieze of lamps involves carving into the slab to a depth of up to five centimeters. Given that marble is a metamorphic rock with fracture and fault lines within the stone, deep carving of this type risks splitting the slab far more than the flat-cut relief carving involved in inscriptions.76This method ensured that any fracture in the slab would be discovered at the start of the carving process, when minimum effort had been expended, rather than toward the end. The standardization of production and the separation of different stages suggest that the workshop (or workshops) was relativelylarge and working in a rote fashion. These clues do not fit, however, with the relativelysmall numbers of headstones and cenotaphs that survive at Khambhat, and they therefore seem to confirm the hunch that large numbers of carvings have been lost over time. In this context, by the fourteenth century Muslim clients probably had little space for personal input. Beyond the details of the epitaph, the only area that appears to have allowed for a certain personalization of the epigraphic program was that immediately above the epitaph. Several gravestones carry Qur'anic verses in this area that are not generally used in the Khambhat repertory; in a number of cases it is possible to decipher the link between the deceased and these atypical verses, indicating that they had been chosen particularly for this individual. The epitaph of 'Umar al-Kazeriini is preceded by verses 169-71 of sura Yd'Sin, verses that speak of the first martyrs of Islam, killed at the battle of Uhud. Ibn Batttita'saccount of his visit to India during the 1340S gives a detailed account of al-Kazeriini's death, recounted to the author firsthand by a witness.77According to Ibn Battuita'ssource, al-Kazeriin1was killed by Hindu bandits while traveling to Delhi. His death at the hands of non-Muslims appears to have been sufficient to qualify him as a martyr. The proximity of these verses to al-Kazeruini'sepitaph appears deliberately designed to compare him to these first martyrs of Islam. Another example of the same phenomenon is seen on the headstone of the abovementioned Ikhtiyar al-Dawla wa al-Din, Khambhat's bahr bek (commander of 121
CARVING
AND COMMUNITIES
fi
_
....-,-
I
26
26
Marbleheadstone of al-Damiri (d. 714A.H./A.D. 1315),Victoria and Albert Museum, London. Photographcourtesy of the Victoria and Albert Museum, London.
the sea) (see fig. 6). Ikhtiyaral-Dawla'sepitaphis precededby verse29 of suraalMu'min7n,a prayerrecitedby Noah beforehe disembarkedfromthe arkafterthe
floodhadsubsided.Therelationshipbetweenthis verseandIkhtiyar'sofficialpost seemsevident. Workingfor ExportProduction Byandlarge,the sameconstraintsthatexistedforlocalproductionappearto have held true for exportproduction.Most tombstonesfound outsidethe port follow the same decorativeschemesand inscriptionprogramsas at Khambhatproper andwereprobablyexecutedusingprecarvedslabs.Whilethe purchaseand delivery processesmust havetakenlonger,the basicstagesmust havebeen the same. Apartfromfourinstances,all the tombstonesexportedfromKhambhatappearto haveleft the portwith finishedepitaphs,presumablyafterthe demiseof the individualsthey were commissionedfor. Only the three tombstonesfound at Dhofar in southernOman and one gravememorialfrom Gresikin easternJavawere shippedwith unfinishedepitaphs,and thus probablyduringthe lifetime of the individualthey weredestinedto commemorate.A recenttechnicaland stylistic studyof the headstoneof al-Damiri(fig.26) has demonstratedthatit wasordered with only the firstfourlines of the epitaph,includingthe introductoryformulas and his name, but withoutthe date of death.8This informationwas addedafter his deathby local stone carversat Dhofar,as can be seen by the changein script and carvingtechniquein the lastlines of the epitaph(seefig. 26). Oneof the three 122
ELIZABETH
LAMBOURN
28
27
27
Grave2, KutaKareueng,epitaph dated 851A.H./A.D. 1447-48, marble. LeidenUniversityLibrary,Legatum Warnerianum,the Netherlands, Or.23.48i,photograph481. 28
Detail of headstone of fig. 27.Leiden University Library,Legatum Warnerianum,the Netherlands, Or.23.481,photograph12.
headstones from Gresik is even less complete and gives only the first introductory formulas (for a female epitaph) without even a name. Until now we have underlined the static, somewhat inflexible nature of Khambhat carving. While this assessment undoubtedly applies to the majority of production, particularly gravememorials, it is important to underline instances of a more creative interaction between purchasers and the Khambhat workshops. In spite of the hieratic nature of production outlined above, Khambhat tombstones evolved significantly between the thirteenth and fifteenth centuries, and there was clearly room within these systems for innovation and change, including the production of one-off pieces. In later centuries Gujarati craftspeople were famed for their sensitivity to the demands of the export market and the varying tastes and fashions of different regions. In his A'in-i Akbari (Mirror of Akbar'sWorks), a text composed in the forty-second year of Akbar'sreign, or A.D. 1595-96, 'Allami noted that at Khambhat "imitations of stuffs from Turkey,Europe, and Persia are also produced."79 While much research has been devoted to interactions of this sort in the applied arts, the example of Khambhat carving suggests the same imitative abilities also found expression in the field of stone carving. Probably the best-documented example of the interaction with overseas clients are two cenotaphs produced for the court of Samudera-Pasaitoward the middle of the fifteenth century-grave 2 in the Kuta Kareueng burial group at Pasai, which carries an epitaph dated 851A.H./A.D. 1447-48 (fig. 27) and grave 4 in the same cemetery (fig. 28). Both graves are magnificent white marble cenotaphs of a type known at Khambhat;however, their headstones differ radically from known examples at the port. The two headstones are relatively small and sit on the lid of the cenotaph rather than being attached at the ends, as at Khambhat. Furthermore, while their material, white marble, along with their epigraphy and interlace border attach them clearly to the Khambhat corpus, their shape is entirely unparalleled at Khambhat and even in northern India more widely. The two gravestones 123
CARVING
AND COMMUNITIES
29
29
Muslim headstone recordinga death in Saka1329 / A.D. 1407, Tralaya, EastJava.FromL. C. Damais, "EtudesJavanaisesI:LesTombeaux Musumanes Dat&esde Tralaya," Bulletinde l'EcoleFran,aise d'ExtremeOrient48.2 (1957).
areslightlyhorseshoe-shaped,toppedby a lotus flower,but come in tightlyat the basebeforeflaringoutwardand upwardin an almost Baroquevolute.The shape and particularlythe voluteterminalsdistantlyrecallarchtypes seen in southern Indian architecture.The closest parallels,however,are actuallywith Southeast Asian and particularlyJavanesearch forms, notablythe so-called kalamrga,or deer-arch,form. WhetherSouth Indian or SoutheastAsian in inspiration,the form is entirelyunknown in westernIndia. Significantly,the form is well documented in a large numberof early Muslim gravestonesfound at Tralayaand Trawulanin easternJava,manufacturedfor Muslimcourtiersat the kingdomof Majapahitbetween1407and1475(fig. 29).8o While we still haveno clearindication of whythis gravewasorderedfromKhambhatwith apparentlyEastJavanesestyle gravestones,the extraordinaryfusion betweenKhambhatcarvingand nonlocal elementsdemonstratesthe enormousversatilityof this carvingtradition. The two grave memorialsfrom Kuta Kareuengare not unique within the Khambhatcorpus.Two of the three headstonesorderedfor Dhofarintroducea new programof Qur'anicversesquite differentfrom that in use at Khambhat. While all havea clearfuneraryusageor relevance,theiruse for the Dhofarcommissionseemsto pointto a specificrequestbytheirpatrons,perhapsbasedon local Dhofariepigraphictraditionsor personaltaste.8'Thesameadaptivecapacitiesare seen in the groupof architecturalelementsfound in the Fakhral-Din mosquein Mogadishu.While the mihrabniche therecloselyfollowsKhambhatmodels,the marbleportalfacingwithin the same complex(fig.30) is anotherextraordinary fusion of Khambhatmaterials,techniques,and stylisticelementswith EastAfrican Islamicelements.82 Thevolumeof exportproductionandthe instancesof stylisticfusionobviously beg the questionof whethermarblecarverseverleft Khambhatto workon their commissionsin situ. SouthAsian craftspeopleare sometimesseen as immobile even thoughthereis a considerableamountof evidencefor theirmobilitywithin India,andit is worthconsideringwhetherthis mobilityextendedoverseasaswell. Clearly,when dealing with ordersfor individualheadstones,the commission 124
ELIZABETH
LAMBOURN
!~~~~~~~~
A-
do'
*
30
30
Plastercast of the marbleportal facing of the Fakhral-Din mosque, Mogadishu, fourteenth century. Formerlyin the Museo della Garesa,Mogadishu (present status unknown). FromCatalogo, Museo della Garesa, 1934. Reggio Governo
della Somalia Museo della Goresa Cataloyo (Mogadiscio:Reggio Governo della Somalia, 1934), 77.
could be executedentirelyin Khambhat.While we haveno writtenevidencefor the mannerin whichoverseaspurchaseswereorganized,it musthavebeen either visitedKhambht- or throughpersonalcontact-the client,or a representative, via letter.Eithermannerwouldhaveallowedforthe majorityof informationto be passeddirectlyfromclientto workshop.Thelettersof Jewishtraders,preservedin the Genizadocuments,offernumerousexamplesof sometimescomplexobjects beingorderedacrossvastdistancesdirectlybyletter.83 Commissionsoflargecenotaphgraves,however,andindeedarchitecturalcommissionssuchasthe decorationof the Fakhral-Dinmosquein Mogadishu(seefig. 30) mayhavedemandedchangesto workingpractices.It seemsmorethan probablethatKhambhatstonecarverswouldhaveneededto accompanysomecommissions simplyfor technicalreasons.Khambhatcenotaphsare complexstructures, held togetherby a systemof dowelsand metal pins. Althoughit is impossibleto providecompletelyaccuratefigures,largecenotaphgravesarecomposedof more than seventyseparatepiecesof marble,with overfour-fifthsor 80 percentof these belongingto the plinth (seefigs.9, 27).Thereis no evidencefor piecesbeingindividuallycodedpriorto assembly.Furthermore,the excavationof the foundations to supportthese cenotaphs,as well as the assemblyof the cenotaphsthemselves, all requiredspecialistknowledge.It thereforeseemsunlikelythat largecenotaph graveswouldhavesimplybeen shippedoff to be assembledby local craftspeople. 125
CARVING
AND COMMUNITIES
The largenumberof memorialsorderedat Pasaiwould certainlyindicatethat specializedcraftspeoplefromKhambhataccompaniedthe carvedstonesfortheir properassemblyuponarrivalattheirdestination. TheideathatmarblecarversmovedoutsideKhambhatmightalso explainthe extraordinaryfusionof Khambhatelementswith non-Indianelementsseenboth in certainKhambhatmemorialsat Pasaiand in the carvingof the Fakhral-DIn mosquein Mogadishu.Thecorpusof carvingthatsurvivesatKhambhatcertainly showsevidenceforthe importanceof directobservationand copyingin the production of new carvings,with decorativedetails and epigraphicprogramsregularlyborrowed(sometimescenturieslater)from earliercarvings.Againstthis background,it becomespossibleto envisagestylisticinteractionbasedon direct observationand experienceratherthan via written instructionsor sketches. CouldKhambhatcarverserectingthe threecenotaphgravesat Gresikhaveseen local tombstonetypes at nearbyTralayaand Trowulanand eventuallyreintroducedthe deer-archformforgraves2 and4 in the KutaKareuengburialgroupat Pasai?I havearguedrecentlythatthe structureandepigraphyof the Fakhral-Din portal could havebeen copied directlyfrom the portal of the nearbycongregationalmosqueatMogadishu.84 Althoughit is not currentlypossibleto answersuchcomplexquestions,as this articlehasshown,evenwithoutwrittendocumentsaboutworkshopactivitesand commissions,the analysisof Khambhatcarvingcanprovidea significantamount of dataaboutmanyaspectsof the productionof fine marblecarvingfor Muslim patrons,its market,and the relationshipbetweenproducersand consumersin westernIndiaduringthe latethirteenththroughmid-fifteenthcenturies.Evenif the case of Khambhatcarvingremainsuniqueforthe moment,this articledemonstratesthe amount of informationthat approachesgroundedin the study of materialculturecan coax from materialthat has generallybeen seen as the preserveonlyof epigraphers. Conclusions:StoneandMaritimeTransportin WesternIndia At present,Khambhatappearsto representa near unique case of the seaborne exportof stonecarvingin the Islamicworldandin SouthAsia.Othertombstones producedat coastalsitesor on the islandsof the PersianGulfand RedSeaappear not to havebeenexportedmuchbeyondthe siteof theirmanufacture.Ofthe more than two hundredgravemarkersmanufacturedat Dahlakin the RedSea,only a handfulhavebeenfoundbeyondthe island,andnoneof the finecenotaphscarved at Sirafon the PersianGulfhavebeenidentifiedoutsidethe port. Nevertheless,there are some clues that the Khambhatexamplebelongsto a broaderwesternIndianmodel.Constructionmaterialsotherthanstoneweredef126
ELIZABETH
LAMBOURN
initelytransportedbyseaaroundthe coastof westernIndiaandacrossthe Indian Ocean.SinceRomantimesteakhadbeenshippedfromsouthernIndiato thecoasts Thereis also one recordedinstanceof of Gujaratand beyondto the NearEast.85 finishedarchitecturalelementsbeingshippedby seaaroundthe coastof Saurastra describes in the earlythirteenthcentury.A passagein the Prabandhacintamani an incidentthattook placeduringthe constructionof the Nandisvaratempleon in v.s.1277/A.D. 1220-21. A groupof carvedcolumnsmadefrom MountSatrunijaya when "Kantheliyastone"was being unloadednear a site called Samudrakantha one fellout of the boatandlodgeditselfin the mud.Thecolumnwasfinallyrecoveredthe followingyear"owingto the changeproducedby the tide of the sea."86 The sourceof Kantheliyastone may correspondto the site of KanthKutaor the MahIKantharegionin north Saurastra,oppositethe Rann.of Kachh,suggesting thatthe carvedcolumnsweretransportedby seaaroundthe coastof Saurastrato a coastalsitenearSatrun~jaya. Whateverthe finalidentificationof the site,the passage clearlyindicatesthat in the earlythirteenthcenturyarchitecturalelements werecarvedattheirquarrysiteandthen transportedin a finishedor semifinished stateby seato the siteof theirfinaluse. Forthe momentthis passageis unique,but furtherresearchon othertextsof the periodmayyieldadditionalexamples.87 The peninsulaof Saurastrahas enormousreservesof stone coupledwith a geographythatencouragessea transport.In actualfact,Saurastrawas evenmore suitedto sea transportbeforethe greatearthquakeof 1819that raisedthe peninsula.Numeroussources,of whichde Varthema,Baldaeus,and AlexanderHamilton in additionto contemporarymaps,recordthat Saurastrawas a half-island separatedfromGujaratby a lagoonknownas the Nl.88 Duringthe monsoonthe Ndlflooded,effectivelylinkingthe Gulfof Khambhatwiththe Ran.nof Kachhand transformingthe peninsulainto an islandforsix monthsof the year.89 Thisflooding mayhavebeenespeciallysignificantin openingup the centralplainof Gujarat, the main centerof settlementand locus of the majorcities such as Ahmadabad and Dholka,to Saurastranstone.The quarriesat Dhrangadhrawouldhavebeen locatedon the "coast"of the NMlwith easy accessto these majorcities.Against this background,H. G. Briggs'sslightlysuperiorobservationthatcertainpersons naivelybelievedthat the stone used in the constructionof the congregational mosqueat Ahmadabadwas "agrey-wackeprocuredfrom the contiguityof the Rannof Katch"maybe true.90 Priorto 1819,then,Gujaratandits architecturewere intimatelylinked to the sea, and the phenomenonof Khambhatmarbleexport mayfindits placein this largerwesternIndianpicture,linkingin turnto the webs of stoneexchangearoundthe SouthChinaSeaandIndonesianarchipelago.9'
127
CARVING
AND COMMUNITIES
NOTES
i.
PeterRockwell,TheArtofStoneworking: A ReferenceGuide(Cambridge:Cambridge University Press, 1993),
2.
2-3.
Theliteratureis extensivebut see, for example,Nusin Asgari,"TheStagesof Workmanshipofthe CorinthianCapital in Proconnesusand ItsExportForm,"in ClassicalMarble:Geochemistry,Technology,Trade,ed. Norman HertzandMarc Waelkens(Dordrecht,Boston,and London: Kluwer, 1988),115-25; Patrizio
Pensabene,"ACargoofMarbleShipwreck atPuntaScifo,nearCrotone(Italy)," InternationaljournalofNauticalArchaeologyand UnderwaterExploration 7.2 (1978): 105-18; J.B. Ward-Perkins, "Nicomedia
andthe MarbleTrade,"PapersoftheBritish SchoolatRome48,1980. Seealso F.J. Clayton,ed.,AncientMarbleQuarrying and Trade,BARInternationalSeries453 (Oxford:B.A.R.,1988). Wolfgang Frankeand Ch'enT'ieh-fan, "A 3. Chinese TombInscriptionof A.D. 1264, DiscoveredRecentlyin Brunei,"Brunei Museum Journal3.1 (1973):91-99; Ch'en T'ieh-fanand WolfgangFranke,"710 YearsOld Chinese TombstoneDiscovered in Brunei,"in Chinoisd'outre-mer,ed. Denys Lombard,Actes du XXIXe Congresinternationaldes Orientalistes (Paris: L'Asiatheque, 1976),1-2; Dasheng
Chen, "Unepierretombale du debut du XIVes. retrouveea Brunei,"Archipel42 (1991):47-52. The Filipino examples have
been publishedonly cursorilyin local publications,for example,C. A. Majul, Muslimsin thePhilippines(Quezon City: Universityof the PhilippinesPress,1973), unnumberedplateswithin the text. Their location in the southern Philippinesnow makes furtherstudy extremelydifficult. It is perhapsnot surprisingthat these export stones haveneverbeen tracedback to the corpus from which they originate in FujianProvincein China, which has itself been publishedin fragments.For 128
ELIZABETH
LAMBOURN
stone carving producedlocally for Buddhistcommunities, see GustavEcke and PaulDemieville, The TwinPagodasof Zaytun:A Studyof LaterBuddhist Sculpturein China(Cambridge,Mass.: HarvardUniversityPress,1935).The corpus of Islamicinscriptions is presented most accessiblyin Chen Da-shengand LudvikKalus,Corpusd'inscriptions arabesetpersanesen Chine.ist ser. ProvincedeFu-Jian(Paris:Geuthner, 1991).A production also existed for local Hindu and Nestorian Christiancommunities, describedin, e.g., GustavEckeand Ananda Coomaraswamy,"Hindu Sculpturesat Zayton,"Ostasiatische Zeitung,2nd ser.,9 (1933):5-11; Eckeand Demieville, TwinPagodas,pls. 6ga-d, 70b. The Nestorian materialhas recently been the objectof a projectbased out of MacquarieUniversityin Australia,see in particularthe relevantchaptersin lain Gardner,SamuelLieuand KenParryeds., FromPalmyrato Zayton:epigraphyand iconography(Turnhout: Brepols;NSW, Australia:Ancient History Documentary ResearchCentreMacquarieUniversity, 2005). Forthe workfor Hindu patronssee especiallyrecentworkby JohnGuy, "TamilMerchantGuilds and the Quanzhou Trade"in AngelaSchottenhammer ed. TheEmporiumofthe World.
MaritimeQuanzhou, 1000-1400 (Leiden: Brill, 2001),283-308. 4. Forthe firstattemptat recordingthis diffusion, see H. M. Ambary,"L'art funerairemusulman en Indonesiedes origines au XIX siecle,"(PhD diss., Ecole des HautesEtudesen SciencesSociales, Paris,1984),map on 340. Fora first attemptat mappingdiffusion acrossthe MalaccaStraits,see Othman Mohammed Yatim,Batu-Aceh:EarlyIslamicGravestonesfromPeninsularMalaysia(Kuala Lumpur:Museum Association of Malaysia,1988),maps 1,3.
5. Fourgranitestatueswerefound atthe port ofKotaCina.SeeE.E.McKinnon,"Tamil ImageryinNortheastSumatra,"Oriental
ian Remainsin theBombayPresidency, ArchaeologicalSurveyof India New ImperialSeries14(Bombay:Government CentralPress,1897), 8:313-21, by the same author.The gravehas since been published extensively.Forthe main publications, see Z. A. Desai, "SomeFourteenth CenturyEpitaphsfrom Cambayin Gujarat,"EpigraphiaIndica,Arabicand
Art40.3 (1994):17.Forabroaderview, see
also E.E.McKinnon,"MediaevalTamil Involvementin Northern Sumatra,cilC14(The Goldand ResinTrade),"Journalof theMalaysianBranchoftheRoyalAsiatic Society69.l(1996): 85-99; see9g,wherethe authornotes that a bodhisattvatorso carvedfrom redgranitewasdiscoveredat Baruson the westerncoast of Sumatraat the turn ofthe lastcentury. 6. ElizabethLambourn,"'ACollection of MeritsGatheredfrom Different Sources': The Islamic MarbleCarvingand Architectureof Cambayin Gujarat between 1200 and 1350A.D." (PhD diss., University of London, 1999).
7. ElizabethLambourn,"TheDecoration of the Fakhr-al-DinMosque in Mogadishu and Other Pieces of GujaratiMarble Carvingon the EastAfrican Coast," Azania:Journalof theBritishInstitutein Eastern Africa 34 (1999): 61-86; Lambourn, "FromCambayto SamuderaPasaiand Gresik:The Exportof Gujarati GraveMemorialsto Sumatraand Javain the Fifteenthcentury C.E.," Indonesiaand the Malay World31.90
(2003):
Persian Supplement (1971):39-43; Wayne
Begley,MonumentalIslamicCalligraphy fromIndia (Villa Park,Ill.:Islamic Foundation, 1985), inscription 15;
ElizabethLambourn,"'ThoseSlainin God'sway... ':The Martyrdomof'Umar al-Kazaruniand His Graveat Cambayin Gujarat(India),"in InscriptionasArt in the Worldoflslam, ed. HabibeRahim, HofstraUniversity,LongIsland, (Conference proceedingsforthcoming). 10. In total, about ninety inscriptions and gravememorials, dating to between 615 A.H./A.D.
221-89;
Lambourn,"Carvingand Recarving: Three RasulidGravestonesRevisited," NewArabianStudies6 (2004): 10-29. 8. Rockwell,Art of Stoneworking,5. 9. JamesBurgess,On theMuhammadan ArchitectureofBharoch,Cambay,Dholka, Champanir,andMahmudabadin Gujarat, ArchaeologicalSurveyof WesternIndia 6 (London:W. M. Griggsand Sons, 1896), 27-29 and plates.The gravewas first publishedwithout illustrationsin Burgess,ListsoftheAntiquarianRemains in theBombayPresidency, Archaeological Surveyof WesternIndia (Bombay: GovernmentCentralPress,1885),267-75 and the laterRevisedListsoftheAntiquar-
1218and 1297 A.H./A.D.
1879-80,
havebeen documented at Khambhat.The best idea of the rangeof materialcan be gleaned from the lists of Islamicinscriptions from Khambhatin variousvolumes oftheAnnualReportson IndianEpigraphy.A number of pieces carryingdated epitaphsor foundation texts havebeen publishedin more detail in articlesin the seriesEpigraphiaIndo-Moslemicaand its successor,EpigraphiaIndica,Arabicand PersianSupplement.Fora detailed study of the production spanning
1200
tO1350,
see Lambourn,"Collectionof Merits." i. The patronageof the Jainamerchantministers Vastupala(d. v.s. 1298/A.D. 1241), at one time governorof Khambhat, and his brother Tejahpala (d. v.s. 1308/A.
may have been critical. For a useful summary of patronageat Khambhat,see A. K. Chatterjee,A ComprehensiveHistoryof Jainism,1000 AD to 1600 AD, 2 vols. (Calcutta:Firm KLM, 1978), 2:31-32; see also M. A. Dhaky, "The D. 1251)
129
CARVING
AND COMMUNITIES
Chronology of the Solanki Templesof Gujarat,"Journalof the MadhyaPradesh Itihasa Parishad,no. 3 (1961): 66-68. 12. SeeAnnualReportson IndianEpigraphy listings, also bibliographyfor alKazerini's gravememorial (n. 9 above), particularlyDesai, "SomeFourteenth CenturyEpitaphsfrom Cambay,"and relevantinscriptions in Z. A. Desai, "SomeArabicInscriptionsof the Rajput Periodin Gujarat,"EpigraphiaIndica, Arabicand PersianSupplement(1961):124. Also see GhulamYazdani,"Inscriptions of the KhaljiSultansof Delhi and Their Contemporariesin Bengal," EpigraphiaIndo-Moslemica(1917-18):3738.A numberof fifteenth-century headstonesin and around Khambhatare publishedfor the firsttime in Lambourn, "FromCambayto Pasaiand Gresik." 13. See Mahdi Husain, "SixInscriptionsof SultanMuhammadbin Tughluq," EpigraphiaIndica,Arabicand Persian Supplement(1957-58):29-34; Z. A. Desai, "Khaljiand TughluqInscriptionsfrom Gujarat,"EpigraphiaIndica,Arabicand PersianSupplement(1962): 1-40; Z. A. Desai, "Inscriptionsof the Gujarat Sultans,"EpigraphiaIndica,Arabicand PersianSupplement(1963):5-50; Z. A. Desai, "Inscriptionsof the Sultansof Gujarat,"EpigraphiaIndica,Arabicand PersianSupplement(1974):1-57. 14. Z. A. Desai, "APersianInscriptionof KarnaDeva Vaghelaof Gujarat," EpigraphiaIndica,Arabicand Persian Supplement(1975):13-20. 15. Architecturalfragmentsareprobablythe most neglected group of carvingsfrom Khambhatsince many do not carrydated foundation texts and werethus ignoredby epigraphers.A number of unpublished pieces arepresentedin Lambourn, "Collectionof Merits;"Lambourn,"The EnglishFactoryor 'Kothi'Gatewayat Cambay:An UnpublishedTughluq
Structurefrom Gujarat,"Bulletinof the Schoolof OrientalandAfricanStudies65.3 (2002): 495-517.
i6. Fora first outline of the history of these techniques in Gujarat,see V. S. Parmar, Haveli:WoodenHousesand Mansionsof Gujarat(Ahmedabad:Mapin, 1989), historical surveychapter.Forbrickand timber construction in the Muslim context in Gujarat,see relevantsections of ElizabethLambourn,"OfJewelsand Horses:The Careerand Patronageof an IranianMerchantunder ShahJahan," Iranian Studies 36.2
(2003):
213-58.
17. There is practicallyno literatureon
historical quarryingin westernIndia. Petrographicanalysisof quarrysites and surviving sculpturesto match sources of stone, as carriedout for antique marbles, has neverbeen attemptedin India, and attributionrestson logic more than firm proof. However,some details about quarryingin the earlytwentieth century are given by JamesBurgess,and Henry Cousens, TheArchitecturalAntiquities of NorthernGujarat,ArchaeologicalSurvey of WesternIndia 9 (London:Bernard Quaritch,KeganPaul,TrenchTrubner and Co. and Luzacand Co., 1903), 28-29. 18. Similarmotifs can be found on roughly contemporaneousfunerarymemorials and shrinesin Egypt,Syria,eastern Anatolia,the northernJazira,at Dahlakin the RedSea,and in Afghanistan.SeeNuha Khoury,"TheMihrabImage:CommemorativeThemesin MedievalIslamic Architecture,"Muqarnas9 (1992):11-28. 19. See Desai, "SomeFourteenthCentury Epitaphsfrom Cambay,"53-58. 20. Fora more in-depth study of titles and names on Khambhatinscriptions between 21.
1200
22.
Fourteenth Centuries (Leiden: Brill, 2004),
53. Most accessiblypublishedin Yazdani, "Inscriptionsof the KhaljiSultansof Delhi,"37-38; Begley,Monumental IslamicCalligraphyfromIndia, inscription 12. 24. Desai, "SomeFourteenthCentury Epitaphsfrom Cambay,"35-38. 25. Fora biographyof al-Biutihari,see Lambourn,"ArchitecturalInfluences," 23.
123-25.
26. See Patel, Building Communities, 49-50,
52-55. 27.
and 1350,see Lambourn,
"Collectionof Merits,"chap.5, pt. 7. Fora biographyand study of his patronage at Khambhat,see Lambourn,"'A Collection of Merits
Influencesin the FridayMosque and KazaruniTomb Complexat Cambayin Gujarat,"SouthAsian Studies17(2001): 131-32; Lambourn,"Martyrdomof'Umar al-Kazaruni,"forthcoming. Desai, "SomeArabicInscriptionsof the RajputPeriod,"22-23; and "Some FourteenthCenturyEpitaphsfrom Cambay,"46-47; Alka Patel,Building Communitiesin Gujarat:Architectureand Societyduringthe Twelfththrough
. . .':Architectural
130
Justbeforegoing to pressI became aware of a furtherthree elements of the Khambhatcorpus along the westerncoat of India and in the Maldiveislands. While it was too late to add this information to the tables and main text, readersshould be awareof the following new additions to the corpus. A damaged14th-centurycenotaphside in two parts from Quilon in southern Kerala(publishedin MehrdadShokoohy, MuslimArchitecturein SouthIndia:the SultanateofMa'barand the traditionsof maritimesettlerson theMalabarand CoromandelCoasts(London:Routledge Curzon, 2003), 139). One 14th-centuryheadstoneblank also reachedMale in the Maldiveswhereit was inscribedlocally for use as a mihrab dedicatedby SultanJalalal-Dunyawa alDIn 'Umaribn Salahof the Maldivesin ELIZABETH
LAMBOURN
1322. The stone was publishedfor the firsttime by the Departmentof Informationand Broadcastingof the Maldives(Department of Informationand Broadcasting, MosquesofMale (in Divehi) (Male: Departmentof Informationand 722 A. H./A.D.
Broadcasting, 1980),
23)
and has recently
been discussedby LudvikKalusand ClaudeGuillot (L. Kalusand C. Guillot, "Inscriptionsislamiquesen arabede l'archipeldes Maldives,"Archipel,70 Kalusand Guillot's (2005):19-27). analysisis largelyepigraphicand assumes the stone to be a purpose carvedmihrab principallyinscribedat Khambhat; however,it representsa farmore complex and interestingappropriation.Firstof all, comparisonwith the Khambhatmaterial makes it abundantlyclearthat this is a tombstone ratherthan a Khambhat mihrab;its transformationinto a mihrab was effectedin the Maldivesthrough its positioning as a mihrib and the carving of the dedicatoryinscriptionwhich describesit as such. Thereis also no doubt that the dedicatorytext was carvedin the Maldivessince it is fartoo crudeto be a productof the Khambhatworkshopsof the period but is typical of non-specialist carvers(here local Maldiviancoral carvers)strugglingto workwith a much harderand more complexstone. A very similar processhas alreadybeenexamined in the two Dhofarstones recarvedfor the Rasulidgovernoral-Wathiq(see Lambourn"Carvingand Recarving"). The only epigraphicelementcarvedat Khambhatwould havebeen the elaborate geometricand floralbismillih alongthe shoulderof the stone. Finallya tombstone pair from RaaAtoll in the Maldivesare also of Khambhat manufacture,although unfortunately only the headstonehas been illustrated and describedin any detail ("Inscriptions
islamiquesen arabede l'archipeldes Maldives,"47-51). Both aremarblestones and the headstonerecordsthe death of one Abfi Bakran'urfShaikhJuinairiwho died in 885A.H./A.D. 1480. The headstoneis a unique "collage"of familiar Khambhat motifs (a flatterreinterpretationof the hanging lamp) and inscriptionalpanels, such a departurefrom more standardised decorativeand inscriptional systems that is typical of latermid-fifteenth century workand thus fits with the date of the epitaph.Fromphotographsat least it appearsthat this stone was carved in toto at Khambhatand shipped as a finished piece;however,there is no doubt that it would repayfurther study.Togetherwith the eight marblecolumns from Juban, this tombstone pair representssome of the latest evidence for the continued export of Khambhatimarblecarving around the Indian Ocean rim. 28. Neville Chittick,Kilwa:An Islamic TradingCityon theEastAfricanCoast,2 vols. (Nairobi:BritishInstitute in Eastern Africa, 1974), 2:262, pIs. 1O5a,1o0b, io6a, io6b. 29. Reggio Governo della Somalia,Museo della Garesa:Catalogo(Mogadiscio: ReggioGovernodella Somalia,1934), 6971, nos. 12, 14; Enrico Cerulli, Somalia: Scrittivari editi ed inediti,2 vols. (Rome: Istituto poligraficodello StatoP.V., 195759),1:9-lo, inscription13,fig.5; J.S. Kirkman,Men andMonumentson the EastAfricanCoast(London:Lutterworth Press,1964), 44; P. S. Garlake,TheEarly IslamicArchitectureof theEastAfrican Coast(London:OxfordUniversityPress, 1966), 49-50, figs. 65, 82A; Michele Inzerillo, LeMoscheedi Mogadiscio: Contributoalla Conoscenza dell'Architettura Islamica(Palermo: Ilapalma,1980), pls. 17-18, 23, 27, 28,30; most recently,Lambourn,"Decorationof the Fakhr-al-DinMosque,"61-86.
30. An unpublishedtombstone in the British Museum (acc. no. OA 2307) clearly
belongs to the Khambhatcorpus and probablydates to the late thirteenth century.The stone was found in Aden and donatedto the BritishMuseum in 1840. My thanks to MichaelWillis and Venetia Porterof the BritishMuseum for informing me of the existence of this tombstone and for organizing access. 31. R. Guest, "Zufarin the Middle Ages," Islamic Culture 9 (1935): 408-1o; Venetia Porter,"TheArt of the Rasulids,"in Yemen:3000 YearsofArt and Civilisation in ArabiaFelix,ed. WesnerDaum (Innsbrucket Frankfurt/Main:Pingiun Verlagand UmschauVerlag,1987), cat. nos. 52-54; Porter,"ThreeRasulid Tombstonesfrom Zafar,"Journalof the RoyalAsiatic Society 1 (1988):32-37; and Lambourn,"Carvingand Recarving." 32. RobertHoward,"TheLarMihrab,"Art andArchaeologyResearchPapers9 (1976): 24-25.
33. AnnualReportson IndianEpigraphy (1954-55), inscription Ci68. 34. Publishedby MuhammadNazim, "Inscriptionsof the BombayPresidency,"
Epigraphia Indo-Moslemica (1933-34):
36-37, inscription3335,but not identified as part of the Khambhatcorpus. 35. RosaM. Perezet al., Hist6riasde Goa (Lisbon:Museu Nacional de Etnologia, 1997), cat. no. 165. 36. AnnualReportson IndianEpigraphy, (1965-66), inscription D 95. 37. Z. A. Desai, "AFourteenthCentury Epitaphfrom Konkan,"EpigraphiaIndica, ArabicandPersianSupplement(1965):9 lo. The two stones from Trincomaleehave recentlybeen re-evaluatedby Claude Guillot and LudvikKalus(LudvikKalus and ClaudeGuillot, "Reinterpretationdes plus anciennes stees funerairesislamiquesnousantariennes:III.Sri Lanka," Archipel, 72 (2006): 47-59). There
131
CARVING
AND COMMUNITIES
is no doubt aboutthe readingof the first epitaphwhich recordsthe death of a certain Qadi 'Afifal-Din 'Abdallahal'Alawiwho died in 808 A.H./A.D. 1405. However,debatehas ragedaboutthe readingof the date on the second fragmentaryand weatheredstone which recordsthe death of the daughterof a certainAmIrBadral-Din H'asanalH'alabi.In awranglefamiliarto Islamic epigraphists,debatehas centeredaround the common problemofwhether to read sevenhundredor nine hundredhijri,since the two areeasilyconfused. In this recent publicationKalusand Guillot opt for929 A.H./A.D. 1523.However,this interpretation overlooksthe stylistic evidenceof the scriptitself,which points firmlyto a fourteenthcenturydateand thus the reading729 A.H./A.D. 1329. This is the date retainedin my discussions of the stone. 38. SomasiriDevendra,"ArabicGravestone from TrincomaleeDockyard,"Journalof theCeylonBranchofthe RoyalAsiatic Society14(1970):28-35;AnnualReportson IndianEpigraphy(1972-73),inscriptions D 284-85; SomasiriDevendra,"NewLight on SomeArabicLithicRecordsin Sri Lanka,"in SriLankaand theSilkRoadof theSea, ed. SenakeBandaranayake,Lorna Dewaraja,RolandSiva,and K.D. G. Wimalaratne(Colombo:SriLanka National Commission for UNESCO/ CentralCulturalFund,1990), 212-13. 39. J.P.Moquette,"DeGrafsteenente Paseen Griseevergelekenmet dergelijke monumenten uit Hindoestan,"Tijdschrift voorIndischeTaal-,Land-en Volkenkunde 54(1912):536-48; G. L. Tichelman,"Een MarmerenPraalgrafte KoetaKareueng (Nordkustvan Atjeh),"CulturOIndie2 (1940):205-11;H. K. J.Cowan,"APersian Inscriptionin North Sumatra,"Tijdschrift voorIndischeTaal-,Land-en Volkenkunde 8o(1940):15-21;Lambourn,"From Cambayto Pasaiand Gresik."
de Malik 40.Andre Cabaton,"L'Epitaphe Ibrahima Gresik,"Revuedu Monde Musulman 13 (1911): 257-60; Moquette,
"DeGrafsteenente Paseen Grisee,"53648 and plate;Lambourn,"FromCambay to Samudera-Pasaiand Gresik." 41. Among which the cenotaphside from Kilwa,the mihrdbfrom Lar,one of the three headstonesfrom Dhofarand grave io at KutaKareuengand grave2 at Candi Said Syariffrom Pasai.The reverseof the Kilwacenotaphside is publishedin Chittick,Kilwa,pls. 1o0b,io6b; for grave io at KutaKareueng,see Moquette, "Grafsteenente Paseen Grisee,"pl. 3. A 42. graveat Pasaiwas comparedto that of al-Kazeruiniby J-PMoquette in his "Grafsteenente Paseen Grisee"(for a summary in French,see Andre Cabaton, "Lestombes musulmanes de Pase [Sumatra]et de Grisee [Java]compares aux monuments de la meme esp&e de l'Hindoustan,"Revuedu MondeMusulman
23 [1913]: 112-16).
Porter paralleled
the three Dhofarstones to known examplesfrom Khambhatin "Three RasulidTombstonesfrom Zafar,"34-35. 43. Selma al-Radi, The "Amiriyain Rada". TheHistoryand Restorationof a Sixteenth-CenturyMadrasain the Yemen, Oxford Studiesin IslamicArt 13(Oxford: Oxford University Press, 1997), figs. 6, 52. 44. al-Radi, Amiriya, 100-101, parallels given in nn. 95,96.
45. A marblecenotaphside panel in the style commonly seen atAhmadabadwas sent to Goa in the mid-sixteenth centuryfor the graveof Dona Caterinaa Piro, the wife of the PortugueseviceroyGarciade Sa,in the Churchof Our Ladyof the Rosary. However,the stone bearsno relationto cenotaphsides producedat Khambhat and should not be included in the corpus. 46. Similartombstones in Aden were mentioned to me by GeoffreyKing of the School of Orientaland African Studies,
Universityof London. 47. The three Khambhattombstones now held in the Mahdawimosque in Vadodara were transportedthere in the early twentieth century and thus belong to the home corpus (unpublishedbut listed in AnnualReportson IndianEpigraphy [1978-79], inscriptions D 32-34). A
tombstone commemoratinga death in 675 A.H./A.D. 1276and formerlyheld in the Pir MuhammadShahdargahat Ahmadabad may havebelonged to the corpus but has since disappeared.See M. A. Chagatai, "MuslimMonuments of Ahmadabad through Their Inscriptions,"Bulletinof theDeccanCollegeResearchInstitute3.2 (1942): 25-26, pl. 3. 48. See Guest,"Zufarin the Middle Ages," 409, citing Ibn Battata, Travelsin Asia and
Africa, 1325-1354,trans. H. A. R. Gibb
(London:Hakluyt Society,1929): 2:201. Ibn Battutaapparentlyvisited the shrine of this sheikh and also met his sons. 49. A bilingual inscription in Sanskritand Persiandescribesthe construction of a mosque at Somanatha-pattanain 662 A. H./1320 v.s. (equivalentto a.d. 1264). The mosque was constructedby the merchant Firuizibn AbuiIbrahimibn Muhammad ibn 'Alial-'Iraqiin memory of his father AbuiIbrahim.A comparisonof the nasabs (lines of descent) given in our Khambhat tombstone and this inscription suggests that the Hasan of our epitaphwas Abfi Ibrahim'sbrother.Forthis inscription, see Desai, "SomeArabicInscriptionsof the RajputPeriod,"10-15, inscription 4. 5o. Also, we haveno evidence to suggestthat the Khambhatworkshopsalso workedfor the many non-Muslim Gujaratissettled outside India. 51. Desai, "FourteenthCenturyEpitaphfrom Konkan,"9-1o. In this case the slabwas found at the Aidruslyyakhanqahin Rajpuri,and there is no guarenteethat 'Ala'al-Din was actuallyburied here. 132
ELIZABETH
LAMBOURN
52.
ForMalik Mufarrah.'s careerin Gujarat, see Desai, "Khaljiand TughluqInscriptions from Gujarat," 9-14,
19-21.
53. The term is also found in a variant spelling, naina, on a number of fourteenth-centuryMuslim tombstones from Fujianin easternChina. See Chen Dasheng and LudvikKalus,Corpus d'inscriptionsarabesetpersanes,87-88, 232-33, 241-43.
54. Quoted in Ahmad al-Qalqashandi,Subh al-a 'sha'fisina 'atal-inshd' An Arab Accountof India in the14thCentury:Being a Translationof the Chapterson Indiafrom al-Qalqashandi'sSubhal-A'sha',trans. and ed. Otto Spies (Stuttgart:W. Kohlhammer, 1939), 32.
55.al-Qalqashandi,ArabAccountof India, 27-28.
56. Ruth Barnes,IndianBlock-PrintedTextiles
in Egypt:TheNewberryCollectionin the AshmoleanMuseum,Oxford,2 vols. (Oxford:ClarendonPress,1997),2: cat. nos. 354,944;the firstis datedA.D. 1420, +/- 40, and the second
A.D.
1410, +/- 40.
57. The motif of single silhouetted tree was certainlyvery fashionableon latersilks. See StephenCohen, "AGroupof Early Silks:The TreeMotif,"in The WovenSilks ofIndia, ed. JasheenDhamija(Bombay: Marg Publications, 1995), 19-36.
58. Fora comparisonwith materialavailable locally along the coast of EastAfrica,see G. S. P. Freeman-Grenvilleand B. G. Martin, "APreliminaryHand Listof the ArabicInscriptionsof the EasternAfrican Coast,"Journalof theRoyalAsiaticSociety, no.
2 (1973): 98-122.
For Dhofar, see
Giovanni Oman, "PreliminaryEpigraphic Surveyof Islamic Materialin Dhofar," Journal of Oman Studies 6 (1983): 227-89.
Forcontemporarylocal alternativesat Samudera-Pasaiin northern Sumatra,see ElizabethLambourn,"TheFormationof the batuAcehTraditionin Fifteenth CenturySamudera-Pasai,"Indonesiaand
the Malay World 32.93
(July
2004),
211-48.
ForeasternJava,see L. C. Damais, "EtudesJavanaisesI. LesTombeaux MusulmanesDatees de Tralaya,"Bulletin de l'EcoleFranfaised'ExtremeOrient48.2 (1957):353-415.
59. Rockwell,Art of Stoneworking,166. 60. These ideas are developedin more detail in a forthcoming article,Elizabeth Lambourn,"Aedicules,mihribs, Minaretsand Chawris:Experimentsin IslamicArchitecturalDecoration in Fourteenth-CenturyGujarat." 61. The Larmihrabshows tracesof ocher paint in its uppersection, and the possibility is not to be excluded. See the discussion of painted scuplturein Kirit Mankodi, TheQueen'sStepwellatPatan
69. Inscription from Kapadwanj,dated A.D. 1370-71, signed by 'Uthman [ibn]Kamal Khattat.See Desai, "Khaljiand Tughluq Inscriptionsfrom Gujarat,"17-18. 70. Desai, "Khaljiand TughluqInscriptions from Gujarat,"23-24. 71. Desai, "Khaljiand TughluqInscriptions from Gujarat,"32-37. 72. Z. A. Desai, "Inscriptionsof the Sultansof Gujaratfrom Saurashtra,"Epigraphia Indica,Arabicand PersianSupplement (1953-54):
61-62.
73. Desai, "Inscriptionsof the Sultansof
in the 113os, gives detailed instructions for the commission of a bronze lamp. See S. D. Goitein, ed., LettersofMedievalJewish Traders:TranslatedfromtheArabicwith Introductionsand Notes (Princeton: Princeton UniversityPress,1974),195, letter39. 84. See Lambourn,"Decorationof the Fakhr al-Din mosque,"71-76. 85. MarkHorton, "Colonisationof the East African Coast:The MandaEvidence," Journalof theRoyalAsiaticSocietyno. 2
(1986):414-15. 86. Merutunga,Prabandhacintamani,159. 87. Though much later,we also know that in 74. Merutunga,ThePrabandhacintamani,or the earlytwentieth century "near Wishing-StoneofNarratives,trans. C. H. Porbandara valuabledescriptionof Tawney(Calcutta:Gilbertand Rivington building stone is extractedfrom the hills for the Asiatic Society, 1gol), 165. and sent to Bombayin largequantities," 75. Desai, "Inscriptionsof the Sultansof (Bombay: Franco-Indian Research, 1991), ImperialGazetteerof India, Provincial Gujarat from Saurashtra," 65-67. 9-12. Series,BombayPresidency,2 vols. 76. The same approachis seen on two half62. The majorpublicationon Dabhoi in (Calcutta:Superintendentof Governfinished cenotaphlids from Pasai (grave2 Englishis still JamesBurgessand Henry Printing, 1909), 2:349,356. ment 6 at and grave Syarif Said at Candi Town of the of The Antiquities Cousens, Gazetteerof India, Bombay 88. Imperial Sareh). Teungku Waterstone, George Dabhoi (Edinburgh: 2:346.Forthe maps, see a map Presidency, Battuta: Rehla ofIbn The Battita, Ibn 77. i888). PetrusBertiusfrom the by of Gujarat trans. Ceylon, and Islands Maldive India, 63. See Khoury,"MihrabImage." of 1612or a map of Indiaby Thresoor Caert Oriental Mahdi Husain (Baroda: 64. See Alka Patel,BuildingCommunitiesin YearesTravelsinto Some in Herbert T. Institute, 1976), 67-69. Gujardt:Architectureand Societyduring andAfriqueof 1665, ofAsia Parts Divers 78. Forfurther analysis,see Lambourn, the TwelfththroughFourteenthCenturies in both reproduced SusanGole,A Series "Carvingand Recarving." (Leiden:Brill, 2004), 51ff. for another of EarlyPrintedmapsof India in Facsimile, 79. 'Allami, A'7n-iAkbari,trans. H. F. point of view. 2nd rev.ed. (New Delhi: Sanskritiand Blochmann, D. C. Phillott, H. S. Jarrett, 65. See discussion and referencesin LamHarnold-HeinemannPublishers,1984), and J.Sarkar,3 vols. (Delhi: LowPrice bourn, "Collectionof Merits,"155-60. map lib and title page,respectively. Publications,1989), 2:247. 66. See Ghulam.Yazdani,"Inscriptionsin the 89. ImperialGazetteerof India, Bombay 8o. Damais, "EtudesJavanaisesI,"353-415. Tombof BabaArjunShah,Petlad," Presidency,2:340-46. 81. The fact they arealso carvedon both Epigraphia Indo-Moslemica (1915-16): 15sides, a practicerarelyseen at Khambhat, go. H. G. Briggs,TheCitiesof Gujarashtra 16;and for my re-analysis,Lambourn, (Bombay:Times Press,1849), 202. may also reflectlocal Dhofariusage. For "Collectionof Merits,"94-96. This topic is explained more fully in E. 91. influences local of possible details further in this 67. The verb katabais ambiguous "Brick,Timberand Stone: Lambourn, Recarving." and "Carving Lambourn, see context and could havethe simple and the Construction Materials Building fusion the of discussion in-depth For an 82. out wrote meaning "towrite" (AbuiBakr History in Architectural of Islamic "Decoration Lambourn, see influences, of write the text) or the more complex "to 191-217. 23 (2006): Muqarnas Gujarat," 71-76. Mosque," al-Din Fakhr of the and and compose" (AbuiBakrwrote out in a merchant from a letter For example, 83. composed the programof inscriptions on Aden to the owner of a bronze foundry at the tombstones). Dahbattanon the Malabarcoast, written 68. Lambourn,"CollectionofMerits,"222-27. Gujarat from Saurashtra," 66-67.
133
CARVING
AND COMMUNITIES
ML
At
PZ
LA .1A
Nor
A
RUTH BARNES
INDIAN TEXTILES FOR ISLAND TASTE GujaratiClothin EasternIndonesia
Indian textile (detail);Gujarat, tradedto Sulawesi,fourteenth century;cotton, block-printed, resist applied, dyed;warp310cm, weft 97 cm. TapiCollection, Surat, Gujarat,01.196.
Abstract The articlediscussesspecificallythe Indiantradetextilesfoundin easternIndonesia, but also attemptsto place them in the wider context of Indian Ocean trade. The main focus is on the wide rangeof painted and block-printedcotton fabricsthat havesurvivedin Sulawesiand the southernMoluccas.Recently discoveredtextiles with large figuralrepresentations,hand-paintedwith mordantsand resist,add importantnew evidenceto this material.Not only do they expand our understandingof export textiles, but they add considerablyto the historyof Indianpaintingpriorto the Mughalperiod.A secondissueaddressed is the Indianproductionof cloth, possiblymade on orderto suit easternIndonesian taste. It is shown that these distinctlyun-Indiandesignswere probably made in Gujarat,as similar materialwas also exported from northwestern Indiato Egypt.
Coarseclothfrom Cambayis of valuein theMoluccas;andfor thefinersort,all the enroladoclothfromBonuaquelim, withlarge,mediumorsmallladrilho,patolas,all thecoarseandwhitecloth... buttheprincipalmerchandise is clothfromCambayand thetailsofwhiteoxenandcowswhichtheybringfromBengal. TomePires,writingaboutTernate,ca. 1515 When Europeanexplorersand merchantsbecame involvedwith the markets of Southand SoutheastAsia in the earlysixteenthcentury,manyof them commentedon the richnessof the Indiantextiletradeandon its importanceforAsia's maritimecommerce.ClothfromGujarat,Bengal,andthe CoromandelCoastwas exportedto manypartsof the IndianOceanto be used as dressand for furnishings,butin someregionstheseforeigntextileswerealsointegratedinto a complex politicalandsocialsystemof gift exchanges.Thisintegrationoccurredespecially in SoutheastAsia,wheretextilesfrequentlymovedfroma secularinto a religious context and could eventuallybecome elevatedto the high status of heirloom items.The regionapparentlyhad an insatiableappetitefor cloth. Despitehaving a highlydevelopedweavingtraditionof its own,with a time depththatgoesback to the prehistoryand earlyhistoryof Austronesiansocieties,therewas a considerabledemandfor importedtextiles, especiallyfrom India.,These foreignfabricshada formativeinfluenceon indigenoussocieties,both as prestigeitemsthat could eventuallytakeon a significantritualroleand as inspirationforthe developmentof localdesign.ForeasternIndonesiatwo typesof clothwereof particular importance:double-ikatsilkpatofriand mordant-and resist-dyedcotton fabrics. Bothtypeswereat leastinitiallyproducedin NorthwestIndiaandwereexported 135
via the ports of Gujarat.But as the epigraphfrom TomePiresindicates,textiles fromCentraland easternIndia,exportedvia the CoromandelCoast,werealso of importance. AlfredBuhlerwas the firstto point out the effect of Indianpatola textileson Indonesianikat design.2Severalstudieshave followedon the receptionof these textilesin SoutheastAsiansocieties,in particularin Indonesia.3 Therenow is reasonablygoodknowledgeof theirhistoricalandsocialsignificanceandtheirinfluence on indigenousdesigns.So farthe only extensiveethnographicdiscussionof the local function and use of the resist-and mordant-dyedcotton textiles exists, however,forthe Sa'danToraja,publishedby HettyNooy-Palm.4It is basedon her field researchamongthis particularethnic groupinhabitinga part of the wider Torajaregionin Sulawesi.ToosVan Dijk and Nico de Jongealso havepresented evidencefor the use of Indiancotton textilesas ceremonialcloths and heirlooms in the southernMoluccas.5 The silkpatola are technicallyand aestheticallyextremelyaccomplished,but they follow a fairlystrict and limited design repertoire.The cotton textiles are far morevariedin their iconographyand technicalexecution.They rangefrom meticulouslyhand-paintedresist-and mordant-appliedfabricsto textiles with ratherstatic,repetitiveblock-printedimages;manyof them combineblock-printing with hand-appliedmordantsand resist.This articleis concernedmainlywith the mordant-and resist-dyedcotton textilesproducedin Gujarat,and it focuses on theirreceptionin easternIndonesia,althoughit beginsby placingthe material into the widercontextof IndianOceantextiletrade.Thetradeof textilesfromthe CoromandelCoastto SoutheastAsia,which increasedin prominencein the seventeenthand eighteenthcenturies,is not a majorconsiderationhere,as it moves beyondthe chronologicalrangeof this volume and was intendedprimarilyfor the westernIndonesianmarket.6Nevertheless,a particulartype of clothexported from SoutheastIndia,the palampore with an elaboratefloweringtree as its main motif, will enterinto the discussion,as it maycontributeevidencefor the transmissionof design,not onlyfromIndiato Indonesiabutpossiblyalsoin the reverse direction.7 136
RUTH BARNES
Textilesfrom Gujaratfor an InternationalMarket It was in particularGujarat,or possiblymorewidelyNorthwestIndia,that suppliedthe earliestmajorsurvivinggroupsof textiles.Smallfragmentsof probably Indianoriginhavecometo lightboth at CentralAsianandRedSeaarchaeological sites, datingfrom the third to fifth centuriesA.D.8 But the firstsubstantialfinds all come from Egypt.As these textilesarevisuallyand technicallyrelatedto the resist-and mordant-appliedcloths tradedto easternIndonesia,they need to be introducedhere as evidencefor the geographicalexpansionof the Gujaratitextile market.Theyareplainweavecotton,usuallyblock-printedwith resistand/or mordantand dyed red and blue. Their Indian origin was first identifiedin the The textiles obviouslyhad a utilitarian 1930s by the textile historianR. Pfister.9 function in the Near East,as there are numerousremainsof hems and seams, indicatingthat they were sewn into garmentsor used for pillows and curtains. The fabricssurviveas fragments;they often seem to havecome fromwastedisposalsites,the mostfamousat Fustat(southernCairo).Somemayalsohavecome fromgraves,possiblyassociatedwith the cemeteriessouthof the Fustatsite.More than two thousandof thesetextilefragmentsfirstappearedin the art marketsof CairoandAlexandriain the earlytwentiethcenturyand endedup in privateand publiccollections,of whichthe largestis the NewberryCollectionof morethan Itneedsto be stressed,though, 1,200 textilesin theAshmoleanMuseum,Oxford.10 from sites;eventuallysimilarmatenone of the came provenanced fragments that rialwas found at GebelAddaand QasrIbrim,both in UpperEgypt,but the first archaeologicalexcavationto pay seriousattentionto similarfragmentswas carriedout in 1978and 1980by DonaldWhitcomband JanetJohnsonat the RedSea Thatmaterialwasdatedby contextto the thirteenthto portof Quseiral-Qadim.11 fifteenthcenturies. On stylisticgrounds,Pfisterhad arrivedat a similardatefor the Indo-Egyptian tradetextiles.He comparedtheirdesignsin particularto motifsfoundin the architectureof Gujarat,wherehe sawstrongparallels.Thedatedarchitecturewas therebyusedas a meansto providea tentativechronologyforthe textiles,relating themmostlyto the thirteenthto sixteenthcenturies.As partof recentresearchon the NewberryCollectionin Oxford,however,radiocarbondatingshowedthatthe earliestof thesetradeclothsweremadein the tenthcenturyA.D.,withthe majority datingbetweenthe eleventhand fifteenthcenturies.This periodcoincidedwith the activeeconomicmaritimecontactbetweenFatimid,Ayyuibid,and Mamluik EgyptandIndia, via Aden and the Red Sea. Thedesignsfoundon theseIndo-Egyptianfragmentsoftenarestrikinglysimilarto thoseseenon the largecottontextilesthatweretradedto SoutheastAsiaand survive in various parts of eastern Indonesia, so far documented in particular for 137
INDIAN
TEXTILES
FOR ISLAND
TASTE
ip
~ ~ ~~ ~ ~ ~ ~ ~
le, AL AV,
V-
iL
2
3
2
Gujarattextile for the Egyptian market,calibratedradiocarbondate 1260to 1330;cotton, block-printed resist, dyed;35x 23.5cm. Ashmolean Museum, Oxford University, NewberryCollection, iggo.logg.
Sulawesi, the southern Moluccas, and Timor. Until recently, it had been assumed that in this tropical climate textiles could not have survived for more than two to three hundred years, at the very most. Here again the use of radiocarbon dating has furnished surprising results.12 It is now known that many of the cloths found in eastern Indonesia were made as early as the fourteenth or fifteenth century.
3 Ganesaof Bara,Blitar,East Java (A.D. 1239),view of back with linked circle design. FromN. J. Krom, Inleidingtot de Hindoe-Javaansche Kunst,3 vols. (The Hague:Martinus Nijhoff, 1923),3:pl. 56.
Indian Textiles in Indonesia: The Early Evidence Many parts of Indonesia, particularly the eastern part of the archipelago, are still historically and archaeologically understudied. At present there is little evidence in written documents or other media for a date post quam for the arrival of Indian textiles in the island cultures. We can establish a certain link between Indian textile prototypes and Indonesian production, but at the moment we cannot saywhen this adaptation first took place. Regarding local production in Java,we have Jan Christie's analysis of the Central Javanese sima charters, which record tax and labor rights granted, in particular, to religious foundations from the ninth to the fifteenth centuries. From these documents we know that imported cloths were a desired commodity to the island societies from the ninth century onward, when Indian textiles, called buat klingputih, i.e., "white cloth made in India/ Kalinga," are first referredto as gifts in Javanesetax edicts.13The Southeast Indian Kalinga domain (shortened to Kling in the Javanese source) had economic and political contacts with Javaat the time. Javanesesculpture of this period pays considerable attention to representing the patterns of dress and ornament, and by the thirteenth century we see designs that may have had their source in an Indian textile prototype; for examples of designs available to connoisseurs of Indian textiles, it is helpful to look at the Indo-Egyptian material (figs. 2, 3), although the cloth exported to Javamay have been patterned with more costly materials. Christie has pointed out that the sculptural evidence between the ninth and the thirteenth centuries shows a development from band-arranged designs to an overall, continuous pattern, and she has tentatively related this shift to the increased impact of Indian designs on 138
RUTH BARNES
4
4
Gujarat,traded to Sulawesi, calibratedradiocarbondate fourteenth century; cotton, blockprinted resist, mordant dyed;warp 406 cm, weft 99 cm. Ashmolean Museum, Oxford University,EA 1995.61.
Indonesian aristocratic fashion.'4No actual textiles of this date have so far been discovered in Java.The earliest surviving cloths come from eastern Indonesia, and they are slightly later,with radiocarbon dates that point to the late thirteenth and fourteenth centuries for the earliest examples. They generally do not come from a court or temple context but were found in the highland interior of Sulawesi, inhabited by the Toraja.They have become sacred heirloom cloths and are essential ceremonial paraphernalia for certain annual and life-cycle rituals." Coastal settlements with access to international trade would have been the initial recipients of the fabric. For the fourteenth-century textiles that survive from the Torajaregion, the importing principality would possibly have been Luwu, at the time the major economic and political power in southern Sulawesi and probably the oldest ofthe Bugis kingdoms.'6Forcomparative purposes one may consider Leonard Andaya'saccount of the spice-rich and wealthy northern Moluccas at the time of first European contact in the early sixteenth century. Especially relevant is his discussion of the relationship between the powerful sultans of Ternate and Tidore and village societies on nearby Halmahera and in the Bird'sHead region of New Guinea, where Indian cloths played an important role. He shows that cloths were passed on by the ruler to his dependencies on the periphery of the realm, as essential gifts to establish and confirm local alliances.'7To the people of the interior the presentation of Indian textiles was a significant indicator of the coastal ruler's legitimacy. By offering and accepting these donations, the relationship of interdependence was emphasized and reinforced. It is possible that on Sulawesi the relationship between coastal and interior regions may have followed a similar pattern. There is considerable historical and ethnographic evidence for the social 139
INDIAN
TEXTILES
FOR ISLAND
TASTE
-
lw
. ....W, ..Aifimi
L -
..v
'WV
.1 -%--- .
5
5
Gujarat,tradedto Egypt, fourteenth-fifteenth century; cotton, block-printed resist, dyed;warp30 cm, weft 12.5 cm. Ashmolean Museum, Oxford University,NewberryCollection, 1990.105.
significanceof cloth in Indonesiansocieties,and the emphasison gift exchanges involvingtextilesis welldocumented."8 Clothis presentedat all life-cycleceremonies, with its presenceoverwhelminglyin evidenceat weddingsand funerals,in particular.Duringritualceremonies,the sacredspaceis often definedby textiles, and lineagehouses may be drapedwith them, as is documentedfor the Toraja, who use theirIndianimportedclothsfortheseoccasions. Indian Productionfor LocalTaste Were the island recipientsof Indian textiles satisfiedwith whateverstyle and designthey received,or is thereevidenceforpreferencesof particulartypes?Was there evidencefor textiles being tradedthat were specificallyin line with local taste?The cotton clothsthat survivefromeasternIndonesiatend to havesimilar, ratherlarge,dimensions:typicallythey areapproximatelyone meterwide,with a lengthof fiveto six meters.Theyareusuallykeptat their full length,ratherthan cut;this is trueevenforclothsthatarevisuallydividedinto severalpanels(seefig. ii, below).The designscan varya greatdeal,but the majorityof clothshaveoverall floraldesigns.Animalsand humanfiguresarealso found;amongthe earliest areredtextileswith resist-and mordant-dyedgeese (fig.4). Also amongthe early piecesaretextileswith largefemalefigures;this type will be discussedin further detailbelow. The similaritiesbetween the Indo-Egyptianfragmentsand the textiles for easternIndonesiahavebeen stressedabove.It is just as revealing,though,to recognize the differences.Amongthe Indiantextilesfor the Egyptianmarket,there is a largenumberthathasbeenresist-dyedblueonly;apparentlytheseindigo-blue cloths weremuch in demand.Manyof them havecontinuousdesignfieldswith small rosettesor star shapes (fig. 5). Similarpatternsalso survivefrom eastern Indonesia,but they are inevitablydyedred (fig.6); up to now not a singletextile hasbeen foundin the regionthatwasdyedblueonly,althoughmanycombinered andblue.Thereasonforthis preponderanceis likelyto be culturalandprobablyis relatedto the local significancegivento the colorred.Redis of particularimportance in many SoutheastAsian societies, and is often associatedwith spiritual 140
RUTH BARNES
*
0
0~~~~~~~~~~~~~~~~~~~~~~~~~2
6
6
Gujarat,tradedto Sulawesi, radiocarbondate late fourteenth to earlyfifteenth century; cotton, block-printedresist, mordant dyed; half-length, warp 242 cm, weft 95 cm. Ashmolean Museum, Oxford University,EA1998.223.
powerand fertility.This associationis apparentin the ninth- and tenth-century Severaltwentieth-century Javanesedocumentswith whichChristiehasworked.19 ethnographicreportsalso confirmthe associationfor easternIndonesia.2Many of the Indiantextilesexportedto Indonesiaareused locallyin a ceremonialcontext, wherethe use of a particularcolormaybe prescribed.Colorsymbolismis a difficulttopicto interpretin a cross-culturalmanner,as thereis rarelyone single answerto the meaningof any color.Whateverthe specificmeaning of red in a SoutheastAsiansocietyor culturalcontext,certainassociatedpropertiesrelateto powerandfertility,butalso to danger.Red,as the colorof blood,can standforthe positiveaspectsof fertilityandchildbirth,but also forthe negativeone of warfare. However,even the latteris an ambiguousconnection,as warfarein manySoutheastAsiansocietieswas formerlyassociatedwith head-hunting,a practicethat in turn fulfilledthe cycletowardfertility.The reasongiven for the takingof heads involvedthe gainingof well-beingandfertilityforthe community. arerareamongthe Indo-EgyptiantextilesbutarecomFiguralrepresentations mon forthe clothsfoundin easternIndonesia.Ratherthan relatingtheirabsence in the formergroupto the oftenoverstressedprohibitionof showinghumanforms in Islamicart (a prohibitionthat does not applyto secularart in the Islamiccontext),this absenceis morelikelyto haveits explanationin the functionforwhich the textileswereintendedin Egypt:large-scalefiguraldesignswerenot suitableto be cut up andtailoredinto garmentsand domesticfurnishings.TheIndiancloths fromeasternIndonesiadid not havethis function,and amongthe earliestradiocarbon-datedtextiles from the regionare cloths that show largerepresentations of femaledancers,often accompaniedby attendants.These fourteenth-century piecesareblock-printedand dyedredandblue;the indigo dye is often so crudely appliedthatmuchof the designis obscured.2Butwhereit is possibleto makeout the designof the women'sdress,we discoverthey arewearingfloralpatternsthat 14
INDIA
TETIE
FOILNDTAT
21
A.TMA
....... .... .......
...'
.
~ ~~ ~~ ~~~~~~~~~~~~~~~~~~~~~ .
.....
....
. ......
~~~~~~~~~~~~~~.
Di'
8~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,4
7 Gujarat,traded to Sulawesi, radiocarbondate ca. 1500-50; cotton, mordant- and resist-painted; warp 550cm, weft 102 cm. Tapi Collection, Surat,Gujarat,01.51. 8 Gujarat,traded to Sulawesi, radiocarbondate ca. 1500-50; cotton, mordant- and resist-painted; warp 531cm, weft 99 cm. Tapi Collection, Surat,Gujarat,01.28.
Textiles as Painting: New Evidence for the History of Jaina Painting Against the background of these ratherunrefined images, I want to bring attention to two remarkabletextiles that were made in an Indian Jaina context, most likely in Gujarat.Although obviously made for an Indian client, they entered Gujarat's international trade network, and they must have been received with acclaim, as they were carefully preserved for five centuries. Radiocarbon analysis dates them to the late fifteenth or early sixteenth century, and they were recently collected in Sulawesi. Both are now in the private Tapi Collection in Surat, Gujarat(figs. 7, 8).23 It is true that most Gujaraticotton textiles for the Indonesian market were blockprinted. These pieces, however, are rare examples of hand-painted cloths, where mordant and resist were applied with a fine brush before the textile was dyed. The line drawings and subsequent color applications are of superb quality. In the first painted textile, we are looking at a festive entertainment scene, probably from a courtly context (fig. 9). A richly dressed couple at the center of the cloth possibly represents a prince and his consort, or at least enacts that role. Between them is a small flowering tree; he holds a long fly whisk or flower stalk, while she has small single disks in her raised hands, no doubt meant to be cymbals. On either side of the couple are five females; they all move toward the two central figures, some with dancing steps, others bringing offerings of refreshments. On the consort's side, four of the attendant ladies are playing musical instruments. The figures' movements are lively and highly expressive. The figurative style and depiction of facial features are familiar from Jaina manuscript paintings of the fifteenth and sixteenth centuries, with the characteristic three-quarter profile including a protruding second eye. No further attempt is made to add a depth of perspective; instead, the figures remain strictly in the two-dimensional space of the cloth's surface, even when their bodies turn to suggest movement. This convention is maintained throughout. However, the attention given to details of individual features beyond this convention of perspective, 142
RUTH BARNES
NO! 4CW4AA91 Z1,04! ~ ~ ~ ~ aq I
~
~
~
~
~
~
4-
~ ~ ~~ ~~~~~~~~~~~1
9
Detail of fig. 7. 10
Detail of fig. 8.
particularlyto dressandpersonaladornment,is unusual.Thehalflife-sizefigures aredrawnwith suchcarethatit is possibleto identifytextilepatterns,jewelry,and coiffure.Severalof the dancershavelargepatkas(sashes)with swirlingspiralstied aroundtheirwaist.Theprinceis wearingCentralAsian-style boots, althoughhe is otherwisenot depictedin a mannerreservedfor Mongolsin Jainapaintings.24 We can thereforeassumethat he is supposedto be Indianratherthan of Central Asian origin. The backgroundis denselyfilledwith floweringbushes,small elephants,horses,gazelles,birdsand two lions, as well as textilesdrapedfrom ornamentedhorizontalpolesor shelves.Someof these motifsreappearon othertrade textiles;comparethe floweringtreesandsmallgeeseof thebackground,whichcan becomethe dominantfeatureon clothssurvivingfrom easternIndonesia.Three shieldsand swords,as well as a bow and quiver,aresuspendedfromledgesabove the figures'heads.The entirescenethereforecombinesentertainment- dancing andmusic-with imagesof nature,hunting,andweaponry. Thesecondcloth (fig.1o; seealso fig.8) showsa prominentladywith herentourageof attendants,againincludingdancersand musicians.Onceagain,the technicalachievementis remarkable.Despitethe formulaicrenditionof facialfeatures andthe exaggeratedgestures,the figuresaredrawnwith considerableindividuality.This is particularlyapparentin the majorfigureherself,who is distinguished fromherattendants:sheis slightlytallerthanthe otherwomen,with a plumpface and a prominentdoublechin, and she is furtherdifferentiatedby havinga small servantgirl facingherwith the offeringof a bowl;althoughplacedoff-center,all other figuresare turning towardher.The iconographicmeaningof the scene is uncertain;it mayrepresentthe preparationsfor an entertainment,maybeevena weddingceremony. As most fifteenth-centuryJainapaintingsfromGujaratsurviveonly in smallscalemanuscriptform,thesehalflife-sizeimagesprovideimportantnewevidence for the artisticlevel achievedin paintingof the time. Apartfrom these cloths in 143
INDIAN
TEXTILES
FOR ISLAND
TASTE
Ai/
Er
11
.W"ir; l ~S!''
e a
r} >
<>1X;> h
e
i
s
' ts,X;itt
%
%
"
t'
W
;;
i
.8~
11
Gujarat,traded to Sulawesi, seventeenth-eighteenth century (?);cotton, painted and blockprinted resist, mordant and resist dyed;warp 473cm, weft go cm. Tapi Collection, Surat,Gujarat,01.337.
12
Detail of fig. ii.
13 Gujarat,tradedto Sulawesi,fourteenth century; cotton, mordant-and resistdyed;warp 525cm, weft 107cm. Tapi Collection, Surat,Gujarat,01.335.
the Tapi Collection, only one other textile of this type has so far been published; it is in the Abbott Collection in the National Gallery in Canberra and is dated ca. 1500.25The Canberra textile is also entirely hand-painted rather than blockprinted. Compared to the two cloths in the Tapi Collection, it displays a slightly mannerist stiffness, as the figures are further elongated and move with a stilted stylization. But all three textiles have a quality and liveliness about them that is otherwise not known from fifteenth- and sixteenth-century Jaina painting. The history of Indian manuscripts usually discusses Jaina painting as formulaic and stylized, although not without charm. These characteristics are true for the manuscript illustrations that have survived. With the discovery of these exceptional textiles, however, there is evidence for a painterly tradition that is monumental and of outstanding quality. How do these painted textiles fit into the story of Indian trade textiles? This is a difficult question to answer. They were recently bought in Bali or eastern Indonesia, regrettably without further detail of family or clan affiliation other than that they had once been heirlooms in Sulawesi. It is often repeated in the literature on Gujarati export textiles that the cotton goods were typically of medium quality, block-printed, and mass produced. That description is generally borne out by the material surviving from both Egypt and Indonesia. But the picture emerging now is more complex. There also was the economic connection and incentive to send high-quality cotton cloths to eastern Indonesia from Gujarat, in some cases as early as the fourteenth century, as can be seen in an exceptionally fine cloth of flowering trees, also entirely painted resist and mordant (fig. 13). These examples raise an important issue: they represent scenes and settings that are purely Indian in context. There is no reference to any non-Indian taste, style, or iconography. Yet they were clearly appreciated and treasured at their receiving end in eastern Indonesia. From India to Sulawesi-from Sulawesi to India? To return to an earlier question: is there evidence for textiles being imported especially to suit local taste or ceremonial requirements? Figure ii is relevant for this consideration. It shows a version of the "floweringtree" motif that was widely accepted at an early date in Sulawesi; among the most common fourteenth-century textiles are elaborate versions of this design, usually block-printed. This particular example is likely to take us chronologically beyond the fifteenth century. The cloth's length is divided into two panels, each with a flowering tree design that has its origin in ornate tree images typically found on palampore textiles from 144
RUTH BARNES
VA
14
the Coromandel Coast,tradedboth to Europeand SoutheastAsia from the sev-
14
Coromandel Coast, traded to Sumatra,seventeenth-eighteenth century; cotton, painted mordant and indigo; warp 173cm, weft 107
cm. TapiCollection, Surat,Gujarat, 99-1949-
tre pmge2
Cooane
utas
Cos.aapr
Toaa,a te da thAV
hv
enaalbet
h
thesutCorajamandel CaThe trdelothha tworetnulrop cremna
aiao
maindfields,east Asiarmthwie sevmgso hiw ocsioal s"reo-ie
a'txie.7I
n
nepesteesnl-evd
ag
ders at the narrow ends; the direction of the tree patterns in the two panels are reversed. The reversal of viewpoint is not uncommon among the textiles traded to the Torajaregion. The panels are variations of the same design, with trees or branches growing out of a triangular base, partly with joined trunks that emerge from two roots. The crowns have single leaf shapes with concentric lines repeating
145INDANEXTLESFORISLN TAST
the outlinesK(fig.12). The triangles are here in place of the mound or rocky hill from which the ("classic"palampore tree grows (fig. 14);the two roots that merge into a single trunk or branch are a standard detail that has its source ultimately in Chinese
15
15
Gujarat,traded to Egypt, seventeenth century (?);cotton, mordant and resist applied, length 30 cm, width 22 cm. Ashmolean Museum, Oxford University, Newberry Collection, 1990.1136.
bushes as growing out of one source-the triangle that is the equivalent of the rocky hill of the palampore-then we may be looking at the representation of a single tree. This understanding adds a local dimension to the interpretation of the representation. Both panels show seven trunks or branches, with seven large leaves. The Toraja call this textile a daun bolu (betel pepper leaf) cloth, and its presence is part of the merok ceremony, an annual thanksgiving ritual and evocation of future prosperity. On that occasion a buffalo is sacrificed, but first there is a recitation of the passomba tedong (laudation of the buffalo), a poetic litany that consecrates the animal. In this recital the "heavenly sendana tree" (sandalwood tree) is evoked, a "tree with the life-fluid of the people of the earth": it is called the tree with "seven branches and seven leaves."28What is the evidence that this textile was made in India for Torajarequirements? And if it is Indian, where was it made? We can be fairly certain that the Indian fragments from Egypt were exported from Gujarat,rather than the Coromandel Coast. If this provenance is accepted, then it also becomes relevant for the origin of this Toraja cloth, as designs with large, single-leafed trees or branches also occasionally appear in Egypt and survive in the Ashmolean Museum's Newberry Collection (fig. 15).They may be among the later (sixteenth- or seventeenth-century) part of the collection. The technical knowledge that went into patterning the cloth also strongly suggests an Indian, probably Gujarati, source. The large motifs are hand-drawn resist and mordant, but the very wide border patterns are block-printed. However, they are similar to the borders of the large painted figural textiles discussed earlier, although they 146
RUTH BARNES
aremorecrudelyexecutedhere.Themannerof transferforthe "primitive" design the has to resonance interpretationthat had particular remain among Toraja hypothetical,although one may speculatethat a locally producedceremonial cloth,calledsarita, mayhavebeensentto Indiato initiatean Indianreproduction for the Torajamarket.Ratherthan assumingthat the easternIndonesianrecipientsof Indiantradeclothsplayeda passiverolein this exchangeof goods,it seems that they may haveplacedspecificordersto fulfill their own requirementsand tastes. Conclusion What made Indian cloth so very desirablefor easternIndonesiansocieties?A majorattractionwas certainlythe qualityof dyesused in decoratingthe fabrics. The brillianceand colorfastnessof the designswas unrivaledat the time, and remainedso until well into the nineteenthcentury.Indiantextile tradewas initiallyin the handsof Araband Indianmerchants,but eventuallywas dominated by the Europeantradingcompanies.The Portugueseand laterDutch, English, and Frencharrivalin Asia was drivenby economic factorsand their desireto gain controlof certainmarketproductsof Southand SoutheastAsia.The spices and aromaticsof India and maritime SoutheastAsia were a potential source for hugewealth,and Indiantextilesprovedto be a majorcurrencyin exchange transactionsinvolvingeasternIndonesiancloves, nutmeg, and sandalwood.It was quicklyrealizedthat a successfulmerchantcompanyneededto haveaccess to and preferablya monopolyoverthe distributionof Indiancloth. Patoldand block-printedcotton textiles surviveto this day in easternIndonesian societies that were, even if in an indirect way, connectedto the wider tradenetworkof the IndianOcean.It is clearfrom the materialdiscussedhere that eastern Indonesiancommunities were discerning customerswho could appreciatethe qualityof Indianpaintedtextilesand may havesent ordersto be fulfilled to their own specifications.Nowadaysthe Indiantextiles are typically kept either as personalbelongingsof high social prestigeor as lineage-owned heirlooms.Their designs may have influencedthe design of indigenousweaving. Textileshavebeen-and still are-a majortransmitterof designand technology, and they tend to convey considerablesocial meaning. Of course they share that role with other manufacturedgoods that have historicallymoved between societies,both as prestigeitems and as utilitarianobjects.Few products, though, can claim the same convenientportability;while in the long term consideredfragile,textiles are initiallyfar more durableand easierto transport than, for example,glassand ceramics.Theywere,therefore,primarysourcesof cross-culturalinfluences. 147
INDIAN
TEXTILES
FOR ISLAND
TASTE
NOTES
The epigraphis from TomePires, The SumaOrientalof TomePires,ed. Armando Cortesao,2 vols. (London:Hakluyt Society, 1944),1:216. Bonuaquelim means "Land of
the Kling,"a referenceto the Coromandel Coast;see also n. 3, below.Apartfrom the referenceto patola, the meaning of the textile terms is uncertain. 1. Thereis both linguistic and technological
evidence for the earlypresenceof weaving in SoutheastAsia. Otto Dempwolff, vol. 3, AustronesischesWorterverzeichnis, LautlehredesAustroneVergleichende sischenWortschatzes,Beiheft zur Zeitschriftfur Eingeborenen-Sprachen19 (Berlin:Reimer,1938),has Austronesian tenun,meaning "weave."This definition was refinedby RobertBlustto refer specificallyto loom weaving ratherthan the wider categoryincluding, e.g., basketry.RobertBlust, "Austronesian CultureHistory:Some Linguistic Inferencesand Their Relationsto the ArchaeologicalRecord,"WorldArchaeology8 (1976):34. He argues,"Theconclusion seems inescapablethat the loom was known to speakersof a language ancestralto at least Malay,the Batak languagesand various languagesof northern Luzon.A minimum time depth of 4,000 yearswould seem to be implied." This linguistic interpretationis confirmedby the backstraploom technology sharedby the Aboriginesof Taiwan,who areAustronesianand peoples of Indonesiawhose ancestorsleft Taiwan close to 3,500 years ago. 2.
Alfred Buihler,"PatolaInfluencesin SoutheastAsia,"Journalof Indian Textile History 4 (1959): 4-46; Buhler and
EberhardFischer,ThePatolaof Gujarat: DoubleIkatin India (Basel:KrebsVerlag, 1979).
3. MattiebelleGittinger,SplendidSymbols: Textilesand Traditionin Indonesia 148
RUTH BARNES
(Washington,D.C.:TextileMuseum, 1979);Gittinger,MasterDyersto the World:Techniqueand Tradein Early IndianDyed CottonTextiles(Washington, D.C.:TextileMuseum, 1982); Ruth Barnes,TheIkat Textilesof Lamalera:A Studyof an EasternIndonesianWeaving Tradition(Leiden:E. J.Brill, 1989); Marie-LouiseNabholz-Kartaschoff,"A SacredCloth of Rangda:KambenCepuk of Bali and Nusa Penida,"in ToSpeak with Cloth:Studiesin IndonesianTextiles, ed. MattiebelleGittinger(Los Angeles: Museum of CulturalHistory,University of California Press, 1989): 181-97;
Nabholz-Kartaschoff,"TheSame-Yet Not the Same:Similaritiesand Differences in Some TextileTechniquesand TextileProductsfrom India and Bali,"ed. Ruth Barnesand RosemaryCrill, special issue on South and SoutheastAsian textiles, Textile History 30.1 (1999): 46-
68;RobynMaxwell, Textilesof Southeast Asia: Tradition,Tradeand Transformation (Melbourne:OxfordUniversityPress, 1990); JohnGuy,WovenCargoes:Indian Textilesin theEast (London:Thames and Hudson, 1998).
4. See Hetty Nooy-Palm, TheSa'dan Toraja: A Studyof TheirSocialLifeand Religion, vol. 1,Organization,Symbolsand Beliefs, Verhandelingenvan het Koninklijk Instituutvoor Taal-,Land-en Volkenkunde 87 (The Hague:MartinusNijhoff, 1979);Nooy-Palm, TheSa'dan Toraja,vol. 2, Ritualsof theEastand West,Verhandelingenvan het KoninklijkInstituut voor Taal-,Land-en Volkenkundei18 (Dordrecht:ForisPublications,1986); Nooy-Palm, "TheSacredCloths of the Toraja:UnansweredQuestions,"in Gittinger,ToSpeakwith Cloth,163-80. 5. See Toosvan Dijk and Nico de Jonge, "Bastasin Babar:ImportedAsian Textilesin a South-EastMoluccan Culture,"in IndonesianTextiles:
Symposium1985, ed. GiselaVolgerand Karinvon Welck,EthnologicaN.F. 14 (Cologne:Rautenstrauch-Joest-Museum, 1991), 18-33.
6. Partlyor entirelyhand-paintedcotton textiles werepredominantlytradedto Sumatra,and so far only a few havecome to light from the easternpart of the archipelago,despite Tome Pires's referenceto them. 7. The termpalamporeis an Anglicized combination of the Persianwordspalang, "bed,"andposh, "cover." 8. JohnPeterWild and FelicityWild, "The Textiles"in Berenike 1995, ed. StevenE. Sidebothamand Willemina Z. Wendrich (Leiden:ResearchSchool CNWS, 1996),
12.
(1936): 161-64; Pfister, Les Toiles
Imprimeesde Fostatet l'Hindoustan (Paris:LesEditions d'Artet d'Histoire, 1938). 1o. See Ruth Barnes,IndianBlock-Printed
TextilesfromEgypt:TheNewberry Collectionin theAshmoleanMuseum, Oxford,2 vols. (Oxford:ClarendonPress, 1997),for the full catalogueof the collection. ii. Donald S. Whitcomb and JanetH. Johnson,Quseiral-Qadim 1978:PreliminaryReport(Cairo:American Research Centerin Egypt,1979);Whitcomb and Johnson, Quseir al-Qadim 1980: PreliminaryReport,American ResearchCenter in EgyptReport7 (Malibu:Undena, 1982). The textile finds from India were publishedby Gillian Vogelsang-
20.
1990).
21.
Ruth Barnes,"IndianTradeTextiles," Hali 87 (1996): 80-85; Barnes, "From
India to Egypt:The NewberryCollection and the Indian Ocean Trade,"in IslamischeTextilkunstdesMittelalters: AktuelleProbleme,ed. KarelOtavsky, RiggisbergerBerichte5 (Riggisberg: Abegg-Stiftung, 1997), 79-92; Guy,
WovenCargoes. Jan Wisseman Christie,"Textsand 13. Textilesin 'Medieval'Java,"Bulletinde l'1colefranfaised'Extreme-Orient8o.1 (1993):181-211, esp. 199. 14. JanChristie,"Ikatto Batik?Epigraphic
Data on Textilesin Javafrom the Ninth to the FifteenthCenturies,"in Weaving PatternofLife:IndonesianTextile
245-56; Sophie Desrosiers, Corinne
Debaine-Francfort,and Abdurassul Idriss, "TwoResist-DyedCottons RecentlyFound at Karadong,Xinjiang," in TheRomanTextileIndustryand its Influence:A BirthdayTributetoJohnPeter Wild,ed. PenelopeWaltonRogers (Oxford:Oxbow Books, 2001), 48-55. 9. R. Pfister,"Tissusimprimees de l'Inde medievale,"RevuedesArtsAsiatiqueslo
Eastwood, TheResistDyed Textilesfrom Quseiral-Qadim,Egypt(Paris:AEDTA,
29-43.
See Guy,WovenCargoes,112, fig. 146. 23. They haverecentlybeen publishedin RuthBarnes,StevenCohen and Rosemary Crill, Trade,Templeand Court: Indian Textilesfromthe TapiCollection (Mumbai:India Book House, 2002). 24. See KarlJ.Khandalavalaand Moti Chandra,New Documentsof Indian Painting:A Reappraisal(Bombay:Prince
22.
of Wales Museum, 1969), fig. 3, where a
Symposium, 1991,ed. Marie-Louise
Nabholz-Kartaschoff,Ruth Barnes,and David Stuart-Fox(Basle:Museum of
25.
Ethnography, 1993),11-29, esp. 15. 15. See Nooy-Palm, TheSa'dan Toraja,vol. 2.
16. See David Bulbeckand Ian Caldwell, Landof Iron:TheHistoricalArchaeology of Luwuand the CenranaValley(Hull: Centrefor South-EastAsian Studies, Universityof Hull, 2000). The Bugis are an ethnic and linguistic population based in coastal southern Sulawesi,historically and to the presentdaywith wide-ranging seafaringcontacts. 17. LeonardAndaya, The WorldofMaluku: EasternIndonesiain theEarlyModern Period(Honolulu:Universityof Hawaii Press, 1993) 66,106-7.
i8. ForeasternIndonesia,see, e.g., Barnes, Ikat TextilesofLamalera;RoyHamilton, TheGiftof theCottonMaiden (Los Angeles:Universityof CaliforniaPress, FowlerMuseum of CulturalHistory, 1994). 19. Christie, "Ikatto Batik?"in Weaving Pattern, 11-29, esp.12.
149
INDIAN
TEXTILES
FOR ISLAND
TASTE
Fora brief summarysee Barnes,Ikat Textilesof Lamalera,93-94. See Ruth Barnes,"Womenas Headhunters:The Making and Meaning of Textiles in a SoutheastAsian Context,"in Dress and Gender:Makingand Meaningin CulturalContexts,ed. Ruth Barnesand JoanneB. Eicher(Oxford:Berg,1992),
distinction is made in the facial features between an Indian and two Mongols. See RobynMaxwell, "TextileContributions,"in TraditionsofAsianArt Traced throughthe Collectionof theNational GalleryofAustralia,ed. MichaelBrand (Canberra:National Galleryof Australia, 1995); Guy, Woven Cargoes, 112, fig. 147.
26. See Guy, Woven Cargoes, 1o6. 27.
28.
Both terms referto rituallyimportant cloth used by the Toraja.Saritausually arelong, narrowresist-dyedcloths producedlocally,while maa'are Indian tradetextiles;but occasionallythese terms areinterchanged. Nooy-Palm, Sa'dan Toraja,1:220.
PHYLLIS GRANOFF
LUXURY GOODS AND INTELLECTUAL HISTORY TheCaseofPrintedand WovenMulticoloredTextilesin MedievalIndia
Abstract Advancesin the technologyof cloth designfascinatedmedievalIndianpoets and intellectuals.Aftera briefsurveyof the manywaysin which cloth figuredin the imaginationof the medievalpoet and storyteller,this articlefocuseson one particulardiscipline-philosophy. It arguesthat innovationsin cloth designchallengedtraditionalunderstandingsof howtheworldwasconceived.Asan example, it discussesin detail a particularphilosophytext of the tenth centurybelonging to the realistschool,or the Nyaya-Vaisesika, to showhow the extravagantmulticoloredcloth that was being producedcompelledphilosophersto rethinktheir understandingsof some fundamentalcategoriessuch as substanceand quality. Cloth-makinghereprovidesan excitingcasestudyof the waysin whichtechnologicaladvancesin the artsstimulateddebateandprogressin philosophy,a seeminglyunrelatedarea.
Clothin the MedievalImagination It may come as somethingof a surpriseto learn that it was not particularlyon accountof theiraestheticqualitiesthatthe magnificentproductsof India'smedieval weaversand dyerssparkedthe imaginationof contemporaryintellectuals. Poets,to whomwemightlookexpectantlyforanhomageto the delicatedesignswe know frompaintingsof textilesand actualremainingtextilesthemselves,wrote lavishdescriptionsof the richandfamousof theirday,andof fictionalheroesand heroines,in which they all but ignoredthe finerythat their subjectsmust have worn.Forthe medievalpoet, clotheswerean obstaclethatpreventedthe observer fromseeingwhatwasreallyimportantabouta person-his orherphysicalbeauty andthe specialmarkson the humanbodythatproclaimeda person'sworthiness, just as centuriesearlierit was thoughtthatthe extraordinarymarksor laksanaof the Buddhaproclaimedhis futuregreatnessas eithera worldemperoror a great sage.Whereclothingis mentionedbythe poet, it seemsas if its specialqualitieslie in the factthat it is virtuallyinvisible.Insteadof drawingattentionto the details of the cloth, its colorsand patterns,the poet praisesit becauseit seemsnot to be thereat all;the finestcloth is diaphanous,a shimmeringcurtainthroughwhich the beautyof the bodyis tantalizinglyrevealed. Clothis oftenlikenedto the elusiveraysthatareemittedby the gemsthe hero or heroinewears;indeed often the rays of these gems alone are deemedsufficient coveringfor the body.Forexample,the eleventh-centurySanskritpoet, Sri Harsa,in his Naisadhiyacarita, perhapsthe most famousof all medievalSanskrit poems, paid almostno attentionto whathis heroinewas wearingas she entered the hall wherethe kings wereassembledon the occasionof her self-choice,the 151
ritualby which she would selecther husband.In one versethe poet tells us that the flawless radiance of her jewels served her as an upper garment (10.94), while
in anotherwe are told that the rays of the gems she wore veiled her from the eagereyesof the kingsin the assembly(io.ioo).Manyof the versesin the chapter describeherjewels;none describesher clothing.In the versesin whichthe kings praiseher beauty,it is of coursethe beautyof her body that they single out and not her clothes.,It may not be an exaggerationto say that in medievalIndian poetry it is in factjewelsand the marksof the body itself that serveas the outwardsigns of a person'sspecialstatus,a rolethat clothesmorefrequentlyplayed in medievalEurope.2 That Sri Harsawas adheringto well-establishedpoetic conventionis clear fromevena cursorysurveyof Sanskritliterature.Bana,in his Kadambari,a prose taleof the seventhcentury,similarlydescribesthe heroineKadambariin termsof the auspiciousmarkson herbody,the radianceof herjewels,andthe beautyof her nakedlimbs. Indeed,thereis a sumptuousdescriptionof her body,her feet, her toes, thejewelson heranklets,the girdlearoundherhips,herwaist,hernavel,her breasts,herarms,herearrings,herhands,herchin, herlips and cheeks.Theonly mentionof clothis of the blueclothcoveringhercouch.3Littleattentionis paidto in the frametalewears;all we aretoldis thathis garmentis whatthe kingSuidraka whitelike the frothof the drinkof immortalitythatcamefromthe oceanwhenit waschurnedby the godsanddemonsandthatit hada borderon whichgeesewere paintedwith the pigmentgorocand,or yelloworpiment.Bycontrast,the outcaste womanwho comesto theking'scourtwearsa darkrobeandredscarf.Whatseems most importantaboutthe clothesthese characterswearis in facttheircolor;the fineclothesof the heroineandthe kingareluminousanddiaphanous;lowercastes wearcolorslike darkblueand red.4Thelowerthe status,the moredeeplycolored andobviouslyconcealingthe clothing. In his other famous prose tale, the Harsacarita, an account of the deeds of his patron King Harsa,Ban.atakes little note of what the king wears. For example,in a long and complexdescriptionof the king there is only one spare phrasethat describeshis lowergarment,while much is made of his jewelsand his body. As we might have expected,the king'slower garmentis describedas luminouslywhite and glisteningwith the reflectedrays from the jewels of his girdle.Whateverpatternit has is thus extrinsicto it, a reflectionof the rich and variegatedjewelssurroundingit. Indeed,it is like a jewel itself in its abilityto capturereflections.5 The emphasison jewelsand a lackof interestin clothingarecommonto texts of differentgenresandreligiousgroupsin medievalIndia.In a Jainaversionof the story of Nala and Damayanti,widelyknown from the Mahabharata, the medi152
PHYLLIS GRANOFF
evalpoet similarlydescribesDamayanti'sbody and herjewels,while her clothes arelikenedto merewisps of cloudsthatveil the beautyof the heavens.6The same conventionscouldalsogoverndescriptionsin Buddhisttexts.An earlyMahayana Buddhisttext,the Gandavyaha,in describingoneof thewomenwhomitshero,the pilgrimSudhana,meetson his searchforteachers,describesherjewelsatlength;it saysnothingof herclothing.7 Thesefewexamplescould easilybe multiplied.Whilethey aremeantto show that medievalpoetic conventionwas little interestedin describingthe clothing thatpeoplewore,they arenot meantby anymeansto implyeitherthatthe value of fine cloth was unrecognizedor that its existencewent entirelyunnoticedin the literatecirclesof medievalIndia.We know from othertexts that cloth could cloth is often amongthe ritualgifts to be given be highlyvalued;in the Pur@nas to Brahman.apriests,along with other obviouslyvaluablethings such as gold and cows. One of the most popularJainastories,a storyaboutthe wealthymerchantSalibhadra,has the merchantexcitethe jealousyof the king and his queen when he is ableto afforda fabulouslyexpensivecloth and proceedsto use it to wipe his feet.8A Buddhiststory told of the famous discipleof the Buddha,the monk Mahakasyapa,suggeststhat clothingwas not the only use for fine cloth in the worldof these stories.Mahakasyapain his past birth asks for something to wear and receivesthe finest cloth there is, a divine cloth from the heavenly wishing tree. But insteadof using it for clothing,he makesit into a canopyfor the Buddha.9 Suchfine cloth canopieswerein factmarkersof royalstatus.A tenth-century textaboutthe accoutrementsof kingshipascribedto the ParamarakingBhojahas muchto sayaboutclothobjectsas a necessarypartof the ceremonialsignsof kingship.Clothis an importantcomponentof the royalstandardand the royalparasol,forexample,andKingBhojaemphasizesthatthe particularcolorof clothused in royalobjectsannouncesthe statusof a personand has the powerto procure certain magicalbenefits.10Well beforeBhoja,the sixth-centuryauthorVaraha had describedvariouskinds of pennantsand their Mihira,in his Brhatsamhita, colors;specificcolorsarelinkedto specificbenefitsand misfortunes,with white leadingto victoryandyellowto illness." Storyliteraturealso took note of someaspectsof the processof preparingtexfamiliarto all studentsof Sanskritbecause tiles.An earlystoryin the Panicatantra, of its inclusionin the Lanmanreader,told of an indigo dyerwho left his vat of indigounattended.A jackalfallsinto it andendsup dyedblue.He is so impressed with his new colorthathe sets himselfup as king of the forest,only to come to a of the eleventh-centuryKashmiriauthor disastrousend!In theAvadanakalpalata Ksemendra,a Buddhistmonk is dyeinghis monasticrobe red, but becauseof 153
LUXURY GOODS AND INTELLECTUAL
HISTORY
his previousbad karmathe cloth turns into meat and the dye into blood. He is accusedof murderinga cow.'2A Jainastoryfromthe fourteenth-centurysuggests that behind the Buddhiststorymaylie a deep suspicionof the craftof the dyer. Thestoryis briefandI giveit herein full: Therewasa certainmonkof the rankof masterwhowasextremelylearned. Throughthe forceof his karmahe camedownwith leprosy.Whenhe saw that his diseasecould not be curedby variousmedicines,he made a pilgrimageto SrIserisaka and havingrenouncedall solid food he satdown in frontof the god. He did this for sevendays.Whenthat was of no availhe next renounceddrinkingas well. Now some of the superintendinggods of the main deityhadbeen awayandwhenthey got back,the one who had remainedbehind askedthem, "Whydid you take so long in the distant worldof Mahavideha?" Theyreplied,"Thedaughterof the perfumemerchantBhjma,whowasto be the wifeof the ministerTejahpala,hadreached marriageable age.Butsherefusedto be marriedandinsteadbecamea nun. TheJinaSimandhara,wholivesandpreachesin Mahavideha,himselfinitiatedherinto the order.Herfatherspentall thathe hadplannedto spendfor her weddingon the ceremonyof her becominga nun. We tarriedso long therebecausewe wantedto see the greatfestivitiesat the time of her consecration."And then they turnedto the leprousteacherand said, "Seven births ago you were a cloth-dyer.Youkilled seven babymongeesein an orchardone daywhentheyfell into yourvatof boilingdye.Becauseof that baddeed,in this birth,sevenbirthslater,you havebecomeill with leprosy. Youdo not havelong to live.Yourkarmais almostexhausted.If you want, we can cureyou. Butin the nextbirthyourkarmawouldreassertitselfand you wouldhaveto live throughits results."Whenhe heardthis, the monk took his leaveof the lay devoteesand in his leprousconditioncontinued justas he was.13 The craftof the dyerseemssomehowto be connectedto the deathof animals in the literaryimaginationof both BuddhistsandJainas.Artisansfigurerarelyin these stories;thesetwo storiesareunusualand intriguingglimpsesinto the attitudes towardsomeonewho dyes cloth,whetherprofessionaldyeras in the Jaina storyor the monkdyeinghis own robeas in the Buddhiststory. Thesediversetexts showus thatthereweremanywaysin whichcloth and its preparationfiguredin medievalliterature.Butthereis oneparticulargenreoftexts that I wouldlike to highlightin the restof this article.If we aresurprisedby the observationthatpoetsseemindifferentto the marvelousaestheticqualitiesof the 154
PHYLLIS GRANOFF
textileartist'swork,it canonlybe of stillgreatersurpriseto learnthatphilosophers wereintriguedbyaparticularcategoryof cloth,namelycloththatwaseitherwoven fromthreadsof morethan one coloror cloth that hadbeen dyedwith designsof manycolors.India'sphilosophers,who in theircommentariesto the foundational worksof theirdifferentschoolsendeavoredto understandthe natureof realityand ourperceptionsof reality,foundin the existenceof clothwith morethanone color somethingextraordinary,somethingthat defiednormalexpectationsof the way thingsoughtto be. Committedto the presuppositionthattheirtheoriesaboutthe natureof our worldmust coincidewith whatwe actuallyobserve,one particular groupof philosophersfoundin the veryexistenceof such cloth of manycolorsa challengeto any simpleunderstandingof the world aroundus. The mere possibilityof weavinga clothof differentcoloredthreadsto createa variegatedproduct or dyeingpatternsonto finishedclothto createcitrapatas, or chintz,gavethese Indianphilosophersmuchto ponder.14 As one popularsaying,citedoften in such philosophicaldiscussionsofthenatureofvariegatedcloth,putit, "Thatacloth[ora color]shouldbeoneandmulticolored[many](citra)atthesametime,isthegreatest marvelofall (citratara)."Thesaying,hardto capturein English,is aplayon theword citra,whichcanmean"many," and"marvelous" "multicolored," or"fantastic." One clothcannot,afterall,beboth"one"and"notone"or"many"simultaneously. The argumentin whatfollowsis that the existenceof variegatedcloth of differenttypes challengedthe earlymedievalIndianunderstandingof the world.I wouldlike to suggestfurtherthat in the case of at leastone philosopher,knowledgeof the differentkindsof patternedcloththe philosopherwastryingto explain mayhelpus to understandwhyhe developedhis somewhatidiosyncraticdoctrine. Here,I will proposethat,with Bhasarvajna in the tenthcentury,we havea casein whichthe technicalvirtuosityand the varietyof textileshe could see beforehim led the philosopherto searchfor a generaltheoryof howwe perceivethe colorof variegatedclothsthatcouldaccountforthe manydifferenttypes of multicolored cloth he knew.His contemplationof multicoloredcloth in turn determinedthe somewhatunusualformthathis theorywasto takeand indirectlyinfluencedthe courseof subsequentphilosophicaldiscussions. In what followsI look brieflyat severalphilosophersand their responsesto patternedcloth.I concentrateon one schoolof philosophers,those of the NyayaVaisesika,or IndianRealists,who analyzedthe worldand its complexobjectsby breakingeverythingdown into a numberof constituents.Our concernwill be with their categoriesof substanceand quality,for these are what the existence of cloths of many colorschallenged.Colorwas regardedby the Nyaya-Vaisesika as a quality.All qualitiesexist in substancesby meansof a separaterelationship that the Nyaya-Vaisesika called inherence(samavaya). In addition,substances 155
LUXURY GOODS AND INTELLECTUAL
HISTORY
weresaidto be constitutedfrompartsthatweredifferentfromthe substanceitself, whichwas consideredto be a "whole."The whole restedon its partsby meansof the relationshipof inherence.It couldbe in contactwith otherobjectsbymeansof anothercategoryof relationshipcalledcontactor samyoga. In addition,the color of thewholewasthoughtto havebeencausedbythe colorof the parts.Toputthese conceptsback into our discussionof cloth, an ordinarycloth, say a blue cloth, wouldbe analyzedas a wholethatresidedin its parts,the bluethreads.The color of the cloth, "blue,"is causedby the colorof the threads.Therewasanothertenet of the Nyaya-Vaisesika that we need to consider.The Nyaya-Vaisesika described in some detailhow largeobjects-wholes-are built up from their constituent parts,frominvisibleatomsto units of two atoms,to visiblewholes.It was a given in this systemthat the partsthat combineto yield wholes must be of the same kind. It is possible,I suspect,to intuit the challengethat India'smarvelouslypatterned or woven cloths providedto this ontology.These problemsdid not go unnoticedby the primeopponentsof the Nyaya-Vaisesika, the Buddhists,and it is in this arenaof debatethat patternedcloth assumedits considerableimportance in the medievalintellectualworld.Technologicaladvancesin textile art challengedthe oldworldview,not onlybythe changeswroughtin societythrough trade,increasedwealth,and foreigncontact.An equallystartlingchallengewas one thathasyet to be noticedin the historyof the medievalIndiantextiletrade:it wasthe challengeto receivedunderstandingsof howthe worldis constructedand how we as perceivingsubjectscan know thatworld.In the end, considerationof cloththathadpatternsappliedto it or waswovenof manycoloredthreadswould forcethe Nyaya-Vaisesika to abandonsome of their basictenets, a processthat maywellhaveunderminedtheirabilityto withstandthe onslaughtsof theirBuddhistcritics.As an exampleof the effectsluxuryitemscouldhaveon suchrarefied fieldsas philosophicalspeculation,the caseof multicoloredclothraisesthe larger questionof the potentialthat new advancesin technologyand the availabilityof new goodsthroughtradehadfundamentallyto alteror enricholdwaysof seeing. As a starttowardunderstandingthis largerquestion,I focus hereon the case of multicoloredcloth, for which the writtendocumentationis ample.In what follows I brieflyreviewa few examplesof Nyaya-Vaisesika speculationon multicoloredcloth. Variegated Cloth and the Indian Philosopher A careful reading of select passages in Nyaya-Vaisesika writings indicates that
by the tenth centurythe philosopherswere trying to understandthe natureof two types of multicolored cloth: cloth woven of threads of different colors and 156
PHYLLIS GRANOFF
clothwovenof threadsof a singlecolorto whicha patternwasappliedafterweaving. Thetexts giveus little or no indicationof the processby whichpatternswere addedto wholecloth,butthereis a briefstatementin one text of the tenthcentury suggestingthat the authorknew of a processsimilarto that in which mordants wereappliedto the clothandthe clothwasthenputinto a bathof dye.Theremark, madein the Nyayabhuisana of Bhasarvajnia, occursin a contextthathasnothingto do with cloth or variegatedcloth.'5The contextis the proofof the existenceof the mind as a senseorganin additionto the fiveorgansof sight,smell,taste,hearing, andtouch. The standardNyaya-Vaisesika proofforthe existenceof the mind is an inference made from the observationthat our perceptionsare alwaysin sequence, one at a time. Perceptionis thoughtto occurthrougha chainof connections.For example,in the caseof sight,the eyeis in contactwithits objectandin contactwith the soul, the locus in whichknowledgewill arise.The philosopherobservedthat in factour senseorgansmaybe in contactwith a multitudeof stimuli at anyone point in time;he thereforepostulatedthe existenceof an additionalcausalfactor to explainwhyat anygivenmomentwe see one thing and not another,or we only see somethingbut do not also tasteit at the sameinstantin time. This additional causetheycalledthe mind,andtheyarguedthatit mustbe atomicin sizeandthus capableof beingconnectedto only one objectat anygivenmoment.The causeof perceptionis thus the contactof the object-eye-mind-soul.TheNyayabhuisana is uniqueamongNyayaand Vaisesikatexts in the wayit presentsits argumentfor the existenceof the mind. Hereis its inferenceto provethe existenceof the mind: "Thesenseorgansandtheirobjectsdependupon anotherfactorwith whichthey come into contactat a particularmoment,becauseeven when they are all connectedwith eachotherthey still only producetheirproductsat a givenmoment, justlikethe coloringagentsthatproducepatternson cloth."'6 Theinferenceis clear:the senseorgansandtheirobjectsarein contactvirtually all the time whenwe areawake;however,we do not perceiveeverythingat every moment.Weperceiveonlycertainthingsandthatoccurswhenwe arepayingparticularattentionto them.Staringinto spaceabsentmindedly, we do not noticethe personin frontof us. It is not sufficientfor thereto be contactbetweenthe sense organand the object;the mind also has to be engaged,and its engagementis not guaranteed.We cannotbe "absentminded." This is whatit meansto saythat the senseorgansandobjectsdependon the existenceof a factorthatis connectedwith them only occasionally.Everyinferencerequiresa valid example,in which the reasongivenforwhatyou wish to proveand the factorthatyou wish to proveare knownto exist together.Thecoloringagentsarethe examplein this inference.In Indianlogic, an examplethus showsthe presenceof both the inferentialmark, 157
LUXURY GOODS AND INTELLECTUAL
HISTORY
which we may paraphraseas "producingproductsfrom time to time although they arecontinuouslyin contactwith eachother,"andthe attributeto be proved, in this case"dependinguponanotherfactorwithwhichtheycomeinto contactat a particularmoment."Thuswe aretold that the coloringagents,varnaka,which aresaid to producea pattern(citra)that is addedto a cloth (utksipta),do not act to producethatpatternas soon as theyareappliedto the cloth;they dependupon contactwith yet anotherfactor,whichis not alwaysthere.I understandthis statement to mean that the varnakais firstappliedto the cloth and then the combination of cloth-varnaka(like the combinationof sense organ-object)must be broughtinto contactwith anotherfactor,whichcouldbe the dyebath.Onlythen is the patternproduced.This processis also reminiscentof the use of mordants to applythe designandthe subsequentimmersionof the clothin a dyebath.'7The mordantcould also be analogousto the mind;it wouldbe thatadditionalsomethingthatmustbe presentforthe processof dyeingto succeed,justas "presenceof mind"is requiredforperceptionto occur. It was in the tenthcentury,the time of the Nyayabhuisana in whichthis inference is found,thatNyaya-Vaisesika authorsseemto havebeen preoccupiedwith the implicationsfortheirphilosophyof the existenceof richlypatternedandmulticoloredcloth.,8We shall see that these two types of cloths-the cloth woven of threadsof differentcolorsand the cloth wovenof threadsof a single colorto whicha patternwasappliedafterweaving-raised somewhatdifferentproblems for the philosophers.In both cases,of course,therewas the question,whatis the colorof the finalproduct?It was generallyagreed,at leastin the caseof the cloth madeof threadsof differentcolors,thatthe colorof the clothcouldnot be simply blueor redor anysinglecolor.Althoughthey differedon its exactnature,NyayaVaisesikaauthorstendedto agreethattherehadto be a separatecolor,whichthey calledcitra,"variegatedcolor."In addition,the case of the cloth madeof threads of differentcolorschallengedthe rule that things can only be made from componentsthat are alike;some Nyaya-Vaisesika philosopherswould ultimatelybe forcedto rejector greatlymodify this key rule. The case of patternedcloth was evenmorecomplicated;the patterndoesnot coverthe cloth,andproblemswould arise in trying to understandhow a single cloth could be both coloredand not colored.In answerto the questionNyaya-Vaisesika philosopherswouldoffertheir particularunderstandingof what actuallyhappenedto a cloth in the processof dyeingor paintingit and how we perceivedthe final result,whetherour perception of the clothas havingthe patternis in factan erroror a trueperception.The answersto thesequestionswouldhaveramificationsboth forthe developmentof their own philosophicalsystemand fortheirdebateswith theirarch-opponents, the Buddhists. 158
PHYLLIS GRANOFF
Severalof the discussionsof variegatedcloth can be found in commentaries to the sixth-centurycommentaryof Prasastapadaon the Vaisesikasitras.'9The tenth-centurycommentaryby Sridhara,the Nyayakandali,has an extendeddiscussionon the natureof the color of cloth woven of threadsof differentcolors. The contextof the discussionis the natureof the substanceearth.Substancesare dividedinto six categories.The categoryearth is distinguishedfrom other substancesin that all individualsbelongingto that categorypossessthe universal, "earthness."All individualinstances of "earth"also share certain featuresin common;one of themis thattheyall possesscolor.Prasastapada hadsaidthatthe earthhas manycolors.This leadsSrIdharainto a lengthydiscussionof whathis predecessormighthavemeantbysucha statement.He commentsthatit canmean both thatthe manydifferentinstancesof earthtakentogethercanbe saidto have manydifferentcolors(onecoloreach)andthata singleinstanceof earthmayhave manycolorsif it is madeup of partsof differentcolors.Thislaststatementin turn occasionsfurtherdiscussion,since it is a clearviolationof the rulethatproducts can onlybe producedbypartsthataresimilar.Hereis Sridhara's explanation: Justas a wholeis producedfromits parts,so is the colorof the wholeto be producedfromthe colorof the parts.Now the colorof the partsin sucha case is not justwhite nor just black.It is black,white,yellow,and so on. It makesno sense to saythat only one of those colorsproducesthe color of the wholeandnot the other,forin everyothercasewe haveseenthatanyof thesecolorsin thepartsof somethingis perfectlycapableof producingcolor in the product.Noris it reasonableto saythatnoneof the colorsin the parts wouldproduceanycolorin the productsincetheywouldall be in conflict with eachother.Afterall we do see the whole,whichis many-colored,and a substancethathasno coloris notvisible.It is alsonot logicalto saythatin sucha caseit is the colorsof the partsalone,takentogether,thatareseenand understoodas "multicolored" andthat this accountsforthe perceptionof the whole.If you admitthatin this casethe colorof the partsis sufficientto accountforthe factthatthe wholecanbe perceived,thenyouwouldhaveto extendthatlogic to the perceptionof everywholeandyou wouldneverbe ableto saythatanywholehas its own color.Thereforeyou mustadmitthat togetherthe colorsof the partsproducethe colorof the whole.20 The argumentis clear.Partsthataredifferentin colorcan acttogetherto producea colorin the product;that coloris calledcitra,"variegated"> or "many-colored."Whatmaybe less clearis why it is unacceptablefor Sridharato allow the color of the partsto accountfor the visibilityof the whole. This positionwould 159
LUXURY GOODS AND INTELLECTUAL
HISTORY
weakenhis caseagainsthis Buddhistopponent.Muchof the debateaboutcolored cloth needsto be evaluatedin the light of the Buddhist/Nyaya-Vaisesika debate. The Buddhistwasvehementlyopposedto the Nyaya-Vaisesika ontology,particularlyto its notionof wholes.Insofarasthe Buddhistbelievedin objectsthatexistin the outsideworld,apartfromthe awarenessof the perceivingsubject,he believed only in the existenceof the "parts"; the notionof a wholewasjustthat- a notion, an abstractionmadeon the basisof our perceptionsof rawsense datalike color, taste,etc.21Toallowthatour perceptionof the colorof the wholeis not dueto the existenceof the colorof the wholebut is the resultof the colorof the partswould underminethe entire structureof Nyaya-Vaisesika ontologyby admittingthat whatwe attributeto the wholeis not thereat all;by extension,then,all of ourperceptionsof attributesof the wholecouldbe saidto haveas theirobjectsthe parts. It wouldbe difficultindeedin such a case to arguethat wholesexist. The problems posed by variegatedcloth are thus not trivial;it was deemedmoreprudent to abandonthe principlethatproductsmustbe producedby partsof a like nature andthusallowthatthreadsof differentcolorsgaveriseto the color"multicolored" in thewhole,ratherthanto allowthatthe colorof the partswasresponsibleforour perceptionof colorin the whole.ButSridharain factendedup sayingalmostjust this, aswe shallnowsee. Sridharacontinueswith anotherproblem,this one raisedby the existenceof a plainborder(orside)to our cloth.An objectornow saysthatif you admitthatthe colorof the clothis "many-colored" andthatit is one,thenit shouldbevisibleeven on thoseplacesof theclothwherewein factonlyseetheplainclothitself.Hereis his reply:"No,you arewrong,becausewe knowfromexperiencethatseeingtogether the variouscolorsof the partsis also a causefor seeingthe color'many-colored.' Andwherethis doesnot happen,we do not seethe color'many-colored."'22 Sridhara'sanswermakesthe perceptionof the color of the partsthe causeof ourperceptionof the colorof thewhole.Itentailsa numberof problems-for one, the assumptionthat the colorof the cloth is presentin its plainborderalthough we cannotsee it there.Thisproblemcouldleadto anynumberof absurdities;one couldalsoassertthata rabbitis presentbutwe simplydo not seeit. Particularlyfor the Nyaya-Vaisesika, in whichseeingis believing,to assertthe existenceof somethingwe do not see is a riskystep.Sridharahasalso nowassertedthatthe perception ofthe colorofthe partsis requiredforthe perceptionofthe colorofthe whole. An objectorraisesthe obviouspoint thatin sucha caseone wouldneverreallybe ableto see the colorcitra; at some point the partsof the threadsaretiny diatoms. Wecan neverseethe colorof theirparts,sincethe colorof atomsis invisible.If the colorof atomsis invisibleandthe colorof diatomsalsonowinvisible,the sequence of invisiblecolorswould continuerightup to the whole!But the worstproblem i6o
PHYLLIS GRANOFF
is left unstated.It is not difficultto see thata Buddhistmightwelljumpin at this pointandaskwhatnecessity(orindeedproof)couldtherethenbe ofthe existence of the colorcitrain the wholeat all?Whynot just saythatthe colorof the partsis sufficient,since perceptionof the colorof the partsis now requiredfor us to see the wholeandits color.And it is not sucha big stepfromthereto assertthatthere would also be no proof for the existenceof the whole;whatwe see are the parts and their color.Indeed,this is what the most seriousBuddhistcritic of NyayaVaisesikatheories,the seventh-centuryphilosopherDharmakirtidid say in his Pramanavdrttika(2.202).23
Ultimatelythe sheervisual complexityand technicalvirtuositydisplayedby the clothin his examplehasdefeatedSridharaby raisinga numberof problems.It wasnot enoughthatthe clothwascomposedof differentcolorthreads;it also had a borderof a singlecolor.Theissueof a cloththatdoes not havea consistentcolor throughoutwouldcontinueto plaguethe medievalIndianphilosopher. Thegenerallinesof argumentusedby Sridharawerefollowedin the latertexts. Thusthe sixteenth-centuryauthorSamkaraMisraarguedthat differentcolored threadsactedtogetherto producethe colorcitrain variegatedcloth. He triedto maintainthe principlethat likenessof parts is necessaryfor the productionof the productand said that it was sufficientfor the threadsto sharethe attribute "havingsome color"in orderfor them to havethe requiredsimilarity.It was not necessaryfor them to be so similarthat they sharedthe intermediateuniversals "blueness"or "redness."24 Anothercontextin whichthediscussionaboutvariegatedclothoccurredwasin the contextof the Buddhistattackon the notionof wholes.Wehavealreadyanticipated some of the discussion.Dharmakirtihad arguedin his Pramanavarttika that the notion of a whole was fraughtwith contradictions.As one examplehe citesa cloththatis both dyedandnot dyedin differentplaces(2.85-86).Thestandardreplyto this argumentafterDharmakirtiis interesting.Sridhara,it is to be remembered,consideredin his discussionof color reallyonly one type of cloth, clothwovenof threadsof differentcolorswith a plainborder.He arguedthatthe clothhasa singlecolor,butthatthe coloris not alwaysseen,sinceit requiresforits perceptionthe perceptionof the manycolorsof the parts.Wehaveseen thatthis explanationwaspotentiallydangerousforaNyaya-Vaisesika philosopherto make. A similarconundrumarisesin the caseof clothto whicha patternis appliedafter weaving.On suchclothstherecould clearlybe distinguishedthe patternand the backgroundofthe cloth.What,then,wasthe colorofthewhole?Thiswasa distinct problemfor philosopherswho believedthat thereis such a thing as a wholeand thatit shouldhaveits own color.Theanswerlayin an assumptionthatan applied patternwasnot to be regardedas the colorof the clothin the samewayin whicha 161
LUXURY GOODS AND INTELLECTUAL
HISTORY
blue cloth is saidto be blue.The patternwas regardedas an applique,something addedonto the cloth, which had its own color that had nothing to do with the pattern.Sridharaand Vacaspatimisra both consideredDharmakirti'sobjection andgavejustthis answer.25 Thisline wasalso followedin the Nyayabhusana: "The wholeis not saidto be dyedor not dyedon the basisof whetherits partsaredyed or not dyed,forthewholeis an entitydifferentfromitsparts.Tosaythatthewhole is dyedjust meansthat it is in contactwith a substancethatbearsthe dye;in the absenceof that contactthe cloth is saidto be not dyed.Now contactis a category thatdoesnot existeverywhereon a substance,andso thereis no contradictionfor a singleclothto be both dyedandnot dyed."26 The answeris telling.Applyinga patternto a cloth does not producein the clotha newcolor,calledcitra.Color,as a quality,existsin its locusbymeansof the specialrelationshipsamavaya,or inherence,andit wouldindeedexisteverywhere on the locus. If a patternedcloth has one color,citra,we would haveSridhara's dilemmawith the border,whichwasto explainwhythe colorcitrais perceivedin some placesof the cloth (wherethe patternexists)and not in otherplaces(where thereis only the background).The Nyaya-Vaisesika authorswould seem to have agreedthat in the case of a cloth onto which a patternis dyed,the dyeingprocess does not producea newcolorcalledcitra.Whathappensis that certainparts of the cloth come into contactwith a substancethat has color,the dye-stuff(the termused, rakta,can meandyedin generalor red),while otherpartsof the cloth do not come into contactwith sucha substance.The contacthereis the relationwhich is the normalconnectionbetweendifferentsubstancesas ship sa.myoga, opposedto the relationshipof inherencebetweena substanceand its qualities. What is specialaboutsamyogais that it does not need to exist everywhereon a substance.Thestandardexampleis the monkeyin a tree;the monkeyis in contact with one branchof the tree, but not elsewhere.The use of this explanationfor cloth on whicha patternhasbeen dyedafterweavingallowsthe Nyaya-Vaisesika to explainwhythe patternis seenin someplacesof the clothandnot otherswithout resortingto a processwhich requiresthat the perceptionof the color of the partsis necessaryto the perceptionof the colorof the whole,a stepin eradicating the existenceof wholesthemselves.27 Sridharaand Vacaspatimisraboth considerthe case of the dyed cloth in their defense of the existenceof wholes and apart from the discussionof the colorof a cloth madefromdifferentlycoloredthreads.The distinctivenessof the is that it considersthese two types of cloth togetherin its efforts Nyayabhusania to developa generaltheory of how we perceivea cloth to be of variegatedcolor. in the Nyaiyabhusania directlyfollowshis discussionaboutthe dyed Bhasarvajnia cloth with a discussionof a secondtype of cloth,one in whichthe cloth is woven 162
PHYLLIS GRANOFF
from threadsthat areboth dyedand not dyed (or red and not red).Hereis what it says:
Well then, what is the color of the whole that has been made up of parts thatareboth dyedand not dyed?The colorof the wholewill be impossible to designateas any particularcolor, since the parts of differentcolors, being in conflictwith each other,and being the cause of the color of the whole, cannot togethergive rise to any one particularcolor. But since blue, yellow,and so on all sharethe common generalpropertyof being "somecolor,"they will give rise to "somecolor"in generalthat cannotbe statedto be blue, or red, or any particularshade.That this must be the case is clearfrom the fact that we can see the whole;thus it must haveits own color, since a substancewithout color is imperceptibleand to see one thing by virtue of the fact that somethingelse has colorwould result in generalconfusion of things. Thereforeit must be acknowledgedthat "color"in general,incapableof being specifiedas blue,red,or some other particularcolor has been producedin the whole. The appearanceof the cloth as of "variegatedcolor"is dueto the connectionwith partsof different hues,just as the appearanceof blue,etc., in a crystalis due to conjunction with things of these colors.28 This is one answerthat the text gives;the whole must havesome color,since we can see it, butit saysthatit is impossibleto designatethatcolormoreprecisely, sinceit cannotbe anyone of the singlecolorsof the parts.It thenproceedsto offer a secondexplanation29: "Hereis anotherexplanation.Thecolor'many-colored'is producedin thatcloth,forthatis whatwe perceiveby meansof a validperception thatis not contradictedby anyothermeansof knowledge."30 The authorof our text thus givestwo possibilitiesfora clothwovenof threads of differentcolors:eitherit hassomecolorthatcannotbe namedby anyone of the componentcolors,or it has one distinctivecolorof its own, whichis calledcitra, "variegated" or "many-colored." Bhasarvajnia thenproceedsto raiseandrefutean objectionto the verynotionthatsomethingcanbe one andyet perceivedas many, howeverthat perceptioncomesabout.As noted earlier,the objectionis basedto some extenton a playon the meaningof the wordcitra, whichcan indeedmean "many."To saythat the cloth is one and "citra" (many-colored)is to saythat the clothis bothone andmanyatthe sametime,an obviousself-contradiction. Tosay that the coloris one and citra involvesthe same contradiction.ForBhasarvajnia, the two optionsarenot thatfarapart;in the firstcase,the cloth cannotbe designatedas havingone particularcolor;in the secondcase,it is designatedwith the 163
LUXURY GOODS AND INTELLECTUAL
HISTORY
termcitra,which,ashe explains,meansliterallyof variouscolors:it meansthatthe one colorcitrahasthe universalsforblue,red,green,etc.,allatonce.Bhasarvajnia's replyto the objectionis interesting.He returnsto the caseof the patternedcloth, whichhe nowusesas an example.He remindsthe objectorthatsucha clothis one objectand is indeedperceivedto be variegatedor of manycolors,since it is connected to many substances,themselvesof differentcolors (the applieddyes).In the presentcase,the colorof the clothcouldsimilarlybe perceivedto be variegated since it is the locus of the differentuniversals,"blueness,""redness,"and so on (all of which inherein the colorcitra, accordingto Bhasarvajna). As we shallsee whatis commonto both casesis the connectionof the one entity(clothor color)to manycolors/coloruniversals.The text continuesto arguethatthereis absolutely no contradictionin the factthatthesemanycoloruniversalscoexistin the single locus, the color citra.
If you shouldsay,"Athing cannotbe both one andcitra, 'many';this is blawe replythatit is perfectlypossibleforone color tantlyself-contradictory," to be perceivedas citra, thatis, as "many-colored," beingas it is the locus of multipleuniversalssuchas "blueness"and "redness," etc. Wehavealready seen how one piece of cloth, being in contactwith differentdye stuffs, is itselfperceivedas of variegatedcolor.Shouldyou arguethatthe universals "blueness,""redness,"etc., aremutuallyin conflictand thus cannotexist in one locus at one time, I replythatit is not provedthattheyarein conflict with eachother.And shouldyou sayto this, "Indeedit is, becausethey are neverin fact seen to occurtogetherin any othercase,"then I would reply thatthe samelogicwouldprovethat"blueness"and"lotusness"arein conflictwith eachother.And shouldyou then saythereis no conflictbetween themsincetheyarefoundtogether,well,myanswerto you aboutthe caseof bluenessandrednessis thatthe samethingappliesto themaswell.Theyare foundtogetherpreciselyin the case of the colorwe arecalling"many-colored."Reasoning,too, establishesthatthe colorof the whole,causedas it is by colorsin the partsthatarecharacterizedby differentuniversals,should indeedbe the locus of morethan one universal,forwe observethe general rulethatthe colorsof the partsgiverisein the wholeto colorspossessedof the sameuniversal.31 finallyconcludesby sayingexplicitlythat the color citra is one Bhasarvajnia colorandthatit is the locusof manyuniversals,suchas "blueness," "redness," and so on.32I wouldliketo emphasizethe roleplayedin this argumentby the example of the cloth to whicha patternhasbeen applied:this givesBhasarvajnia an exam164
PHYLLIS GRANOFF
ple,to whichallpartiesmayagree,of oneentitythatthroughits contactwithmany colors/coloruniversalsappearsas variegated,or many-colored,citra.In the case of the dyedcloth,the cloth is in contactwith the applieddyesand throughthem It is this relationshipthat allowsthe one with their colorsand their universals.33 cloth to appearas citra,or many-colored.Similarly,he arguesin whatI havejust translatedabovethatin the caseof the colorof a clothwovenof threadsof different colors,the colorof the cloth,calledcitra,is one andappearsas manybecauseit andso "redness," is in contactwiththe manydifferentcoloruniversals"blueness," colorcitra.Theclothon whicha pattern forth,all of whichinherein Bhasarvajnia's with a paradigmin which if one entity has been dyedhas providedBhasarvajna is in contactwith many colors (and throughit with the universalsthat belong to those colors)it can appearas many-colored.This paradigmmayhavehelped to formulatehis uniquenotion of the colorcitraas itself something Bhasarvajfia multiple,havingmanydifferentuniversals,ratherthan as a singlecolorwith its own single,universal"many-coloredness."34 what To summarize,in my understandingof this passage,for Bhasarvajfia ensuredthatthe one colorcitracouldbe perceivedas manywaspreciselyits connectionto themanydifferentuniversalsthathearguedinheredin it.Thepatterned clothhadgivenhim a specificcasefromwhichhe couldderivea moregeneralrule, namelythatthe perceptionof many-colorednessin a unitaryentityis occasioned by the connectionof thatunitaryentityto multiplecolorsor universalsof colors. commitmentto realismseemsto haverequiredthatthe perception Bhasarvajfna's of manifoldnesscome from manifoldentities;the exampleof cloth on which a patternhadbeen appliedby dyeingafterthe weavingprocessgavehim the clueto understandhowthis couldbe possible. It is intriguingto ponder the significanceof the fact that, of our authors, seemsto havebeen the only one to integratethe case of dyedcloths Bhasarvajnia into the discussion of color. I noted above that Vacaspatimisraand Sridhara both consideredit separatelyin theirdiscussionsof Dharmakirti'sattackson the conceptof the whole. Forthose philosophers,the perceptionof Nyaya-Vaisesika the cloth that has been dyedafterweavingas a cloth of variouscolorsis false;it is an errormadewhen the perceiverattributesthe colorof the dye to the cloth.35 wantsto explainthe is thatBhasarvajnia Myunderstandingof the Nyayabhuisana perceptionsof many-colorednessin sucha wayas to includethe colorof the dyed clothaswellasthe clothmadeof threadsof differentcolors.Itmaybe recalledthat in his firstsuggestionaboutthe natureof the colorof the cloth wovenof threads of differentcolors-namely, that it is some color that cannot be specified-he says:"Theappearanceof the cloth as of 'variegatedcolor'is due to the connection with partsof differenthues,just as the appearanceof blue, etc., in a crystal 165
LUXURY GOODS AND INTELLECTUAL
HISTORY
is due to conjunctionwith things of these colors."36 In otherwords,the operative principleis the sameone thatwe havealreadyseen:the objectthatis to be seen as variegatedis in connectionwith thingsthatarevariegatedandthroughthemwith theirmanycolors/coloruniversals.Dyedcloth,whichappearsto us as variegated in color,is similarlyin conjunctionwith manydye-stuffsof differentcolorsand thus with manydifferentcolorsand coloruniversals.If we considerthe colorof the clothwovenof threadsof differentcolorsto be someunspecifiedcolor,thenwe cansaythatthe clothis in contactwith its parts,the threadsthataremanyin color, andthus in contactwith theirmanycolors/coloruniversals;we havejustseenthat if we considerthe colorof the clothto be a newcolorthatwe designateas citraand in whichvariousuniversalsof colorsreside,then the cloth that appearsto us as variegatedor many-coloredis in contactwith the unversalsof manycolorsthat residein its own color,citra,and throughthose universalswith the manycolors themselves.If I am correct,then,Bhasarvajina wouldhaveregardedthe perception of manifold-colorin all threecasesto bevalidanddeterminedbyadequatecauses. If my reconstructionof the argumentsis valid,Bhasarvajnia hastakenashis starting point the dyedcloth that his contemporariesdid not includein their discussions of color to build a comprehensivetheory of how we perceive"variegated color."Perhapsthis point tells us somethingof the culturein whichBhasarvajna worked,wheredyedclothswereso abundantandimportantthathewasprompted to considerthemfirst.Unfortunately, we lackthe informationthatwouldconfirm or refutesucha suggestion. In summary,the Nyayabhuisana seemsto be the text most consciousof the differentwaysin whichcloth maybe multicolored.In its argumentit differentiates clearlybetweencloth to whicha patternhasbeen appliedthat does not coverthe entireclothandcloththathasbeenwovenof threadsof differentcolors.Bothcases challengedthe Nyaya-Vaisesika philosopher,but othershad consideredthe cases as presentingtwo separatechallengesthat requiredtwo separateresponses.The cloth madefrom threadsof manycolorswas a challengeto the Nyaya-Vaisesika understandingof how the qualitiesof a whole aremadefrom the qualitiesof its parts.The dyed cloth was a challengeto the notion of the existenceof a whole as one unitaryentity;havinga color and its absenceis an exampleof mutually contradictoryqualities.Itwasagreedthatthe existenceof mutuallycontradicting attributeson a locus meantthat the locus in questioncouldAot be one thing. In otherwords,thereis no suchthingas a "whole"aboveandbeyondthe partsout of whichit is constructed. In this way,the existenceof cloth on which a patternhas been dyedgavethe Buddhistopponent a powerfulexamplein which the Nyaya-Vaisesika understandingof wholes could be shown to be contradictory,while the existenceof i66
PHYLLIS GRANOFF
clothwovenof threadsof differentcolorsunderminedthe Nyaya-Vaisesika analysis of howproductsandtheirqualitiescomeinto being,as a gradualbuildupfrom partsof similarnature,with the colorof the wholebeingcausedbythe colorof the parts.37 seemsto haveattemptedan analysisof the natureandmechaBhasarvajnia nism of our perceptionsof a clothas multicoloredthatwouldencompassthe differenttypes of cloth that he no doubtknew firsthand.It has been instructiveto contrasthis treatmentof the subjectwith that of his contemporaries.In all their in that they cases,the analysislacksthe richnessof that offeredby Bhasarvajnia, confine themselvesin their discussionof color to the one type of multicolored cloth, cloth wovenof threadsof differentcolors.38Nonetheless,it is clearthat for all of these medievalthinkersthe problemsthatvariegatedcloth posed for these philosophersweresubstantial. Conclusions In this articleI havetried to use the exampleof textile arts to illustratehow the creationsof artistscould havean unexpectedinfluencein widerintellectualcircles.Fromthe startthe earlyVaisesika,whichwaslaterto mergeinto ajointschool with the Nyaya,had taken into considerationman-madeobjectsin buildingits understandingof the world.Itsdiscussionof colorin earthsubstances,forexample, includedthe possibilitythat colorwas producedby heatin orderto account forthe changeof colorthatis seen to occurwhenpots arefired.It wouldseemto havemaintainedits interestin describinga worldthat could explainboth manmade and naturalobjects.By the sixth century,the NyayaauthorUddyotakara hadraisedthe issueof cloththatis variegatedin colorbecauseit is madeof threads of differentcolors.39 Shortlythereafterin his attackon the Nyayanotionof wholes, theBuddhistphilosopherDharmakirtiraisedtheissueof a cloththatis raktarakta, both redand not red,or both dyedandnot dyed.Wecannotbe sureexactlywhat he had in mind, but for the tenth-centuryphilosopherswho refutedhim, it was assumedthat Dharmakirtiwas referringto cloth that hadbeen dyedafterit was woven.Whilethe detailsof the debatesin the philosophicaltexts mightnot be of particularinterestto scholarsof Indiantextiles,the factthat they occursuggests thatboth patternedand variegatedtextileswerewidelyknownby the tenth century.If the poets ignoredthem,the philosopherscouldnot. Theseargumentstell us thatadvancesin makingobjectscouldhavefar-reachingeffects. In closing I would like to add that this was not the only context in which his knowledgeof cloth and the process of its manufacturehelped the NyayaVaisesikaphilosopherto formulatehis theories.In anotherBuddhisttext, the of Santaraksitawith the commentaryby the eighth-century Tattvasa.mgraha authorKamalasIla,a Nyayaauthoris strugglingto establishthatthe wholeandits 167
LUXURY GOODS AND INTELLECTUAL
HISTORY
partsaredifferentfromeachother.He formulatesan inferencethat relieson his knowledgeof how clothis made.Hereis the inference:"Theclothandthreadsare different,becausetheyhavedifferentcapabilitiesanddifferentmakers"(10.560). In his commentary,Kamalasilaexplainsin whatway cloth and threadshave differentcapabilitiesandmakers,exist at differenttimes andareof differentsize: "Womenarethe makersof the thread,whilethe maleweavermakesthe cloth.The cloth has the abilityto wardoff the cold, while the threadsdo not. In addition, threadsare seen beforethe cloth is made,while the cloth is only thereafterthe weaverhas donehis work.The clothhas a width andbreadththatarenot therein the threads;this constitutesa differencein size."40 Thereare no doubt many more examplesof how the medievalphilosopher lookedat the technologyof the objectsin his environmentas he soughtto formulatehis theoriesaboutthe natureof the physicaluniverse.Furtherresearchwill, I hope,addto theseremarkableexamples.
168
PHYLLIS GRANOFF
NOTES
All translationsareby the author. SriHarsa,NaisadhiyaCarita,ed. NarayanaRamaAcarya(Bombay: NirnayaSagaraPress,1952). 2. N. B. Harte and K. G. Ponting, Clothand Clothingin MedievalEurope:Essaysin MemoryofProfessorE. M. Carus-Wilson, PasoldStudiesin TextileHistory 2 (London:Heinemann Educational Books, PasoldResearchFund,1983).See also MalcolmVale, ThePrincelyCourt: MedievalCourtsand Culturein NorthWestEurope(Oxford:OxfordUniversity Press,2001); Raymondvan Uyten, "ShowingOff One'sRankin the Middle Ages,"in ShowingStatus:Representatives ofSocialPositionsin theLateMiddleAges, ed. Wim Blockmansand Antheun Janse (Tournhout,Belgium:Brepols,1999),1935. 3. Bana,KadambariofBcinabhatta,ed. KdsinathaPandurangParab(Bombay: NirnayaSagaraPress,1896), 358- 65. 4. Bana,Kadambari,17,22. The importance of color is clearnot only in a prosetext like the Kidambari;the color of a deity's garmentis alwaysstatedin iconographical texts, and a text to which I will return-the Yuktikalpataru of the tenthcenturyKingBhoja,even classifiesthe insects from which threadsarederivedby caste accordingto the color of the thread that they produce.The Brahmanainsects, at the top of the scale, producepurewhite thread.It should be noted that the word for "white"can also mean transparent. This is threadthat to the eye reallyis not even there. See Bhoja,Yuktikalpataru of Bhoja,ed. IsvaraChandraSastri (Calcutta:SiddehesvaraMachineYantra, 1917),8i. Other sections of the same text, for example,on canopies and pennants, also emphasizethe importance of color as an indicatorof status and conveyorof specificbenefits. 1.
169
LUXURY GOODS AND INTELLECTUAL
5. Bana,Harsacarita,ed. Kasinnatha PandurangParab(Bombay:Nirnaya SagaraPress,1897), 72. I havewritten on jewels and their rolein poetic descriptions in ThreeMountainsand Seven Rivers:ProfessorMusashiTachikawa's FelicitationVolume,ed. Shoun Hino and ToshiroWada(Delhi: Motilal Banarsidass, 2004), 331-355.
6. I havetranslatedthe story of Nala and Damayantifrom the twelfth-century Kumarapalapratibodha in TheForestof Thievesand theMagicGarden:An AnthologyofMedievalJainStories(Delhi: Penguin Books, 1998),177-224.
7. Gandavyuha,ed. P.L.Vaidya,Buddhist SanskritTexts5 (Darbhanga:Mithila Institute of Post-GraduateStudiesand Research in Sanskrit Learning, 1960),
chap.io. 8. The Salibhadra-Dhanna-Carita:The Tale of the Questfor UltimateReleaseby Salibhadraand Dhanna, A Workin Old Gujarati,ed. and trans. Ernest Bender (New Haven, Conn: American Oriental Society, 1992). 9. Theragdthatthakatha-, vol. 3, ed. F.L. Woodward(London:Pali TextSociety, 1974), 124.
10. Bhoja,Yuktikalpataru, e.g., chap.37.
il. Varahamihira'sBrhatasamhita, ed. and trans. M. RamakrishnaBhat(Delhi: Motilal Banarsidass,1981-82), chap.43, 26-27.
12.
Ksemendra,Avadanakalpalata, ed. P.L. Vaidya,BuddhistSanskritTexts22, 2 vols. (Darbhanga:Mithila Institute of PostGraduateStudiesand Researchin
Sanskrit Learning, 1959), 2:558. 13. Puratanaprabandhasamgraha, ed.
JinavijayaMuni, Singhi JainSeries2 (Calcutta: Singhi Jaina Jfiana Pitha, 1936), 114, par. 249. 14. I take the term "chintz"and much of my
HISTORY
understandingof the technical aspects of medieval Indian cloth from Mattiebelle
1
contention hereis that the Nydyabhuisana is acutelyawareof the two types of variegatedcloth:cloth woven of threads of differentcolors and the cloth woven of threadsof a single color to which a pattern was appliedafterweaving.I am arguing that this awarenessis what motivates Bhasarvajnia to formulatehis unique theory;he is trying to explain why in both caseswe perceivethe cloths to be variegatedin color.He is in searchof some generaltheory that will explain both cases. Bycontrast,Udayanaseems, if anything, to conflatethe two types of cloth or to ignorethe case of the cloth to which the patternhas been addedafter weaving.His theory-that there is a single color citrawith its own universal, citratva-is applicableonly to the cloth that has been woven of threadsof differentcolors. It leavesunexplained why we see both types of cloth as citra,or of many colors. Fora very readableoverviewof the Buddhist-Nyayadebate,see D. N. Shastri, CritiqueofIndianRealism(Agra:Agra
Gittinger,MasterDyersto the World: Techniqueand Tradein EarlyIndianDyed CottonTextiles(Washington,D.C.: TextileMuseum, 1982). Bhasarvajfia,Nyayabhuisana,ed. Svami Yogindrananda(Varanasi:Saddarsana Prakasana Pratishana, 1968), 440. Little is
known about Bhasarvajnia beyond the broadoutlines of his dates.The editor of the text comments that he is generally thought to havecome from Kashmir, although there is no positive proof. i6. Indriyarthadaya h kramasa.myogisahakaryantrapeksahsannihitanim api kramakaryotpadakatvad vastropaksiptacitrotpddakamivavarnakadayah. Bhasarvajfia,Nyayabhu-sana, 440. 17. lam basing my understandingof the processon Gittinger,MasterDyersto the World,19-23. It is possible that the author of the text intended some other similar process;in any case, it is clearthat he knows a sophisticatedprocessof producingpatternson cloth that requires severalsteps. It is interestingto note that has this inference; only the Nyayabhuisana of our texts it also has the most elaborate discussion of patternedand variegated cloth. Bhasarvajnia seems to havehad a particularinterestin patternedand variegatedcloth and some knowledgeof its production and to haveeagerlyused it in formulatinghis philosophical
University, 1964), 256-59, which gives a
summary of some of the argumentsabout variegatedcolor.The Nyayabhuisana had not yet been discoveredat the time he wrote his book. JohannesBronkhorsthas also written brieflyon the debateoverthe color citrain the Vaisesika."Studieson Bhartrhari,pt. 5:Bhartrhariand Vaisesika,"Proceedingsof the First InternationalConferenceon Bhartrhari, Universityof Poona,January6 -8,1992, AsiatischeStudien/ltudesAsiatiques47.1
discussions.
i8. Much of the materialthat I reviewhere has been discussed elsewhere.The most detailed discussion, including translations of the relevanttext passages,is in Otto Grohma,"Theoriezur BuntenFarbe im alterenNyayaund Vaises.ikabis die Udayana,"WienerZeitschriftfuir
(1993):75-94. 19.I am relyingfor my datesof texts on Karl
Kunde Sudasiens 19 (1975):147-82. While
Potter,TheEncyclopediaofIndian Philosophies,vol.1 (Delhi: Motilal
Grohmareviewsmany of the texts that I discuss in this article,he does not note that differenttypes of cloth figureas the objects of the differentrefutations.My
Prasastapadabha.yawith Sridhara's commentaryis edited by VindhyesavaraprasadaDvivedin,
Banarsidass, 1971).The
170
PHYLLIS GRANOFF
20.
21.
VizianagramSanskritSeries6 (Benares: E. J.Lazarus,1895), 30-32. Prasastapadabha.yawith Sridhara's commentary,30-32. Forthe Buddhistson perception,see MasaakiHattori,Dignagaon Perception (Cambridge,Mass.:HarvardUniversity Press, 1968).
Acitrepdrsvepatasyatadasrayasya citrarfipasyagrahanaprasangah tasyaikatvaditi cenna samadhigatasaanvyavavyatirekabhyam marthasyavayavandndruipadarsanasydpi citrarupagrahanahetutvat tasyaca pdrsvantare'bhavdt.Sridhara Nyayakandali,30. 23. Dhamakirti,Pramdnavarttika,ed. Svami DvarikadasShastri,BauddhaBharati Series3 (Varanasi:BauddhaBharati, 22.
1968), 161. 24.
SamkaraMisra, Vaisesikopaskara, ed. SriNarayanaMisra,KashiSanskritSeries 195(Varanasi:ChowkhambaSanskrit Series Office, 1969), 368.
Sridhara,Nyayakandali,42; Vacaspatimisra,Nydyav1rttikatutparyatika, ed. TaranathaNyayaTarkatirthaand AmrendramohanTarkatirtha(Calcutta: MunshiramManoharlal,1985), 474. 26. yad apyuktam:raktecaikasminnityddi. Tadapyayuktam,nahyavayavaraktataraktatabhyamavayavinastathdbhavastasya tebhyo'rth,antaratvad. Api tvavayavina eva raktadravyasamyyogo raktatai, tadabhavascairaktateti samnyogasya cavyapyavrttitvena raktataraktatvam apy ekasyavayavinovirudhyate.Bhasarvajnia, iii. Nyayabhuisana, 27. The same explanationis given by the tenth-centuryauthorVacaspatimisra, Tatpdryatikd,ed. TaranathaNyaya Tarkatirthaand Amrendramohan Tarkatirtha(Calcutta:Munshiram Manoharlal,1985), 473. 28.yas tarhiraktaraktairevavayavairavayavi ni.paditastasyakimp ru4pamiti? 25.
Visesanarambhdd anirdesyamevavisesato viruddhavisesandmvisesdrambhetu virodhona sdmdnyarambhe nilap7tadisu sarvatraruipatmanahsa,mbhavat nilidivisesarahitamapi rupamutpannam ityavavavidarsanatevagamyate niruipadravyasya darsanayogad anyaruipendnyadarsane catiprasanigdt. Tasmadvisesato'nirde.ya,mru4pamatram eva tatrotpannamiti citrapratibhdsastu tatracitravayavasambandhatsphatike ntldidipratibhdsavat. Bhasarvajna, Nyayabhuisana,112. The tenth-century authorVyomasiva, Bhasarvajnia's contemporary,has a lengthy section on citrain his commentaryto the Vaisesikasutras,the Vyomavati, ed. SrigaurinathaSastri(Varanasi: SampuirnanandaSanskrit
niyamendvayavinisamanajdtiyaraparambhakatvadarsaniditi. Bhasarvajnia, Nydyabhuisana, 113. 32. Bhasarvajfia, Nyayabhuisana, 113. This
view was explicitlyrefutedby Udayana, whose datesaregiven as A.D. 975-1050. Udyanaand Bhasarvajnia maywell have been contemporaries.See theAtmatattvaviveka,ed. PanditDhundirajaSastri (Benaras:ChowkhambaSanskritSeries Office,1940), 274. Udayanaarguesthat citra,like all the other colors, must have its own universal.It is clearthat the 29. existence of cloth woven from threadsof differentcolors challengesthe fundamental principlesof the Nyayasystem. Udayanahas abandonedthe principle that like gives rise to like, in having the colorsblue, red,and so on give rise to Visvavidyalaya, 1983), 63-66. He something entirelydifferentfrom them. combines the two possibilities:that there Bhasarvajfiahas preservedpart of that is color in generaland that there is a principlebut has allowedthat many specific color known as citrainto one. universalscan exist in the same substraFirst,he argues,one provesto a doubter tum. Therearealso other problemsin that the variegatedcloth must havesome both theories;in both cases consideration color since it is perceived.One then of these cloths has led to a re-evaluationof deducesthat this color must haveits own basicphilosophical principles. name, since it cannot be any of the colors 33. To the Nyayaphilosopherthese arereally we alreadyknow.Thatname is citra. the same thing; anything connected to a 30. Atha va citrameva tadruipamutpannam color is also indirectlyconnected to its tathaivabadhitapratibhdsat. Bhasarvajnia, universal. Nyayabhuisana, 113. 34. As did, for example,his contemporary 31. Ekamtaccitra,mcetivirodhadayuktamiti Udayanain theAtmatattvaviveka,274. cet, na;ekasyapyanekanilatvadidharmaid- 35. Vacaspatimisra,Tatpdryatika,474. hikaranatvenacitrapratibhasavisayatvas- 36. Bhasarvajfia,Nyayabhuisana, 112. ambhavat;yathagairikadyanekavarnasa- 37. It may also havestimulatedBuddhist mbaddhamvastra,mcitramitipratiyate. philosophersin their developmentof Nilatvadijatinamviruddhatvid their doctrine known as citradvaita ekatrasamavayonayukta iti cet, na; (variegatedmonism), in which convirodhasydsiddhatvdt. Anyatra sciousness is saidboth to be one and to sahadarsanattatsiddhiriti cet, na; havemany forms or appearancesas the nilatvotpalatvayorapi virodhaprasaingat. various externalobjects.Bythe same Tayorekatradarsaandd avirodhaiti cet, token the difficulties the Buddhistsfound samanametat. Yuktyapibhinnajattyaruipin the Nyayawillingness to attribute drabdhasyavayaviuipasyanekajatyadhika- variegatedqualitiesto any one thing ranatvamsambhavyate,avayavaruipa na,m could be used by the Jainasagainstthe 171
LUXURY GOODS AND INTELLECTUAL
Buddhists.On the centralityof the issue of citrafor Dharmakirti,see Grohma, "Theoriezur BuntenFarbe,"155,citing Tilmann Vetter,Erkenntnisprobleme bei Dharmaktrti(Vienna:UniversitatWien, 1964),66. ForJainarefutationsof the BuddhistsbyAkalanka,see the Siddhiviniscayatikd,ed. Mahendrakumar Jain,JfianapIthaMuirtideviJaina Granthamala22 (Kashi:Bharatiya Jfianapitha,1959),157.Dharmakirti himself could use the inabilityof the Nyaya-Vaises.ika satisfactorilyto explain how variegatedcloth can be perceivedto show that all philosophical investigation leadsto one point- the conclusion that the worldas we see it cannot be ultimately real.See the Pramdnavarttika3.2o8-9. The exampleof variegatedcloth was also used by the philosopherBhartrharito illustratehow wordscombine to make sentencemeanings. See Bronkhorst, "Studieson Bhartrhari."While these briefcomments arein no wayadequateto an understandingof the philosophical discussions in these texts, they should make clearthat how variegatedcloth was conceivedbecame a crucialstandpoint that could be used to undermine a realist understandingof the world. 38. This also includesVyomasiva,another tenth-centuryauthorwhom I havenot discussed in any detail and who also dealt with some of the problemsthat cloth of variegatedcolor presentedin his Vyomavati,65. 39. See Uddyotakara,Nydyavarttika,in the Nyayadarsanam,ed. TaranathaNyayaTarkatirtha(Repr.,Delhi: Munshiram Manoharial,1985),1053. 40. Santaraksita,Tattvasamgraha,ed. Swami DwarikadasShastri,BauddhaBharati Series1,2 vols. (Varanasi:Bauddha Bharati,1968),235.
HISTORY
CAROL BIER
PATTERNS IN TIME AND SPACE Technologies of Transferand the CulturalTransmission ofMathematicalKnowledgeacrosstheIndianOcean
I
Woodcut from GregorReisch, MargaritaPhilosophica(Freiburg, 1503).Arithmetica is shown instructing an algorist and an abacist,labeled Boethius and Pythagoras,respectively. FromClaireL. Parkinson,ed., A Chronologyof Breakthroughs: GreatAchievementsin Scienceand 1200-1930 (Boston: Mathematics,
G. K. Hall, 1985),figure i.
Abstract This articleexploresthe potentialroleof textilesin the transferof mathematical knowledgefrom the Indiansubcontinentto the centralIslamiclands and westwardto an emergingmodernEuropethroughan inquiryinto prospectivetechnologiesof textilemanufactureandpattern-making.Ikattextilesof the ninthand tenthcenturies,foundin Egyptbutpresumedto be fromYemen,serveas a means to explorepossibilitiesof numerationandtreatmentof the spatialdimension.An initialattemptis madeto separatepatterningfromthe technologyof textileproductionin an effortto treatthe mathematicalpossibilitiesthatpatterningoffers forthe applicationof mathematicalknowledge.This articleproposesan ontology of pattern,distinct from the categoryof a textile itself,which raisessignificant questionspertainingto the transmissionof mathematicalknowledgein relation to expandedtraderoutesin the eighththroughtenth centuries,coincidentwith Islamicdevelopmentsin the understandingof two-dimensionalspace.
IN 1202 A MAN BY THE NAME of LeonardoPisano, later known as Fibonacci, publisheda work that introducedthe Hindu numeralsto his countrymen.' Leonardoof Pisa had grown up in North Africa in the town of Bugia, today in Algeria,where his fatherwas engagedin mercantileactivities.Bugia was a Pisan tradecolony in an Arabic-speakingenvironment.2Leonardoin his youthwas regularlyexposedto methodsof calculationusingthe numeralsi to 9, plus o as a placeholder,which werecommonplaceamongthe Arabswith whom he grew up along the southernshoresof the MediterraneanSea. In countries to the north, the practicewas to use fingercalculationat the time his book was written.3By using the Hindu numerals,he carefullyexplained,one could perform arithmeticcalculationsin a mannerpreviouslyunknown in Italy.Hindu numeralshad enteredthe Islamicrealmseveralhundredyearsearlier,whenthey causedan immediatestir amongthe intellectualelite, precipitatingthe writing of mathematicalworksin Arabicthatexplainedthe new prospectsofferedby the methodsof "Hindureckoning."4 Booksandoraltransmissionrepresenttwo technologiesbywhichmathematical knowledgewas transferredacrosstime and space.Lookingat variouscategoriesof evidence,this articleexaminesanotherpotentialtechnologyof transferby exploringthe rolethattextilesmayhaveplayedin the transmissionof mathematical knowledgefromthe Indiansubcontinentto the centralIslamiclands-Iraq, Iran, Syria,Egypt, and North Africa-and westwardto an emergingmodern Europeas earlyas the thirteenthcentury,a transferthat occurredin both directions severaltimes. 173
Textilesaregenerallytreatedas physicalobjects,visualworkscreatedin specifictimesandplacesto servea multiplicityof practicalfunctions,fromproviding shelterandwarmthto expressionsof groupandindividualidentitywith respectto class,rank,status,ethnicity,andreligiousbeliefs.,As physicalobjects,textilesare usuallyconsideredas integralanddiscreteunits,althoughthey areoftenmadein multiplesdrawingon technologiesthat facilitaterepetitionof formsand designs to createpatterns.In manycultures,patternsthemselvesoccurin multiplicities, with manypatternsoften presentin anygiventextile.In general,formandstructure, design and composition,are treatedas integralcomponentsof the whole object.While understandinghow a textilewas maderequiresan externalfamiliaritywith technologicalprocessesand often dependsupon interpretationbased on diagnosticstructuralfeatures,techniqueis not inherentin a textilein the mannerthatstructureinheres.6In this study,an initialattemptis madeto separatethe patterningprocessfromthe technologyof textileproductionin an effortto treat the possibilitiesthat patterningoffersfor the understandingand applicationof mathematicalknowledge. TheTransmissionof MathematicalKnowledge Theworkof Leonardoof Pisaconveyedthe capabilityof calculatingusinga base10 set of numeralsby meansof a writtentext. The text itselfreflecteddirectlythe firsthandexperienceof a man who had lived and traveledextensivelythrough Arabic-speakinglands. The fingercalculationmethod, advancedby Leonardo, enabledquick computationsof relativelysmall sums but was problematicfor anythinglarge.In an emergentEuropeof the thirteenthcentury,computation usingRomannumeralsdid not lend itselfto long divisionor rapidmultiplication exceptby fingercounting,with its attendantlimitations.The mentalgymnastics requiredfor calculationrelyingupon numbersdesignatedby lettersof the alphabet, whichwerethen commonboth in the Greektraditionand in contemporary Arabicusage,also did not lend itselfto the manipulationof largenumbers.The introductionof the nine-digitsystemplus zero as a placeholdermadecomputation accessibleto all who knew just ten numbers,to sums as largeas the mind could comprehendand beyondto infinity.7The initial encounterin the Islamic worldwith the nine digitsof the Hindusystemhadalreadygeneratedconsiderable scholarlyexcitementfor the potentialthatwas offeredin reckoningand calculation.8An earlysixteenth-centurywoodcut(fig.i) visuallyrecordsthe competitive advantageofferedby the new method of calculation,which eventuallytransformedmethodsof mathematicalthinkingin earlymodernEurope.9 Thisprefatoryconsiderationof whattodayin schoolsin theWestarestillcalled Arabicnumeralsmight seem simplisticand prosaicsincewe takethem so much 174
CAROL BIER
for granted.Butthe transmissionof this mathematicalknowledgefromIndiato Europehas profoundimplicationsfor questionsconcerningculturaltransfers acrossthe IndianOcean.Mathematicians todayreferto thewrittennumbersfrom i to io as Hindu-Arabicnumerals,in recognitionof theirdualheritage.10 Indeed, thereis clearandincontrovertibleevidencefortheirintroductionfromIndiainto the Islamicworldat leastby the ninth centuryof our era,their influencereaching Europeseveralcenturieslaterwith the publicationof Fibonacci'sLiberAbaci (Bookof Calculation). Textualtransmission,as in the case of Hindu-Arabicnumerals,is the mode by which we generallyassume intellectualunderstandingcrosses spatial and temporalboundaries.In this instance,the technologyof transferalso involved translationbaseduponthe directencounterof an individualwith othercultures. Fibonaccihimselfdescribeshis personalexperience,articulatingin writingwhat he absorbedduringhis travelsand exchangeswithin Araband Islamiccultures." This exampleprovidesus with a documentedcase of an identifiableindividual who traveledin foreignlands,observedculturalpractices,and then wrotedown his observationsand understandingin a book format,which powerfullyand effectivelyservedto transmitnewknowledgefromone cultureto another.Indeed, todaywhenone considersthe meansof transmissionof mathematicalknowledge, it is typicalto considerthe roleof the individual,includingpersonalobservation andtravel,as recordedin booksaswellasby oraltransmission.2 These "technologiesof transfer"were recognizedin the fourteenthcentury by the Arabhistorian,essayist,and encyclopedistIbnKhalduin, who compiledan introductionto historicalthought, TheMuqaddimah(An Introductionto History), which exploresprocessesof changeover time in the political and social In this extensivework,Ibn Khalduinaddressesthe organizationof humankind.13 questionof how knowledgeis transferredamongindividuals,acrossgenerations, andto differentcultures.He discussesseveralmodesof the transferof knowledge, includingbooks, observationand travel,and instruction.To these he adds the crafts,recognizingthe valueof experientiallearning. Ibn Khaldfinwas himself a travelerand a criticalobserver.Bornin Tunisin 1332, he traveledthroughoutthe Islamicworld,thinking everywherehe went, as reflectedin his treatmentof the largeissuesof historyandprocessin TheMuqaddimah.Writingaboutwriting,Ibn KhaldAin highlightsthe productionof books, which"preservethe thingsthatareof concernto man andkeepthem frombeing He also saysthatwriting"enablesthe innermostthoughtsof the soul forgotten."''l to reachthose who are far and absent."Writingenablesthe "intention(of one person)to be carriedto distantplaces,and,thus,the needs(ofthatperson)maybe executedwithouthis personallytakingcareof them.It enablespeopleto become 175
PATTERNS
IN TIME AND SPACE
t~~~~~~
*
2dtEXSl~~ul 10Bi -1
tgl
L~~~~~~~~~~~1L. 14#. Jo-
2?an;othetechn learning, tofMb
A
1
4
M'
g on
intr
o
wite
J1~~~~~~~~~~~~~~~~~~~~J
2
Extractfrom the Sine Tables composed by Ibn Yuinus.MS Berlin Staatsbibliothek,Ahlwardt5752 fols. 13v-14r (Lbg1038) reproduced in J.L. Berggren,Episodesin the Mathematicsof MedievalIslam (New York:Springer,1986),pl. 5.1.
"Thetransformationof writingin man frompotentialityto actualitytakesplace throughinstruction."''6 This point brings us to anotherset of technologiesby which knowledgeis transmittedacrosscultures:variousformsof instructionthat takeplacethrough the interactionsamongindividualsor groups.With the benefitof hindsight,and highlightingthe contributionsof the GreeksandPersiansofpre-Islamictimes,Ibn Khalduinrecountsseveralsourcesof learningandmethodsof instruction.Among the latter,he highlightsthe importanceof hands-onexercisesin additionto book learning,identifyingcraftsas a sourcefor knowledge.'6He emphasizesthat the individualhumanbeing cannot exist withoutthe cooperationof others,for "to makeall the thingshe needs,a manby himselfwould requirelongerthanthe time he can keepalivewithoutthem. The abilityto think . .. enableshumanbeingsto cooperate."As for crafts,he saysthat "Themind does not ceasetransformingall kindsof [crafts]. .. frompotentialityinto actualitythroughthe gradualdiscovery of one thing afteranother,until they areperfect.This is achievedin the courseof time and of generations."'h While emphasizingthatthe craftsrequireteaching,he also notesthat"thecraftsandtheirhabitalwaysleadto the acquisitionof scientific norms, which resultfrom the habit.Therefore,any experienceprovidesintelliHe arguesthatthe craftsresultfromman'snaturalabilityto think andto gence."''8 determineoutcomes,the causesandeffectsof reality.19 Other technologiesof transferof mathematicalknowledgefrom the Indian subcontinentto the Islamic world, centeredat Baghdadas the capital of the 'Abbasidempire,maybe enumeratedbasedon availabledocumentarysources.20 In the middleof the eighthcenturyA.D., Baghdadwasfoundedas the capitalof the 'Abbasidempireand soon becamea flourishingintellectualcenterunderpatronageof the 'Abbasidcaliph.21 Al-Mansur(reigned754-75),the founderof Baghdad, 176
CAROL BIER
3
3
Cotton ikat textile fragment with Arabicinscriptions in gold outlined in ink, tenth century,Yemen (Full view, see fig. 6). Textile Museum 73.59,acquiredby George Hewitt Myers in 1931.
received a delegation from Sindh among whom there was an Indian familiar with astronomy. This individual facilitated al-FazarI'stranslation of a Sanskrit astronomical text. The result was Z-j al-Sindhind (based on an Indian astronomical table), which contains astronomical traditions, including mathematical methods using the sine function.22 The use of the sine function is earliest attested in an Indian astronomical handbook of the fourth or fifth century, the SuryaSiddhanta, written in Sanskrit verse. The term zijrefersto an astronomical table (fig. 2). According to J.L. Berggren, it is a Persian word taken into Arabic, which originally signified a "thread" or a "chord"and then a set of such threads "as in the warp of a fabric."23 He relates by the of a whole set of parallel analogy, "astronomical tables, presenting appearance lines separating the columns, came to be known by the same word," an allusion that is visually apt (fig. 3) and relates to the discussion below. Sindhind, similarly, is a perplexing term that may reflect local usage and common parlance. In a sense, sindhind may be considered a genre of Arab mathematical literature, but the word itself as used in Arabic is a toponym, combining geographic references to Sindh and Hind, vaguely referring to lands of the northwestern Indian subcontinent.24Aside from the Sindhind of al-Fazari, many other astronomical texts were translated into Arabic under the patronage of al-Mansur and his successors; several were called Sindhind. The Sindhind of al-Khwarizmi would seem to have been constructed from direct observation of the stars. In the Fihrist al-Nad7m (A Tenth-Century Survey of Muslim Culture), which serves as a guide to what was known and read in tenth-century Baghdad, the entry for alKhwarizmi, one of the "masters of the stars,"states, "Both before and after [confirmation by observation], people relied upon his first and second astronomical tables known as the Sindhind."25 177
PATTERNS
IN TIME AND SPACE
Muhammadibn Musa'al-Khwarizmi(fl. 820) is best known for his workin Arabicthat explainsthe Hindu method of calculation,Kitab Hisabal 'AdadalHindi (The Bookof Additionand Subtractionaccordingto Hindu Calculation), which introducedthe decimalpositionsystemwith ten ciphers,as developedby the Hindus.This book was the firstarithmeticalworkin Arabictranslatedinto Latin.26 Althoughthe book itselfdoes not survive,not only do we havenumerous referencesto it in contemporaneoussourcesbut its significanceis referredto also by Fibonaccimore than three centurieslater,confirmingthe notion that much knowledgewas transferredby textualtransmission.27 Khwarizmihad also written a workcalledKitabal-Jabrwa'lMuqibala(Bookof RestoringandBalancing), dedicatedto al-Ma'muin (reigned813-33).Thetitle couldnot be readilytranslated into Latin,so fromit we havegainedthe word"algebra." Todayit is oftenreferred to as the Bookof RestoringandBalancing,whichadequatelyconveysits approach to the developmentand treatmentof equations.Khwarizmiis often considered to haveevolvedin his thinkingon thesematters,basedupon influencesthatdrew fromboth Babylonianand Hindumathematicalmethods.In succeedinggenerations, Kushyaribn Labban(fl. 970) also wrote a treatiseon Hindu reckoningKitabfi'usil hisib al-Hind(Principlesof HinduReckoning).28 Fromthe perspectiveof reconstructingthe transmissionof knowledge,the curriculumof the Fihristal-Nadim is one of the most importantArabictexts to survive.With a full enumerationof workscirculatingin the tenth centuryand givingsummariesof contentsas well as informationaboutthe authors,this is the best mapwe haveby whichto assessthe intellectualhorizonsof the Islamicworld at thattime. Section2 of chapter7 providesaccountsof scholarsandthe namesof the books they composed,including"menof learningwho weregeometricians, arithmeticians,musicians,calculators,astrologers,makersof instruments,and personsinterestedin mechanicsand dynamics."29 Amongthose listedareEuclid, Archimedes,Apollonius,and numerousothers in the Greektradition,as well as Ptolemy'sAl-Magest(Geography).30Then are namedseveralindividualswho are giventhe epithet"theIndian"(Kankah,Judar,Sanjahil,and Naq) and others referredto as "scholarsof India"(Bakihur[Bhagahara], Raha [Raja],Sakah [Saka],Dahir [Dahara],Ankfu[Indu],Zinkal [Ranakala], Araykal[Arikala],Jabhar,Inda,and Jabara[Jonaror Jitar]).31 Theirworks,for the most part,dealwith the stars,or with medicine.The factthatthesenameshavenot, to my knowledge, yet been linked with known Indianmathematiciansis less significantthan the recognitionin this tenth-centuryworkof the directinfluenceof Indianthinkers within centralIslamiclands. TwogreatIslamicmathematicaladvancesof the tenthcenturyareconsidered to havebuiltuponan understandingof Indianmathematics-the use of decimal 178
CAROL BIER
fractionsand the furtherdevelopmentof trigonometricfunctions.In the middle of the tenth century,Ahmadal-Uqlidisi(fl. 950) used decimalfractionsto solve manyproblems,suchas findingthe rootsof numbersand expressingthe ratioof the circumferenceof a circleto its radius.32 He alsowroteaboutHinduarithmetic. Earlier,$amadibn 'Alial-Yahiidiabtual-Tayyib(fl. 820), a Jewwho convertedto Islamand servedunderthe patronageof al-Ma'muin in Baghdad,builtthe observatory.Amonghis workswasalso a bookon Indianarithmetic.33 Whilethereis ampleevidenceforthe directinfluenceof Indianmathematical thinking in Baghdadunderthe patronageof the 'Abbasidcaliphatein the ninth and tenth centuries,thereareseveralearlierpoints of contactthat suggestintellectualexchangeanddevelopmentbetweenIndiaandlandsto the west.Beforethe Arabconquestof Iranin the mid-seventhcentury,a researchcenterexistedat the town of Jundishapuir in westernIran,whereit is thoughtthatmuchknowledgeof Indianmathematicshad been gatheredunderthe patronageof Sasaniankings.34 Thepersoncreditedwith makingthe firstreferenceto Hindunumeralsoutsideof Indiais BishopSeverusSebokht,who residedin Keneshraon the upperEuphrates Riverin the middleof the seventhcentury.He describedthe extraordinarypossibilitiesof thesenumbers: I will not sayanythingnow of the scienceof the Hindus,who arenot even Syrians,of their subtlediscoveriesin this scienceof astronomy,whichare even more ingeniousthan those of the Greeksand Babylonians,and of the fluentmethod of their calculation,which surpasseswords.I want to sayonly that it is donewith nine signs.If those who believethat they have arrivedat the limit of sciencebecausethey speakGreekhad known these things they would perhapsbe convinced,even if a bit late, that thereare otherswho know something,not only Greeks,but also men of a different language.35 The date and location of this admonishment,duringthe time of 'Umayyad rule in Damascus,suggeststhe possibilitythat Hindu numeralsmay havebeen transmittedto the ArabsthroughSyriacsources,which accountedfor so much othermathematicalknowledgetransferredfromthe Greekcorpus.36 Forearlierperiodsin the developmentof astronomicalcalculation,the influencesofboth GreeceandIndiaareoftencited.Ptolemy'sAl-Magestofthe mid-second centuryA.D. iS thoughtto reflectthe workof both Indianand Greekauthors. Accordingto DavidPingree,the earliestmanualin Indiato addressmathematical astronomy,however,was dependenton earlierMesopotamiansources,transmitted by Persiansduringthe Achaemenidempire(sixth-fourth centuriesB.C.). The 179
PATTERNS
IN TIME AND SPACE
manualoffersa method"fordeterminingthe (mean)times forperformingthose Vedicsacrificeswhich are bound to the seasons,to the months, or to a specific time of day,"basedon the use of a waterclockof a type known fromBabylonian sources.37 Pingreealso refersto laterSanskrittextsthatdescribegnomonscasting noontime shadows,which were computedusing a Babylonianzigzagfunction, suggestingthatthe introductionof the gnomonto Indiamayalsohavecomefrom Mesopotamia."8
Betweenthe Hellenisticperiodand the fourteenthcenturyA.D., and specificallyat the heightof Islamichegemonythroughoutthe NearEastfromthe eighth to twelfth centuries,thereis a largeamountof evidencefor the transmissionof mathematicalknowledgeto centralIslamiclandsfromthe Indiansubcontinentas wellas throughtranslationsof ancientGreekandHellenisticworks.39 TheFihristal-Nadimconfirmsourunderstandingof the roleplayedby textual transmissionin the transferof knowledgeacrossthe Indian Ocean,and it also affirmsthe rolesof firsthandexperiencethroughtraveland observation.Booklearningwas clearlyrevered,to judgefrom contemporaneoussources,but seminarsalso servedas a forumforthe developmentandspreadof knowledge.Several 'Abbasidcaliphsendeavoredto establishlibrariesin which all the knowledgeof the worldwas gathered.40 Harunal-Rashid(reigned786-809) gatheredworksin Sanskrit,Persian,and Greek;subsequentlythe caliph al-Ma'muin(reigned81333) establishedan advancedresearchinstitute called Bayt al-Hikma(House of Wisdom)that employedbinders,translators,and copyists.Al-Ma'muin also sent scholarson missionsto collectworksto translateinto Arabic;one storyrelatesthe presenceof an Indian delegation,includingan astronomer,who was invited to the court in Baghdad,wherethe delegatesmet with Islamicscholars.41 It is possible that al-Khwarizmiwas presentat such a meeting.42The sharingof experience and the generationof knowledgethroughseminarsand conversationsis at the heart of the work of architecturalhistorianAlpayOzdural,who addressed the relationshipof mathematicalthinking to artisanalproduction,recognizing the involvementof mathematiciansin the designof architecturalmonuments.43 Suchhigh-levelmeetingsamongthe intellectualelitemayhaveoffereda standard forum for the exchangeof knowledge.Accordingto Ozdural,therewere meetings amongmathematiciansand artisans,conversationalseminarsat which the intellectualunderstandingof abstractconceptson the part of mathematicians may havebeen transferredto artisans,who put such ideasinto practice.In one specificcase,Ozduralsuggeststhatthe mathematicianOmarKhayyammayhave been directlyresponsibleforthe designof the northdomeof the masjid-ifjami' in This dome,unlikeanyotherdomein Islamicarchitecture,makesvisuIsfahan.44 ally explicitseveralof the proportionalrelationshipson which OmarKhayyam 18o
CAROL BIER
was workingin a theoreticalmanner.The argumentsput forwardby Ozdural offerpotentialsupportfor the oral transmissionof mathematicalknowledgeat the levelsof both theoreticalformulationand practicalapplicationby meansof conversations,seminars,and discussionsamong mathematiciansand artisans workingtogether. Thereareyet other possiblemeans of transmission,not the least of which is the observationof natureitselfandthe constructionof whathascometo be called naturalknowledge.45 This processis akin to Ibn Khaldun'sexplanationof experientiallearningand to his understandingof how one arrivesat the recognition of a scientificnorm.If one considers,for example,whathappensin the layingof bricksone by one in the constructionof masonry,one maydiscoverby doing,or by observationand analysis,the formulasby which a unitaryprocessis at once systemic.Suchis the originof theoryfrompractice,whichmayhavecontributed furtherin this periodto the evolutionof numbertheoryandpatterntheoryfrom the observationof numbersandpatterns. Textilesas a Technologyof Transmission Thereis an additional,potentiallysignificanttechnologyfor the transmissionof mathematicalknowledgeacrosstime andspace,one thathasnot to myknowledge been treatedin this manner.It is the embodimentof mathematicalknowledgein the physicalstructuresof textilesand in the patterningof textiles.Bothphysical structuresand patterningaretechnologiesof transferthat arevisual ratherthan verbal;they both involvestructuraldimensionsand surfaceappearance.In generalparlance,we do not often conceptualizethe ontologicaldistinctionsbetween a book andits contents(orwe takeit forgrantedsuchthatthis differenceis unarticulated);the sameontologicaldistinctionsmaybepertinentto the consideration of textiles.Considerthata bookis a physicalobject,butwhatit transmitshaslittle if anythingto do with the physicalformatof the codex itself.Likewise,consider that a textile is a physicalobjectwith a three-dimensionalstructure,but whatit transmitsmayconveya differentset of valuesthanits physicalityas a textile.Textiles, as a mediumof transfer,havemost often been consideredin termsof their portability.46Whatis oftencitedwhenconsideringtextilesasa meansof transferis the transmissionof stylesandthe transferof culturalformsof expressionthrough designsandmotifs.Historically,textileshavecertainlyservedin this manner,and this significantaspectof theirrolein culturaltransmissionshouldnot be underestimated.But at the sametime, textilesmayalso haveconveyedotherinformation thatis at onceboth tangibleandabstract,concerningthe relationshipamong number,shape,and the natureof space in the formalrelationshipsof patterns that embodya unit and its iterationaccordingto particularprinciplesof organi181
PATTERNS
IN TIME AND SPACE
zation.This information,conveyedwithin the textile medium,is distinct from the physicalformof the textileitselfand ontologicallydifferentfromthe styleor contentof its decoration.Justas a book is a "technologyof transfer"of information, so maytextileshaveservedin this mannerapartfromtheirphysicalorvisual aspects.The term "embodiment"is intendedto conveythe separateidentityor entityof patternsfromphysicalaspectsof the textileaswellas fromthe technologies utilized forthe formationof the textile,its structure,and its patterns.What I seekto emphasizeis that the patternsthat appearon a textile are ontologically distinctfromthe textileitself,evenwhenthey area partof the textile'sstructure. Accordingto the paradigmI wish to develop,the textile servesas a mediumto conveypatternsand theirembeddedmathematicalknowledge,thus servingas a technologyof transfer.Textiles,throughtheirpatterns,"embody"mathematical knowledge.As such,textiles-in the plural-may haveplayeda significantrole in the transmissionof mathematicalknowledgeand otherculturalvaluesacross the IndianOcean. In lookingat textiles,one may considerstyle and imageryand portability,as wellas culturalmeanings,but one also needsto considerseveraldistincttechnologies andprocessesthatoperatecollectivelyand interactto producea finalproduct that may be transported.The technologiesassociatedwith textiles include the domesticationof animals (forwool), cultivationof the silk worm (for silk), and horticulturalactivitiespertainingto the cultivationof plants(forflax/linen, and for cotton, amongothervegetalfiberssuitablefor textile production).They also entailmethodsof yarnpreparationandthe constructionof fabric,dependent upon the technologiesof weaving.Dyeing technologiesrelyupon an empirical familiaritywith naturalsubstancesand theirpreparationand treatmentto yield dyes.47
To explorethe possibilitiesof the kinds of knowledgeor informationthat might be storedin a textile, I havesomewhatarbitrarilyselecteda fragmentary textile from the collectionsof the TextileMuseum(fig. 4) that fits broadlyand specificallywithin the culturaltransfersacrossthe IndianOcean.Attributedto Yemenin the tenth century,it sharesseveralfeaturesthat link the Indian subcontinentwith the Islamicworld:it is cotton, dyed using indigo, and patterned by a warp-resistmethod calledikat. The cotton warpsof the textile are dyedin variouscolorsusing resisttechniquesbeforeweaving;it was embroideredafter weaving(seefig. 4). Amongthe colorsis a rangeof bluescharacteristicof indigo and its normalvariationsthat resultfrom the strengthof the dye solution,the length of time the yarnsweresubmergedin the vat duringthe dye process,and the numberof times the yarnswere submergedin the dyebath.Variationsalso resultfrom the variabilityof the oxidationprocessduringdryingand exposure 182
CAROL BIER
a
Ab
4
4
Detail, embroideredcotton ikat with Arabicinscriptions, tenth century,Yemen.Textile Museum 73.494, acquiredby GeorgeHewitt Myers in 1940.
to air after dyeing. Although fragmentary, this textile is indicative of several textile technologies (yarn preparation, the extraction of dyes from plant materials, indigo dyeing, resist-dyeing methods, weaving, embroidery) as well as the cultivation of cotton.48 The textile under consideration (see fig. 4) has a striped appearance;it is woven with the simplest interlacing of warp and weft, referred to as plain weave.49Its visual aspect depends primarily upon two distinct categories of repeated designs. One set of patterns results from warp-resist dyeing methods before weaving; the other set of patterns was executed in embroidery after the fabric was woven.50The narrow chevrons and lozenges aligned vertically in a rhythmical arrangement give the appearance of stripes. This pattern was created by a selective warp-resist dyeing process executed before weaving took place. The warp yarns, once dyed, were arranged side-by-side when the loom was dressed to yield the resulting patterns of color after weaving was completed. The second set of patterns was achieved by means of embroidery. As seen here, the repeats of small units create tessellations, patterns composed of a single shape, which when repeated cover the plane with no gaps and no overlaps.idThese geometric patterns were embroidered with undyed cotton, offering a contrasting embellishment without color, which was added after weaving was complete. In addition to the embroidered patterns, there are two embroidered inscriptions written in Arabic, each an excerpt from the Qur'an (chap.
i, 112).52
The warp-resist process yielding a pattern with variable blocks of color is worth exploring in greaterdetail. The selective binding and dyeing, combined with careful placement of dyed and resisted sections of yarns, result in a pattern of chevrons and lozenges. These rhythmic stripes that characterize Yemeni ikat textiles represent a high level of achievement of the dyer's art. The play of color results from selectively binding and dyeing the warp yarns prior to weaving; the bound sections 183
PATTERNS
IN TIME AND SPACE
k~~~~~
5
5
Diagram illustrating steps in the process of resist-dyeing:sequential binding of groups of warp yarns to resist the dye, one set of bindings for each color. Drawing by Erin Roberts.
resisted the dye. One method of binding utilized a standardized selective length of yarn; after dyeing, warps would have been positioned at different heights when dressed on the loom (fig. 5). Repeating the process of binding and dyeing would result in the observed color sequences of blues, yellow, and brown. The brown may have once been red, but analysis of this color has not proved conclusive. The white areas presumably would have remained bound, thereby resisting all colors of dye. Another method of binding would have segregated groups of warps much as they appear in the finished textile, dividing each stripe into smaller divisions and selectively binding and dyeing at differential heights. From our vantage point today, if we look critically at the arrangements of forms, we may identify in each pattern a relationship to number, as well as to single units that are repeated according to particular algorithms to create specific, intentional patterns to fill the plane-conceptually filling space in two dimensions, as perceived visually. Yet the textile itself was executed in three dimensions, relying upon the interlacing of warp and weft, each element spun to form a yarn before its manipulation in the process of weaving. The patterns themselves are finite, repeated in a manner that may imply infinite extent, yet they are arbitrarily cut off to fill the field or to fit within borders or not repeated beyond a particular specific extent. By such an analysis, we may relate the patterns on this textile to concepts and modes of abstract thinking that pertain to number, shape and the nature of space.53In this manner, our viewing of the textile allows us to relate tangible aspects of the patterns to intangible mathematical concepts. Such visual thinking actually relates to philosophical and mathematical ways of thinking that can be documented among the widespread peregrinations that define a global history of mathematical thought as it developed over more than two thousand years from Babylonian, Greek, and Indian origins to the Islamic world and the Latin West. This textile is part of a fairly large group of similar textiles (fig. 6; see also fig. 3), woven in plain weave, patterned by warp-resist processes using indigo, bearing Arabic inscriptions in gold leaf and ink or embroidered, attributed to Yemen in the tenth century. This fragment and others of its kind are found in museums throughout the world; they seem to have been purchased from dealers in Egypt 184
CAROL BIER
s
/
!
;
I/( ;(,),/4
/ ./
6
6 Full view of fig. 3.
in the firsthalf of the twentiethcentury.Examplesare presentin the collections of the TextileMuseumin Washington,the MetropolitanMuseumof Art in New York,RoyalOntarioMuseumin Toronto,KelseyMuseumin AnnArbor,andelsewhere.TheYemeniattributionhingesin partuponan inscriptionthatappearson one fragmentin the IslamicMuseumin Cairo,whichbearsthe statement"woven in San'a."54 This group of textiles is unique within the Islamic world. Many physical aspectsof thesetextilessuggestlinks to traditionsoften consideredmorefamiliar acrossthe IndianOcean.55 Relationalaspectswith textiletraditionsof the Indian subcontinenthaveto do with both materialsandtechnologiesaswellaswith style. Suchfeaturesas the use of cotton fiberfor the yarns,the use of indigo as a basis for color,and the decorationof stripedpatternseffectedthrough a warp-resist processareall aspectsof textilesoften associatedwith the Indiansubcontinent.56 Yetthe natureof the links with regardto this groupof textilesremainssomewhat elusiveandinconclusive. These textiles representearlyexamplesof the techniquetoday called ikat, a Malayword generallyattributedto its use in Indonesia.But 'aqadais an Arab Semiticrootfortheverb,"totie"or "tobind."It is possiblethatArabtradersintroduced ikat-dyeingtechniquesto SoutheastAsia as earlyas the eighth century.57 MedievalIslamic sources referto madder(Arabic,fuwwa [Rubiatinctorum]), indigo (Arabic,nila [Indigofrrsp.]), and a yellow dyestuff (Arabic,wars [Mem185
PATTERNS
IN TIME AND SPACE
ecylon tinctorum]), exportedfromAdenandYemen.58 Sincethesesourcesof color
werereadilyavailablelocally,it is likelythey wereused for these textiles;indigo dyeingis still practicedin Yementoday.59 The designson these textile are generallyexaminedfrom the perspectivesof textiletechnologies(warp-resistpatterning,embroidery)or epigraphy.60 Another approachto thesepatterns,however,wouldencompassthe mathematicalaspects of whatthey presentby way of structureand pattern,exploringthe impactthat suchpatternsmighthavehadeitherupon the generationof mathematicalthinking or upon the culturaltransmissionof mathematicalknowledgefromIndiato the Islamicworldandbeyond. Clearly,to achievethe repeatedpatternsof chevronsand lozengesin stripes, countingwasnecessary;to formthediscretegroupsbycolor,divisionwasrequired. The mirrorrepetitionacrossa verticalaxis requiredhalvinga group;the stepped appearancealso dependedupon divisionsof the warps.Once an algorithmwas established,the unit so determinedcouldbe iterateda sufficientnumberof times to completethe full patternedareaof a completewidth of cloth.As with anypattern,a unitwasdeterminedalongwith a methodof iteration.Forthe embroidered patterns,similarly,a unit was determined,its reiterationforminga tessellation. In each case, the repetitionof the designunit dependedupon a particularalgorithmforthe patternto emerge.Conceptually,suchpattern-makingcouldgo on forever;in realspace,however,it is boundedby the spatialityof the textileitself, if not arbitrarilyboundedby borders.Patternsin textiles implyinfinity,yet are executedwith boundariesthat serve as limits to their extent. In this manner, two-dimensionalpatterns(evenif, as in the caseof ikatand embroidery,they are three-dimensionalin structure)playwith spacein two dimensionsand relateto a conceptualunderstandingof two-dimensionalspace.The patternsresultfrom countingand repeatingsequencesof actions,acted upon objects.In the case of ikat, the objectsare the individualdyes and the warpyarnsbeforeweaving;in the caseof embroidery,the objectsarethe embroiderythreadandthe fabric.The visual aspectof the completedpatternspreservesthe sequenceof operationsof theircreation;the warpsarecountable;the interactivemovementof the embroiderythreadthroughtime andspaceis observable.6' The questionis whethersuch mathematicalaspectsof patternsand patternmaking were or were not consciouslyunderstoodby either contemporaneous mathematiciansor by artisansengagedin the processesof pattern-making.We can look to architecturefor comparableevidencethat patternswere held as a high culturalvaluewithin the centralIslamiclandsby the eleventhcentury.Lisa Golombekhasadvancedthe notionthatbythis timetherewasa conscious"textile aesthetic."62 Sheattributesthe prevalenceof interlace,aswelloverallpatterningin 186
CAROL BIER
architectureandthe decorativetreatmentof surfaces,to a "textilementality"that is reflectedin "bookbindings,wood-carving,architecturalfaience,and Koran pages"that"alllooklikecarpets."Golombekpresentsaworkinghypothesis,which is followedby numerousscholarsof Islamicart history,that "textilesin Islamic societyfulfilledfarmorethan the functionsnormallyexpectedof them in other societies."And she ascribesmanycharacteristicsof Islamicart to whatshe hesiButthe sameevidencemaybe adduced tantlycalls an "obsessionwith textiles."63 in supportof a "patternaesthetic"havingnothing directlyto do with textiles.64 Towhatextentcanwe knowwhetherthis appreciationof patternswasan applied mathematicalunderstandingof the spatialdimension,ora visualexpressionfrom whichmathematicalunderstandingwasderived? To addressthis question,a cursorylook at the contemporaneousIslamiccultural context in the 'Abbasidempirerevealsa distinctionbetween theoretical mathematicsand practicalapplicationsof mathematics.Practicalmathematics pertainedto landsurveyingandcivilengineering,determinationof times of daily prayerand qiblaorientation.It also pertainedto calculatingdivisionsof inheritance.Butduringthe sameperiod,theoreticalconsiderationswerebeginningto addressthe formulationof algebraandthe extensionsof mathematicsof the Late Antiqueworldto quadraticandcubicequationsandtheirpracticalapplicationsin art and architecture.Towhatextentmightwe be ableto discernparadigmshifts from the transmissionof such knowledge?To what extent might we judgethat theoreticalformulationmayhavein factderivedfromworksof art and architecture?Or that technologicaldevelopmentsprefiguredor postfiguredintellectual advancesin theoreticalor abstractthinking?Canwe look at an object,such as a patternedtextile, and hope to determinewhat might havehappenedearlier,or whatmight havehappenedafterwardin termsof contemporaneousintellectual developments? Thenexusof intersectionsbetweenartandmathematics,thoughoftenpresent, is difficultto situate.One recognizableexceptionis the developmentof perspective in the ItalianRenaissanceof the late fifteenthand earlysixteenthcenturies, when both the mathematicsand artisticapplicationswere exploredsimultaneously,as articulatedin the treatisesof Leonardoda Vinci,LeoneBattistaAlberti, andAlbrechtDurer.Thesemen weretheoreticiansas well as painters,sculptors, and draftsmen- contemporariesof LucaPacioli,who was a teacherof Leonardo da Vinci.65In this unique set of historicalcircumstances,experimentationand empiricalstudiesled directlyto theoreticalunderstandinganditsfurtherapplication in the arts. For tenth- and eleventh-centuryBaghdadand the further reachesof the 'Abbasidcaliphateand its rivalcaliphatein FatimidEgypt,emergentquestions 187
PATTERNS
IN TIME AND SPACE
focus on whetherand how the knowledgeof pattern-makingthat is so evident in both architecturaldecorationandtextilesrepresentedadvancementsin mathematicalknowledgeandthe understandingof two-dimensionalspace.66 Whether andhow suchknowledgeof patternswassharedamongculturesto the east,west, andnorthof the IndianOceanremainsa subjectforfutureinvestigation.Fornow, it is sufficientto describethe extentto whichweknowthe stripedtextilesof Yemen werefamousthroughoutIslamiclands.As is wellknowntoday,ArabandPersian geographersof the ninth and tenth centuriesdescribedin fair detail the goods tradedin manycities,andthey givea senseof the characterof manyports,listing resourcesand theirproducts.The stripedclothsof Yemen,referredto by several names,weresoughtfarandwide throughoutcentralIslamiclands.6'Butthereis no directlink to identifythe stripedikatcottontextilesknownin museumcollections todayto thesetextualreferences. Yet to look at the textiles themselvesreveals additionalprospectivelinks betweenartof the physicalobjectandmathematics.Thelayeringof patternsupon patterns,as remarkedabove,is significant.The embroideryrepresentsa process embarkedupon afterthe textile was woven,whereasthe variegatedstripedpattern with chevronsand lozengesresultsfrom warp-resistdyeingprocessesthat took place priorto weaving.In its finishedappearance,then, this textile bears marksor patternsthataredistinguishedfromone anothertemporallyin the constructionof the object,and they aredistinguishedfromone anotherspatiallyby the repetitionof differentdesignelements,althoughboth arepresent,visuallyand conceptually,withinthe spaceoccupiedby the physicalobject. Wemayalso look to architecturein whichthe conceptsof unit and repeatare equallyapt. In both architectureand textiles, the processesof constructionof both structureand designconstitutea systemor systemsthatareat once unitary and systemic.In arts and architectureof the Islamicworld,patternedsurfaces becomeparticularlysignificantin the ninth and tenth centuriesand continueto be of primaryvisualimportancethroughthe fifteenthcenturyand later.Several specificexamplesmaysufficeto providea generalappreciationof the highcultural valueplacedon patternsin architecturethroughoutthe Islamicworldin the tenth throughtwelfthcenturies.Thetombof the Samanidsin Bukharaof the tenthcentury is often consideredthe earliestmonumentto utilize an overallpatterning determinedby the placementof bricks.68 In the eleventh-centurytomb towersof Kharraqanin westernIran(fig.7), morethan a hundredpatternsexecutedin cut brickcombineto covereverysurfaceof the structurebeneaththe domesbut for the bandsof inscription,whicharealsoexecutedin cutbrick.69 Tombmonuments in the followingcenturyintroducethe useof glazedceramicmosaictile in the executionof geometricpatterns,as forexamplethe tombsatMaraghaalso in western 188
CAROL BIER
e~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~
I..
~ ~ ~ ~ ~ ~ ~ ~
ap
a
4.
-4.~~~~~~~~~~~
7 Bricktomb tower,eleventh century, Kharraqan,Iran. Photograph courtesy Ann C. Gunter.
Iran.70Farthereast, the stellatedthree-dimensionalstructureof the minaretof Mas'uidof Ghaznais coveredwith panelsof geometricornament,combinedwith floriatedKuficinscriptions,and relatedby its all overgeometricpatterningto the minaretsof Kalayan,Damghan,and Jamof the late eleventhand twelfth centuries.71 The Qut.bMinarin Delhi of the twelfth centurycontinuesto advancean aestheticthatis geometricallybasedin its three-dimensionalform.7 In all of these examples,whetherthe patterningis executedin brick,glazed ceramic,tile, or a textile medium, the variousprocessesby which designs are y.73 In each manipulatedto formpatternsareconstrainedby the lawsof symmetr instance,the algorithmicrepeatof a designto coverthe planeis dependentupon the relationshipof numberto shapeand the natureof two-dimensional space.To ~ explorethe phenomenonof pattern-makingin artsandarchitectureof the Islamic worldin the ninth and tenth centuriesleads to the speculationthat within this time frametherewerenew empiricaldiscoveriesaboutthe natureof two-dimensional spaceas articulatedthroughpatternsthat led to theoreticalformulations It is conceivablethat sucha discoveryrelied concerningtwo-dimensionalspace.74 a new of Greek upon synthesis geometryand Indian arithmetic,both of which One might werein evidenceat the courtof Baghdadby the time of al-Khwairizmi. thus interpretthe extraordinaryproliferationof patternsand pattern-makingin all mediaas derivativeof this new,or renewed,understandingof the planein the tenth centurythat reverberatedthroughoutthe arts in the Near Eastand India duringthe followingcenturies.75
PATENS
189~~~~
mathematicalprocessesfindsvisualexpressionin the patterningof textiles.Further,they relateto the processesof pattern-makingin a varietyof mediathat are at once unitaryand systemicby virtue of the fact that a predeterminedunit is repeatedaccordingto at least one organizingprinciple.Within any given pattern in the plane,the numericalrelationshipsamongthe units repeatedare also expressedgeometricallyin the relationshipof shapes,relatingan understanding of geometryof the time to the emergingunderstandingof algebrain the workof al-Khwarizmi.Thepatterningthatwe see in architecturalornament,andthat in textilesaswellas in othermedia,mightwellreflectthe developmentof mathematicalthinkingof this era,yet this connectionremainsspeculativeatthis time. Withinthe worksof mathematiciansof the Islamicworld,numeroussubjects are broached,cumulativelyaddressingthe practicalconcernsof land surveying andthe divisionsof inheritanceaswellas astronomicalcalculationandastrological interpretations.The abstractconceptof numberitself is also explored.In a philosophicalvein, Al-Farabi,who studiedmusicas well as mathematics,distinguishesin his Ihsa' al-'ulum (Enumerationof the Sciences)betweentwo sciences of arithmetic-practical arithmeticandtheoreticalarithmetic: Practicalarithmeticinquiresinto numbersinsofarasthe numbersarefixed to things numbered,i.e. bodies and otherthings whose enumerationit is necessaryto have,like men, horses,dinars,dirhams,or otherthings possessedof number.Thisis the arithmeticthatthe peopleuse in the commercialtransactionsof the marketandthe city.Theoreticalarithmeticinquires into numbersin the abstractinsofarastheyareseparatedin the mindfrom bodies and everythingelse which is numberedby them.... And it is this latterarithmeticwhich investigatesnumbersin an independentmanner (i.e., in the abstract)with regardto i) everythingwhich happensto them in their simpleessenceswithout mutuallycomparingthem, for example: their being even or odd, and 2) everythingwhich happensto them when they aremutuallycompared,for example:theirbeing equal;or one being morethan another(i.e., theirbeingunequal);or one numberbeinga part or partsof anothernumber,or doubleit, or equalto it with the additionof a part or parts;or like numbersbeing proportionalor not being proportional, similar or dissimilar,commensurableor incommensurable.Furthermore,theoreticalarithmeticinquiresinto whathappensto numbers upon additionandtheirresultingsum,theirsubtractionandresultingdifference,theirmultiplication,andtheirdivision.Thisis likenumbersbeing squared,or representingsurfacesor solids, or being perfector imperfect. Thus theoreticalarithmeticinvestigatesall of these things and whatever 190
CAROL BIER
happensto numberswhen they aremutuallycompared.It teacheshow to find [unknown]numbersfrom known ones, and, in short, everyway of findingnumbers.77 Similarly,NasirKhusraudistinguishestwo kinds of matter,one absoluteand the otherrelative:"Therelativeis likewood to a chair;thatis, wood whichis preThe absolutematteris intelligible,but not paredand arrangedto forma chair."7" perceptiblewithout form;he relatesmatterto length, width, and depth. When prime matteracceptslength,width, and depth, it takes on a form in which the absolutematterresides.79 Evenmoreprecisedocumentationwith respectto textilesis broachedin Indian philosophictexts of the tenthcenturyin whichdistinctionsarisebetweenmatter and substrate.I0 The questionpertainsto the colorof a multicoloredtextile, distinguishingcoloras ontologicallyseparatefromthe textile itself.Variegatedand stripedtextilesareontologicallydistinguishedfromthe colorstheybear. Conclusions This inquiryinto prospectivetechnologiesof transferand culturaltransmission acrossthe IndianOcean,particularlywith referenceto mathematicalknowledge, offersan initialattemptto distinguishan ontologyof patternthatis distinctfrom an ontologyof textiles.As such,it raisesseveralquestionsto whichtherearenot yet definitiveanswers.The ikattextiles,foundin Egyptbut presumedto be from Yemen,selectedas examplesseekconclusiveconfirmationasto actualoriginsand a factualspatialandtemporalplacement.Butthe largerconceptualproposalof an ontologyof pattern,distinctfromthe categoryof a textileitself,raisespotentially significantissues with respectto the transmissionof mathematicalknowledge acrossthe IndianOceanandbeyond. Toconsideroverallrepeatpatternsin textilesis to considerspace-fillingcurves in mathematicslong beforethese werecategorizedwithin the scope of Western mathematics.Suchpatternsare typicalin physicaltime-spacein the designsof woventextilesfromWestAfricaeastto India,Borneo,andthe islandsofthe Indonesianarchipelago,spanningAfrica,the NearEast,the Indiansubcontinent,and SoutheastAsia. It is temptingto proposethat such patternsreflecta historical mathematicalknowledge,the originsof whicharetodayshroudedin lost memory that pertainsto the expansionof traderoutesin the eighththroughtenth centuries,coincidentin time with Islamicdevelopmentsin the understandingof twodimensionalspace. The historicalvalidity of such a proposal- suggestingthe spreadof mathematicallybasedpatternsthroughthe mediumof textiles-must awaitfurtherstudyalongthe outlinesof researchsuggestedhere. 191
PATTERNS
IN TIME AND SPACE
NOTES
1.
Fibonacci,Fibonacci's"LiberAbaci".-A TranslationintoModernEnglishof LeonardoPisano'sBookof Calculation, trans. LaurenceE. Sigler,Sourcesand Studiesin the History of Mathematics and PhysicalSciences (New York:
London:KluwerAcademicPublishers, 2000),
Introductionto History,trans. Franz Rosenthal,ed. N. J.Dawood, Bollingen Series(Princeton:Princeton University
Springer, 2002). Sigler,introduction to Fibonacci,Liber Abaci,3-4; Fibonacci,LiberAbaci,15. 3. Fibonacci,LiberAbaci,21-22. 4. These were initially discussed by alKhwarizmiand laterby Kuishyaribn Labban,whose worksare cited by J.L. Berggren,Episodesin theMathematicsof MedievalIslam (New York:Springer, 1986),6-9,31-36. 5. CarolBier,"Textilesand Society,"in WovenfromtheSoul,SpunfromtheHeart: TextileArtsof Safavidand QajarIran (i6th-l9th Centuries),ed. CarolBier (Washington,D.C.:TextileMuseum, 1987),1. 6. IreneEmery,ThePrimaryStructuresof Fabrics(Washington,D.C.:Textile Museum, 1980),xi, xv. 7. Fibonacci,LiberAbaci,17. 8. Berggren,Episodesin Mathematics,7. 9. The woodcut of Arithmeticais often reproduced.See, for example, CarlB. Boyer,A HistoryofMathematics,rev.Uta C. Merzbach,2nd ed. (New York:John Wiley & Sons, 1991), 253. 1o. See, for example,Victor J.Katz,A History ofMathematics:An Introduction,2nd ed. (Reading,Mass.,and Menlo Park,Calif.: Addison-Wesley,1998), 230-32; or Claire L. Parkinson,ed., Breakthroughs: A Chronologyof GreatAchievementsin Scienceand Mathematics,1200-1930 (Boston:G. K. Hall, 1985),1. 1i. Fibonacci,LiberAbaci,15-16. 12. LeighN. Wood, "Communicating MathematicsacrossCultureand Time," in MathematicsacrossCultures:The HistoryofNon-WesternMathematics,ed. Helaine Selin (Dordrecht,Boston and 2.
192
CAROL BIER
1-12.
13. Ibn Khalduin,TheMuqaddimah:An
Press, 1967). 14. Ibn Khalduin, Muqaddimah, 319.
Ibn Khalduin,Muqaddimah,327. 16. Ibn Khalduin, Muqaddimah,314-15. 15.
17. Ibn Khalduin, Muqaddimah, 314.
18. Ibn Khalduin,Muqaddimah,331. 19. Ibn Khalduin, Muqaddimah, 319-21. 20.
A summary of such sources is presented in Berggren,Episodesin Mathematics,210.
21.
JacobLassner,The Topography of Baghdadin theEarlyMiddleAges (Detroit:WayneStateUniversity,1970); GuyLe Strange,Baghdadduringthe AbbasidCaliphate(Oxford:Clarendon Press, 1goo).
22.
Berggren,Episodesin Mathematics,2.
23. Berggren,Episodesin Mathematics,132. 24.
See EdwardWilliam Lane,An ArabicEnglishLexicon,pt. 4 (1872;reprint, Beirut: Librarie du Liban, 1968),1443;
GuyLe Strange,TheLandsof theEastern Caliphate(London:FrankCass &Co., 1966), 331. 25. Al-Nadim, The "Fihrist" ofal-Nadtm:A
Tenth-CenturySurveyofMuslimCulture, ed. and trans. BayardDodge (New York: ColumbiaUniversityPress,1970), 652. 26. Berggren,Episodesin Mathematics,7. 27. Fibonacci,LiberAbaci,631, n. 7,8. See also Wood, "CommunicatingMathematics acrossCultureand Time,"lo-ii. 28. Kfishyaribn Labban,Principlesof Hindu Reckoning,trans. and ed. Martin Levey and Marvin Petruck(Madison and Milwaukee:Universityof Wisconsin Press, 1965). 29.
Al-Nadim, Fihrist,634.
30. Al-Nadim, Fihrist, 634-40.
31. Al-Nadim, Fihrist,644-45.
following articles:"OmarKhayyam, Mathematicians,and Conversazioniwith Artisans,"Journalof theSocietyof Architectural Historians 54.1 (March 1995): 54-71; "OnInterlockingSimilaror CorrespondingFiguresand Ornamental Patternsof CubicEquations,"Muqarnas 13 (1996): 191-211; "Mathematicsand Arts: Connections between Theory and Practicein the MedievalIslamicWorld," HistoriaMathematica27 (2000), 171-201; "TheUse of CubicEquationsin Islamic Art and Architecture,"in Nexus IV: Architectureand Mathematics,ed. Kim Williams and JoseFranciscoRodrigues (Florence:Kim Williams Books, 2002),
32. Berggren,Episodesin Mathematics,7.
33. Al-NadIm,Fihrist,652-53. 34. This centeris referredto, for example, in SeyyedHossein Nasr,IslamicScience:An IllustratedStudy(London:Worldof Islam Festival, 1976), 11.
35. Quoted in Berggren,Episodesin Mathematics, 30,166.
36. Berggren,Episodesin Mathematics,166. But compareDimitri Gutas, Greek Thought,ArabicCulture(London and New York:Routledge,1998), 20-22. 37. David Pingree, "TheMesopotamian Origin of EarlyIndian Mathematical Astronomy,"Journalforthe Historyof Astronomy 4 (1973):1-12, quoted in Pingree, "Legaciesin Astronomy and Celestial Omens,"in TheLegacyof Mesopotamia,ed. StephanieDalley (Oxford: Clarendon Press, 1998),130, n. 77.
38. Pingree,"Legaciesin Astronomyand Celestial Omens," 135-36, also mentions
the construction of Arabiccompendia of astronomicaltables (zijes), which present descriptionsof the Uighur calendar. These "indicate that the Uighurs ... also
used the 248-day period for the moon," but Pingreenotes that it is not clear whetherthis is an influence absorbed from India or more directly from Mesopotamia. 39. See, for example, GeorgeGheverghese Joseph,TheCrestof thePeacock:NonEuropeanRootsofMathematics(London: Penguin Books,
1991),
215-348. Math-
ematical sciences in earlyIslam are also discussed by E. S. Kennedy,"TheExact Sciencesin Iran under the Saljuqsand Mongols,"in TheCambridgeHistoryof Iran,vol. 5, TheSaljuqand Mongol Periods,ed. J.A. Boyle (Cambridge: University Press, 1968), 659-79. 40. Gutas, GreekThought,ArabicCulture. 41. Berggren,Episodes in Mathematics, 31. 42. Berggren,Episodes in Mathematics, 31.
165-79.
44. Ozdural,"OmarKhayyam." 45. Recentdiscussions of naturalknowledge focus on the historyof science. SeeJan Golinski,MakingNaturalKnowledge: Constructivismand theHistoryofScience (Cambridge:CambridgeUniversityPress, 1998);JohnV.Pickstone,WaysofKnowing: A NewHistoryofScience,Technology, and Medicine(Manchester:Manchester UniversityPress,2000). Neitherbook considers,however,the embodiment of knowledgein craftedproductsor craftsas a sourceofknowledge.A poetic description of the ontology of naturalknowledge surviving from classicalIslamictimes is the storyof Hayyibn Yaqzan,related (differently)byAvicennaandbyIbn T ufail. See IbnT ufayl,Ibn Tufayl'sHayy ibn Yaqzan:APhilosophicalTaleTranslatedwithIntroductionandNotes,ed. LennEvanGoodman (NewYork:Twayne Publishers,1972). 46. See, for example,LisaGolombek,"The DrapedUniverseof Islam,"in Content and Contextof VisualArtsin theIslamic World,ed. PriscillaSoucek (University Park:PennsylvaniaStateUniversity, 1988), 25-49; Oleg Grabar, The Mediation
43. Alpay Ozdural'sworksinclude the
of Ornament,Bollingen Series35.38 193
PATTERNS
IN TIME AND SPACE
(Princeton:PrincetonUniversityPress, 1992), 119-54.
47. Foran overviewof diverseperspectives on textiles of ancient Mesopotamia,see CarolBier,"TextileArts in Ancient WesternAsia,"in Civilizationsof the AncientNear East,ed. JackM. Sassonand KarenRubinson (New York:Charles Scribner's Sons, 1995),1567-88.
Art: 48. See CarolBier,A Calligrapher's InscribedCottonIkatfromYemen, exhibition brochure(Washington,D.C.: TextileMuseum, 2001). 49. The Yemeniikatstend to be balanced plain weaveor warp predominant;see Emery,PrimaryStructures,76, fig. 85,86. 50. The firstpublisheddescriptionof the fabricappearsin ErnstKuhneland LouisaBellinger,Catalogueof Dated Tiraz Fabrics:Umayyad,Abbasid,Fatimid (Washington,D.C.:TextileMuseum, 1952), 89-go, with parallelscited. 51. Basicdefinitions of grids and tessellations, discussed in relationto Oriental carpetsand Islamic art, appearin Carol Bier,"Symmetryand Pattern:The Art of OrientalCarpets,"http://mathforum. org/geometry/rugs/symmetry/grids.
html (accessedMay14,2004). 52. Kuhneland Bellinger,Dated Tiraz,89. 53. See CarolBier,"Number,Shapeand the Nature of Space:An Inquiryinto the Meaningof Geometryin IslamicArt,"in How to TalkaboutReligion,ed. James BoydWhite (South Bend, Ind.:University of Notre Dame Press, 2006), 246-77.
54. Illustratedin GastonWiet, Tissuset tapisseriesdu Muse Arabedu Caire (Paris:LibrarieOrientalistePaul Geuthner,1935) (repr.Syria: revue d'art oriental et d'archeologie 1935),278-go, pl. 48, whereit is incorrectlydescribedas linen. The fragmentis embroideredin a style similar to that illustratedhere in figure4, the Arabicinscription showing tall letterswith triangularfinials;the
fragmentinscribed "wovenin San'a"also bearsthe name of the caliph al-Muqtadir and a date of 311 A.H./A.D. 923. 55. See, for example,RosemaryCrill, "Ikatin India and the TradeWestward,"in International"Ikat"WeavingForum: Kuching,Sarawak,Malaysia,June11-16, 1999 (SocietyAtelierSarawak,[2000]), 17; Crill, IndianIkat Textiles(London: Victoriaand Albert Publications,1998). 56. Foran illustration,see Benoy K. Behl, TheAjantaCaves:AncientPaintingsof BuddhistIndia (London:Thames and Hudson, 1998), esp. 82,83, 85,87-96. 57. CarlJohanLamm, Cottonin Mediaeval Textilesof theNearEast (Paris:Librarie
PaulGeuthner,1937),144. Orientaliste 58. R. B. Serjeant,IslamicTextiles:Materials for a Historyup to theMongolConquest (Beirut:Librariedu Liban,1972), 131-32. 59. See JennyBalfour,Indigoin theArab World(Richmond, Surrey:Curzon,1997); Balfour-Paul,Indigo(London:British Museum Press,1998). 60. See for example,Kuhneland Bellinger, Dated Tiraz,which treatsof structures, techniques, and inscriptions;Alfred Buhler,IkatBatikPlangi,vol. 3 (Basel: Pharos-VerlagH. Schwabe,1972),which exploresdecorativeprocesses;Sheila Blair,"LegibilityversusDecoration in Islamic Epigraphy:The Case of Interlacing,"in WorldArt:Themesof Unityin Diversity,ed. Irving Lavin,Acts of the 26th InternationalCongressof the History of Art (Pittsburgh:Pennsylvania StateUniversityPress,1989), 329-34, which addressesissues of epigraphy. 61. Fora similar relationshipbetween patternsand pattern-makingin pile carpets, see CarolBier,"Choicesand Constraints:PatternFormationin OrientalCarpets,"in Proceedingsof the 2nd InternationalKatachiU Symmetry Symposium,Tsukuba,1999,Forma(Journal of the Society for Scienceon Form,
Japan) 15.2 62.
IslamicArchitecture(Washington,D.C.: Smithsonian Institution Press,1972), pIs.
(2000):127-32.
Golombek,"DrapedUniverseof Islam,"
17-26.
25-49.
63. Golombek,"DrapedUniverseof Islam," 36,25
64. See CarolBier,"Geometryand the Interpretationof Meaning:Two Monuments in Iran,"in Bridges: MathematicalConnectionsin Art, Music, and Science,conferenceproceedings,July 2002, ed. RezaSarhangi(Winfield, Kans.: BridgesConference,2002), 67-78. See also Grabar,Mediationof Ornament;Bier, reviewof Mediationof Ornament,by Grabar,MiddleEastStudiesAssociation Bulletin28.1 (1994): 28-30. 65. Useful summariesof the relationshipof artistic workand mathematicaltheory areprovidedunder the entries for "AlbrechtDurer,III.Theoretical writings,""Leonardoda Vinci, III. Theory,"and "LucaPacioli"in TheGrove DictionaryofArt Online,ed. L. Macy, http://www.groveart.com(accessedJune 10, 2003). Fora welcome revisionist approach,advocatingacknowledgment of strongArabinfluence and numerous contributionsto science in the Renaissance, see GeorgeSaliba,"WhoseScience Is ArabicSciencein Renaissance Europe?"at http://www.columbia. edu/-gasi/project/visions/casei/
sci.i.html (accessedMay14, 2004). 66. An initial inquiry into this dimension is addressedin Bier,"Number,Shapeand the Nature of Space." 67. Serjeant,Islamic Textiles, 123-34. 68. Illustratedin JonathanBloom and Sheila Blair,IslamicArts,Arts and Ideas (London:Phaidon,1997), pl. 84. 69. David Stronachand T. CuylerYoungJr., "ThreeSeljukTombTowers,"Iran4 (1966): 1-20;
Bier, "Geometry and the
Interpretationof Meaning;"for color illustrations,see S. P. Seherr-Thossand H. C. Seherr-Thoss,Designand Colorin 194
CAROL BIER
70. Illustratedin Seherr-Thossand SeherrThoss, Design and Color, 30-36. 71. Illustratedin RichardEttinghausen,Oleg Grabar,and MarilynJenkins-Madina, IslamicArtandArchitecture,650-1250 (New Havenand London:YaleUniversity Press,
2001),
figs. 231-34.
Illustratedin Bloom and Blair,Islamic Arts,pl. 80. 73. Fordiscussion of symmetry and pattern in relationto Islamicarts, see Bier, "Symmetryand Pattern." 74. Bier,"Number,Shape,and the Natureof Space." 75. Alternativeexplanationsfor the dramatic proliferationof geometricforms have been offeredby GtilriiNecipoglu, The TopkapiScroll:Geometryand Ornament in IslamicArchitecture(SantaMonica,
72.
Calif.: Getty Trust Publications, 1995);
YasserTabbaa,The Transformation of IslamicArtduringtheSunniRevival (Seattle:UniversityofWashingtonPress, 2001); J.M. Rogers,"Thelith Century:A TurningPointin the Architectureof the Mashriq?"in lslamic Civilization,95oi15o, Paperson IslamicHistoryIII,ed. D. S. Richards (Oxford: Cassirer, 1973),211-49.
76. Kuishyaribn Labban,Principlesof Hindu Reckoning,34-35. 77. Al-Farabl,quoted in Kuishyaribn Labban, Principlesof HinduReckoning,x. 78. Nasir Khusrau,Knowledgeand Liberation:A Treatiseon PhilosophicalTheology, ed. and trans. FaquirM. Hunzai (London and New York:I. B. Tauris,1998):56. See also note 80 below. 79. Nasir Khusrau,Knowledgeand Liberation, 57. 80. PhyllisGranoff,"LuxuryGoods and IntellectualHistory:The Case of Printed and WovenMulticoloredCloth in MedievalIndia,"in this volume, 157.