Islam, Archaeology and Slavery in Africa Author(s): J. Alexander Source: World Archaeology, Vol. 33, No. 1, The Archaeology of Slavery (Jun., 2001), pp. 44-60 Published by: Taylor & Francis, Ltd. Stable URL: http://www.jstor.org/stable/827888 . Accessed: 23/06/2013 04:57 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp
. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to World Archaeology.
http://www.jstor.org
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
Islam, archaeology and slaveryin Africa J.Alexander
Abstract were andIslamicreligions, bytheChristian thosepermitted Twodifferent typesofchattelslavery, slavetradeto theAmericashas been studiedby intoAfricabutonlytheChristian introduced (over1000years)oftheIslamicslavetradeto Asia and The muchlongerduration archaeologists. andeyewitness knownonlyfromliterary oftheDar el IslaminNorthandEastAfricais atpresent willhaveto be and newtechniques archaeologically to recognise accounts.It willprovedifficult whichexisted ofslavery forms willbe theindigenous to recognise developed.Evenmoredifficult and Islam,andtheinteraction at thecomingofbothChristianity in manypartsofthecontinent conceptson whichtheywerebased. betweenthethreedifferent
Keywords Dar el Islam;Dar el Mu'haa;Dar el Harb;Bilades Sudan;Zanj;Jihad. Chattelslavery;
Slaveryis a termused so looselyin European languagesand Christiansocietiesthatonly by carefuldefinitioncan it be used in studyinghuman relationshipsthroughoutthe world. A correct identificationof the word's correspondencewith terms in other languages is essential,especially when an attempthas to be made, as it must be by archaeologists,to identifyslaveryby a studyof materialremains.The problem is a one in AfricawhereChristianand Islamic conceptsofslavery,both difficult particularly and complexlegal systemsand imposedfromoutsideon large areas definedin different of the continent,interactedwitha varietyof indigenousAfricanconceptstransmitted onlyby oral traditionsbeforecontactswithArabic or European visitorsadded a new source. The interactionof Islamic and indigenousconcepts resultedin a numberof partial assimilations more easily recognized from literarythan material cultural evidence. The presentinabilityof archaeologiststo recognizeslaveryand its effecton societieswithoutusingliteraryevidenceremainsone of thelast major fieldproblemsof the discipline(Insoll 1998).
Frar5
ofSlavery WorldArchaeologyVol.33(1):44-60 TheArchaeology online print/1470-1375 C 2001Taylor& FrancisLtdISSN 0043-8243 10.1080/00438240120047627 ~~~~~~~~~~~DOI:
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
Islam,archaeologyand slaveryinAfrica 45 Here only the strictestdefinitionof one kind of slavery,'chattel slavery',will be considered.'A slave is a human being who is the propertyof, and entirelysubject to, anotherhumanbeing underthe religious,social and legal conventionsof the societyin whichhe or she lives.'Being 'the propertyof' meansthatan owner,restricted onlyby the conventionsof his society,is able to buy,sell,free,adopt,ill-treator killhis slave whose childrenbelongto theirownerand can be treatedin thesame way.A slave has no freedom or personalrightsand can become one voluntarily, by a legal decisionor by force.The generalpatternof Islamicslaveryin Africahas been mostrecentlysummarizedby Kelly (1996) and Levtzionand Pouwels (2000), butsee also Willis(1985). Beforeaddressingthe evidencefromAfrica,a generalappreciationof the Islamic conceptof slaveryis necessaryforit differsfromthe Christianone. Quranic teachingfromthe firstdistinguished between the Dar el Islam (the land where its inhabitantshave made theirsubmission (Islam) to theMuslimfaith)and the Dar el Harb (wheretheyhave not). Slavery,including chattelslavery,was permittedby the Holy Quran and furtherdefinedin the Hadith (traditionsof the Prophet Mohammed's lifetime).It became codifiedin the Shari'a (sacred) law codes administeredby membersof the Ulama (those learned in its interpretation).In general,its conceptof slaveryfollowedthe practiceof the Roman Empire and itspredecessorsin WesternAsia (Snowden 1970), the main difference beingthat,in the Dar el Islam, Christianity and Judaismwere accepted as permissibleif incomplete religions.Believers in them,if theymade their(non-religious)submissionto Islamic rulers,paid the requiredtaxes and accepted Shari'a law, were citizens(Dhimi) and not subject to chattelslaveryunless condemnedto it for crimescarryingthat punishment underShari'a law in the same way as Muslimswere. The distinctionbetween the Dar al Harb and Dar al Islam had profoundeffectsin Africa.Inhabitantsof the Dar el Harb (Figs 1-3), whichMuslimswere underobligation to conquer and incorporateinto the Dar el Islam, could be enslaved,although,if individualsvoluntarilyaccepted the Muslimfaith,theycould mitigatetheirstatus,although, unlessmanumitted, theyremainedslaves (Fisher and Fisher1970). DuringthefirstMuslimpenetrationof Africain theseventhcenturyAD theseconcepts were put into practice(Lovejoy 1983). The sedentarypopulationsof Northand Northeast Africa had long professedChristianity withinthe provincesof the Late Roman Empire,whichstretchedfromEgyptto Morocco and outsideit in theMiddle Nile Valley and in modernEritrea/Ethiopia. Nomadic transhumant Berber and Beja communities southofthenortherncoastalplainsor awayfromtheNile Valleyhad been a littleaffected by Christianity and Judaismbutremainedlargelyanimists(Brettand Fentress1996;Paul 1954). The submission(Islam) of the various Roman provincesto the Muslim Arab invadersmeantthattheirChristianinhabitantswere accepted into the Dar el Islam and not subjectedto chattelslavery.Desert-dwellinganimistswere part of the Dar el Harb, and could be enslavedso thattheboundarybetweentheDar el Harb and Islam in North WestAfricaroughlyfollowedthe old Roman (Christian)frontier. A slave tradebringing Saharans and sub-Saharansthroughthe desertto NorthAfrica,whichexistedin Roman times,continuedand documentary evidencein theNile Valleyshowsitto have been regulated thereby treaty.In succeedingcenturiesthe desertrouteswere increasingly used as camel nomadismbecame commonerand the frontierof the Dar el Harb was pushed further and further southwardsuntilin the twelfthto thirteenth centuriesit had reached
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
46 J.Alexander theWestAfricanforestzone and thesub-Saharansavannahsbecame knownas theBilad es Sudan (the land of the blacks) and the source of slaves. The natureof thistrade and the use of slaves in thesavannahswillbe discussedbelow. The second Muslimpenetrationof Africafolloweda different course. There was no militaryconquestby Arabs of the Africancoast of the Red Sea southof Egypt,but here and beyondtheHorn thepre-Islamicsea-bornetradecontinuedalongtheSomali,Kenya, Tanzania and Mozambique coaststo Madagascar (Alpers 1975;Chaudhuri1990). Muslim tradingpostsexistedfromat least theeighthcenturyAD on thiscoast (Horton 1996) and, except on the Eritrean/Ethiopian coast, Muslims were in direct contact with Negro animistswho could legallybe subjectedto chattelslavery.Small-scaleslave tradingmay well have begun at thistime.The developmentof a coastal MuslimKi-swahili-speaking societyand its slave tradein theinteriorwillbe consideredbelow.
Slaveryand the slave tradein NorthAfrica Afterthe Arab conquest of the NorthAfricanprovincesof the East Roman Empire in the seventhto eighthcenturiesAD, chattelslavery,whichhad alwayshad legal existence underthatEmpire,continuedto be legal underIslamicreligious(Shari'a) law. The new social and governmental systemsintroducedchangedthepatternof slave recruitment as did thespreadofIslam amongpeoples,especiallyBerbersand Bejas, who had neverbeen withinthe Roman Empire (Brett and Fentriss1996). There continuedto be a need for slaves in the traditionalcategoriesin whichtheyhad been employed:domesticservice, artisanindustriesand mineralextraction.In thesucceedingcenturiestheuse agriculture, of chattel-slavesas soldiers and in agricultureamong the nomadic pastoralistsof the Sahara were importantnew developments.The importanceof the camel-keepingArab, Beja- and Berber-speakingtribeswho came, at least six centuriesafterthe introduction of thecamel (Shaw 1979),to filltheemptyecologicalnichesin thesahels northand south ofthedesertand theoases withinitcannotbe over-emphasized. For overa thousandyears aftertheiracceptanceof Islam theydeveloped and controlledthedeserttransitroutesof the slave tradeand by theirspread throughthe sub-Saharansahel steadilyadvanced the boundarybetweenDar el Islam and theDar el Harb southwards,untilthebordersof the forestswere reachedin the thirteenth to fifteenth centuriesAD. Non-Muslimenclavesin the savannahscontinuedto be raided and the sourcesof slaves enlargedinto the eighteenthand nineteenthcenturiesby 'Jihads'undertakenforreligiousreasons. The boundarybetweentheDar el Islam and theDar el Harb,in termsoftheslave trade, was moreimportantthanin Roman timesbecause,in religiousterms,itmarkedthezones in whichtherecruitment of slaves by forcewas eitherpermittedor forbidden.Although sometimesignored,thisdistinction was alwaysmaintainedand resultedin thepenetration of the sub-Saharansavannahsand so greatan increasein the tradein slaves throughthe desertsthat,althoughthenumberscannotbe preciselydefined,theyare thoughtto have reached,overthethousandyearsofthetrade,millions(Manning1981;Lovejoy 1983). Its effecton the sedentarynegroidpopulationsof the savannahsand on the Dar el Islam rivals that of the Christianslave trade to the Americas after1600 AD. It can best be analysedbyconsideringthesix typesofslave requiredfordifferent typesofemployment,
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
Islam,archaeologyand slaveryinAfrica 47 Tunis
Miles 0
100 200 300 400
i
\
A) ? Salt
,P .^Fez Marrakech Mogador -.la / Trudn
~
W
/
Arul ts
Chinguetti
--------Bakel.
toot
/
804
|
S
X WEZZX
--TEH
FEZZAN Murzuk
Ghat AHAGGAR
/
, Iferuan
t
I--
say
v
/
Tripoli
in Salah
--...4
? /,Taodeni
'Waden A..--
~ ~~
TUAT
Tenduf ITaghaza '
Tuggurt Gdme
Q~~~~~~~~~~\\ZZ
ldjil~
l
Ylemcen, Laghouat~
AD
fdAgades Ia1300 , KANEM
X
a
->-> ~'''K"" BORNUX
Figure]1NorthAfrica:The routesoftheslavetradeandtheapproximate boundary oftheDar el IslamandtheDar el Harb.
thewaysof obtainingthemand theeffecttheyhad on different partsof theDar el Islam (Fig. 1) at different periods. Typesofslave required Domesticservice The social organizationof the Islamic familyled to an increasein the numberof slave women and childrenin households.The richerhouseholdmightnow includeup to four wivesand theirchildren,slave concubinesand theirchildrenand manyslave servantsfor the more onerous duties of cereal-grinding, water-carrying, cooking, etc. (Klein and Robertson1983). As in earliertimes,male slave attendantsand guardsin largenumbers added to the prestigeof the male householder.The house plans of the sedentarypopulationsdemonstratethisthroughto thepresentdate,withthe Hoch (enclosed courtyard) containingthe Diwan (male entertaining rooms) and the Harim (privateand women's quarters).Quranic injunctionrequiredthatdomesticslaves be kindlytreated,theirchildren recognizedas legitimateand theireventual manumissionencouraged. Although theiracceptance of Islam was permitted,thatdid not removetheirchattel-slavestatus.
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
48 J Alexander Among the Muslimnomads and sub-Saharanindigenoussedentarysocieties,who from the tenthcenturyincreasingly accepted Islam, the rigorousenforcement of fullreligious (Shari'a) law was long, if passivelyresisted,women remainingfreerand unveiledand polygamywidespread. Agriculture Large numbersof chattel-slaves(as opposed to variouskindsof semi-slavery) were used in cultivationand in animalhusbandry.The kindof employment variedthroughtimeand place. In thenortherncoastal plainsthesedentarycommunities used slavesbothin smallholdingsand on largerestates;here men,preferablystrongyoungmen,were required. More were probablyrequiredthereforesince the Arab/nomadicdisdainforagricultural workbecame widespread.From thefifteenth centuryAD onwards,plantationagriculture can be recognizednorthand south of the Sahara, while in the northmore slaves were obtainedto exportto plantationsin SouthernEurope and theLevant.A special developmentin the Islamic periodwas the cultivation, particularly of date palms and cereals,in theoases oftheSahara made availablebycamelpastoralismand essentialbothfornomad diet and for feedingcaravans,especiallyslave caravans. Controlledby nomad tribes, whose males especiallydespised manual work,the oases were dependentupon slave labour,requiringagile and strongyoungmen.Theywere also used in animalhusbandry, especiallysmall stockbut also camels. Anotherdevelopmentwas in plantationagriculturein thenorthplains,especiallysugarcane forexportin thefifteenth century. Mineralextraction Whilefreemen mighttake partin mining/quarrying, especiallyin theeasterndesertsfor gold and preciousstones,the bulk of the workwas done by chattel-slavesin both state enterprisesand privateones. Here too activemaleswere required.More wereemployed, in the Sahara, in salt extraction,especiallyat sites like Taghata and Bilma whereslave colonies providedsalt forthe extensivetrade throughthe sub-Saharansavannahpopulations,especiallyin the sixteenthto nineteenthcenturiesAD (Levtzion and Hopkins 1981). Soldiersand sailors of thearmiesof some Islamicrulersand Slave soldiersbecame a particularcharacteristic mustbe distintook variousforms(Pipes 1981). In NorthAfricatheuse of chattel-slaves nomads and their and from established with Arab Berber client guished relationships employmentas mercenaries.Small numbersof chattel-slaves(al Sudani) fromsouthof the Sahara seem to have been employedfromthe eighthcenturyAD, but in the tenthto forceused in eleventhcenturiesAD theFatimiddynastydeveloped a largeslave infantry to seventeenthcenturiesAD theywere used in theMahgreb Egypt,and in thefourteenth (Johnson1992). The developmentof the Mamluk Turkishsystemof recruitmentin WesternAsia in the tenthcenturyAD was a modifiedformof chattelslaveryas was the Devsirme systemof the fifteenthto seventeenthcenturiesin the Ottoman Turkish sultanate,but,sinceneithersystemrecruitedin Africa,no tradein sub-Saharanmales for soldierstook place in the OttomanTurkisheyaletsbeforethe nineteenthcentury.The attemptof MohammedAli Pasha of Egyptin 1823-40AD to recruita chattel-slavearmy
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
Islam,archaeologyand slaveryinAfrica 49 by conqueringin the Fung Sultanatein the Middle Nile Valley and reachingthe slaving groundsof the Dar el Harb failed,although,untilthe 1880s AD, it resultedin a greatly increasedprivateslave trade. Muslim fleetsin the Mediterranean,like Christianones, relied mainlyon oars and requirednumbersof galley slaves. While criminalsand war captiveswere used, chattel-slaves,active males, whiteor black,were also used (AbunNasr 1987). Industryand commerce Continuingearlierpractices,chattel-slaveswere trainedand used in artisanworkshops, males in metalworking, woodworkingand potteries,femalesin textileones. Males were also employedas agentsand assistantsin commerce. Administration Male slaves, includingeunuchs,were used aftertrainingin many states as officials althoughthenumberswere neverlarge.
Muslimslavery inNorthandWestAfrica The recruitment of chattel-slaves forN W Africa The acceptanceof Islam,howeverincomplete,by Saharan nomadsin the eighthto ninth centuriesAD took the boundariesof the Dar el Harb to the sub-Saharansahel and gave access to the sedentarynon-Muslimpopulationsof the savannahs.Three phases in the tradein slaves northwardscan be recognized. From the eighthto the fifteenth centuries,althoughraidsby Muslimnomadsforslaves took place, morewere obtainedby tradewiththe indigenouskingdomsof the savannahs, notablythelong-established Ghana and thesubsequentMalian Empires.Captivestakenin theirlocal warsweremarchedto themarketslike Gao or Aghordator thecapitalofGhana in or nearthebordersof thesahel wheretheywereboughtbyMuslimmerchants fromthe north.The situationis bestreportedon by Ibn Battuta(1962) in theninthcenturyAD and Ibn Khaldun(1986),thewell-built Muslimsettlements excavatedat KumbehSaleh (Berthier 1997) and Al Bakri (1913) showingthenorthern connection(Levtzionand Hopkins1981). An increasingacceptanceof Islam in the indigenousstates,culminatingin conversion in the SonghayEmpire (eleventhto fifteenth centuriesAD), led to an increasinguse of chattel-slavesin the savannahs. The Moroccan invasion of the Songhay Empire in 1583 and its success led to an increasedtrade northwardsin slaves withnew routesbeing opened across the deserts. This seems to have maintainedforsome two hundredyears and even increasedin the nineteenthcenturywhenEuropean and Americanfleetsfirstreducedand thenputan end to NorthAfricanpiracyin theMediterranean(Hogendorm1993).This stoppedthesupply of whiteslaves,capturedor kidnappedfromtheDar el Harb of SouthernEurope, which had long supplementedthe black slaves. Slaves sold in the marketsof the sub-Saharansahel now faceda 1,000kmmarchnorthwardsthroughdesertsin whichtherewere fewwateringplaces or food sources,an ordeal
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
50 J Alexander as thesea passagefromWestAfricato theAmericasand causinga similar quiteas traumatic to nineteenth are availablebeforetheeighteenth rate(Devisse 1988).No statistics mortality and thenonlypartialones,buta 50 percentdeathrate,moreinthecase ofwomen centuries, seemsto havebeen normal(Fisher1975;Baet 1967) and wholecaravanscould and children, or a sandstorm(Levtzionand Hopkins1981). be lostbymiscalculation Controlof travellingacross the desertswas in the hands of Berber and Arab camel pastoralistsfor whom raidingor protectingcaravans was one of theirfew sources of income.The oases were undertheircontroland agricultureat themcarriedout by their times,thosefromDarfur,Kanem slaves,dependingon thesecurityof routesat different and Bornu to Egypt (O'Fahey 1973) and Cyrenaicain the east, throughAwjila to the Fezzan and Bilma,and fromTimbuctu,Gao and theMiddleNigerValleyto Morocco and theMahgrebvia Ghar and Tuat (Bovill 1958) in thewest.The routesremainedin use into thenineteenthcentury(Cordell 1985). forNE. Africa of chattel-slaves The recruitment At the comingof the Muslim Arab armyin the seventhcenturyAD slaverycertainly existedin the Roman provinceof Egyptand in the independentChristiankingdomsof Nobatia-Maqurraand Alodia (Alw'a) in theMiddle Nile Valley.There is no evidenceof theirnumbersbutthesmallannualtributeof350 slavesdemandedofMaqurraintheBaqt treatyof thatcenturysuggeststhatit was consideredas Dar al-Mu'haa (compromise),or Dar al-Sulhas it is oftencalled in theSudan, and not theDar al-Harb.Slaves were probably obtained fromAlodia, whichwithits capital near modernKhartoumwould have 'pagans' ofthesavannahs.The position been able to raidamongthesedentaryagricultural fromthatin NorthWestAfricafora directlinkwiththesavannahswas was verydifferent alreadywell establishedbeforecamel nomadismopened up the desertroutes,although dividedthe journey,at least fortydaysof travelling, the same problemof a 1,000km-long sourcesof slaves fromthe Egyptianmarkets. The firstMuslimrulersof Egyptpreservedthissituationby concluding,aftera failure treatywithNobatia-Maqurra.Its name,theBaqt (Pactum) to conquerit,a non-aggression and in effectadmittedthe suggestsa continuationoftheRoman policyofclient-kingdoms on submission.The legalityofthisdecision kingdomintotheDar el Islamwithoutinsisting was muchdebatedbylaterMuslimjurists(Spaulding1995). Since thismeantthattheDar el Harb lay farto the southand depended on tradewithand throughone, perhapstwo Christiankingdoms,theBaqt tributeof 350 slaves musthave been, by thetenthcentury, supplementedbyprivatetrade,forblack slave troopswerebythena powerin Egyptand in the eleventhcenturyAl Mustansirreported30,000 Fatimid'black' troopsin Cairo. These probablyincludedthemercenaryBerber(Katama) cavalryfromtheMahgreb,but were describedas 'al Nuba' or 'el Sudani' and musthave come fromthe the infantry Middle Nile Basin savannahssince theywere said to 'have come froma regionsouthof Nubia withlarge pasturesand strongpeople' (Hrbek 1977:70). The presenceof Muslim merchantquartersin thecapitalsof bothMaqurra and Alodia (Alw'a) probablymeanta privateslave tradeexistedduringtheFatimiddynasty(973-1090) in Egypt(Brett1978), althoughthecommercialcorrespondenceofthisperiodfoundat Qasr Ibrim(Sartainpers. com.) containedno mentionof it.
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
Islam,archaeologyand slaveryinAfrica 51 45Y
40Y
35Y
30Y
25Y
Metres C~~~airo
- -
30Y -
2000
1000
*-.cjoP'D-?
~~~~~~~~~~~~~500
-
-- ---Marsh
LIBYAN DESERT
500
0 Miles
25YF 1~~~~~~~~~sCat
c
00 AD
-
P~~~~~~~/
XN
Dongla.
20Y/
I th
i
/ ~~~~~~~~~~~6thc~at/-
/
ni Kh~~~~~~~~~~~~artoum
/
-
/
Elobei
/J.Marra
acceptance sta
.-.(
4 Cat-AbuHamed
Cataract. Aftertheninth-century
15
C
~
1400m ie sr ~th ~opt
~
o
eeepotdb ~ swa(a~ ~~5t 00 X
/
ulmeternus(AD Cat
tgii19)
oftheDar boundaries Figure2 TheNileValley:theroutesoftheslavetradeandtheapproximate el Islam,Dar el MuhaAandtheDar el Harb.
wereneeded,especiallyfromtheninthto thirteenth Chattel-slaves centuries,in thegold and emerald (carbuncle)minesof the Wadi Allaqi in the desertseast of the Nile's 2nd AD victoriesof al-Umariover Beja nomads and their Cataract.Afterthe ninth-century acceptanceofIslam,themineswereexploitedbyMuslimentrepreneurs (Castiglioni1998). Slave labourwas used and littlecould be obtainedlocally,especiallyas nomadicmigration southwardsin search of pastureswas takingplace (Paul 1954). These migrationstook
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
52 J.Alexander camel-keepingMuslim pastoraliststhroughthe ChristianKingdomof Alodia (Al'wa), perhapsitselfnow partof theDar el Harb, intothe sahel and savannahsof Kordofanand Darfur.Here sedentaryNubian-speakinginhabitantscould be enslavedand marchedfor fortydaysthroughthe desertor riverroutesto Egypt.An alternativeroute,the 'Tariq el Sudan' whichran west-eastto the Red Sea coast throughthe savannahs,was muchused centuryAD and also by slave caravans,especially bypilgrimsto Mekka afterthefifteenth fromEthiopiawhichalso lay in theDar el Harb (Abir 1985). Fromthecoast slavescould be shippedto Egypt,Arabia and beyond.The Nile Valleyrouteseems to have been little of Maqurra into a series of small AD disintegration used afterthe thirteenth-century Muslimmekdoms(kingdoms)untilthesixteenthcenturyAD. Politicaland technologicalchangesin thesixteenthto nineteenthcenturiesAD brought yearsthe substantialchangesto slave tradingin theNile Basin. For twohundredand fifty Nile valleyas farsouthas 3rdCataract(600kmnorthofpresent-dayKhartoum)was part sultanateand theregionsouthofittheindigenousFungsultanate oftheOttoman-Turkish withits capital at Sennaron the Blue Nile (Alexander2000). The Dar el Harb was now, as inWesternAfrica,in thesouthernsavannahsofKordofanand Darfur,and slave traders penetratethe greatswampsof the Sudd and century, did not,beforethe mid-nineteenth Lake No and reachthe forestsof Uganda. Ottoman and Fung demand for slaves differedgreatly.Ottoman/Mamlukmilitary ofAfricans,a prohibitionwhichlastedintothe policyforbadetheenlistment recruitment nineteenthcenturyAD. As a result,the tradewas in those suitablefordomesticservice, mainly young females and children and males for agriculturalwork. In the Fung sultanate,on whose bordersthe raidingforslaves mostlytook place, an armyof slaves was maintainedforuse in local warsand slaveswere settledin villagesto undertakefood production(Bruce 1798). A trade to the northalso continued.Althoughfirearmswere known,theiruse was restrictedbefore1823 AD when MohammedAli Pasha, the virtuallyindependentgovernorof OttomanEgypt,sentan armywell-equippedwithfirearms to conquer the Fung sultanate,its main aim being to obtain youngmale chattel-slaves fromDar el Harb forhis army(Prunier1992). For the firsttimefirearmsappeared in large numbersin the easternsavannahsand in the nextthirtyyearsmanythousandsof men were obtainedforthe state and privateslave-raidingincreased(O'Fahey 1973). In the 1840s the swampsof the southwere penetratedand the vast regionbetweenthem and theforestsofUganda opened up to slave raiding;itwas in Uganda thatMuslimslaveraidersfromNorthand East Africametin thelaternineteenthcentury(Grey 1961). The of transportation exportof slaves to the northwas limitedby the increaseddifficulties and ended withthe suppressionof the trade between 1880 and 1930 (Spaulding 1988; Toledano 1982).
Muslimslaveryin East Africa markets This existedthroughthe same centuriesas in NorthAfricabut serveddifferent times(Alpers ways at different in the Dar el Islam and had to be organizedin different 1975; Chaudhuri 1990; Chittickand Rotberg 1980). There was no land connection traveller'stale is to be betweenthe northand the east, unless a dubious ninth-century
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
Islam,archaeologyand slaveryinAfrica 53
z
|
}
Bildd
al-Habasha> al-Habasha
A
Q\+
/1 \t
'v
Z~~Iah/Zaia ila ZAlagh/Za Qaljur
~
Aqant Bdqati Bata* -
Bilad al-Hawiya
4'
Marka
/ *~nga.ja
r
/>9
v
c~~~
~~. S
Dandama
n-Najd *Qarnua
\Baz a
Baduna(Mo dishb| )
1300A
0-.
(Gd/
i/
/~~~~~~Kh
Jabal Ajrad
WaqWa
I,
B9wari
(Gedi?)rlsls~a
Manbasa
Tuhna
sbtShrbaJzrt g~~~~~ " ........... *\&Bukha Muan-Ndama -~n ,./ Ka~ 130 ADZ A/
I
IslaJmaAndad te DanrbUlH
Bukha
alKh|r
lva
,~~jnipOPOR,
an-Naddma*Danam
Isla an th
Jabael Ha
~~~~~~~Soqotrd
Barma Hafu Ras Hdfi~n
Juwa ,~a
&
/
c:
Qarfuna
Jaost
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
'VG
(
54
. Alexander
believed (Buguryibn Sharriar,quoted in Lewis 1977),no Christianstatesto subdue and incorporateintotheDar el Islam,no camelpasturessouthofSomalia suitableformigratingArab nomads and, insteadof immediatecontactwiththe Dar el Harb called Zanj, it could be reached onlyby long sea journeys,utilizingthe monsoonwinds,fromthe Red ofZanj weresedenSea and theeast Arabiancoastsand thePersianGulf.The inhabitants tarynegro animistswho could, under Shar'ia law, be enslaved, as Al Idrisi's twelfthcenturyaccountdescribes(quoted by Lewis 1977: 117-20). Maritimetrade withthe coast had existedformanycenturies(Chittickand Rotberg 1980;Galaal 1980),butMuslimmerchantsappear to have been thefirstto establish,from the eighthcenturyAD onwards,permanenttradingposts which soon developed into townswithmosques,law courtsand thepalaces ofsmallsultanates(Horton 1996). Their strongestlinkswerewithMuscat,Oman and thePersianGulf.The tradingpatternin the ninthcenturyAD includedslaves takenfromSofala at the mouthof the Zambesi River to Pemba and the Lamu archipelago(Sirajal at Muluk,quoted by Lewis 1977: 212-13). The mostpowerfulstatewas Kilwa,largenumbersofmale slaves beingshippedto Basra to work in the irrigationprojects in southernIraq, theirrevolt there causing much destruction(Wright1993). Others,mainlywomenand children,were takenthereand to Arabia fordomesticservice.They appear to have been broughtto the coast by indigenous rulersin the immediatehinterlandand to have been obtainedin local wars.There is no mentionof long-distanceMuslimpenetrationof the interior(Sperling2000), but on thecoast a seriesofmixedArab-Zanj communitiesspeakingKi-swahili,an Arabic/Bantu language,developed, the language becomingthe lingua francaof the whole coast and the Indian Ocean (Prinz 1968). They,ratherthanthe Arabs, became themerchantsand agentsbringslaves to the coast and sellingthem.This state of affairslasted untilc.1500 and reduced by the arrivalin 1543 of the Portuguese, AD but was abruptlyinterrupted establishedtheircontrolof the whole coast who,withtheirsuperiorshipsand firearms, and destroyedor occupied many towns (Duffy1963). Slave employmentand trading coast,especicontinued,theMuslimtradebeingconcentratedon theKenyan/Tanzanian ally on the Lamu archipelago.Portuguesepowerweakened in the seventeenthto eighteenth centuries,the whole Kenyan-Tanzaniancoast and Zanzibar graduallycoming underOmani control,and thisculminatedin the fallof FortJesusat Mombassa in 1837 (de Cardi 1970). The transferof the capital of the Omani sultanateto Zanzibar Stone Town (Abdul Sherrif1997) in 1832 led to a big increasein demandforchattel-slaves.On Zanzibar,plantationdevelopmentsforgrowingspices,especiallycloves,restedon slave labour and therewas an increasedexportofslaves to Arabia and beyond(Cooper 1977). parties,well The tradein slaves and ivorywas now suppliedby Muslimraiding/trading who penetratedfarintotheinteriorofthecontinent(Abungo and suppliedwithfirearms, Matturo1993), Arabs fromTabora in thesouthreachingBuganda in the 1820s (Posnansky1977:218) and Rwanda in the1870s.Those capturedwerefirstused as portersto carry ivoryto the coast and then sold on in the urban marketsthere;a contemporaryArab proverbclaimed that 'when you play the flutein Zanzibar, Africaas far as the Lakes naval patrolsreducedthesea-bornetradefrom dances' (Lodhi 1974). Britishanti-slavery the 1840s onwardsand it was formallyended in the laternineteenthcenturyAD (Abdul Sherrif1990).
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
Islam, archaeology and slavery in Africa
II~~~~~~~~~~~~~q
Figure4a Slave caravan,East Africa
Fi g ure _bSl a ve
E ast A frica
Figulre4b Slave dhow,East Africa
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
55
56 J.Alexander
Therecognition ofIslamicchattel-slavery fromarchaeological evidence The near-impossibility, in the presentstate of fieldtechniques,of recognizingchattelslaveryfrommaterialremainsunassociatedwithdocumentary evidencehas alreadybeen commentedon and can be summarizedas follows. The archaeologicalevidenceforIslamicslavery In theDar el Harb Proofofslave takingor raiding?
Arguments against
Destruction ofindigenous settlements. Possibly dueto otherreasons. Reduction ofdensity ofsettlement inregion. Increasedfortification ofsettlements orchange to defensive sites. Possiblecamps/slave collection pointswith Notfirmly attributable to slavers. Islamicartefacts. ofMuslimartefacts inindigenous Discovery sites. Notfirmly attributable toslavetrade.Even neck,hands,waistandankleironscouldbe forwarcaptivesorcriminals. In Dar el Islam Proofsofslave-trading or employment?
Slavecaravans: nightly campsorenclosures, possiblewith humanremains. intownsnorthof Slavequarters andmarkets theSaharaoronEastAfricacoast. Moreelaboratequarters inseaports. Domesticemployment ofslaves. ofslaves. Military employment ofslaves. Industrial/mining employment Slavecemeteries.
inNorthWestAfricaif Mightbe recognized datable. Notlikelytobe recognizable archaeologically Notarchaeologically or recognizable distinguishable fromprisons. Archaeologically unrecognizable. ThosewhobecomeMuslims buriedinMuslim otherscasuallyburied.No wayof cemeteries, freemigrants fromslaves. distinguishing
This means thatmajor problemsin the studyof the threeculturalconceptsof slavery in Africa(indigenous,Muslimand Christian)have yetto be solved.In thecase ofindigenous slaverythisis particularly difficult since it is knownto have existedin manyregions whenMuslimsand Christiansmade theirfirstcontactswiththe inhabitants. But, sincein oral sub-SaharanAfricathese were non-literate, only archaeologicallyunsubstantiated withsucha tradition traditions suggeststhatslaveryhad longexisted.Since theinteraction is essentialto underby the incomingMuslimand Christianconceptsof chattel-slavery in archaeologicalfield itcan onlybe solvedbyimprovements standinglaterdevelopments, techniques. in the Dar el Harb in Africalies in the The importanceof indigenouschattel-slavery partitplayedin supplyingtheDar el Islam withslaves.Stateformation longprecededthe
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
Islam,archaeologyand slaveryinAfrica 57 arrivalof the Arab armiesin NorthAfricaor Muslimmerchantsin East Africa.Beyond the Roman Christianprovincesof the formerand the Muslim towns and Ki-swahili communities ofthelatterlaythestateswhosewarsor Muslimraidscould providethenonMuslimslaves required. The failureof archaeologicalevidence at presentto recognizethisrelationshipor to identify withinthe Dar el Islam has been analysedabove and means that, chattel-slavery whilein the Dar el Harb the exportof small numbersof men,women and childrenmay well alwayspass unnoticed,large-scaleextractionmightbe recognizedin thefuturefrom the widespread and contemporarydestructionof settlements;withoutdocumentary evidencethiscannotat presentbe confirmed. The assemblingofcaravansand theirmarch to Muslimslave marketsat the boundariesof theDar el Islam,even whentheroutesare well known,cannot at presentbe linkedto slavery,althoughtheycould be to trade of some kind.The findingof artefactsfromthe Islamic worldmightalso indicatetrade,or conversion,but no evidenceof slavery. InsidetheDar el Islam routestakenbyslave caravansand thelocationofslave markets are known fromliterarysources but cannot be supportedfromarchaeological ones, althoughdhowsdesignedfortransport mightbe recognizedfromwrecksand desertroutes by the concentrations of humanbones along them. The same insufficiency of archaeologicalevidenceto recognizeslavery,especiallythe uses made of chattel-slavesin the Dar el Islam,mustbe accepted.Slave soldierscannot be distinguished fromfreeones, agriculturalor buildingprojectscould not be linkedto slave labour withoutdocumentaryevidence,whiledomesticslaverywillprobablynever be identified fromhouse plans or artefacts.Even theuse ofcemeteryevidence,whichhas been so successfulin Americancontexts,is not available in the Dar el Islam since excavationis, quite naturally,forbiddenforMuslimgravesand thereis no evidence of nonMuslimslave cemeteries.New evidence is most likelyto come frommore studyof the Ottoman governmentarchivesin Istanbul,Egypt and Tunisia, and the many private familyarchivesknownto exist,especiallyin the Republic of Sudan. St John'sCollege,Cambridge
References Abir,M. 1985.TheEthiopian slavetradeanditsrelation totheIslamicWorld.InSlavesandSlavery inMuslimAfrica(ed. J.R. Willis).NewJersey: Totowa,pp. 123-36. AbdulSherrif 1990.Slaves,SpacesandIvoryinZanzibar.London:Currey. AbdulSherrif 1997.TheHistory and Conservation ofZanzibarStoneTown.London:Currey. Abun-Nasr, J.M. 1987.A History of theMaghribin theIslamicPeriod.Cambridge: Cambridge Press. University AbungoG. H. 0. andMatturo, H. W. 1993.Coast- interior settlements andsocialrelationships in theKenyacoastalhinterland. In TheArchaeology ofAfrica:Food,Metals,Towns(eds T. Shawet al.). London:Routledge, pp.694-705. Africanus, Leo 1956.Description de l'Afrique. Paris:Adrien- Maisonneuve. Al-Bakri1913.Description de l'Afrique septentionale. Paris:Adrien- Maisonneuve, 1965.
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
58
1 Alexander
in theMiddle Nile Valley: Alexander,J.2000.The archaeologyand historyoftheOttomanfrontier AH 910-1233 (AD 1504-1820).AdumantuI. Riadh. Alpers,E. A. 1975.Ivoryand Slaves in East Africa.London: Heinemann. Austen,R. A. 1992.Egypt-Sudanslaverylists.In TheHuman Commodity(ed. E. Savage). London: SOAS. Berthier,S. 1997.Recherchesarcheologiquessurla capitalede l'empirede Ghana. CambridgeMonographsin AfricanArchaeologyNo. 41. Oxford. Bovill,E. W. 1958. The Golden Tradeof theMoors. London: Macmillan. Brett,M. 1975-6.The journeyof Al-Tijanito Tripoli.Societyof LibyanStudies7thAnnual Report, pp. 41-51. In The CambridgeHistoryofAfrica,Vol. Brett,M. 1978.The FatimidRevolutionand itsaftermath. 2. Cambridge:CambridgeUniversityPress. Brett,M. 1979. Ibn KhaldUnand the Arabisationof NorthAfrica.MaghrebReview,4. Brett,M. 1983.Islam and tradein theBilad al Sudan: 10-11thcenturies.JournalofAfricanHistory, 24: 431-40. Brett,M. and Fentriss,E. 1996. The Berbers.Oxford:OxfordUniversityPress. Bruce,J.1798. Travelsto Discover theSource of theNile. Edinburgh. Castiglioni,A. 1998. L'Eldorado dei Faraoni:alla Scopertadi BerenicePancrisia.Novara: Instituto GeograficoDe Agostini. Chaudhuri,K. N. 1990.Asia beforeEurope: Economyand Tradein theIndian Ocean fromtheRise of Islam to 1756. Cambridge:CambridgeUniversityPress. Chittick,K. H. and Rotberg,R. I. (ed.) 1980. East Africaand theOrient:CulturalSynthesesin Precolonial Times.New York: Holmes & Meier. Cooper, F. 1977.PlantationSlaveryin East Africa.Newhaven,NJ. Cordell,D. D. 1985.Dar el Kutiand theLast Yearsof theTrans-SaharanSlave Trade.Madison,WI: Universityof WisconsinPress. in easternUbangi-Shari. Cordell,D. D. 1986. Warlordsand enslavement:a sample of slave-traders In Africansin Bondage: Studiesin Slaveryand theSlave Trade (ed. P. E. Lovejoy). Madison,WI: Universityof WisconsinPress,pp. 336-65. 44: 288-95. de Cardi,B. 1970.TrucialOman in the 16-17thcenturies.Antiquity, Devisse, J.1988.Trade and traderoutesin WestAfrica.In UNESCO HistoryofAfrica,Vol. 3. Paris: UNESCO. Duffy,J.1963. Portugalin Africa.London: PenguinAfricanLibrary. Dunn, R. E. 1986. TheAdventuresof Ibn Battuta:A Muslim Travellerof the14thCentury. Encyclopaediaof Islam 1954-, 2nd edn. Leiden: Brill. Endam,Y. H. 1966.Slaveryin theOttomanEmpireand Its Demise: 1800-1909.London: Macmillan. Fisher,C. B. and Fisher,H. J.1970. Slaveryand MuslimSocietyin Africa.London: SOAS. Fisher,H. J. 1975. CentralSahara and the Sudan: the contributionof slavery.In The Cambridge HistoryofAfrica,Vol. 4, Cambridge:CambridgeUniversityPress,pp. 97-105. Fisher,H. J.and Conrad,D. C. 1983.The conquestthatneverwas: Ghana and the Almoravids.In UNESCO HistoryofAfrica,Vol. 4, 9: 21-59; 10: 53-78. Paris: UNESCO. Galaal Musa, H. J. 1980. Historicalrelationsbetween the Horn of Africa,the Persian Gulf and the Indian Ocean throughIslam. In UNESCO Historyof Africa,Vol. 3, 111: 23-30. Paris: UNESCO.
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
Islam,archaeologyand slaveryinAfrica 59 Sudan1839-89.Oxford:OxfordUniversity Grey,R. 1961.History oftheSouthern Press. Grey,R. (ed.) 1975.TheCambridge Vol.4. Cambridge: History ofAfrica, Cambridge University Press. Hill,R. 1970.On theFrontiers Press. ofIslam.Oxford:OxfordUniversity E. 1993.Slow Deathfor Slavery:The Courseof Abolitionin Northern Hogendorm, Nigeria (1897-1936).Cambridge: Cambridge University Press. Horton,M. 1996.Shanga:thearchaeology of a Muslimtrading settlement on thecoastof East Africa.British inEastern Institute AfricaMonograph VII. London:BIEA. Hrbek,I. 1977.Egypt, NubiaandtheEasternDeserts.In TheCambridge Vol.3. History ofAfrica, Cambridge: Press. Cambridge University J.0. 1977.BlackAfricans intheIslamicworld.Tarika,5(4). Lagos,Nigeria. Hunwick, HuseynEffendi1966OttomanEgyptin theAge of theFrenchRevolution (trans.B. W. Shaw). MA: HarvardUniversity Press. Cambridge, Ibn Battuta1962. The Travelsof Ibn Battuta,Vol. 2 (trans.Gibb). Cambridge:Cambridge Press. University IbnKhaldun1986.PeuplesetNationsdu Monde.Paris. J.1995.Africans: Iliffe, TheHistory ofa Continent. Cambridge: Press. Cambridge University Insoll,T. 1998.TheArchaeology ofIslam.Oxford:Blackwell. D. H. 1992.Recruitment of privateslave armies.In TheHumanCommodity Johnson, (ed. E. Savage).London:SOAS. Kelly,K. G. 1996.Slave tradein Africa.In Encyclopaedia of Pre-colonial Africa(ed. J.Vogel). WalnutCreek,CA: Altamira Press,pp.532-5. J.1977.Someconclusions Kirkman, fromarchaeological excavations onthecoastofKenya.In East Africaand theOrient:CulturalSyntheses in Pre-colonialTimes(eds H. N. Chittick and R. I. Rotberg).NewYork:Holmes& Meier. Klein,M. andRobertson, C. (eds) 1983.Women andSlavery inAfrica.Madison,WI: University of Wisconsin Press. Levtzion,N. and Hopkins,J.F. P. (eds) 1981.Corpusof EarlyArabicSourcesforWestAfrican History. Cambridge: Cambridge University Press. N. andPouwels,R. (eds) 2000.TheHistory Levtzion, ofIslaminAfrica.Oxford:Currey. Lewis,B. (ed.) 1977.IslamfromtheProphet Muhammad totheCaptureofConstantinople, Vol.2. Oxford:OxfordUniversity Press. Lodhi,A.Y. 1974.TheInstitution ofSlavery inZanzibarandPemba.Uppsala:Scandinavian Institute ofAfrican Studies. Lovejoy,P. E. 1983.A History ofSlaveryinAfrica.Cambridge: Cambridge University Press. Lovejoy,P. E. (ed.) 1986.Africans inBondage:StudiesinSlavery andtheSlaveTrade.Madison,WI: ofWisconsin University Press. McDougall,E. A. 1985.The viewfromAwdaghurst: war,tradeand socialchangein thesouth western Saharafromthe18thto 15thcenturies. Journal ofAfrican History, 26:35-63. P. 1981.The enslavement Manning, ofAfricans: a demographic model.CJAS,15(3). P. 1990.Slavery Manning, andAfrican Life.Cambridge: Cambridge University Press. H. T. 1986.TheArabConquestoftheWestern Norris, Sudan.Harlow:The Press. O'Fahey,R. S. 1973.SlaveryandtheslavetradeinDarfur. Journal ofAfrican History, 14:40-75.
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions
60 f.Alexander Paul, A. 1954.A Historyof theBeja Tribesof theSudan. London: Methuen. Pipes, D. 1981.Slave Soldiersand Islam. New York: Posnansky,M. 1975.The Lacastrinepeoples and the coast. In East Africaand theOrient:Cultural SynthesesinPre-colonialTimes(eds H. N. Chittickand R. I. Rotberg).New York:Holmes & Meier. Prins, A. H. J. (ed.) 1968. The Swahili-Speaking Peoples of Zanzibar and the East AfricanCoast. London: InternationalAfricanInstitute. Prunier,G. 1992. Militaryslaveryin the Sudan Turkiya:1820-85. In The Human Commodity(ed. E. Savage). London: SOAS. Renault,F. 1982. La Traitedes esclaves noiren Libye au XVIII secles.JournalofAfricanHistory, 23(2). Rodney,W. 1975.The Upper Guinea Coast. In The CambridgeHistoryofAfrica,Vol.3. Cambridge: CambridgeUniversityPress. Savage, E. (ed.) 1992. The Human Commodity.London: SOAS. Seligman,C. G. and Seligman,B. Z. 1932.Pagan Tribesof theNiloticSudan. London: Macmillan. Shaw,B. W. 1979.The camel in Roman NorthAfricaand theSahara. Bulletinde l'Institutfranqaise de l'AfriqueNoir,41: 30-58. Shaw,T. etal. (eds) 1993. TheArchaeologyofAfrica:Food, Metals,Towns.London: Routledge. Spaulding,J.1988. The businessof slaveryin the CentralAnglo EgyptianSudan 1910-30.African Economic HistoryReview,17: 23-44. of the Baqt Spaulding,J.1995. Medieval ChristianNubia and the Islamicworld:a reconsideration JournalofAfricanHistoricalStudies,28(3): 577-94. Treaty.International Sperling,D. C. 2000. The coastal hinterlandand interior.In The Historyof Islam in Africa(eds N. Levtzionand R. Pouwels). Oxford:Currey,pp. 274-302. Tamrat,T. 1975. Ethiopia, Red Sea and the Horn. In The CambridgeHistoryof Africa,Vol. 3. Cambridge:CambridgeUniversityPress. Temimi,A. (ed.) 1997. La CultureArabo-Islamiqueen Afriquede l'Ouest. Tunisia: Foundation TeminiZaghouran. Toledano, E. 1982. The OttomanSlave Trade and Its Suppression.Princeton,NJ: Princeton UniversityPress. Vogel,J.(ed.) 1997. Encyclopaediaof Pre-colonialAfrica.WalnutCreek,CA: AltamiraPress. AnnalesIslamologique,15. Waltz,T. 1979.Tradingin the Sudan in the 16thcentury. Willis,J.R. (ed.) 1985.Slaves and Slaveryin MuslimAfrica.Totowa,NJ,pp. 182-98. Wright,H. T. 1993.Trade and politicson the East Africanlittoralof Africa(AD 500-1300). In The ArchaeologyofAfrica:Food, Metals,Towns(eds T. Shaw etal.). London: Routledge,pp. 658-72.
This content downloaded from 163.1.62.81 on Sun, 23 Jun 2013 04:57:47 AM All use subject to JSTOR Terms and Conditions