Sei sulla pagina 1di 22

Size and the Structure of Authority in Canal Irrigation Systems

Author(s): Robert C. Hunt


Source: Journal of Anthropological Research, Vol. 44, No. 4 (Winter, 1988), pp. 335-355
Published by: University of New Mexico
Stable URL: http://www.jstor.org/stable/3630504
Accessed: 10-08-2015 20:12 UTC

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.

University of New Mexico is collaborating with JSTOR to digitize, preserve and extend access to Journal of Anthropological
Research.

http://www.jstor.org

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
JOURNAL OF
ANTHROPOLOGICAL RESEARCH
(Formerly SouthwesternJournal of Anthropology)

VOLUME 44 * NUMBER 4 * WINTER * 1988

SIZE AND THE STRUCTUREOF AUTHORITYIN


CANALIRRIGATIONSYSTEMS
RobertC. Hunt
ofAnthropology,
Department Brandeis MA02254
Waltham,
University,
It is widelyassumedthatall irrigationsystemsmusthaveconstituted authorityand that
all largeirrigationsystemsmusthavecentralized authority.Thesmall literaturewhich
teststhesebeliefsis reviewed,and theresultsof a systematic comparativestudyarepre-
sented.Theconcepts of "irrigation systemsize,"and"irrigation
system,""irrigation system
structureofauthority" aredefined.Variables whichmeasuretheseconcepts areconstructed,
and measurements froma purposive samplearedisplayed. A handfulof smallirrigation
systemswithoutauthoritystructureswasfound.An inspectionof the data revealsno
relationship betweensize and thestructureof authorityin systemsrangingfrom 700 to
458,000 ha. Furthermore, an irrigationsystemof 458,000 ha is managedbyfarmers.
Conclusions: (1) irrigationsystemswithoutconstituted existand(2) largesystems
authority
do not requirecentralauthority.

CANALIRRIGATION are generally believed to require an authority


SYSTEMS
structureif constructionandoperationare to be successful,andall largeir-
rigationsystems are widelyassumedto requirecentralizedauthority.These
two largelyunconfirmed plausibleandhavebeenwidely
beliefsare superficially
used by scholarsto interpretarchaeological,historic,andethnographicdata.
Whenthese propositionsare argued(whichis rare),authorityis deemednec-
actionandto controlcompetition:
essary to effect coordinated

controlof irrigation, makesa strongcentral


andeven perhapsinstallation,
power essential. (Linton1939:286,on the Betsileoof Madagascar)
irrigation. . imposedthe necessityfor a closely integrated society, since
anelaboratesystemof canalscanonlybe maintained andthe watershared
out by strictcontrol.(Bushnell1957:56,on the late Formativein Peru;
emphasis added)
The more important of these irrigationsystems must have been con-
structed and administeredby a body of irrigationofficialsdirected by a

335

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
336 OFANTHROPOLOGICAL
JOURNAL RESEARCH

centralizedpre-Incagovernment.(Forbes 1955:8, on Peru; emphasis


added)

Withthe publication of Wittfogel'sOrientalDespotismin 1957, scholarsbegan


to concentrateon the positiverelationship betweencentralauthorityandthe
size of the irrigationsystem (in spite of the fact that Wittfogelclearlywas
discussingall types of hydraulicworks, not merelyirrigationsystems; see
Mitchell1973):

Large-scalehydraulic activity... does makemanyof the technicaland


socialdemandswhichWittfogelclaimsfor it. The mobilization of large
quantitiesof labor for the and
construction maintenanceof such works,
andthe allocationof wateron an equitablebasisamongcompetingcom-
munities,alikerequirethe presenceof a superordinate
authority.(Wheat-
ley 1971:298, on China; emphasis added)
No evidencehas yet appearedfor anyMayasystem of watercontrolso
largeas to requirestatemanagement. (Bronson1978:279,on the lowland
Maya;emphasisadded)

These generalbeliefsare in the formof two nomotheticpropositions:(1)


allcanalirrigationsystems haveconstitutedauthorityformanagement and(2)
alllargecanalirrigationsystems have centralizedmanagement authority.These
propositions containthree concepts which refer to empiricalentities:the ir-
rigationsystem itself, the size of that system, and the authoritystructure
whichmanagesthat system. In this paperI subjectthese propositionsto
confirmation by examiningandhopefullyreducingthe uncertaintyassociated
with them. Insofaras a particular propositionis used to interpretdataor to
serve as a foundationforotherpropositions, thenthe uncertainty whichpertains
to it is as importantas its content.
A majorstrategy to reduce the uncertaintywhichattachesto a general
propositionis to makeits conceptsas unambiguous as possible.As they have
been used in the literature,allthree of the aboveconceptsare less clearthan
they mightbe. No authordefinespreciselywhat is meantby "anirrigation
system"or by "largesize,"andfew haveeven specifiedwhatsize systemthey
are dealingwith.No moreprecisearethe conceptsof centralauthority,strong
centralpower, centralizedgovernment,or state management.Measuresof
these conceptshave also been unclearor have been unreliablein application
to particulardata. These two sources of uncertaintyinhibitourjudgmentof
the internalvalidityof the propositions(see CampbellandStanley1963). In
addition,the most generalformof the propositionshas been challengedby a
case study, since supported by additionalcases, in which Netting (1974b)
proposed that an irrigationsystem can exist and persist without any organized
authority whatsoever.

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
SYSTEMS
IN IRRIGATION
AUTHORITY 337

PREVIOUS
STUDIES

Systematiccomparativeanalysisis the majorstrategy for attemptingto


confirma generalization abouta statisticalrelationship.To datethe relationship
between size and centralization of authorityin irrigationsystems has been
examinedin fourcomparative studies:Millon(1962),Mencher(1966),Kappel
(1974), and Earle (1978). Each of these studies containslarge amountsof
uncertainty.
Millon'ssampleconsistedof seven cases, four of them alreadypublished
(the Sonjo,a smallchiefdomin Tanganyika; the 12-Go system in Japan;Pul
Eliyain Ceylon;andGeertz'searlymaterialon Bali).The threeadditional cases
were derivedfromthenunpublished sources(Fernea'saccountof the El Sha-
banain southernIraq;an accountof a "tribal" system in the Hadramaut; and
Millon'sownmaterialon severalvillagesin the SanJuanTeotihuacan Valleyof
Mexico).Millonmeasuredsize andauthoritystructurefor these seven "rel-
ativelysmall"systems andconcludedthat "thereis no clearrelationship be-
tweendegreeof centralization of authorityandthe size of the irrigationsystem
or the numberof personsit supports.Furthermore,the practiceof irrigation
hasnot, apparently, broughtanysubstantial growthof centralauthority" (Millon
1962:56).Althoughconfinedto "relativelysmallsystems,"these resultschal-
lenge the generalbelief aboutthe relationshipbetween size and authority.
Millon'sconclusionshavebeenacceptedby Wolf(1966:26,n. 8), Price(1971),
Lees (1973), andMitchell(1973).
Mencher(1966)comparedtwo areasof southIndia,focusingon nucleation
of villages,centralizationof kingdoms,andthe ecologyof waterworksandland
transport in Kerala and Madras. Shefoundthathighlynucleatedvillages,highly
developed roads, and more highlyintegratedkingdomswere associatedwith
the areashavinglowerrainfallandlargerirrigation works,thussupporting the
general belief.
Kappel(1974)selected seventeensocietiesfromthe Outlineof WorldCul-
turesandmeasuredthreevariables:(1) size anddensityof population, (2) size
of irrigationfacilities,and (3) degree of centralization of decision-making au-
thority. He concluded that as population size and density increase, so also do
political centralizationand the size of irrigationsystems.
Earle(1978)comparedhis Hawaiian datawithfourothersingle-village small
systems (PulEliya,Ifugao,Moala, and the Sonjo). He concluded that specific
managerial rolesseemedrelatedmoreto the socialmatrixthanto the demands
of irrigation(Earle1978:135).
The resultsof these studiesare equivocalin the cases of MillonandKappel
and weak in the cases of Mencherand Earle. Earle'ssampleexhibitsno
variationin the structureof authorityandvery littlein size. As a consequence,
littlecanbe learnedaboutthe relationship betweenthetwoconcepts.Mencher,
in effect, does a concomitantvariationstudy withinone regionof India.As
Clignet(1970) and PrzeworskiandTeune(1970) have shown,this research

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
338 RESEARCH
OFANTHROPOLOGICAL
JOURNAL

design does not permitidentification of all, or even the major,sources of


variancein the dependentvariable.
The comparativestudiesby MillonandKappelhave weakinternalvalidity.
Millondefinedhis variableof size both as acres irrigatedandas population.
Whenhe actuallyused acres irrigated,however,he in some cases measured
other phenomena:the size of a smallkingdomin Baliandthe area of tribal
politicalcontrolin Iraq.This practiceintroducesmassiveuncertaintyintothe
resultingrelationshipbetweenthe two variables.Millon'sresults have been
questionedby Bennett(1974), who calledattentionto the smallnessof his
sampleandthe limitedsize rangeof the irrigationsystems studied,andby R.
HuntandE. Hunt(1976),who raisedquestionsabouthis measurements.
Kappel(1974:162)labels the variablefor size "extentof irrigation" and
definesit as miles of canal,acres irrigated,or the system's relativesize as
noted by ethnographers.However,his results are expressedin household,
village,district,andgovernmentunits, therebymakingcomparisonwith the
originaldatadifficult.In addition,his variableforauthorityrefersto the whole
society,not the irrigationsystem per se. Finally,if we accepthis measureof
size, Kappel'sconclusionin the text does not agree withthe datapresented
on figure13.2, "Relationships of Sizeof Irrigation andPopulationSize"(Kappel
1974:164).
In conclusion,seriousproblemshavebeen identifiedwiththe conceptsand
theirmeasuresas used in the comparative studies,andthe internalvalidityof
these studiesis thuscompromised. Furthermore, sincethe samplesusedhave
no knownrelationship to any universe,the externalvalidityof these studies
is also extremelylow. As a consequence,littleor no "knowledge" aboutthe
of
relationship irrigationsystem size to the structure of authorityhas been
gained.

THIS STUDY

Thispaperpresentssomeof the resultsof systematiccomparative research


on the authoritystructureof canalirrigation
systems.' In orderto evaluate the
likelihoodthat the nomotheticpropositionsare correct,the concepts of irri-
gationsystem, administrative authoritystructure,andsize were narrowlyde-
fined;measurementsof authoritystructureand size were constructed;and
these variableswere measuredon a smallpurposivesample.

Universeand Sample
systemsin states.
The universeforthis studyis confinedto canalirrigation
Since Wittfogel'sOrientalDespotism,discussionsof irrigationsystems with
centralizedauthorityhave at least impliedthat the state is involved.It is
criticallyimportantthat each case in the sample have the potential to be po-
litically centralized, a criterion which bands and probablychiefdoms do not
meet. States have a politicalcenter (by definition),layers of territorialadmin-
istration, and peripheral organizations(villages, kin groups, ethnic groups,

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
AUTHORITY
INIRRIGATION
SYSTEMS 339

corporations),any of whichcouldconceivablysupplythe authoritystructure


for a canalirrigation
system. In states, then, a choicemustat least potentially
be madeaboutwhichlevel of authoritystructurewillruna canalsystem.
For ease of access to data,the universefor the presentstudymustfurther
be confinedto modernnationstates. Althoughthe worldnow containswell
over a hundredthousandcanalirrigationsystems,2few havebeen described.
The majorityof the availablestudieshave appearedsince 1970anddealwith
contemporary conditions.Datafromhistoricalinvestigationsof irrigationsys-
tems wouldbe a valuableadditionto this study, but since such researchis
very costlyandlengthy,few havebeen done(Glick'sstudyof Valencia[1970]
is the best example;the few others includeKelly[1982], Stone [1984], and
Ludden[1985]).

SamplingUnit-The IrrigationSystem
A persistentproblemin nomotheticstudiesis to finda stableunitof analysis.
In a recent paper Kelly pointedout that "the irrigationsystem"seems to
functionas the masteranalyticconcept,but it is usuallyleft undefined.He
quotes one of the few definitionsof it in the literature:"Anirrigationsystem
is anarrangement by whichwateris conveyedfroma sourceto anareaneeding
water to facilitatethe productionof desiredcrops"(Kelly1983:881,quoted
fromvan der Mere 1968:720).Vander Mere'sdescriptionstates thingsthat
are true of irrigationsystems, but it tells us nothingabouttheirdiscreteness
and is thereforeof no use in decidingwhere the boundariesbetween such
systems are.
Ambiguityandmultivocality in the unitof analysiscontributeso muchun-
certaintyto irrigationsystemstudiesthattheirresultsareseriouslyweakened.
Oneexamplemakesthis clear.Millon'scomparative study(1962)includedthe
El Shabana,a tribalgroupin southernIraqstudiedby Fernea(subsequently
publishedin 1970).Millonmeasuredthe irrigationsystem size by the number
of hectaresthe tribeoccupied.However,Iraqiirrigation engineers,appointed
andhiredby the nation,hadsystem-wideadministrative authority,andvarious
unitsof the tribereceivedwaterfromthe nationalcanalsthey administered.
The territorycoveredby the irrigationsystem andits bureaucracy was thus
vastly greaterthanthe territoryoccupiedby membersof the tribe.This sit-
uationis intolerablein nomotheticstudies,forno cross-systemcomparisonis
possibleif two or moreunitsof analysisare used. In nomotheticstudieseach
variablemustbe measuredon identically definedsystems. Wethereforemust
have a way of identifying particularirrigationsystems.
Since any large-scalecomparativestudy is forced to rely on the extant
literature,the samplingunitused must be identifiable in secondarysources.
Ideally, the unit should be defined so that it is also applicableto archaeological
and historicalsources. The followingdefinitionis an attempt to specify how to
find the boundaries of particularirrigationsystems: a canal irrigationsystem
is composed of (1) a facility (gate, offtake) which takes water from a natural
channel and moves it away from its naturaldownhillcourse and (2) the sub-

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
340 OFANTHROPOLOGICAL
JOURNAL RESEARCH

sequentcontrolworks(canals,gates, fields)thatguidethe waterflowingon


the surfaceto the agricultural
plantsuntilthatwatereithersoaksintothe earth
or flowson the surfaceout of the controlworks.3

The Variable ofAdministrative Authority


All the theoriesaboutcentralization of control,or lackof it, concernthe
organization of power.Inthe longrunwe wouldliketo haveanswersto several
fundamental questionsaboutpowerin the contextof irrigation: Whohaspower
over water,andhow andwhy do they have thatpower?Do those who have
powerover waterhave powerover anythingelse? Do those who have other
sources of power have power over water decisions?And, finally,is there
directionalitybetweenpoweroverwaterandpoweroveranythingelse? Power
has been notoriouslydifficultto define,observe,andmeasure,especiallyfrom
secondaryliterature.The few extendedaccountsof conflictin the irrigation
literatureseem clearlyto reflectpowerdifferencesbroughtto bearonirrigation
decisions(Gallin1966;vander Mere 1971;E. Hunt1972).Whilethese cases
have stronginternalvalidity,no justification exists for generalizingfromthem
to a wholevillage,to a system as a whole,or over any substantialtime span
(i.e., theirexternalvalidityis low).
Since poweris problematic to measureandits measurementis all but im-
possible from existing sources, the emphasisin this paperwill be on the
structureof authorityin the administration of irrigationsystems. A nearly
universalcontextof power,especiallyin states, authoritymay be definedas
the legitimaterightto wieldpower.Authorityis rarelyirrelevantto powerand
at times probablymapsit ratherwell. Authoritycan be measuredrelatively
easily(comparedwithpower)bothin the fieldandin the secondaryliterature.
The phrase"centralized authority"appearsrepeatedlyindiscussionsof canal
irrigation,andyet it is rarelydefinedclearlyor used consistently.In a recent
articleKelly(1983)has begunthe task of segmentingthis muddledareainto
separatedimensions.He notes that two majorconceptshave been involved
in previousdiscussionsof centralizedauthority:(1) the internalconfiguration
of authorityamongrolesof a systemand(2) the externalrelationship of these
irrigationroles to roles in othersocialsystems, especiallythe politicalsystem
of the state. Kelly (1983:883)suggests the terms "centralization/decentrali-
zation"to designatethe dimensionof internalorganization, to indicate"the
degree to whichirrigationroles are hierarchically configuredandauthorityin
irrigationtask performance is concentrated."He suggests"articulation/auton-
omy"to indicatethe degree to whichthe irrigationorganization is linkedto,
or is independent of, the state. Kelly'sseparationof the conceptof centralized
authorityinto these two dimensionsis potentiallyvery useful.But since the
term "centralized" may refer to bothdimensions,I suggest "concentrated /
dispersed"for the end pointsof the dimensionof internalcoherence of authority.
A system without constituted authoritywould have a maximumof dispersed,
as opposed to concentrated, authority and might be labeled "acephalous."A
system with a constituted authority system might be labeled "unified."The

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
INIRRIGATION
AUTHORITY SYSTEMS 341
nomotheticpropositionwe are investigatingstates, in effect, that all canal
irrigationsystems are unified.
The meaningof centralizedauthoritymost prominentin the literaturecon-
cerns the kindof linkagebetweenthe authorityroles of an irrigationsystem
and externalentities, usuallythe state.4This articulation is at the core of
Wittfogel'stheory and is the concept of centralizedthat Millonused in his
comparativestudy.We shouldnote, however,thatallcanalirrigationsystems
in modernstates are articulatedin some way withthe state. Even the most
independentirrigationcommunitiesoften have some sort of legalpermission
to organizeandare frequentlyjuralpersons(R. Hunt1987).5This dimension
(Kelly's"articulated/autonomous") is the focusforthispaper.Onewayto state
the questionis whetherauthorityover the decisionsandactivitieswithinthe
irrigationsystem is locatedwithinthe irrigationsystem or externalto it.
An administrative structuremust containworkto be done andindividuals
to do the work.Inthe administration of canalirrigation,severaltasksandroles
are involved.Severaluniversally foundworktaskshavebeenidentifiedin canal
irrigationsystems, includingconstructionof the physicalsystem, captureof
waterfromthe environment,allocationof wateronce captured,maintenance
of the physicalsystem, conflictresolution,andaccounting.Drainageandritual
tasks are also sometimesfound.If an authoritystructureis responsiblefor
these tasks, then administrative roles must exist to performthem. Systems
withconstitutedauthorityare headedby a chiefexecutiveofficer,6definedas
that officerresponsiblefor allocationat the facilitywhere the system takes
water fromnature.Varioussubordinatestaffmay also be responsiblefor al-
location.Some or all of the other tasks of the irrigationsystem maybe per-
formedby this staff. The CEOusuallymust reportto a personor persons
occupyingsomeroleor set of roles (Ministerof Irrigation, Boardof Directors,
etc.).7 All irrigationsystems contain the role of farmer. For some tasks (e.g.,
maintenance andconstruction), large numbers of workers are required.These
workersmaybe full-timeemployeesof the system, they maybe farmerspart
of whose responsibilityis to do this work, or they may be laborershired
temporarily for the task at hand.
A featureof authoritysystems is a charterfor that authority.All formally
constitutedcanalirrigationadministrative systems have some sort of charter
for their authorityto run the system. The variablecharterof authorityhas
three valueswith empiricalcontentin canalirrigationsystems: nationalgov-
ernment,irrigationcommunity, andprivate."These termsreferto the source
of legitimacyfor the chiefexecutiveofficer'sexercise of authorityover allo-
cationat the headgate.
Witha nationalgovernmentcharter,the centerof the polity(a ministryor
an officialconnected directly to the head of state) has direct responsibilityfor
operating the irrigationsystem, appoints the CEO, and formulates the rules
for operating the system. In modern times the individualswho occupy the
position of CEO in systems with a nationalgovernment charter usually have
formaleducationin civil or agriculturalengineering, are paida salaryin money,

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
342 OFANTHROPOLOGICAL
JOURNAL RESEARCH

andoftenappearin the centralgovernmentbudget.They tendto be urbanin


life-styleand aspirationsand often in residenceas well. The standardsand
sanctionsfor theirperformance are set by the ministryoverseeingirrigation
(R. Hunt n.d.b).
Systems with an irrigationcommunitycharterare often calledindigenous,
traditional,or community-based (Beardsley1963;Coward1976, 1979, 1980;
MaassandAnderson1978). In these cases a corporategroupof water-rights
holders(the irrigation community) elects or appointsthe CEO,whohas direct
responsibility for the operationof the headgate.Thisindividual frequentlyhas
no formaleducationin engineeringandis usuallya memberof the corporate
groupand a directuser of the water himself.He may receive no directre-
compenseat all. The less monetizedthe system, the more likelythat his
recompensewillbe in the formof relieffrommaintenance labor,extrawater
or land,or produce.Onlyvery recentlyhas his recompensebeen in the form
of money.Suchindividuals usuallyareresidentsof villagesor ruraltowns,have
a rurallife-style,andhavefew significant urbanaspirations.The standardsand
sanctionsfor executingthe dutiesof the officeare determinedandcarriedout
locallyby the corporategroup.At leasttwo cases onrecordinvolvewhatmight
be calledconsortiaof irrigation communities: Valenciain Spainandthe King's
River WaterAssociationin Fresno, California.In both cases a numberof
irrigationcommunitiesare locatedon a riverandcombinetheireffortsto deal
withvariousjointmatters.Whilesuchconsortiaare not strictlyirrigation sys-
tems as definedabove,theircharterof authorityproceedsfromfarmers,and
in this sense they are likeirrigation communities(R. Huntn.d.b).
Irrigation
systems withprivate charters are runby individuals
whoin effect
charterthemselves.This commonlyoccurswhenan entrepreneurfundsthe
buildingof the system. Mostof the cases knownto me involverelativelyhighly
capitalizedagricultural enterprisesin LatinAmerica(particularly sugarmills)
(see BarkinandKing1970;Ronfeldt1973).Suchindividual entrepreneurscome
froma widevarietyof backgrounds; they can be merchants,locallygenerated
caciques(politicalbosses), or foreigncapitalists.In almostevery case, the
irrigationsystem is seen as belongingto the entrepreneur, ratherthanto the
nationor a groupof farmers(R. Huntn.d.b).
Amongthese types of charters,considerablevariationis exhibitedin the
numberof tasks the CEOis responsiblefor andin how fardownthe system
that responsibilityextends. Irrigationcommunitiesare the most denselyor-
ganizedfrom this pointof view: the officeresponsiblefor allocationis also
responsiblefor maintenance,for accounting,and for the early (andin some
cases nearlyallof the) stages of conflictresolution(see R. Hunt1978,n.d.b).
These responsibilities extenddownto the farmerlevel. Systemswithnational
governmentchartersarefarmorevariable.Sometakeresponsibility forevery-
thing right down to the farmgate (the Gezirascheme in the Sudanis the prime
example of this [Dishoni 1966; Farbrother 1973]). Others will only take re-
sponsibilityfor allocationand maintenanceat the main-canallevel, leaving all

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
AUTHORITY
IN IRRIGATION
SYSTEMS 343

othertasksandallocation andmaintenance atlowerlevelsto otherorganizations


or to no organization at all (see R. Huntn.d.a for an extendeddiscussion).
The questionof whetheranirrigation systemhasa constitutedauthoritycan
be answeredusingthe conceptsdiscussedabove.A systemwithoutanauthority
system has no CEO.Systemscanbe unified(anauthoritysystemwitha CEO
is present)or acephalous(no authoritysystem andno CEO).
The charterof authorityis the variableused for measuringadministrative
authoritystructurein thispaper.Charteris the sourceof authorityof the chief
executive officer,who is responsiblefor carryingout allocationdecisionsat
the head-facility.
The scale of thisvariableis nominal.Its valuesas isolatedso
far are (1) nationalgovernment,(2) irrigationcommunity,and (3) private.
Irrigationsystems with a national(or provincial) charterof authorityare po-
liticallycentralized(i.e., state-controlled).Systems with an irrigationcom-
munitycharteror witha privatecharterareclearlynot state-controlled; however
they maybe articulatedwithoutsideinstitutions.

TheVariable of Size
Three differentmeasuresof the size of anirrigation systemhavebeen used
in the literature.The populationcontainedwithinthe boundariesof a system
has been used by Millon(1962)andEarle(1978).The lengthof the maincanal
or the totallengthof the canalshavebeenproposedby Kappel(1974).Finally,
the overallextent (area)of the system has also been used by Millon(1962)
and Earle (1978). In addition,distinctionssuch as technologicalcomplexity
havebeen attractiveto some, whileorganizational complexityhas appealedto
others(see Netting1974a;Spooner1974;Uphoff1986).Theliteratureis silent,
however,on where andhow one is to measurethese features.
Use of the populationof a canalirrigationsystem as a measureof its size
is an attractiveconceptformanyreasons,butthe difficulties in conceptualizing
and measuringit are very great and have not been overcome.A primary
problemis decidingwho is to be counted.If one focuseson the farmers,then
one has to decide whetherone means the landowners,the farmoperators
(they are oftennot one andthe same),and/orthe farmworkers(whomaybe
partiallyor totallydifferentfromthe previouscategories).An addedcompli-
cationis that withintheir boundariessome canalirrigationsystems contain
ratherlargetowns(andperhapscities), few of whoseresidentsare connected
directlyto the waterinthe irrigation
canals(e.g., Valencia; Fresno,California).
Censustractsare virtuallynever drawnin termsof hydrological featuresand
thereforedo not differentiatebetween irrigatorsandnonirrigators. The size
of the population withinthe boundariesof an irrigationsystem thus maycor-
relate with many other factors, such as technology,the price of land, and
sourcesof employment,ratherthanwiththe size of the irrigation systemitself.
All of these difficultiesmightbe resolved,given the time andresources.At
the present,however,the population of an irrigationsystem is neithera clear
conceptnor one thatcanbe measuredfromthe secondaryliterature.

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
344 OFANTHROPOLOGICAL
JOURNAL RESEARCH
The lengthof the maincanalandthe totallengthof the canalsof the system
wouldbe very instructivemeasuresto have.At the very least, a ratioof the
extent of the system to lengthof canalwouldtell somethingaboutthe likely
maintenance burden.9It is not clearto me thatthe lengthof the canalsystem
willcorrelatewithvariationin administrative structure,althoughit might.But
the lengthsof canalsystems are only very rarelyreportedin the secondary
literature,andthereforethe concept,howeverinteresting,cannotbe converted
into a usefulvariableat this time.
Overallextent has been the usualmeaningof the size of a canalirrigation
system. Althoughno one has definedit, most authorsapparentlymeanthe
areaof the fieldsirrigatedby the system, measuredin acres or hectares.An
advantageof this conceptis that manyauthors(or governmentdocuments)
publisha figureon the extent of a system. Extentis thus botha usefuland
feasiblemeasure,althoughit is not the onlyusefulone imaginable. Difficulties
neverthelessexist with publishedfiguresof the extent of a given system.
Those responsiblefor a system maybe motivatedto inflateits extent, partic-
ularlyif it is a nationalgovernmentsystem. Oftena system is designedto
reach given size, and althoughit never reachesthat size, the designedsize
continuesto be the officialsize of that system. In some systems the amount
of landthatis irrigatedin any givenyearis a functionof the amountof water
in storage. This is true of the 53-hectaresystem in Pul Eliya(Leach1961)
and of a 100,000-hectaresystem in northwesternMexico (Hunt1982 field
notes). Whatthenis the extentof the system-the largestnumberof hectares
thathaveever beenirrigatedin thatsystem, the averagenumberof hectares,
the designednumberof hectares,or the numberof hectaresunderirrigation
thisseason?To select the averagenumberas representingthe administratively
relevantnumberis tempting,andI wouldputleast emphasison the designed
size, for this is occasionallywildlyinflated.For workingin the secondary
literature,however,the questionis moot,forvirtually noonerelatesthe source
of the numberused, muchless how thatnumberis defined.
Even more usefulthanthe abovefiguresfor measuringthe size of an irri-
gationsystem wouldbe a ratioof the numberof canalgates to the extent of
the whole system. This wouldbe a tellingfigure,for it wouldidentifythe
numberof decisionpointsin the system (eachgate mustbe operated;andthe
moregate operations,the moredecisionstaken)andcouldvery well serve as
an indexof administrative density.However,suchdataare rarelyavailable.
The definitionof the size of an irrigationsystem must be linkedto the
definitionof the system itself. If we are attemptingto correlatetwo or more
variables,those variablesmustbe measuredon the samebasicsamplingunit,
in this case an irrigationsystem. The system was definedin termsof a head-
facilityand the controlstructuresleadingthe water away fromthe natural
drainagesystem. The administrative authoritywas measuredby the type of
charter of authorityresponsible for that facility.The size of the system, in turn,
is the extent (measured in hectares) of the fields which are irrigatedfrom that
head-facility.It has a ratio scale. It is very difficultto know how to relate extent

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
INIRRIGATION
AUTHORITY SYSTEMS 345
to the casualuses of terms suchas "large,""extensive,"and"huge"thatare
oftenfoundinthe literature.Oneimportantjobforthe futurewillbe to calibrate
such intuitivejudgmentsaboutsize to a consciousmeasureof size such as
extent.

RESULTS

SystemswithNo Constituted Authority


The vast majorityof canalirrigationsystems reportedin the literaturedo
havea constitutedauthoritystructure.Mexicocontainssomethingof the order
of fifteenthousandnamedirrigationsystems, each of whichhas a constituted
authority(see R. Huntn.d.a). Ofthe irrigation
systemsreportedforthe United
States, all have constituted and
authority, the same holds(so far)forJapan,
India,andIndonesia.Thetotalrecordednumberof systemswithno constituted
authorityis no morethanfourteen.So to date the evidenceindicatesthatthe
vast majorityof systems have formalauthorityorganization. How far backin
time this conditioncan be extendedis unknown.Netting (1974b, personal
communication 1983)has arguedthat a smallsystem in the Swiss Alpsdoes
not now, and has not for at least eight hundredyears, had any constituted
authorityto runit. Earle(1978)has presenteddatafromcontemporary Hawaii
thatcanbe interpretedin the sameway,andde los Reyes (1980a,1980b)has
presentedbriefaccountsof ten smallirrigationsystemsinthe Philippines which
lack constitutedauthority.These systems are all smallin size (less than20
ha) andin numberof farmers(fewerthanthirtyfarmers).

Relationshipof Size and TypeofAuthorityStructure


Table1 shows the distribution of the variableof chartertype withrespect
to the variableof size. Noteworthyare the smallestsystem with a national
governmentcharter,at 700 ha (in Indonesia),and the largestconsortiumof
irrigationcommunities,at 458,000ha (inthe UnitedStates). Inbetweenthese
two extremes are foundbotha numberof irrigationcommunitychartersand
a numberof nationalgovernmentcharters.
The interpretationof charterof authorityshouldbe straightforward.A sys-
tem witha nationalgovernmentcharteris surelycentralized(state-controlled),
whereasan irrigationcommunitychartermeans that the system is not cen-
tralizedbythestate.The interpretation of size, however,is moredifficult.No
commentator inthe literaturehasso farsaidwhatsize systemwouldbe "large"
or "small."To calibratethis variablewith previousstatementsis therefore
impossible.It does seem certainthat any system of 458,000 ha wouldbe
consideredlargeby everybody.Mysuspicionis thatmostcommentators would
of
also regarda system 10,000ha as large.
The main conclusion to be drawn from Table 1 is that, within a very broad
range of sizes (700 ha to 458,000 ha), size alone does not determine which
type of charteris necessary to operate a canalirrigationsystem. Both irrigation
community(local control) and nationalgovernment (external control) charters

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
346 RESEARCH
OFANTHROPOLOGICAL
JOURNAL
TABLE1
Size and Charter

Size
Nameof System Country (in hectares) Typeof Charter
SanJuan Mexico 600 community
irrigation
Tayuban Java 700 nationalgovernment
ZanjeraDanum Philippines 1,500 community
irrigation
VicenteGuerrero Mexico 1,575 community
irrigation
12-Go Japan 5,500 irrigation
community
Moncada Spain 7,000 community
irrigation
Morelia#2 Mexico 8,000 nationalgovernment
New CacheLa Poudre U.S.A. 15,400 community
irrigation
AngatRiver Philippines 26,890 nationalgovernment
RioMayo Mexico 95,973 nationalgovernment
Fresno U.S.A. 97,000 community
irrigation
Chia-nan Taiwan 150,000 community
irrigation
HindiyahBarrage Iraq 209,000 nationalgovernment
King'sRiver U.S.A. 458,000 community
irrigation
GeziraScheme Sudan 730,300 nationalgovernment

Note:SourcesforthesedataarelistedintheAppendix.

can and do operate these systems. In other words, politicalcentralization(i.e.,


state control) of canal irrigationis not necessary between 700 and 458,000
hectares. Size alone is not determiningwhat form of control the system has
(within this range of sizes).
The internalvalidityof these findingsis relatively high. The definitionsof a
system and of the two variablesmeasured are fairlyclear and easily measured
from the secondary literature. If these measures are reliable and valid, the
range of sizes of canalirrigationsystems whichhave either irrigationcommunity
or nationalgovernment charters can only increase."'
There are limitations on the external validity of the findings due to the
definitionof the universe used for the study and to the design of the study
(see Campbelland Stanley 1963; R. Hunt 1979). The samplingstrategy was
such that the relationshipof this sample distributionto the total universe of
irrigationsystems is unknown.The sample used in this study was purposive,
and the purposes were (1) to have data that are analyzablefrom secondary
literature and (2) to maximizethe range of values for each variable. It is in no
sense a random sample of anythingand therefore cannot be taken as repre-
sentative of the distributionof these variables in any universe. Because the
universe was limited to modern nation states, it is not known to what extent
the results of this study can be extended to other kinds of states, to states in
other time periods, and to nonstate societies.

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
AUTHORITY
INIRRIGATION
SYSTEMS 347
DISCUSSION

It is widelyheldthatallcanalirrigationsystems mustbe unified,and,as we


haveseen, thisis notstrictlytrue.Butthe probability thatanyirrigation system
over, say, 100 ha willbe unifiedis very highindeed.If the size of an archae-
ologicallyor ethnographically describedcanalirrigation systemis greaterthan
100 ha, it wouldseem safe to inferthatit has or hada unifiedadministrative
authoritystructure.For any size under50 ha, this wouldappearto be prob-
lematic.
The existenceof acephaloussystems challengesthe irrigationorganization
theorythat the problemsinherentin sharingwaterandperformingthe work
demanda constitutedauthority.The questionto be answeredwithrespect to
these smallacephaloussystems, then, is how they deal with the tasks that
mustbe performed.Allocation mustbe accomplished, maintenance performed,
accountingdone, andconflictresolved.The most difficultproblemsthey face
wouldprobablybe sharingwater in times of scarcityand dealingwith free
riders. If in fact these problemsdo not exist, then the questionto answeris
whytheyareabsent,whenmuchexperienceleadsone to expecttheiruniversal
occurrence.Targetedfieldworkon one of these systemsshouldbe a veryhigh
priorityfor the socialanthropology of canalirrigation.
Wittfogelarguedthatboththe construction andoperationof largehydraulic
systems demand what in this paperhas been called centralized(i.e.,
politically
state-controlled) management. Table 1 clearlyindicates that very largecanal
irrigationsystems canbe effectivelyoperatedby the farmers. If largesystems
do not need or demandpoliticalcentralization, then its existenceis the result
of a choice,not a generalneed. Howthenmightwe proceedto try explaining,
in a generalway, the distribution of state-controlled andfarmer-managed ir-
rigationsystems?
One of the first tasks is to establishwhatis meantby "large."Up to this
pointthe discussionhasbeenalmostentirelyfreeofnumbers.Oneconsequence
of the studyreportedhere is the realizationthatdefiningandestablishingthe
size of a system is criticalfor a discussionof the relationshipof size and
organization. Intuitively,it seems to me thatanythinglargerthan3,000-4,000
ha shouldbe calledlargefromthe pointof view of operations.At this size and
above, more thanone settlementis virtuallycertainto be involvedin using
the system.A villagecaneasilybe the unitof operationof anirrigation system,
forcomplexrelationships andsystemsof management arealreadyinexistence.
Whena system is composedof two or more villages,then a different(and
probablyspecial-purpose) organization hasto exist, andwithit comeproblems
of disciplinewithrespectto rulefollowingandrevenue(R. Hunt1978, 1987).
A potentiallypromisinglineof inquiryto exploreis whetherthe kindof state
has an effect on the size of irrigationsystems and their type of management.
It is tempting to say thatJapanand the United States, for example, differfrom
Mexico and India. But the concepts we use to capture those differences are

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
348 OFANTHROPOLOGICAL
JOURNAL RESEARCH

vitallyimportantandnot entirelyclear.Whilecolonialism by industrial powers


mightwell have somethingto do withthis situation,it willclearlynot account
for China.Indiahas builtseveralstate-controlled irrigationsystems since in-
dependence,andMexicoreceivedpoliticalindependence in 1821, longbefore
its large modernsystems were even thoughtof (see Kroeber1983). The
conceptof "kindof state"hasbeen withus fora longtime,andwe writeabout
agrarianstates, empires,industrialstates, mercantilestates, etc. But as yet
I see no evidencethatthe dimensionor dimensionswhichcapturethese dif-
ferenceshavebeen clearlyformulated,muchless operationalized. Anyfuture
to
attempt explain some of the variation
in irrigationsystems with a variable
of "typeof nationstate"mustbe precededbythe workof creatingthatvariable.
Some factorsmightfavorthe presenceof politically centralized(i.e., state-
controlled)systems. The degree to which a state permitsstronglegitimate
localorganizations mightbe a factor.Irrigationcommunities of anysize involve
morethana handfulof villages,imposedisciplineon theirmembers,andhave
revenue-raising power.Somestates mightviewsuchentitiesas potentialcom-
petitors powerandthus discouragethemby takingover irrigation
for control
at the state level. Some other factorsmightincludestate controlover agri-
culturalproductionand knowledgeof the dynamicsof ruralpolitics(see R.
Hunt1988).
Otherfactorsmightpermitorencourageirrigation communities. Somestates
mighthave suchvariedsourcesof revenueandpowerthatcontrolover canal
irrigationsystems is not necessaryfor the maintenanceof state power. Or
agricultural developmentmayhavereducedthe uncertainty of the foodsupply
to suchanextentthatthe statehasless needto manageagricultural production.
The presenceof suchfactorsreducesthe importance of state controlof canal
irrigation.
Communication technologyaffectsthe speedwithwhichmessagesandper-
sonnelcanbe movedaroundtheirrigation system.Theoperationsofthe system
mightthereforebe profoundlyaffectedby changesin this technology.Con-
versely, they maynot be affected,at least fromthe pointof view of the social
organization of control.The 12-Go system has been in operation,as an irri-
gationcommunity,since at least A.D.1600 (Beardsley,Hall,andWard1959),
andthe Valenciansystems havehadthe sameorganization backto A.D. 1200
at least (Glick1970;Butzeret al. 1985).Inbothplacestechnologyhaschanged
a greatdealover thistimeperiod,as indeedhas the politicalenvironment; yet
the irrigationcommunityhas remainedthe dominantirrigationinstitutionin
bothsystems. It is not clearto me at thispointthatchangesin communication
technologynecessarilymakeany substantialdifferencein the formof charter
thatis necessaryfor operatingthese several-thousand-hectare canalirrigation
systems.
Some of the systems in Table 1 have water storage facilities under their
control, and some do not. The effect of storage on administrativeorganization
in this sample is not clear, but storage is a very old phenomenon (storage
works existed in medieval Ceylon and earlier still in South India [Murphey
1957; Leach 1959; Ludden1985]). Storage wouldreduce variancein the amount

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
AUTHORITY
IN IRRIGATION
SYSTEMS 349

of wateravailableandwouldextendwaterusageinseasonalterms.Butstorage
also encouragesexpansionof the areairrigated,whichwouldin turnincrease
the pressureof scarcity,therebyincreasingthe administrative load.At this
point it is not clearthatstoragehas visibleeffects the
upon authoritystructure
of canalirrigation.
The conceptof politicalcentralizationusedin thispaperis of courselimited.
It demandsa state, perhapsa "modern" state. Andit willnot easilyallowfor
the accretionof independence by localmagnateson the fringesof fadingstates.
Even in modernstates ambiguoussituationswillarise, wheremanagersof a
state enterprisealso have very stronglocalloyalties.But even withall these
problems,it seems to me that the conceptof politicalcentralization opera-
tionalizedhere is close to Wittfogel'sconcept,is close to the ideas that are
presented by many authors,and so far seems to work well in measuring
particular cases.
The resultsof the studyreportedin thispaperare stronglycounterintuitive.
No canalirrigationsystem oughtto be able to exist withoutconstitutedau-
thority.Yet they do. Andno largecanalirrigationsystem oughtto be ableto
functionif managedonly by farmers.Yet they do, andvery successfully.We
shouldbe cautiouswiththe assignmentof confidenceto ourintuitionsin these
matters.
CONCLUSIONS
Thispaperhasrefinedthe conceptofirrigation system,constructedvariables
for measuringthe administrative authoritystructureandsize of irrigation
sys-
tems, andmeasuredthese variableson a smallpurposivesample.A majorgoal
of this exercise was to reducethe uncertaintyof the nomotheticpropositions
concerningthe relationship of size andauthoritystructurein canalirrigation.
It has been arguedhere that (1) very smallcanalirrigationsystems canbe
operatedwithno constitutedauthority;(2) canalirrigation systems of consid-
erable size (458,000 ha) can be, and are, operatedby localirrigationcom-
munities;and (3) canalirrigationsystems of smallsize (700 ha) can be, and
are, runby nationalgovernments.Theinternalvalidityof these resultsis rather
high.The uncertainty of these results,whencomparedwithpreviousstudies,
is substantiallyreduced.Their externalvalidityis more problematic.These
results clearlychallengethe standardpropositionsaboutthe relationship be-
tween canalirrigationandthe structureof authority.

Appendix
Sources for Measurements
Listedbeloware the namesof the systems,datesof observations,andsourcesused
for the measuresof size andtype of chartergivenon Table1.
SanJuan,Mexico;1963-64
Size: author'sfieldnotes
Charter: author'sfieldnotes
E. HuntandR. Hunt1974;R. HuntandE. Hunt1976

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
350 OFANTHROPOLOGICAL
JOURNAL RESEARCH

Tayuban,Java;1983
Size: Duewel1983:19
Charter: Duewel1983:19,n. 47
ZanjeraDanum,Philippines; 1970s
Size: Coward1979:29
Charter: Coward1979:31,32
VicenteGuerrero,Mexico(Unidadde Riego,Durango);1982
Size: author'sfieldnotes, WorldBank
Charter: author'sfieldnotes, WorldBank
Moncada,Valencia,Spain;1968
Size: MaassandAnderson1978:20
Charter: MaassandAnderson1978:22-23
Morelia#2, Mexico(Distritode Riego#020); 1982
Size: DGDUR1982, #020:28
Charter: author'sfieldnotes, WorldBank
12-Go, Japan;1950-54
Size: Beardsley,Hall,andWard1959:133
Charter: Beardsley,Hall,andWard1959:135,277-79
New CacheLa PoudreIrrigation Company,Colorado,U.S.A.; 1969
Size: MaassandAnderson1978:298,table7.6
Charter: MaassandAnderson1978:289,313
AngatRiverIrrigation System, Luzon,Philippines;1963-64
Size: Takahashi1970:49
Charter: Takahashi 1970:49,51, 120
RioMayo,Sinaloa,Mexico(Distritode Riego#038); 1982
Size: DGDUR1982, #038:5
Charter: author'sfieldnotes, WorldBank
FresnoIrrigation U.S.A.; 1969
District,California,
Size: MaassandAnderson1978:175
Charter: MaassandAnderson1978:175,179-82
Chia-nanIrrigation
Association,Taiwan;1968-69
Size: Pasternak1972:39
Charter: Pasternak1972:27n. 8, 41
HindiyahBarrageCanal,SouthIraq;1956-58
Size: Fernea1970:163
Charter: Fernea1970:122-23
King'sRiverWaterAssociation,California,U.S.A.; 1969
Size: MaassandAnderson1978:147
Charter: MaassandAnderson1978:255-56
author'sfieldnotes 1987
GeziraScheme,Sudan;1963-64
Size: Dishoni1966:90
Charter: Dishoni1966:90

NOTES

1. Severalinstitutionshave providedsupportfor variouspartsof this study.The


AmericanCouncilof LearnedSocietiesawardedme a Fellowshipin 1975-76, andtwo

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
INIRRIGATION
AUTHORITY SYSTEMS 351
sabbaticalsfromBrandeishave been particularly helpful.I am gratefulto ArjunAppa-
durai,SandraBarnes,andespeciallyRobertNettingfor editorialandsubstantivehelp.
IreneWinterhas as usualbeen of greatassistance.
2. This numberof systems is an educatedguess. Mexicois now knownto contain
aboutfifteenthousandnameddiscreteirrigationsystems, irrigatingabout5.0 x 106
ha (DGDUR1982).Manynationshavea largenumberof irrigation systems, including
Japan,the Philippines, Indonesia,India,China,Russia,Spain,France,Italy,the United
States, Columbia, andPeru. Someof these nationshavelargesystems, butallhavea
large numberof smallsystems. If each of these nationshas aroundtwo thousand
systems, whichseems conservative,thena figureof a hundredthousandsystems for
the worldas whole seems reasonable.The numberof irrigatedhectaresin the con-
temporaryworldis knownwithconsiderably less uncertainty thanthe numberof irri-
gationsystems (Highsmith 1965).
3. Thereare somepotentialproblemswiththisdefinition. Somesystemshavemore
than one headgate,such as the 12-Go system in Japan(Beardsley,Hall,andWard
1959).Howthis situationaffectsthe definition is notknown.Anotherpotentialproblem
is that irrigationsystems as here definedshouldbe clearlyisolatedfromeach other.
However,inmanypartsof the world,largeareasof irrigatedlandareservicedby many
irrigationsystems. The irrigatedcoast of Valencia,for example,has 120 continuous
kilometersof irrigation,withmanysmallnamedirrigation systemscoveringtheterritory.
A questionof interestis whetherwhathappensat the boundariesof these systems
challengesthe definitionof a system presentedhere.
4. Oneof the meaningsof centralizedauthorityhas been whetheror not an admin-
istrativestructureexists. This is betterconceptualized as the presenceor absenceof
constitutedauthority. As pointedoutabove,anotherof the meaningsis thatthe internal
administrative structureis dense, or highlyorganized,etc. (Kelly'sdimensionof cen-
tralization).This latterdimensionhas not been systematically studied.
5. In the articlepublishedwith Eva Huntin 1976, I arguedthat SanJuanandPul
Eliyahad centralizedirrigationsystems because the localelites were in controlof
irrigationandwere closelyconnectedto the state. I now believethatto be an error.
6. This namefor the principal role is not a very goodone, but nonebetterhas yet
been devised.The nameI havechosenconveys(1) thatthe officeis concernedwith
the executionof tasks, (2) thatit is the highestexecutiveoffice,and(3) thatit is an
office.
7. Policyis alwaysassociatedwith unifiedirrigationsystems. The policy-making
bodyis oftenthe sourceof the CEO'sauthority.Thisbodymayor maynot delegate
some or allof the decisionsto be madeto some smallersubsetof people.Manyof the
smaller"indigenous" systems makedecisionsbased on "tradition," whichis usually
some policydecisionmadein the past andawardednearlyunchangeable status. Little
has been publishedaboutthese policy-making bodiesor abouthowpolicyis made(An
exceptionis Valencia,where muchattentionhas been paidto the WaterCourt[see
FairenGuillen1975]).Mostof the literatureis concernedwithcarryingout policy,not
withhow it is made.
8. Anotherpossiblevalueforthisvariableis provincial charter.InIndia,forexample,
the charterfor the administrative authorityover manycanalsystems is vested in the
provincial(calledstate) governments,not in the nationalgovernment.These are not
thestate, buttheyhavemanystatelikefunctions.Ifthe questionis whetherthe farmers
or an externalpoliticalauthoritycharterthe system, thenclearlythe Indianprovinces

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
352 OFANTHROPOLOGICAL
JOURNAL RESEARCH
(states) do so andnot the users of the system. It mightbe arguedthat the central
governmentdelegatesthe authorityto the provincesvia the constitution.Exactlyhow
this situationshouldbe measuredis not yet settledin my mind.
9. It wouldbe betterto measuremaintenance burdendirectly,of course,anda major
problemwithstudiesof canalirrigationsystemsis thatthe amountof maintenance work
requiredandperformedis rarelystated.
10. I wouldexpect that the minimalsize for systems with nationalgovernment
charterswoulddecrease. I do not expect to findsubstantially largersystems with
irrigationcommunity charters.

REFERENCESCITED

Barkin,D., andT. King,1970, RegionalEconomicDevelopment:The RiverBasin


Approachin Mexico.Cambridge, Eng.:Cambridge UniversityPress.
Beardsley,R., 1963, Ecologicaland SocialParallelsbetweenRice-Growing Com-
munitiesin JapanandSpain.Pp. 51-63 in Symposiumon Community Studiesin An-
thropology (ed. byV.Garfield
andE. Friedl).Proceedings,1963AnnualSpringMeeting,
AmericanEthnological Society.Seattle:AmericanEthnological Society.
Beardsley,R., J. Hall,and R. Ward,1959, VillageJapan.Chicago:Universityof
ChicagoPress.
Bennett,J., 1974, Anthropological to the CulturalEcologyandMan-
Contributions
agementof WaterResources.Pp. 34-81 in Man and Water(ed. by L.D. James).
Lexington:The UniversityPress of Kentucky.
Bronson,B., 1978,Angkor,Anuradhapura, Pranbauan, Tikal:MayaSubsistencein
an AsianPerspective.Pp. 255-300 in Pre-Hispanic MayaAgriculture(ed. by P.D.
HarrisonandB.L. Turner).Albuquerque: Universityof New MexicoPress.
Bushnell,G.H.S., 1957, Peru.New York:FrederickPraeger.
Butzer,K.W.,J. Mateu,E. Butzer,andP. Kraus,1985, Irrigation Agrosystemsin
EasternSpain:Romanor IslamicOrigins? Annalsof the Associationof AmericanGeog-
raphers74:479-509.
Campbell,D.T., andJ. Stanley,1963,Experimental andQuasi-Experimental Designs
for Research.Chicago:Rand-McNally.
Clignet,R., 1970,A CriticalEvaluation of Concomitant VariationStudies.Pp. 597-
619 inA Handbook of MethodinCultural Anthropology (ed. by R. NarollandR. Cohen).
GardenCity,N.Y.:NaturalHistoryPress.
Coward,E.W.,Jr., 1976, IndigenousOrganization, Bureaucracy andDevelopment:
The Case of Irrigation.Journalof DevelopmentStudies13:92-105.
Coward,E.W.,Jr., 1979,Principlesof SocialOrganization inan IndigenousIrrigation
System. Human 38:28-36.
Organization
Coward,E.W.,Jr., 1980, Management Themesin Community Irrigation
Systems.
Pp. 203-18 in IrrigationandAgriculturalDevelopmentin Asia(ed. by E.W. Coward,
Jr.). Ithaca,N.Y.:CornellUniversityPress.
de los Reyes, R.P., 1980a,Forty-sevenCommunal GravitySystems:Organizational
Profiles.QuezonCity,Phil.:InstituteofPhilippine
Culture,Ateneode ManilaUniversity.
de los Reyes, R.P., 1980b,ManagingCommunal GravitySystems:Farmers'Ap-
proachesandImplications for ProgramPlanning.QuezonCity,Phil.:Instituteof Phil-
ippineCulture,Ateneode ManilaUniversity.
DGDUR(Direcci6nGeneralde Distritosy Unidadesde Riego),1982, Estudiopara

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
SYSTEMS
INIRRIGATION
AUTHORITY 353
el Mejoamiento de la producci6n:Distritode Riego#020, Morelia;Distritode Riego
#038, Sinaloa.MexicoCity:Secretariade Agricultura y RecursosHidr'ulicos.
Dishoni,S.A., 1966,The Impactof DevelopmentonThreeIrrigated Areasof Sudan.
Ph.D. diss., Universityof Wisconsin,Madison.
Duewel,J., 1983, MountainStreams,SacredRocks,andLowlandCanals.Unpub.
ms., author'sfiles.
Earle, T., 1978, Economicand SocialOrganization of a ComplexChiefdom:The
HaleleaDistrict,Kaua'i,Hawaii.Museumof Anthropology, Universityof Michigan,
Anthropology Papers no. 63. AnnArbor.
FairenGuillen,V., 1975,El tribunal de las aguasde Valencia,y su proceso.Valencia:
Cajade Ahorrosy Montede Piedadde Valencia.
Farbrother,H.G., 1973, WedMedani,Sudan,GeziraResearchStation.Pp. 36-99
in CottonResearchReports,Republic of Sudan,AnnualReportfor1970-1971(byH.G.
Farbrother).London:CottonResearchCorp.
Fernea,R., 1970,ShaykhandEffendi.Cambridge, Mass.:Harvard UniversityPress.
Forbes,R.J., 1955, Studiesin AncientTechnology,vol. 2. Leiden:E.J. Brill.
Gallin,B., 1966, ConflictResolutionin ChangingChineseSociety:A Taiwanese
Example.Pp. 265-74 in PoliticalAnthropology (ed. by M.J.Swartz,V.W.Turner,and
A. Tuden).Chicago:AldinePress.
Glick,T., 1970, Irrigationand Society in MedievalValencia.Cambridge,Mass.:
HarvardUniversityPress.
Highsmith,R.M., 1965,IrrigatedLandsof the World.Geographical Review55:382-
89.
Hunt,E., 1972, Irrigationandthe Socio-Political Organization of the CuicatecCa-
cicazgos.Pp. 162-259in Prehistoryof the TehuacanValley,vol. 4 (ed. by F. Johnson
andR.W.MacNeish).Austin:Universityof TexasPress.
Hunt,E., andR. Hunt,1974, Irrigation, ConflictandPolitics:A MexicanCase. Pp.
127-58 in Irrigation's on
Impact Society (ed. by T. DowningandM. Gibson).Tucson:
Universityof ArizonaPress.
Hunt,R.C., 1978, The LocalSocialOrganization of Irrigation
Systems:PolicyIm-
plicationsof Its Relationshipto Production andDistribution. USAID- AID/OTR- 141-
77-79.
Hunt,R.C., 1979, The Comparative Methodandthe Studyof Irrigation SocialOr-
ganization.CornellRuralSociologyBulletinno. 97. Ithaca,N.Y.
Hunt,R.C., 1987,SocialOrganization The Irrigation
of Irrigation: Community. Paper
presentedat the 1987AnnualMeeting,American AnthropologicalAssociation,Chicago.
Hunt, R.C., 1988, Size andAuthorityStructurein CanalIrrigation.Proceedings,
InternationalConference onIrrigation
SystemEvaluation andWaterManagement, Wuhan,
Hubei,People'sRepublicof China.
Hunt,R.C., n.d.a, Appropriate SocialOrganization? Water-UsersAssociations.Hu-
manOrganization. In press.
Hunt,R.C., n.d.b, WaterWork:Community andCentralization in CanalIrrigation.
Unpub.ms. in author'spossession.
Hunt, R.C., and E. Hunt, 1976, CanalIrrigationand LocalSocialOrganization.
CurrentAnthropology 17:389-411.
Kappel, W., 1974, IrrigationDevelopmentandPopulation Pressure.Pp. 159-68 in
Irrigation'sImpacton Society(ed. by T. DowningandM. Gibson).Tucson:University
of ArizonaPress.
Kelly, W., 1982, WaterControlin TokugawaJapan:IrrigationOrganization in a

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
354 RESEARCH
OFANTHROPOLOGICAL
JOURNAL

JapaneseRiverBasin,1600-1870.EastAsiaPapersSeries,no. 31. Ithaca,N.Y.:Cornell


China-Japan Program.
Kelly,W., 1983, Conceptsin the AnthropologicalStudyof Irrigation.
American An-
thropologist85:880-86.
Kroeber,C.B., 1983,Man,LandandWater:Mexico'sFarmlands Policies,
Irrigation
1885-1911.Berkeley:Universityof California Press.
Leach,E.R., 1959, Hydraulic Societyin Ceylon.Past andPresent15:2-26.
Leach,E.R., 1961, Pul Eliya.Cambridge, Eng.: Cambridge UniversityPress.
Lees, S., 1973, Socio-Political
Aspectsof CanalIrrigationin the Valleyof Oaxaca,
Mexico.Museumof Anthropology, Universityof Michigan,Memoirno. 6. AnnArbor.
Linton,R., 1939, Tamal.Pp. 251-90 in The IndividualandHis Society(ed. by A.
Kardiner).New York:Columbia UniversityPress.
Ludden,D., 1985, PeasantHistoryin SouthIndia.Princeton,N.J.: PrincetonUni-
versityPress.
Maass, A., and J. Anderson, 1978, ... and the Desert Shall Rejoice. Cambridge,
Mass.: MITPress.
Mencher,J., 1966, KeralaandMadras:A Comparative Studyof EcologyandSocial
Structure.Ethnology5:135-71.
Millon,R., 1962,VariationsinSocialResponseto thePracticeofIrrigatedAgriculture.
Pp. 56-88 in Civilizations
in Desert Lands(ed. by R. Woodbury). Universityof Utah,
Anthropology Papersno. 62. SaltLakeCity.
Mitchell,W., 1973,The Hydraulic Hypothesis:A Reappraisal.
CurrentAnthropology
14:532-34.
Murphey,R., 1957, The Ruinof AncientCeylon.Journalof AsianStudies16:181-
200.
Netting,R.M., 1974a,Agrarian Ecology.AnnualReviewof Anthropology 3:21-56.
Netting, R.M., 1974b, The SystemNobody Knows: VillageIrrigationin the Swiss
Alps.Pp. 67-76 in Irrigation's
Impacton Society(ed. by T. DowningandM. Gibson).
Tucson:Universityof ArizonaPress.
Pasternak,B., 1972, Kinshipand Community in Two ChineseVillages.Stanford,
Calif.:StanfordUniversityPress.
Price,B., 1971,Prehispanic Irrigation inNuclearAmerica.LatinAmerican
Agriculture
ResearchReview6(3):3-60.
Przeworski,A., andH. Teune,1970,The Logicof Comparative SocialInquiry.New
York:Wiley.
Ronfeldt,D., 1973,Atencingo:The Politicsof Agrarian Strugglein a MexicanEjido.
Stanford,Calif.:StanfordUniversityPress.
Spooner,B., 1974, Irrigation andSociety:The IranianPlateau.Pp. 43-57 in Irri-
gation'sImpacton Society(ed. by T. DowningandM. Gibson).Tucson:Universityof
ArizonaPress.
Stone, I., 1984, CanalIrrigationin BritishIndia.New York:Cambridge University
Press.
Takahashi,A., 1970, Landand Peasantsin CentralLuzon.Honolulu:East-West
CenterPress.
Uphoff,N., 1986, Gettingthe Process Right:Improving WaterManagementwith
FarmerOrganization andParticipation. Boulder,Colo.:WestviewPress.
vanderMere,C., 1968,Changing WaterControlina TaiwaneseRiceFieldIrrigation
System.Annalsof the Associationof AmericanGeographers 58:720-47.
van der Mere, C., 1971, WaterThieveryin a Rice IrrigationSystem in Taiwan.

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions
INIRRIGATION
AUTHORITY SYSTEMS 355
Annalsof the Associationof AmericanGeographers
61:156-79.
Wheatley,P., 1971, The Pivotof the FourQuarters.Chicago:Aldine.
Wittfogel,K., 1957,OrientalDespotism.New Haven,Conn.:YaleUniversityPress.
Wolf,E.R., 1966, Peasants.EnglewoodCliffs,N.J.: Prentice-Hall.

This content downloaded from 168.176.196.138 on Mon, 10 Aug 2015 20:12:16 UTC
All use subject to JSTOR Terms and Conditions

Potrebbero piacerti anche