The Past and Present Society

Hydraulic Society in Ceylon

Author(s): E. R. Leach
Source: Past & Present, No. 15 (Apr., 1959), pp. 2-26
Published by: Oxford University Press on behalf of The Past and Present Society
Stable URL: http://www.jstor.org/stable/649829 .
Accessed: 21/06/2014 22:08

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.

Oxford University Press and The Past and Present Society are collaborating with JSTOR to digitize, preserve
and extend access to Past &Present.

http://www.jstor.org

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON
THIS ESSAY FALLS INTO TWO PARTS. THE FIRST IS A CRITICAL
commentary on thefactthat Karl Wittfogel, in developinghis twin
conceptsof "hydraulicsociety"and "OrientalDespotism"in his
recentstudyOrientalDespotism, has entirelyignoredthe evidence
fromCeylon.The seconddiscussesthe "servicetenureofland" as it
occursin Ceylonand a numberofctherregionsin Asiaandsuggestsa
connection betweenthistypeoftenureand thepuzzlinghistorical fact
thatmanyofthemostsuccessful archaicsocietieswerelocatedin flood
riddenaridterrain in whichagriculturewasonlymadepossiblethrough
theuse ofirrigation engineering.Most readersare likelyto findthe
second(and shorter)partof the argumentmoreinteresting thanthe
first,but, since Wittfogel's thesis,whichI criticise,is also, in part,
an hypothesis as to whyearly"hydraulicsocieties"wereeconomically
successful, my commentaryon Wittfogelforms a necessary
preliminary to thepresentation of myownpositivecontribution.

I
ANCIENT SINHALA AND "ORIENTAL DESPOTISM"

Karl Wittfogel,besidesbeingone oftheworld'sleadingauthorities

on the politicaleconomyof ancientChina,remainsan unrepentant
Marxist- or perhapsone shouldsayan Ur-Marxist. He maintains
thataroundthe year I853 Marx and Engels held views about the
natureof AsiaticSocietywhichwereessentiallycorrect,but thatas
time went on both masters "retrogressed"from their original
positionsof insight. Furthermore in latertimesLenin and Stalin
and theirdiversefollowershave,forpoliticallyopportunist reasons,
wilfullypervertedthe historicaland sociologicaltruthsenshrined
in theearlywritings oftheirmasterMarx. Wittfogel's volumeis an
expositionof thisthesison a grandscale. The argumentis highly
condensed and often excessivelypolemical, but unquestionably
stimulating.
Very roughlythe main thesisruns as follows:Marxismentails,
amongotherthings,a theoryof unilinearsocial evolution. In the
firstdraftsof Das Kapital Marx distinguished four major socio-
economicorders,or modes of production, labelled Asiatic,Ancient
(Classical),Feudal and Capitalist2
forming a sequence. Leninlisted
thesamefourstagesin an articlepublishedin 1914.

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON 3

The special characterof "Asiatic Society"is thatland is owned

by smalllocalisedcommunesbut "climateand territoria
collectively
conditionsmake artificialirrigationby canals and waterworks the
basis of Oriental agriculture". Furthermore water control
"necessitatedthe interference of the centralizingpower of the
government".3Power is thus in the hands of a bureaucracyeven
thoughownershipof the meansof productionis in thehandsof the
people. For Marx,India was thetypical"AsiaticSociety";Chinais
atypicalbecause of the developmentof privatepropertyin land.
TsaristRussia, despitethe absence of an irrigationbase, is "semi-
Asiatic". Engels, in his critique of Diihring,is unambiguous:
"the ancientcommunes,wherethey continuedto exist,have for
thousandsof yearsformedthe basis of the mostbarbarousformof
state,OrientalDespotism,fromIndia to Russia".4
Althoughthe conceptof "Asiatic" or "Oriental"societyretained
itsplaceintheMarxistcanonformanyyearsitdoesnotfitsatisfactorily
into a general evolutionary scheme based on propertyand class.
Engels in his revampingof Morganmanagesto dispensewiththe
"Asiatic"stageofdevelopment altogether:
"Withslavery, whichis civilisation
developedmostfully,thereoccurredthe
firstgreatsplitofsocietyintoan exploitingand an exploitedclass. The
cleavagelastedthroughout thewholeperiodof civilisation.Slaveryis the
firstformof exploitation, whichis specificforthe ancientworld; it was
succeededby serfdom in theMiddleAgesand wagelabourin morerecent
times. These are the threegreatformsof servitude, characteristic
of the
three great epochs of civilisation"5
Elsewherein the same book Engels maintainsthat the stage of
developmentwhichimmediatelyprecededslaverywas one in which:
"the landed proprietor,whose tenant the farmeris, is still regarded by the
latteras a kind of chiefof the clan, whose dutyis to manage the land in the
interestsof all, while the farmerpays tributein the formof rent,but has a
claim upon him forassistancein timesof necessity"6
This lattermode of organisation, whichwould clearlyincludethe
Sinhalese systemof servicetenure describedat the end of this
essay,is placedby Engelsin thestageof "HigherBarbarism". The
archaic-despotic-bureaucracyhas droppedout.
Later Marxistshave fromtimeto timerevivedthe notionthat
Asiatic Society constitutesa distinctformbut latterly,with the
adherenceof CommunistChina to the Soviet fold,it has seemed
politicto communist theoriststo arguethat,in the immediatepre-
revolutionphase, both Russia and China werein a feudalstageof
development. In I950 an officialreporton recentSoviet Oriental
studieslisted as an outstanding achievement in the field"the rout
ofthenotorioustheoryofthe'Asiaticmodeofproduction'".7

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
4 PAST AND PRESENT

In Wittfogel's view this reclassification

is deeplysignificant.It
constitutes an abuse of historicalmethod,a "mystification" designed
to obscurethegrimtruththatthetotalitarian statesofthecommunist
world,farfrombeing very"advanced" societies,are, like Tsarist
Russia beforethem,modemsurvivalsof "the mostbarbarousform
ofstate,OrientalDespotism".
Ignoringaltogetherthe question of evolutionary sequence and
"degreesof barbarism",Wittfogel'sargumentposes the question:
Is therea particularclass of societieswhich can be usefullydis-
tinguishedunderthe label "Asiatic Society"? And, if so, are the
characteristics commonto this class sufficiently definiteto makeit
clearly distinguishablefrom another class of societies, "Feudal
Society"? By "feudal" I understanda politicalstructurewhich
resemblesthatof medievalEurope in a fairlypreciseway; it is a
typeof society,nota stageof socialevolution. It is, of course,easy
enoughto distinguish thesetwo categoriesas ideal constructs. But
will the distinctionhold good in the case of actual societiesof the
presentday and ofestablishedhistory ?
Wittfogel's own expertisebeingin the fieldof Chineseeconomic
history he has naturally
tendedto see in classicalChinaitselftheideal
typeof "AsiaticSociety". By contrastMarx took his modelfrom
India. It is a seriousdefectin Wittfogel's reasoningthathe ignores
theveryradicaldifferences ofpoliticalideologybetweenthetraditional
Chinese and the traditionalIndian systems. Like Weber, he
picksoutas thekeyelementin Chinesepoliticalstructure theexistence
of a self-perpetuatingbureaucracy whichruledby owningtherights
to powerand office ratherthanbyowningrightsto property. Unlike
Weber,he failsto noticethatin southern Asiathedominating political
modeloverthepast2,500yearshasnotbeen"Chinese"but"Indian",
and that the dominatingtype of politicalauthorityhas not been
"bureaucratic"but "charismatic".8In southernAsia the King did
notadminister througha staffofoffice holders,nordid he rulemerely
by hereditary right. He achievedpower by personalascendency
and he held authority in the capacityof a livingBuddha or of a
Chakravartin.
Wittfogelsees in the classicalbureaucraticstructureof Chinese
societythe archetypalpatternof all modern totalitarianstates,
especiallythose of the Soviet variety. In his analysisthe basis
of the classical Chinese economy was hydraulic engineering.
He claims that in China, centralizedgovernmentcontrolof the
major irrigationworks was the key factorwhich permittedthe
growthof a centraliseddespotismsupportedby a highlyeducated

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON 5

bureaucracy. He thennotesthatthe mostsuccessfulancientstates

both in Asia and elsewherehave veryfrequently been "hydraulic
societies" - that is systemsdependentupon major engineering
worksof watercontrol,centrally controlled. And finallyhe argues,
more or less a priori, that these societieswere also bureaucratic
despotisms. The thesisis deceptively circular. It was bureaucratic-
despotismwhich made the archaichydraulicsocietiessuccessful;
it was the managerialrequirements of hydraulicagriculture which
made the archaicsocietiesinto bureaucratic despotisms. In theory
the correlation should be perfect;but was it reallyso ? Wittfogel
is verydiscreetin hisselectionand handlingoftheevidence.
He presentshisinitialargument in theformofa seriesofsweeping
generalisations whichareevidently intendedto be validfor"hydraulic
society"considered as an idealtype. Butwhenhe comesto exemplify
his thesishe splitsup his conceptinto a greatvarietyof sub-types
so that,intheend,formsandvariants ofOrientalDespotismhavebeen
discoveredall roundtheworldfromHawaiito Peru.9 Whennegative
evidenceappears it is simplyslipped in as a new morphological
variant.
For example Japan, despite its irrigationbased agriculture,is
foundto have a "feudal"ratherthanan "Orientaldespotic"political
structure;thus it is describedas "not hydraulicin termsof our
enquiry". Even so Japanis given a place in the total schemaas
belongingto the"SubmarginalZone oftheHydraulicWorld".10
In briefWittfogel's methodof demonstration is thatof Herbert
Spencer and the very numerouslater exponentsof nineteenth-
century"comparativemethod". The investigator looks only for
positiveevidencewhichwillsupporthis thesis;thenegativeinstance
is eitherevaded or ignored.
Now I admitthatto marshalanydetailedevidenceon thistopicis
difficult.Archaeology certainlyindicatesthat most of the archaic
civilisations- in Egypt,Mesopotamia,theIndus Valleyand so on
were"hydraulicsocieties",in thattheyexistedin areas wherelarge
populationsand urbanorganisation could onlybe supportedby the
systematic maintenance ofirrigationworksofconsiderable sophistica-
tion. Archaeology also tellsus a good deal about the natureof the
irrigation worksin question. Butit tellsus verylittleaboutthekind
of societywith which theywere associated. Even in the case of
Egypt,whereliteraryevidenceis extensive,we have no information
concerningthe organisationof the governmental bureaucracythat
is in any way comparableto the evidencewhichexistsconcerning
classical and mediaevalChina.

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
6 PAST AND PRESENT

If Wittfogel is rightin selectingclassicalChinaas an exampleof a

generaltypethenthenatureof thatgenerality is notilluminatedby
enumerating an endlessvarietyofsubspecies. Comparative evidence
mustbe genuinelycomparative and reasonablydetailed. Moreover,
in,view of Wittfogel'sexplicitborrowingof the classical Marxist
concept"OrientalDespotism",it is onlyreasonableto expectthat
he should take special note of the societiesfromwhich Marx's
concept was originallyderived,namelyHindu India and l'etats
hindouises ofsouthern Asia.11 A numberofthesestateswerefocussed
aroundhydraulicengineering worksof the most formidablekind,
and theseshouldsurelyprovidea testcase ofwhether or nothydraulic
engineering underarchaicconditionscan reallybe correlatedwith
any particularformof centralisedauthority.
Burma(Pagan), Cambodia,and Dry Zone Ceylon(Sinhala) were
all sophisticated statesofrelativelysmallscaleforwhichwe possessa
fairlydetailedpoliticalhistory;theyall flourished at muchthe same
period;theywereall "hydraulicsocieties"and theywereall in regular
contactwithChina. Wittfogel does notmentionanyof them. It is
a curiousomission.
I propose to considersome of the evidencefromone of these
examples,thatof Dry Zone Ceylon. It mightbe thoughtunfairto
challengeWittfogel on groundwhichhe has himselfavoided,but a
recentgeneralsurveyby Murpheyof the varioustypesof available
evidenceforthis region- archaeological, epigraphic,documentary
and ecological- has drawnattention to theperfectmannerin which
Ceylonfitstherequirements ofWittfogel'sthesis.12
The great hydrauliccivilisationsof Egypt, Mesopotamia and
NorthernIndia are phenomenaof remotehistoryabout whichit is
easyto speculatein the grandmannerwithouthavingone's theories
disprovedby the evidenceof inconvenient details. The hydraulic
civilisationof Ceylonwas a muchlaterphenomenonand, thoughits
contemporary historyis defective,the hydraulicsystemitselfhas
survivedin active,thoughattenuated, formrightdownto thepresent
day.13
Let us startwithone ofWittfogel's primary propositions:
"In a landscapecharacterisedby fullariditypermanentagriculturebecomes
possibleonlyif and whenco-ordinated humanactiontransfers a plentiful
and accessiblewater supplyfromits originallocation to a potentially
fertilesoil. When this is done, government led hydraulicenterpriseis
identicalwith the creationof agriculturallife. This firstand crucial
moment maytherefore
be designated as theadministrative
creationpoint".
(p. 109).
Having thus assertedthathydrauliccivilisationcan only startif
thereis an embryonic he thengoes on to claimthatthe
bureaucracy,

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON 7

growthof hydraulicenterprise involvesthe growthof bureaucracy.

Note the assumptionsinvolved,particularly those implied in the
term"fullaridity". WittfogeI's thesisis not simplythatareas exist
in whichcultivation is onlypossibleif theyare artificially irrigated,
but ratherthathis typesystem- "hydraulicsociety"- growsup
wherecultivation callsforlargescaleirrigation; theinitialhypothesis
is thatin such societiesall watersuppliesmustbe broughtfromafar.
But what are the facts? Is it the case thatthe greathydraulic
societiesof the past have grownup in "fullyarid" regionsof this
sort? Chronologicallythe earliest successfulhydraulicsocieties
seem to have been those of SouthernMesopotamiafollowedby
Egyptand the Indus Valley. In each case thenaturalenvironment
whichfaced the originalinhabitantsmay be summarisedthus: an
alluvialand fertilesoil, an arid climate,a terrainof flatswampland
subjectto periodicflooding. In such a contexturbandevelopment
calls forformidable and resourceful engineering butthefirstessential
requirement of the agriculturalist is notirrigation on a grandscale
butsimplya littlemodestconservation oflocalwaterresources.
"Irrigation"in Sumeriadid not requirea despoticmonarchto
build vast aqueductsand reservoirs;it simplycalled forelementary
and quite localiseddrainageconstruction and perhapsthe diversion
of riverfloodwaterinto the flatlands on eitherside of the main
stream.
Thus Wittfogel's keyphrase"fullaridity"is itselfan "ideal type"
concept. Doubtlessit is truethatin a genuinearea of "fullaridity"
bureaucratic actionwouldbe necessarybeforeagricultural lifecould
be established. But if we ask thequestion:"Is it an historicalfact
thatmostof thehydrauliccivilisations of the past have grownup in
'fullyarid' regions?", the answeris "No!" Dry certainly;but not
"fullyarid".
Let us considerthe ancientCeyloneseCivilisation(Sinhala) from
thispointofview.14
AncientSinhalawas locatedexclusively in thatpartoftheNorthern
Dry Zone of Ceylon which is now knownas the North Central
Province. It is not an area of fullaridityin Wittfogel's sense,but,
despitean averagerainfallof fifty to seventy-five inchesper annum,
regularcultivation is impossiblewithoutirrigation. The reasonsfor
this apparent paradox have recentlybeen expertlyanalysed by
Farmer.15
The regionis not an area of "potentiallyfertilesoil". On the
contrary thesoilis strikingly infertileand has no capacityforholding
water. Everywhere thenaturalwatertableis verylow; in dryweather

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
8 PAST AND PRESENT

any well whichis not in the immediatevicinityof some artificial

reservoir is almostinvariablydry.
The techniqueof agriculture does not relyon bringingin water
fromelsewhere(as Toynbeeerroneously supposes)but is based on
the local storageof local rain waterfor use throughout the year.
Major irrigation worksare numerousand immensein scale but are
everywhere supplementary to thesmallscale localisedsystem.
Thus while it is true that permanentagricultureis impossible
withoutirrigationengineering, and while major government con-
trolled irrigationworks undoubtedlyimprove the efficiencyof
agriculture, it cannotbe maintainedthatlarge scale administrative
action is essentialfor the workingof the system. In termsof
Wittfogel'sthesis this distinctionis crucial,but I suspect that a
similarqualificationmightbe made with regardto almost every
practicalinstanceof an archaichydraulicsociety.
We havegoodarchaeological evidencethatthegeneralarrangement
of theirrigation systemat thepresenttimeis verymuchthesameas
it was in the eleventhcenturyA.D.16when Classical Sinhala had
attainedits territorial peak. The modernirrigationworksconsist
ofseveraldistincttypesand thesebearon Wittfogel's propositions.
The populationis distributed overthemap in smallvillages,most
ofthemcontaining fewerthanfifty elementary families. Each village
depends for its livelihoodon an area of irrigatedrice land watered
froman artificial reservoir. The reservoirs (tanks)werefirstcreated
in ancienttimesby buildingan earthwork transverselyacross the
lineofa naturalstreamand dammingup thewaterbehindit. Since
thegenerallay oftheterrainis nearlyflatmostofthetanksare large
in areabutshallow;andtheearthwork damis longratherthanhigh.
A fewof thesevillagetanks,notablythoselocatedin the vicinity
ofancientBuddhisttemplesare equippedwithscientifically designed
spillwaysand sluicesbutthemajority are rathercrudeaffairs. Tank
construction is, of coursea specialisedjob, callingforspecialskills
and knowledge,but the buildingand maintenanceof an ordinary
villagetankcallfortheexpertise ofa foreman plumberratherthanthat
of a university graduatein engineering.
From timeimmemorialnormalrepairworkto villagetankshas
been the ordinaryworkof ordinaryvillagers. Major repairsand
new constructions were traditionallyundertakenby a specialised
caste group of Tamil labourers (kulankatti),but these people
workedforthe villagerson directcontract;theywerenotemployees
ofthestate.17 It is onlysinceabout i 86o thata centralised Irrigation
Departmenthas had the rightto interfere in mattersrelatingto the

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON 9

maintenanceand use of villagetanks. I knowof no evidencethat

the ancientcentralstateauthoritiesever concernedthemselveswith
thedetailsofvillagetankmanagement.Villagerswereundera legal
obligationto maintaintheirtankin functioning order,butthereis no
evidenceof any bureaucraticmachinerythat mighthelp to bring
this about. On the contrarythe factthat,in the chroniclesof the
kings,the variousmonarchsreceivepraisefortheirmunificence in
repairingvillagetankssuggeststhattherewas no routineprocedure
forcarrying out such workon a nationalscale.
The monarchsare likewisepraisedfordonatingthe revenuesof
particulartanksto Buddhistmonasteries whichhardlysuggeststhat
the whole systemwas centrallycontrolled. Numerousinscriptions
have survivedwhich specifically prohibitthe king's officersfrom
enteringon the lands of thesereligiousinstitutions.18
However,in additionto the small scale villagetanks,thereare
certainmajorworkswhichare associatedwiththewatersupplyofthe
variousancientcapitalsofthekingdom- particularly Anuradhapura
and Pollonaruwa. Here the engineering is on a grandscale. The
reservoirshave brick cored and masonryfaced embankmentsas
muchas ninetyfeetin heightand up to nine milesin length,with
complexsluiceworksand spillwaysto match. Beforesiltingreduced
the area, the Kalawewatank may have been nearlyfortymiles in
circumference.It leads out into a complex feedercanal system
about fifty-five miles long. For the firstseventeenmiles of its
lengththis canal has a steadyfall of six inches a mile. Another
ancienttankis estimatedby Brohierto have had a surfacearea of
six thousandacreswitha retaining embankment seventeen
containing
millioncubic yardsof earthand stone.
I need not elaborate. Worksof thiskind are not in the plumber
artisan class. The original engineers were very sophisticated
professionalswith large economic resources at their disposal.19
AncientSinhalamayor maynothavebeenan "OrientalDespotism";
it was quiteunquestionably an "hydraulicsociety". And,of course,
it is thiskind of evidencewhichtemptsus to acceptthe Wittfogel
thesis. Let me quote again:
"The effectivemanagement of these works involves an organizationalweb
which covers eitherthe whole, or at least the dynamiccore of the country's
population. In consequence those who control this networkare uniquely
preparedto wield supreme politicalpower" (p. 27).
The thesisis simple and logical: the hydraulicsocietycalls for
managersand professional
professional engineers;it is thesepeople
whobecomeeffectivelythe rulersofthesystem- in muchthesame

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
10 PAST AND PRESENT

wayas somewouldclaimthattheheads of greatindustrialorganisa-

tions are the real rulersof the modernWesternWorld. And sure
enough,if we look at Sinhalesehistory,we findthatthe greatkings
have reputationsas irrigationengineersratherthan as conquerors
or as buildersof cities.
The most celebratedof all the kingsof Sinhala was Parakrama
Bahu I who reignedin PollonaruwafromII64-II97. He was a
usurperand his reign,havingstartedwith civil war, was entirely
givenoverto military conquest. In thishe was strikingly
successful.
In I 80 one of his generalsappearsto have exactedtributefrom
Arakanand at varioustimeshis armieslaid waste large parts of
southernIndia; but, as is the way with militaryconquerors,his
financialexactionsultimately broughtruin and disasterto Sinhala
itself. Ironically,in the Ceylonchronicles,the reignof thistyrant
megalomaniacis recordedas a periodof peace and progress,the
monarchis praisedforhisliberalitytothepoorandforhisbenefactions
to religion,but above all forhis worksof irrigation. It is alleged
that he constructedI470 new tanks,includingthreegiant lakes;
300 further tankswerebuiltforthepriesthoodand 1,395largetanks
restoredbesides960 smallerones; he constructed 534 watercourses
and repaired 3,621.20 His reputed statue at Pollonaruwa stands
to thisday on the embankment lake of all -
of the largestartificial
"the Sea ofParakrama"!
Now all thisis strictly
in accordwithWittfogel's thesis. Wittfogel
maintainsthatit is a characteristic featureof politicalpropaganda
in hydraulicsocietiesto representthe ruleras a benevolentdespot
devotedto theenhancement ofagriculture whereasin actualfacthe is
commonly a ruthlessmilitary tyrantofthemostdiabolicalkindwhose
only interestis the enhancementof his own personalglory. No
doubt Wittfogel's view is strongly colouredby his personalopinion
ofcertaintwentieth-century dictatorsyetit can hardlybe deniedthat
his stereotyped descriptionof the "ideal type" of Orientaldespot
fitswithquitestriking perfection thehistoricalrecordof someofthe
monarchso,fCeylon.
The dryzone "hydrauliccivilisation" of Sinhalaenduredforabout
1,500 yearsbuttheextravagance ofParakramaBahu I finally brought
it to an end. Thereafter, following a halfcenturyof chaos,we find
thatthe focusof the kingdomhas shiftedintothe easternwetzone.
The capitalfromI1246onwardswas at varioustimesat Dambadeniya,
Kurunagala,Gampola,Kotte (Cotta) and Sitawaka,all of whichlie
within fiftymiles of modern Colombo. When the Portuguese
establishedthemselvesat Colombo in I505 the Sinhalese capital
was nearby,at Cotta. It was onlyin I590, afternearlya centuryof

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON II

complexcivilwar thatthe hill cityof Kandyfinallyemergedas the

capitalofan independent kingdomL
RobertKnox'soutstandingly perceptivedescription ofthisKandyan
kingdomis based on a twenty-year period of residencebetween
I659 and I679.21 As I shall show presently, this descriptioncon-
forms to Wittfogel'sideal type in a very strikingway. The
implicationsof thisparallelismdeservecarefulconsideration.
The Kandyan kings thoughtof themselvesas the heirsand
successorsto theancientkingsofAnuradhapura;22 butalthoughmost
ofthelandsofancientSinhalalaywithinthefrontiers oftheKandyan
realm,therehad been no veryclosehistoricalcontinuity betweenthe
kingdomof ParakramaBahu I and the kingdomof Kandy. More-
over while the formerhad been an exceptionally"pure type" of
hydraulicsociety,in Wittfogel's sense,the latterwas not. Knox's
Kandywas notexclusively dependentupon irrigated agriculture.
Clearlywe cannot inferfromKnox's descriptionof Kandy in
1670 the state of Sinhala a thousandyears earlier. Nevertheless
Knox's Kandy demonstrates certainpossibilities;Kandy was what
Sinhalahad becomethroughtheprocessof history.
Now Knox's Kandyanmonarchis an OrientalDespot straightout
of Wittfogel's imagination,but Knox's Kandyan State is a society
organisedon principlescloselyresembling thoseof Europeanfeudal-
ism. Thus, in thisparticularpracticalcase thetheoryof Despotism
and thepracticeof Feudalismwerenotdistinct. Can we interpolate
backwardsand demonstrate convincingly thatthe same was equally
trueof AncientSinhala? If thiswereestablishedthena largepart
of Wittfogel'sthesiswouldfallto pieces; foralthoughWittfogel has
carefullyavoidedall reference to Sinhalait is clearthatthe circum-
stancesoughtto providea perfecttestcase forhis generalhypothesis
that hydraulicbased despotismand decentralisedfeudalismare
polar typesof organisation.
We need to show, fromthe evidenceregardingKnox's Kandy,
not onlytheparallelismbetweenWittfogel's ideal theoryof Oriental
Despotism and Knox's descriptionbut also the way Knox's
seventeenth-century statements fitin withthe "bricksand mortar"
evidencefromtheearlierKingdom.
Firstlyas to Despotism. Here is Wittfogel:
"the myth of an unselfish despotism dramatizes those desiderata which
consciouslyor unconsciouslyare underwrittenby all thoughtfulmembersof
the rulingclass . . . thus agromanagerialdespots may presenttheirregimes
as benevolent . . . actually however, and even under the most favourable
circumstances,theystrivefortheirown ... theyplan theirhydraulicenter-
prises according to what benefitstheirown mightand wealth".
(pp. I35-6).

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
I2 PAST AND PRESENT

And hereis Knox:

"The King oftenemployshis people on vast worksand thatwill requireyears
to finishthat he may inure them to slavery ... he approves not that his
people should be idle but always finds one thing or another to be done
throughthe work be of littleor no purpose". (Knox p. 70).
As an exampleof thisbehaviourKnox thendescribesat lengththe
construction of an elaborateornamental lake at Kandy. This had
necessitatedseveral years of tunnellingand ditching. One con-
sequenceof thisextravagance had been to deprivea formerly fertile
valleyof all the waterwhichpreviouslyservedto irrigateits fields.
The lake in question still exists. Strictlycomparablereservoirs
witha decorativeratherthan an economicfunctionare prominent
also at thesitesoftheancientcapitalsofAnuradhapura, Pollonaruwa,
and Sigiriya. Even the minor details of Knox's pictureaccord
closely with the Wittfogelpattern. His monarchis an absolute
and arbitrary tyrantwhosepoweris focussedaroundtheorganisation
of hydraulicengineering.Despite propagandato the contraryhe
devoteshis main effortto the creationof pleasuregardensrather
thanto the advancementof agriculture.
Buthowfaris Knox a goodwitness? Is nothe too,likeWittfogel,
engagedin politicalpolemic? Plainlymuchof Knox's ethnography
- thatis to say his straightforward accountof Ceylonesemanners
and customs- is excellent. But we need to rememberthat he
himselfwas a citizenof seventeenth-century England,a countryin
whichthe divinerightof despotickingswas stilla primepolitical
issue. Knox's detailedelaborationof the Kandyanking'styranny
is perhaps,in partat least,a slanderousattackagainstmonarchyin
general. If so, the earlierrecordsmaybe propagandist too.
Today, if we simplyobservethe archaeological remainsof the
ancienthydrauliccivilisation,thefactwhichimpressesmostof all is
thesheerscale of whatwas achieved. It is not onlythe enormous
size of theembankments of individualreservoirsbut theirnumber;
and also the correspondingsize and number of non-utilitarian
structuressuch as dagobashrines,templesand abandonedpalaces.
Some of the dagobas are on much the same scale as the
Pyramidsof Egypt - thereare at least two which containover
twentymillion cubic feet of solid brickwork.22The chronicle
recorddeclaresthatParakramaBahu I builtIoI dagobas,476 statues
of the Buddha, 300 image rooms,etc., etc., etc.,23and if we look
aroundtodayat thefabulousruinsofAnuradhapura and Pollonaruwa
we can easilypersuadeourselvesthatthis mightbe true. All this
seemsto fitperfectly withthe Wittfogel-Knox pictureof theancient

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON 13

Ceylonmonarchas a totalitarian tyrantwieldingarbitrarypowerover

countlessthousandsoflabouringslaves.
It is a keypointin Wittfogel's
argument thatthe"agro-managerial"
monarchin an hydraulicsocietymusthave at his disposala huge
unpaid labourforce. It is preciselybecause he feelsthatthisis a
logical necessitythat he is able to assert so positivelythat such
societiesare Despotismsand notFeudalisms.
"Government-managed heavy water works place the large scale feeding
apparatus of agriculturein the hands of the state. Government-managed
constructionworks make the state undisputed master of the most com-
prehensive sector or large scale industry. In both spheres the hydraulic
state levied and controlled the needed labour forces by coercive methods
thatwere invocable by a feudal lord only withina restrictedarea ... The
hydraulic rulers were sufficientlystrong to do on a national scale what a
feudal sovereign or lord could only accomplish within the borders of his
domain". (p. 47).
But if it is a fallacythatthe circumstancesof an hydraulicsociety
logicallyrequirethe centralisedcontrolof largelabourforces,then
the wholethesiscrumbles.
What is the evidence? Can we reallyinferfroma traditionof
"great" kingsthat the kingsin questionwere trulygreat? Might
it not be thatthe "greatness"of the hydraulicmonarchis itselfa
product of propaganda myth? Need we believe the Sinhalese
chroniclesany morethanwe believeKnox? Impressivebuildings
do not necessarilyimply impressivebuilders. The Sinhalese
irrigation systemwas notcreatedin a day,norevenin a lifetime.
I have alreadymentionedthe Kalawewa canal systemwhichnow
has a gianttankat its head whichleads into a fifty-five mile long
watercourse, whichin turnfeedsintothreelargetankswhichprovide
waterforthe ancientcapitalof Anuradhapura. It all lookslike a
colossaland highlyorganisedpieceofbureaucratic planning, thework
of one of Wittfogel'sidealised OrientalDespots. But if so, the
planningmusthavebeendonebya kindofDurkheimian groupmind!
The systemtookaboutI,400 yearsto build. The originalTissawewa
tank at the bottomend of the systemwas firstconstructedabout
300 B.C. The Kalawewatankat the top end of thesystemwas first
constructed about 800 yearslaterand elaborationsand modifications
wenton forat leastanother600 years.
And so too withall the other"stupendous"works. While myth
invariablyattributestheirconstruction to the initiativeof a single
outstandingmonarch,archaeologyshows each has been slowly
developedovera longperiodof time.27 This is truenotonlyofthe
tanks but also of the more blatantobjects of "conspicuouscon-
sumption"suchas thedagobashrines. The giantshrineswhichexist

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
14 PAST AND PRESENT

todayare,in nearlyall cases,enlargements ofstructures whichin the

firstinstanceweremuchmoremodestin scale.
This radicallyaltersthe implicationsoftheobservablefacts.
For example,a dagobaoftwentymillioncubicfeetcontainsaround
200 millionbricks,mostlyroughlaid; this suggestsan enormous
armyofforcedlabourers. Butifweassumethattheworkwas carried
out by corveelabourersworkingnot morethanI00 daysin any one
year,a labour forceof about 600oo men could completethe job in
fiftyyears... whichis probablyjustaboutwhatdid happenin most
cases. An embankmentof twentymillioncubic yards could be
completedby the same sortof labourforceworkingin thesame way
in about twenty-fiveyears.
In otherwords,althoughthe Ceylon irrigationworksand the
associated palace and temple constructionworks do representa
giganticaccumulation froman enormousnumber
of capitalresulting
of man-hoursoflabour,thisfactdoes notin itselfimplyanymassive
controlover labour resourcesby the "bureaucraticrulers". The
extremely highlevelof engineering skilldisplayedin the designand
constructionof these works does clearlyimply the existenceof
professionalengineersbut the indicationsare that such men were
membersof Buddhistmonasticinstitutions and not directservants
of the crown.
As to the natureof the labourforce,Wittfogel writes:
"The hydraulicrulers compelled able bodied commonersto work for them
through the agency of the corvee. Corvee labour is forced labour. But
unlike slave labour which is demanded permanentlycorvee labour is con-
scriptedon a temporaryalthough recurringbasis. Afterthe corvee service
is completed the worker is expected to go home and continue his own
business". (p. 47).
This paragraphis valid forCeylon,both ancientand modern,but
the existenceof such a systemof labour recruitment does not
necessarily was largeor thatits
implythatthe scale of recruitment
imposition was arbitrary.
Fromthe mostancienttimesirrigation workof all kindshas been
knownas rdjakdriya,i.e. king'swork. Everyholderof land holds
in thetitlebywhichthelandis held.
inherent
it subjectto obligations
By farthemostimportant is the dutyto provide
oftheseobligations
corveelabour. But to whomand by whomis thisobligationdue ?
This is thecruxwhichWittfogel's argument tendsto distort. Today,
the obligationis one enforcedby the centralgovernment, but was
thisalwaysso ?
The evidence for the early period consists mainly of stone
inscriptions whicharecapableofmanyinterpretations, buteverything

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON I5

is consistentwith the view that Knox's account of seventeenth-

centuryKandymighthave appliedequallywellto ancientSinhala.26
I haveremarkedthatalthoughKnox assertsthathis Kandyanking
was an absolutedespotofthemostruthlesskind,his modeofgovern-
mentappearsas feudal. Thus:
"The countrybeing wholly his, the King farmsout his land not formoney
but for service. And the people enjoy portionsof land fromthe King and
instead of rent they have their several appointments,some are to serve the
King in his wars, some in their trades, some serve him for labourers and
othersare as farmersto furnishhis house with the fruitsof the ground; and
so all thingsare done withoutcost and everyman paid forhis pains; thatis
theyhave lands forit". (p. 68).
So farthe description servicetenurewiththe
is thatof a centralised
land held in directfieffromthe crown;but now note whatfollows:
"Many towns are in the King's hand, the inhabitantswhereofare to till and
manure a quantityof the land according to theirability,and lay up the corn
forthe King's use. These towns the King oftenbestows upon some of his
nobles for their encouragementand maintenance, with all the fruitsand
benefitswhich beforecame to the King fromthem". (p. 69).
Such a systemmaysurelybe called"feudalism"withoutanystraining
ofterminology.
Knox was an extremely acuteobserverand he recognisedthatthe
allegedlytyrannousauthorityof the king derivedfromthe king's
ultimatetitleover irrigatedland. Yet a man could alwaysescape
fromthe landlord'styrannyby the simple expedientof declining
theroleoftenantwithits serviceobligations:
"Yet all have not watered land enough for their needs ... so that such are
fain to sow on dry land and till othermen's fieldsfora subsistence. These
persons are freefromthe paymentof taxes . . . If any findthe dutyto be
heavy or too much forthem,theymay, leaving theirhouse and land, be free
fromthe King's service,as thereis a multitudedo. And in my judgement
theylive farmore at ease, aftertheyhave relinquishedthe King's land, than
when they had it". (p. 69).
The crucialpoint,whichis validto thisday,is thattheobligations
of servicetenureattachto the soil and to whoeverworksthe soil.
A man who declinesto be a farmerof irrigatedrice land has no
serviceto perform. It is interestingto compareKnox's comment
thattherewas a "multitude"who opted out of corveeserviceand
wereall the betterofffordoingso withWittfogel's pronouncement
on the same possibility:
"When taxation becomes unusually burdensome the peasant may reduce his
cultivatedacreage and when the heavy demands continue,he may become a
fiscalfugitiveabandoning his fieldsaltogether. He may wander in despair,
look forworkelsewhere,or turnbandit or rebel" (pp. 33I-2).
Knox wrote of what he saw; Wittfogeltheorisesabout what he
supposes would be the case.

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
I6 PAST AND PRESENT

DespitemycriticalcommentsI hold thatWittfogel in his concept

of "the hydraulicbased despotism"has pickedout foremphasisthe
criticalfeaturesof an important typeof archaiccivilisation whichis
well representedby the Sinhala example. Where Wittfogelerrs
is thathe does not discriminate sufficientlybetweenideal typesand
case histories. Furthermore he mixeshis analysisof historicalfact
withpolemicabout politicalvirtue. For Wittfogel"the hydraulic
based despotism"is thearchetype policestates,and
ofall totalitarian
in his view any police stateis morallybad fromthe start. In con-
sequence,despitetheastonishing penetration of muchofhis analysis,
histotalpresentation ofhydraulic societyis distorted.
His insistence
thatin hydraulicsocietypowerrelations and property
relationsare distinctis illuminating but the dichotomyis far too
sharp:
Whether hydraulic propertyis large or small or whether it belongs to a
member of the governing class, it provides material advantages. But it
does not enable its holders to control state power throughpropertybased
on organizationand action. In all cases it is not power propertybut revenue
property. (p. 300).
It is illuminatingto be reminded that the magnificentruins
of archaic civilisationimply the existenceof a well developed
machinery of government, but it cannotbe deduced fromthisthat
the bureaucratswereall powerfulor thatthe structure of authority
was radicallydifferentfromthatin less spectacularsocial systems.
One of Wittfogel's themesis that OrientalDespotismis staticand
archaic,a kindofhangoverfromthelast stagesofbarbarism, and he
contraststhiswithFeudalismwhichhad the meritthatit "led to a
limpingand multicentred type of absolutismand, eventually,to
multicentred and private-property based industrialsociety"'.21
But is thisdistinctiona real distinction ?
or an illusionof history
Is it perhapssimplythatwe are ratherbetterinformedabout the
factsin thecase ofWesternEuropeso thatwe can see thattheclaims
of feudalmonarchsto absolutepowerwerefictitious ? In contrast
we are deceivedby theideal theory of orientaldespotism. Afterall,
the Holy Roman Emperorclaimedabsoluteauthorityeven though
he was unableto exerciseit.

II
IRRIGATION BASED "FEUDALISM" IN SINHALA
So far my main purpose has been to show that althoughthe
constructions
engineering of ancientSinhala were of the typethat
indicatesshouldbe typicalforan hydraulicbased Oriental
Wittfogel

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON I7

Despotism, and althoughthe traditionalreputationof Sinhalese

monarchscorresponds to thepurestereotype ofthe OrientalDespot,
neverthelessthe actual Sinhalesesocial systemwas not farremoved
fromEuropeanfeudalism.
I proposenow to give some furtherdetails of this "hydraulic-
oriental'feudalismand to suggestthat this particularspecies of
servicetenureis peculiarlywelladaptedto theecologicalrequirements
of an irrigationbased society. Whereas Wittfogelimplies that
hydraulicsociety and feudalismare mutuallyinconsistentpolar
typesof politicalstructure, my own view is the exact opposite. I
suggestthatit is the mutualcompatability of hydraulicagriculture
and "feudalism"(in the sensenow to be described)whichmayhelp
to explainthepoliticalsuccessofmanyearlyexperiments in hydraulic
civilisation.
Whatactuallyhappenedin the Ceyloncase was thatthekingmade
grantsof land to his provincialgovernorsand to monasteries. The
grantcoveredbothland and population- theground,theirrigation
works,and the people - and the grantincludedthe rightto levy
corvee. Althoughcorveedutywas (and is) knownas "king'swork"
it was in factan obligationdue to thegrantholderratherthanto the
kingin person. The evidenceforthisis notonlyin modempractice
withregardto thelandsofreligiousinstitutions butin thefactthatthe
ancient records include such contradictory expressionsas disdve
rajakdriya- "provincialgovernor'sking's work".
The resultingdelegationof powersproduceda "feudalism"which
differed fromthe Europeanpatternmainlyin that,in Ceylon,the
servicedue froma fiefholderto hislordcouldbe ofvariouskindsand
was notlimitedonlyto military service,or moneyin lieu thereof.
In Ceylon, as in feudal Europe, the monarch'soverridingand
perpetualproblemwas to devise a means of keepinghis feudal
baronsundercontrol. The extremefrequencyof insurrection and
civil war showsthateffective powerusuallylay withthe local land-
lordsratherthanwiththe crown.
In Knox's day the king'sprincipaldevice for maintainingorder
was thatof compellinghis lordsto keep one of theirclose relatives
at courtby way of hostage. He also ratherfrequently changedthe
land grantsaroundfromone baronialfamilyto another. Knox cites
as evidenceof the monarch'styranny the factthatthehostageswere
veryfrequently executed;it mightalso suggestthatthe greatlords
wereveryfrequently in revolt!
Nor was such a stateof affairsa peculiarityof the laterdecadent
phase of the Sinhalesestate. The greatParakramaBahu I himself

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
I8 PAST AND PRESENT

startedlifeas a provincialgovernor
and attainedthethronebyleading
a successfulrebellionagainsthis cousinthe king. And if we credit
theCeylonchronicles (MahavamsaandCulavamsa)withanyhistorical
meritat all, we mustconcludethatthe frequency withwhichkings
were overthrownby their rebelliousrelativeswas at all times
strikinglyhigh.
In theorytheouterprovincesof therealmwereruledbygovernors
(disdva)appointedby the king;in practicethe lordshipof the local
hereditary baron (vanniyar)was virtuallyabsolute. In manycases
therankof disavawas simplythetitularofficeof a courtofficial
who
never went near his domain. When the customarylaw of the
North CentralProvincewas being recordedin I820 it was stated
that: "fromancienttimethe vanniyarhad been deemedto possess
powernearlyequal to thatof the disava,but thathe is restrained
in theexerciseofit whenthedisavais in theprovince".28
Knox,whodiscoveredto hiscosttheextremedifficulty oftravelling
in the remoterpart of the Kandyankingdom,attributes this to the
subtletyof an all powerfulmonarch:
"The King's policy is to make his countryas intricateand difficultto travel
as may be and thereforehe forbidsthe woods to be felled especially those
that divide province fromprovince and permitsno Bridges to be made over
his rivers,nor the paths to be made wider". (p. 70).
A simplerexplanationwould be thatthe provincialbaronswere
beyondcontrol. Certainly theseare notthecharacteristics
we would
expectif government werein thehandsof a centralised bureaucracy.
Finallywe maynoteone further wayin which,in thelaterperiod
at least,the Ceylonesebaronialclass adoptedbehaviourscomparable
to thoseof the Europeanfeudalnobility.
Wittfogel,generalisingfar too easily, assertsthat the rules of
inheritance in hydraulicsocietyalwaysrequiresub-divisionof the
landedproperty amongthechildrenofthedeceased. In thisway,he
explains,the despoticmonarchis able to allow a formof private
ownerslhip of land amonghis people, since he knowsthatno such
holdingcan everbecomeverylarge. Wittfogel contraststhisegali-
tarianinheritancerule with that of the European feudal nobility
who developeda systemof primogeniture.29
Now theinterestingpointis that,in Ceylon,Wittfogel's
generalisa-
tion applies perfectlyto the ordinarypeasantrywhereinheritance
rules do requiredivisionof land to both sons and daughters. But
amongthe aristocracy a differentsystemprevails- the daughters
taketheirsharein theformofa monetary dowerand thesonsinherit
onlywhentheystayat homeand residepatrilocally.Consequently,

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON I9

amongthearistocracy (butonlyamongthearistocracy),theCeylonese
developeda systemofpatrilinealdescentassociatedwiththeholding
of largeundividedestates. This is not quitethe same as European
primogeniture; but it comesverynear.
Having consideredsome of the similaritieslet us considerthe
differencesbetweenCeylonese"feudalism"and the more familiar
European type.
The most fundamentalfeatureof European feudalismwas the
associationofland tenurewithmilitary service. Everyonefromthe
top to the bottomof the feudalhierarchyheld his land underthe
conditionthathe must,whenrequired,rendermilitary serviceto his
immediatelord. The Ceylon systemwas also one of "service
tenure"but withthe difference thattheservicesdue to thelandlord
were of variouskinds and not exclusivelymilitary. Knox having
noted(as above) thatthe kingmade grantsof "towns"(i.e. villages)
thatbeforecameto the
to his nobles"withall thefruitsand benefits
kingfromthem" then continues:
"In eachofthesetownsthereis a smithto makeand mendthetoolsofthem
to whomthekinghathgrantedthemand a potterto fitthemwithearthen-
ware,and a washerto washtheirclothesandothermento supplywhatthere
is needof. And eachone ofthesehatha pieceoflandforthistheirservice
whether it be to theKingortheLord,butwhattheydo fortheotherpeople
theyarepaid for". (p. 69).
Peasant societyin Ceylon is even today divided accordingto
occupationalcastes. The farmercaste which is the highestin
statusis also themostnumerous. Othercasteshavetheoccupational
dutiesof washerman,drummer,blacksmith, potter,spiritmedium
and so on; formerly therewere others- forexample,in the days
beforemotorcars and bicycles,palanquinbearerwas an important
functionalcaste occupation.
The crucialpointto noteis that,as to theirmainoccupation,all
these castes are peasantcultivators, irrigatedrice lands
cultivating
accordingto identicaltechniques. Wherethe castesdiffer fromone
anotheris in the servicedutieswhichtheyformerly owed to their
feudallordsin paymentfortheirland. Thus the farmers rendered
militaryserviceor tax in kind but the lowercastesrenderedtheir
profession:i.e. the washermanwashedclothes,the blacksmith made
tools,thepottermadepots.
Today the feudaldutyto the overlordshas mostlybeen abolished
yet even so the villagersstill feel themselvesunder an obligation
to carryouttheircasteservicesfortheirneighbours;and theyusually
do so withoutanyexplicitpayment.
For example,in an area I knowin NorthernCeylon,a community

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
20 PAST AND PRESENT

of washermancaste providesthe ritualwashermenforabout thirty

differentvillagesof othercastes- usuallyone particularwasherman
householddoingthe choresforone particularpatronvillage. The
washermanreceivesno piece workpaymentbut he receivesvarious
perquisitesfromthefestivals at whichhe attendsin thecourseofhis
duties. So also in thecase ofthedrummers and thespiritmediums.
Pottersand theblacksmiths howeverarepaid fortheactualworkthey
do on a pieceworkbasis.
A veryimportant aspectofthiskindof labourorganisation is that
it permitsthe developmentof quite complex systemsof labour
specialisationwithoutany extensiveuse of a moneymediumor of
creditbanking. In Ceylon,in the "feudal" period,land was so
distributedamong the professionally specialised castes that the
tenantsofanyparticular superiorlandlordformedan overlapping set
ofreciprocalworkteams. In effect each man paid rentforhis land
in the formof specialisedlabourcontributed to the workteamas a
whole. But at the same time,each man, or rathereach sub-caste
group(variga)was separately in clientstatusto thelocal landlord.
In the dryzone regioneach varigaincludedsome ten to twenty
smallsinglecastevillagesand was corporately organisedwithitsown
castecourt. The varigaas a wholewas theclientofthelocal baron
and the baron was ex-officio presidentof the caste courtof each
varigawithinhis domainwhatever its caste.
Each individualheld land privately withinhis own village- but
the villagewas the villageof a particularvarigawhichwas in effect
"owned by" the local baronor sometimesby the local templehigh
priest. Each individualwas thusunderobligationto contribute the
dutiesofhis casteto his varigalord,i.e. to thelocal baron- butthe
lattertransferredthisduty,so that(in practice)varigaservicesbecame
reciprocal. Blacksmiths and mediumsreceiveservicefromwasher-
men; washermenand blacksmiths receiveservicefrommediumsand
so on.
PerhapsI shouldadd thattheseservicedutieswhichattachedto
the lands of each particularvarigagroupwere over and above the
rdjakdriya - "king'swork"- whichwas the generalobligationto
corvee. This now appliesespeciallyto themaintenance ofirrigation
works,but, formerly no doubt, included all the monarch'scon-
structionalventures.
Althoughcloselycomparablesystemsof labour rganisation exist
in other"hydraulicregions"oftheIndiansub-continent,3" thereis no
explicitevidence,otherthanthatof Knox, thatthiskindof "caste-
feudalism"is ancientforCeylon,but it is certainly highlyprobable.

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON 2r

The differentiation of serviceobligationsin termsof caste dutiesis

closelyboundup withtherequirements oftempleservice. Temples,
in theirmodernform,cannotbe maintainedwithoutthe specialised
servicesof washermen,musicians,potters,paintersand the rest
and thereis no reasonto imaginethatthisis in anysense a modern
development. If we couplethiswiththe factthat,fromthe earliest
times,we have reportsof villagesbeinggiftedto templesalongwith
theirgamvara- thatis the "dues" of the tenants;thatthe tenants
of such villagesare in some cases specifiedas carpenters, artisans,
masons,serfs,drummers, washermen, etc.; and thata tenth-century
inscriptionshows that villageswere then administerednot singly
but in clusters(dasagam- lit. "ten villages");thenwe can at least
say thatit seemslikelythata caste organisation of laboursuch as I
haveoutlinedabovegenerally prevailedin ancientSinhala.3
Let me recapitulate and bringtheargument to a conclusion. The
factthattheearliestlargescaleagriculturally basedurbancivilisations
wereso oftenhydraulicsocietiesis an historicalfactofgreatinterest.
Clearlythe historicalfact poses the question: are therecommon
featuresin thesesocieties,economic,politicaland so on ?
In isolatinghis concept of "orientaldespotism"Wittfogelhas
drawnattentionto a numberof commonfeaturesin thesesocieties.
But his analysisis distortedby his anti-totalitarian bias and by his
tendencyto verylargescale generalisations.Since mostofthemore
impressivehydraulicsocieties existed only in remote antiquity,
generalisedstatements concerningthe detailsof theirgovernmental
organisation are difficult
eitherto verifyor contravert.
Sinhalais an exception. It existedas a "pure type"of hydraulic
societydownto thetwelfth century A.D. The hydraulic systemitself
has survivedin attenuatedformdown to the presentday. Many
featuresof presentday socialorganisation in theNorthernDry Zone
are commonlyregardedas directsurvivalsfromthe classicperiodof
hydrauliccivilisation.
Classical Ceylon(Sinhala) was confinedto the irrigateddryzone
fromroughlythethirdcentury B.C. untilthetwelfth century A.D.
Down to the end of the eighthcenturyA.D. the capitalwas con-
tinuouslyat Anuradhapuraand around this city there gradually
developeda vast networkof majorhydraulicworks(in Wittfogel's
sense). This hydraulicsystemserved two functions;it provided
waterforthecapitaland it provideda largeregionaroundthecapital
with a much more reliable source of irrigationwater than that
obtainablefromthesmallscale villagetanksalone.
We may concedethatif the whole of this impressivehydraulic

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
22 PAST AND PRESENT

systemeverfunctioned forany substantiallengthof time

efficiently
then it would have needed a bureaucraticorganisationto run it.
This is Wittfogel's thesis.
But thereis no evidencethatthis was the case, and thereis no
evidenceforthe existenceof an hydraulicbureaucracy whichplayed
a significantrole in the politicsof the Sinhala State.32
The documentary sourcesgo backas faras thefourthcenturyA.D.
(Dipavamsa) and, thoughtheirhistoricalqualityis alwaysdubious
and much more attentionis paid to ecclesiasticalthan to political
historythey can hardly,be dismissedas irrelevant. If hydraulic
engineering was centrally organisedand of centralimportance to the
SinhaleseState thentherewould surelybe some indicationof this
factevenin "fabulous"history ?
But engineering whenmentioned
activities, at all, aretreatedas the
arbitrary actsofmeritofindividualmonarchs, and mostlysuchwork
relatesto dagobashrinesand pleasuregardensratherthanto agricul-
turalirrigation.The politicalbackgroundis one of constantcivil
war and palace murders.
Whenkingsarecreditedwiththeconstruction ofa tanktheyalmost
alwaysgivetherevenuesto a religiousinstitution (vihdra). True,the
very last paragraph of the Mahdvamsa (sixth century A.D.) credits
King Mahasenawiththe construction of sixteennamedtanksand a
canal. But the tanksdo not forma single hydraulicsystemand
severalof themhave alreadybeen mentionedin earlierchaptersas
presentedby previousmonarchsto varioustemples. None of this
suggestscentralisedcontrol,or efficient maintenance.
The onlydirectevidencelinkingthe kingship,as such, withthe
irrigationsystemis that the ritualof accessionto the thronewas
markedbywarlikewatersports heldattheTissa tankatAnuradhapura.
From the eighthcenturyonwardsthe capital oscillatedbetween
Anuradhapuraand Pollonaruwa,citiessixtymilesapart,and there
were repeatedoccasionswhenthe countrywas overrunby foreign
conquerorsor split up into small warringstates. Despite these
politicaltroublesthe irrigation systemsurviveduntilthe end of the
twelfth centuryand evenexpanded. But thereis nothingto suggest
that this expansioninvolvedany kind of centralisedplanningor
control. Many of the modernruinssimplyrepresentreplacements
of stillolderruinsdependentupon the samewatersupply. At least
one highlyexpertjudgeholdsthatthereneverwas a periodwhenthe
wholesystemwas all intact.33
One of Wittfogel's thesesis that,once the hydraulicbureaucracy
has establishedits centralised totaldespotism,the despotwillthere-
afteronlyadvancethe needs of agriculture to providehimselfwith

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON 23

serflabourersand soldiers. In general,he argues,themostimpress-

ive majorirrigation worksof hydraulicdespotismsare createdby the
kingsprimarily to enhancethebeautyof theircapitals. This thesis
fitswellwiththeCeylonfacts. But in theCeyloncase,althoughthe
majorirrigation worksprovidedfoodforlabourersas wellas amenities
forpalaces,thehydraulic systemwas notofcrucialeconomicsignific-
ance forthe societyas a whole. Whenthe centralgovernment was
disruptedand the majorworksfellinto disrepair,villagelifecould
carryon quite adequately;for each villagestill possessedits own
small scale irrigationsystemwhichwas maintainedby the villagers
themselves.
The major workscollectively representa colossal investmentof
laboureffort buttheirconstructionwas haphazardand discontinuous
and spreadovermanycenturies. We cannotinferfroman inspection
of these worksthe existenceof a large labour forceunder central
governmentcontrol;nor can we make inferencesabout the size
of thepopulationwhichwas fedby irrigation system. Stillless can
we make inferencesabout the natureof politicalauthorityin the
ancientstate.
In latertimes,Knox describedthe seventeenth-century monarchy
as an unqualifieddespotism, buthe makesno reference to thekindof
bureaucraticcore which, on my reading of Wittfogel,the state
shouldstillhave possessedif Wittfogel's generalthesiswerecorrect.
What was true of Ceylonwas not necessarily trueelsewhere,but
the Ceylon evidence as a whole clearlycountsagainstthe general
validityof Wittfogel's argument.
Someof the factsof the Ceylonstoryfitthethesiswell; othersdo
not. Let us concede that it is a reasonablepropositionto assert
that it cannotbe a mere accidentof historythat so manyof the
ancientmajor statesshould have startedfroma "hydrauliccore".
Some commonsociologicalprinciplemust be at work. But what
kindof commonfactorshouldwe lookfor? And are we necessarily
dealingwitha universalsocial law, or onlywitha principlewhich
cropsup in somecases and notin others?
Wittfogel claimsthathe has discerneda universalprinciple- the
natural and intrinsicassociationbetweenthe formationof large
scale irrigationsystemsand the formation of bureaucratic-despotic
politicalstructures.
It maybe that,in Chineseeconomichistory, development was as
he says. But did historyrepeatitselfelsewhere?
The Ceylon data remindsus thereis at least one otherkind of
socialfactorwhichis commonto a greatmanyhydraulicsocieties-

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
24 PAST AND PRESENT

namelythat they are, froma structuralpoint of view, peculiarly

static. Wittfogelhimselfrecognises,and indeed emphasises,this
stasis but attributesit to the stagnantconservatismimplicitin
"monopoly bureaucracy"(pp. 422-3). The real explanationis
much simpler.
UnderCeylondryzone conditions, once a villageand itsirrigation
tankhavebeenconstructed, itis thereforeverand sincetheirrigation
area mustalwaysremainthe same size,the populationof the village
itselfcan onlyvarybetweenverynarrowlimits.
Such a situationof enforcedphysicalimmobility lendsitselfto the
developmentof social arrangements wherebythe populationsof
neighbouring villagesstandin fixedsocio-economic relationsone with
another.
Durkheim'sanalysis of the Division of Labour inclines us to
imaginethatcomplex"organic"systemsof labourspecialisation are
intimatelyassociatedwith the existenceof a monetaryexchange
medium,thepaymentof cash wages,and thefreemobility oflabour.
Butin the"castefeudalisms"ofIndia and Ceylon,specialisedoccupa-
tionsareprofessed notbyindividuals butbywholegroups(sub-castes),
and, provided these sub-caste groups are physicallyimmobile,
complex organicstructuresof specialisedlabour divisioncan be
maintainedfromgenerationto generationwithoutany intervention
of a moneymedium. The economiccontractis a permanentone
betweensettledgroupscontinuously resident
in particular localities.
Could it be thatthesociologicalexplanation of whyso manyof the
ancientsocitieswere"hydraulic"societiesis that,in a wide variety
of circumstances, hydraulicsocietylends itselfveryreadilyto the
development ofspecialisedlabouron a non-monetary basis?
I am not puttingforwarda new universallaw but I suggestthat
thereare manycases of hydraulicsocietywherethe characteristic
patternis "Indian" rather than "Chinese" and, in which, in
Wittfogel's terminology, the authority structure is "Feudal" rather
than"Oriental". I believethatthislinksup withthebasicdifference
betweenthe Chineseacceptanceoflegitimate authority as manifested
in patrilinealkinshipand bureaucraticgovernment as againstthe
Indian reverencefor personal Charisma. The Indian type of
hydraulicsociety,of which Sinhala is an example,is cellularnot
centralised in structure; localisedgroupsoftechnicalspecialistsform
a workteam centredin a leader. The majorhydraulicworksare
not createdrationally and systematically but haphazardas pieces of
self advertisement by individual leaders. But once started,such
constructions surviveand can be enhancedbylateradventurers ofthe
same type.

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
 HYDRAULIC SOCIETY IN CEYLON 25

The politicalconditionsin "Indian" hydraulicsocietymay vary

verygreatlyin the courseof a fewcenturiesand forthe analysisof
such changingcircumstancethe labels Oriental Despotism and
Feudalismmaybothprovehelpful. Butto turnMarx'sunelaborated
conceptintothedemonofourtimewillgetus nowhereat all.34

Cambridge E. R. Leach

NOTES
1 An earlier draftof this paper was read to a London School of Economics
Seminar underthe ChairmanshipofProfessorI. Schapera (Social Anthropology)
in the springof I958.
2 K.
WittfogelOrientalDespotism(New Haven I957) p. 373 and 373 note c.
3Wittfogel, p. 374.
4Engels, Herrn Eugen DuhringsUmwalzungder Wissenschaft. Dialektik der
Natur I873-I882 (Moscow I935), p. I65.
5Engels, The Originof theFamily (4th edn. London 1942), p. 201.
6Ibid., p. I50.
Wittfogel,p. 411.
8 In Weber's terminologythe Indian type is labelled "hereditarycharisma"
in contrastto the developed Chinese type which is "patrimonial officialdom
(M. Weber, The Theoryof Social and EconomicOrganization,Edinburgh I947,
p. 34i). The latteris regarded as an historicaldevelopmentfrom an earlier
"hereditarycharismaticfeudal" orderbroughtabout by the ultimateascendancy
of the "irrigation and constructionBureaucracy" (Weber, The Religion of
China, Glencoe I95I, Ch. 2).
9 Wittfogel,ch. 6, passim.
10Ibid., pp. I97, 4I7.
"D. G. E. Hall, A History of South-East Asia (London I955), ch. 2;
Wittfogel, p. 374.
12 Rhoads
Murphey, The Ruin of AncientCeylon,Journalof Asian Studies,
Vol. XVI, No. 2, 1957, p. I94. Murpheyis not concernedwithwhythe hydraulic
economyof Ceylon was successfulbut withwhyit ceased to be so. He assumes
that classical Sinhala was organised in accordance with Wittfogel'stheory.
13
Toynbee, A Study of History,referstwice to the Ceylon evidence (Vol. II,
PP. 5-9; Vol. IV, pp. 45-48). He mentionsno historicalsources but cites onlya
romantictravel book by Still (I930) and Leonard Woolf's distinguishednovel
The Village in theJungle(1913). The latter incidentallyconcerns a part of
Ceylon some hundreds of miles away fromthe Anurahapure "hydrauliczone"
which Toynbee purports to be discussing. Because of his bad sources he
arrives at the completelyfalse conclusion that all dry zone irrigationin Ceylon
depended upon bringingin water fromoutside the area.
14 Cf. R. L. Brohier,AncientIrrigationWorksin Ceylon, Vols. Colombo,
3
I934-35. H. W. Codrington, Ancient Land Tenure and Revenue in Ceylon,
Colombo, 1929. R. W. Ievers, Manual of the NorthCentral Province,Ceylon,
Colombo, I899. H. Parker,AncientCeylon,London, I909. Sir J.E. Tennent,
Ceylon, 2 Vols., London, i86o.
15 B. H. Farmer,PioneerPeasant Colonisationin
Ceylon(London, I955), ch. 2.
6 Brohier; Farmer; (op. cit).
17
Codrington,op. cit., p. 64.
18 S. Paranavitana,Glimpses thePolitical and Social Conditions Medieval
of of
Ceylon, in Sir Paul Pieris PresentationVolume (Colombo, I956), p. 72
19Brohier; Parker; (op. cit.).

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions
26 PAST AND PRESENT

20Tennent, op. cit., Vol. I, Part III, ch. xi.

21
Robert Knox, An HistoricalRelationoftheIsland ofCeylonin theEast Indies
(Glasgow, 1911, originaled. I68I), p. 9.
22 Ibid., p.
52.
23
Tennent, op. cit., Vol. II, p. 623.
24
Ibid., Vol. I, p. 407 n.
25
Actually the same work is often attributed to several quite different
monarchs.
26
Cf. Paranvitana,op. cit.
27Wittfogel, p. 417.
28 Sir
John D'Oyly, A Sketch of the Constitutionof the Kandyan Kingdom
(Ceylon) (Colombo, 1929), p. 27; cf. Ralph Pieris, SinhaleseSocial Organisation
(Colombo, 1956), pp. 233-5, 249-50.
29
Wittfogel, pp. 80-82.
30
Fredrik Barth, The System of Social Stratificationin Swat, North
Pakistan, in E. R. Leach (ed.), Aspectsof Caste, Cambridge Papers in Social
Anthropology,No. ii (in press). And cf. McKim Marriott (ed.), Village
India (Chicago, 1955).
31
Codrington,op. cit., chs. iii and iv.
32
M. B. Ariyapala, Society in Medieval Ceylon (Colombo, 1956), ch. 3;
Paranavitana, op. cit., p. 73 mentions a body called the dolos-maha-va-tana
which was a Government"irrigationdepartment"of some sort but thereis no
evidence at all thatit played a politicalrole.
33
Farmer,op. cit.,p. i88.
34Ram Krishna Mukherjee, The Dynamicsof a Rural Society(Berlin, 1957),
pp. 69-80. It is of some interestthatthis author,who now writesabout India
froma Marxist point of view, appears to findno inconsistencybetween earlier
and later Marxist formulationson this topic.

This content downloaded from 185.44.77.136 on Sat, 21 Jun 2014 22:08:49 PM

All use subject to JSTOR Terms and Conditions