The Ontology of "Political Violence": Action and Identity in Civil Wars
Author(s): Stathis N. Kalyvas
Source: Perspectives on Politics, Vol. 1, No. 3 (Sep., 2003), pp. 475-494
Published by: American Political Science Association
Stable URL: http://www.jstor.org/stable/3688707
Accessed: 14/12/2010 13:28
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=apsa.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
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 [email protected].
American Political Science Association is collaborating with JSTOR to digitize, preserve and extend access to
Perspectives on Politics.
http://www.jstor.org
Violence"
Articles | The Ontologyof "Political
The
and
Ontology of "Political
Wars
Identity in Civil
Violence":
Action
By Stathis N. Kalyvas
of politicalviolence,especiallyas they pertainto actions,
I discussseveralconceptualproblemsraisedby currentunderstandings
motivations,and identitiesin civilwars. Actions"onthe ground"often turnout to be relatedto localand privateconflictsrather
than the war'sdriving(or "master")
cleavage.The disjunctionbetweendynamicsat the top and at the bottom underminesprevailing assumptionsabout civil wars, which are informedby two competing interpretiveframes,most recentlydescribedas
"greedand grievance."Ratherthan posit a dichotomybetweengreedand grievance,I point to the interactionbetweenpolitical
and privateidentitiesand actions.Civilwarsarenot binaryconflicts,but complexandambiguousprocessesthatfosterthe "joint"
actionof local and supralocalactors,civilians,and armies,whose allianceresultsin violencethat aggregatesyet still reflectstheir
diversegoals. It is the convergenceof local motivesand supralocalimperativesthat endowscivil warswith their particularand
often puzzlingcharacter,straddlingthe dividebetweenthe politicaland the private,the collectiveand the individual.
t least 15 people died in Afghanistanwhen gunmen
attackedan isolatedpolice post nearthe country'scapital,
Kabul,in August2002. The identityof the attackerscould
not be ascertained.The chief of police there said that the men
were Taliban and supportersof the terroristorganizationalQaeda. "Otherlocal sources,"however,suggestedthat the men
werethievesand looterslookingto controlthe roadfor revenue.1
This storyillustratesthe poor qualityof informationin civilwars;
it also suggeststhat claimsabout identityand actionmay be selfservingand informationmay be instrumentallymanipulatedby
variousactors.Lessobviously,it hints at a perceptioninformedby
rigid,binarycategorieslinkedto mutuallyexclusivemotivations:
that the attackerscould have been eitherTalibanor thieves,and
their motivationscould have been either"political"(if they were
Taliban)or "private"
(if theywerethieves).But the gunmencould
have been both thievesand Taliban-simultaneouslyor sequentially,dependingon the context. Likewise,their violence could
havebeen both politicallyand privatelydriven.
This story epitomizessome of the problemswith our current
understandingof civil wars,particularlyour interpretationof the
identitiesand actions of the actors,along with their allegiances
and motivations,and our take on the war'sviolence. Prevailing
perceptionsare informedby two competinginterpretiveframes,
typicallyjuxtaposeddichotomously-most recentlyas "greedand
A
grievance."2 The first is Hobbesian in inspiration, stressing an
ontology of civil warscharacterized
by the breakdownof authorand
In
this
view, which can be traced
ity
subsequentanarchy.
StathisN. Kalyvasisprofessor
ofpoliticalscienceat YaleUniversity
He
a manu([email protected]). is currentlycompleting
entitled
The
of
Violence
in
Civil
War.
The
author
script
Logic
thanksMatt Kocherand HarrisMylonas,as wellasJennifer
Hochschildand threeanonymousreviewers,
for theircomments.
backto Thucydides,civilwarsencouragethe privatizationof violence,bringingto the fore,in a virtuallyrandomfashion,all sorts
of motivationsin what is a "warof all againstall."3This thesis
of ethniccivilwars4and so-called
informscurrentunderstandings
"newwars"allegedlymotivatedby greed and loot.5 The other
frame,which we may call Schmittian,entailsan ontologyof civil
wars based on abstractgroup loyaltiesand beliefs,wherebythe
politicalenemy becomesa privateadversaryonly by virtue of a
prior collective and impersonalenmity. The impersonaland
abstractenmity that Carl Schmittthoughtwas the essentialfeatureof politics6echoesRousseau'sperceptionof war,not as "man
to man"but as "stateto state."Individuals,claimed Rousseau,
wereonly enemiesby accident,and then only as soldiers.7In contrastto the Hobbesianthesis,which prioritizesthe privatesphere
at the exclusionof the political,the Schmittianone stressesthe
fundamentallypolitical natureof civil wars and their attendant
of traditional"ideological"
or
processes;it informsinterpretations
civilwars,8aswell as argumentsaboutethniccivil
"revolutionary"
wars and "intercommunalviolence"that stress strong beliefs,
groupenmity,and culturalantipathy.9
Ratherthan posit a dichotomyof greedand grievance,I point
to the interactionbetween political and privateidentities and
actions.I begin by highlightinga simple, though consequential,
observationthat appearsto be as common as it is theoretically
civil warsarenot binaryconflictsbut complexand
marginalized:
ambiguousprocessesthat foster an apparentlymassive,though
variable,mix of identitiesand actions-to such a degreeas to be
definedby that mix. Put otherwise,the widelyobservedambiguity is fundamentalratherthan incidentalto civilwars,a matterof
structureratherthan noise. I tracethe theoreticalsourceof this
observationto the disjunctionbetweenidentitiesand actionsat
the centralor elite level, on the one hand, and the local or mass
level,on the other.This disjunctiontakestwo forms:first,actions
"onthe ground"often seemmorerelatedto localor privateissues
www.apsanet.org 475
Articles I The Ontologyof "Political
Violence"
than to the war'sdriving(or "master")
cleavage;second, individual and local actorstake advantageof the war to settle local or
privateconflictsoften bearinglittle or no relationto the causesof
the war or the goals of the belligerents.This disjunctionchallenges prevailingassumptionsabout the locus of agencyin civil
wars and raisesa series of questions:What is the explanatory
leverageof interpretationsfocusing exclusivelyon the master
cleavage?What do labels and identities really mean on the
ground?Is it reasonableto infer the distributionof individual
and local allegiancesdirectlyfrom the mastercleavage?Is it correct to describeand analyzeall violence in civil warsas "political
violence"?
These questionsforceus to rethinkthe roleof cleavagesin civil
wars and challengethe neat split between political and private
violence. In this article,I point to severalimplicationsand outline an alternativemicrofoundationof cleavagebased on the
interactionof identities and actions at the center and at the
periphery.Actorsseekingpower at the center use resourcesand
symbolsto allywith peripheralactorsfightinglocalconflicts,thus
makingfor the "jointproduction"of action.This microfoundation is fully consistentwith the observeddisjunctionbetween
centerand periphery,which can now be reconceptualizedas an
interactionbetweenvariouscentraland local actorswith distinct
identities,motivations,and interests.
This understandingof civil warsin partcomplementsexisting
ones and in partsubvertsthem:while civilwarsexhibitboth pure
partisanand anomic behavior,they also contain actionsthat are
simultaneouslyboth; moreover,the empiricalbasisof Schmittian
and Hobbesianinterpretationsmay often be an artifactof biased
I emphasizethe
and incompletedata,as well as overaggregation.
it is not
because
evidence
just
pitfallsof overlookingimportant
of
collection
the
easily systematized.In certain researchfields,
reliableandsystematicdataat the masslevelis extremelydifficult,
if not impossible;civil warsareamong the most obviouscasesin
is
point. The requisiteanalyticaland empiricaldisaggregation0?
fineof
the
use
without
unsystematized
typically
impossible
graineddata. Ultimately,the specificationof concepts, models,
and causal mechanisms based on insights derived from this
empiricalevidencewill improvethe theoreticalanalysisof civil
warsand permitinnovativetests that will also assessthis empirical basis.
Complexity and Ambiguity
and Toriesas the occasionserved,were layingwastethe country
almost as much as those who were fighting for the one side or
the other."1lYearslater,AbrahamLincoln describedthe Civil
Warin the AmericanWest as a situationin which "murdersfor
old grudges,and murdersfor pelf, proceedunderany cloak that
will best cover for the occasion."12The Chinese Civil Warwas
often fought by diverseand shifting coalitions of bandits and
local militias;'3for a long time, the Communistswere for the
bandits "only one of several possible allies or temporary
In Manchuria,for instance,it was extremelydifficult
patrons."14
to differentiate between members of the Anti-Japanese
Resistanceand banditsbecausemoving from one to anotherwas
very common: it is estimatedthat 140,000 of a total 300,000
resistancemembershad a bandit background.Common criminalswerealso used extensivelyduringthe CulturalRevolution.15
The determinantsof violence in the provinceof Antioquiaduring the Colombian Violenciawere "farmore complex than any
innate, unavoidabledifferencesbetween monolithic groups of
Liberalsand Conservatives-the traditionalexplanationfor la
Violencia-might suggest";in fact, "the point of la Violencia,
even in supposedareasof 'traditionalsettlement'wherepartisan
objectiveswere the guiding force behind armedinsurrection,is
that it was multifacetedand ambiguous,that politics and economic considerations can never be considered as discrete
forces."16
In short, ambiguityis endemicto civil wars;17this turnstheir
characterizationinto a quest for an ever-deeper"real"nature,
presumablyhidden underneathmisleadingfacades-an exercise
akin to uncoveringRussiandolls. Thus, it is often arguedthat
religiouswarsarereallyabout class,or classwarsarereallyabout
ethnicity,or ethnic wars are only about greed and looting, and
so on.18The difficultyof characterizingcivil warsis a conceptual problem ratherthan one of measurement.If anything, the
more detailedthe facts, the biggerthe difficultyin establishing
the "true"motives and issues on the ground, as Paul Brasshas
nicelyshown in the caseof ethnic riots in India.19An alternative
is to recognize,instead, that the motives underlyingaction in
civil war are inherentlycomplex and ambiguous.At the same
time, just to state this point is as unsatisfactoryas to ignoreit. It
is necessary,instead,to theorizethis more complex understanding of civil warsso as to incorporateit into systematicresearch.
Doing so requires,first,the identificationof the sourceof ambiguity, which turns out to be located in the interactionbetween
centerand periphery.
Civilwarsaretypicallydescribedas binaryconflicts,classifiedand
understoodon the basisof what is perceivedto be their overarching issue dimensionor cleavage:we thus speakof ideological, The Disjunction between Center and
ethnic, religious,or classwars.Likewise,we labelpoliticalactors Periphery
in ethniccivil warsas ethnicactors,the violenceof ethnicwarsas Like in many other places,the occupationof the Philippinesby
turnsout to the Japaneseduring the Second World War generatedboth a
ethnicviolence,and so on. Yetsuch characterization
be trickierthan anticipated,becausecivil wars usuallyentail a resistancemovementand a civilwar,as some Filipinossidedwith
the Japanese.In his researchon the Western Visayas, Alfred
perplexingcombinationof identitiesand actions.
Consider the following descriptionof the AmericanWar of McCoy found that although the country underwentsuccessive
Independencein South Carolina:"There came with the true radicalpolitical changes between 1941 and 1946 (including a
patriotsa host of falsefriendsand plunderers.And this was true U.S. Commonwealth democracy, a Japanese Military
of both sides in this terriblestruggle.The outlawWhig and the Administration,and national independence), provincial and
outlaw Tory,or ratherthe outlaws who were pretendedWhigs municipalpoliticalleaderskept fighting the same parochialfac476
September2003 1 Vol. 1/No. 3
__ __
I
HAV A
4)W1iO
{w,6/W
CHAN67Tir4/T/OVOF
/NO AUL4El/A?NC
.
/IA
POIA/i TO70
\
OPiUNITY,
IN MY COU/N7RY,A5H,'WIHAETA-ROD,,.
I
ONT
V
00
u,
Q
a)
r
0
S!
s
?
oo
-s\
/e.s^.I
69
I MYic HAVV
I7/7SCH
AU1I/ANC0 FOM! T7/M
7T AC TO YAP/C0M70
YRYANAI W
CNNI//i V
w
I" I?-0
U(
a..."'
1Re I ne
' S r_x
_t
.,
1i
II
I
-
Allrightsreserved.
PRESS SYNDICATE.
DOONESBURY?2001 G. B. Trudeau.Reprintedwithpermissionof UNIVERSAL
tional struggles with their local rivals. The region's competing
factions, McCoy points out, were not insensitive to the larger
events emanating from Manila and beyond; in fact, they adapted quickly to each successive regime in an effort to use its
resources to their own advantage and to the detriment of rivals.
Costume and casting directors changed constantly, but actors
and dialogue remained the same. While the context shifted and
factions and their alliances split and realigned, peer rivals
remained in constant diametric opposition and, in so doing,
defined increasingly nominal party labels or categories such as
"guerrilla"or "collaborator."The violence overall was directly
related to these conflicts. McCoy's detailed investigation of the
1942 assassinations of eight prominent men in Iloilo uncovered
that all had their origins in prewar electoral conflicts among rival
municipal factions for control of mayoral and council posts. In
most cases, leaders of opposing factions had been involved in an
intensely personal competition with peer rivals-usually their
neighbors on the town plaza-for a decade or more and thus
took advantage of the new situation to settle local political
accounts. McCoy concludes that wartime factional disputes were
not imposed on Iloilo from above, but sprang spontaneously
from the lowest level of the provincial political system.20A study
of the Filipino island of Leyte during the same period confirms
McCoy's findings. Elmer Lear found that the guerrillas recruited
their supporters from the political faction that had failed to win
out in the previous election, as the winners were drafted into
serving the Japanese:
Neither side necessarilyacted on principle.It was the old case of
tweedle-dumand tweedle-dee-naked rivalryfor the spoils of local
office. Betweenfactions in some municipalities,a long enmity had
existed.It was only to be expectedthat if the factionin office found
the factionout of officewould
itselfrangedon the sideof collaboration,
and proclaimits devotionto resistance."
loudlycondemnits adversary
One may dismiss the Philippines as an isolated case. Consider,
however, the way in which a major and classic ideological conflict, the French Revolution, played out in the French provinces.
It turns out that divisions in the provinces were often highly local
and bore little relation to the Revolution's central issues. For
example, a town that had been denied its request to be the capital of the new administrative districts created by Paris was likely
to feel unsympathetic to the Republic and turn against it. Richard
Cobb provides the following account of the way in which provincial allegiances were shaped:
It wasa questionof chance,of localpowergroups,of whereone stood
in the queue,of at what stageambitionshad been satisfied,of how to
leap-frogover those in front.This is whereexternalevents could be
easily exploited;the Parispolitical labels,when stuck on provincial
backs,could mean somethingquite different.... The labels might
not even come from Paris;they could be of more local origin. In the
www.apsanet.org
477
Articles | The Ontologyof "Political
Violence"
was brought in from the outside, by groups of
Loire, "federalism"
armedmen ridingin from Lyon. But the experienceof "federalism,"
and the subsequentrepressiondirectedagainstthose who had collaboratedwith it, enabledone powergroup-of almostexactlythe same
socialstandingand wealth-to oust anotherin those towns that had
been most affectedby the crisis[emphasismine].22
Cobb is echoed by David Stoll, writing about a very different
time and place, contemporary Guatemala:
When outsiderslook at Ixil country,they tend to see it in termsof a
titanic political struggle between Left and Right. But for most
Nebajefios,these arecategoriesimposedby externalforceson a situation they perceiveratherdifferently.Class and ethnic divisionsthat
seem obviousto outsidersare, for Nebajefios,crosscutby familyand
community ties. Because of their wealth of local knowledge,
Nebajeinosareintimatelyawareof the opacityand confusionof local
politics,farmore so than interpretersfrom afar.... What seem clear
consequencesof nationaland internationaldevelopmentsto cosmopolitanobserversare,for localpeople,wrappedin all the ambiguityof
local life.23
The recent journalistic discovery that Afghanistan is "a world
where local rivalries and global aims seem to feed off each other"
and where "politics are intensely local, with many warlords swapping sides in alliances of convenience that have shifted with the
changing fortunes of the 22 years of war that began with the
Soviet invasion in 1979,"24is but the latest instance of a recurring
pattern. Consider the following anecdotal evidence from a wide
variety of civil wars.
Roger Howell stresses "the persistence of local structures and
rivalries" during the English Civil War, "even in the face of
intense pressuresfrom outside, a persistence that is frequently disguised at first glance because of the patterns by which the labels
of the 'national' struggle-royalist versus parliamentarian, presbyterian versus independent-were taken up by the participants
themselves and super-imposed on the 'local' struggle."25 A
detailed study of Bergen County, New Jersey, during the
American Revolution shows "that the local and bloody battles
between rebel and Loyal militia were related to prewar animosities between ethnic groups, political rivals, churches, and even
neighbors."26The "ferocious"civil war waged in North Carolina
during the American Revolution "involved complexities often
distant from the struggle between Great Britain and the courthouse and statehouse Revolutionaries."27The same was true, later
on, in the context of the American Civil War. In May 1862,
Major General John M. Schofield argued that "the bitter feeling
existing between the border people" was "the result of old feuds,
and involves very little, if at all, the question of Union or disunion."28Roger Gould shows that much of the conflict that took
place in Paris between 1848 and 1872 was related to turf battles
between neighborhoods rather than being a reflection of the class
struggle that is used to describe French politics during this period.29 Local conflicts often trumped ideological ones, writes H. R.
Kedward in his study of the civil war in occupied France, during
the Second World War.30 In his reconstruction of the violent
political battles waged in the region of western Segovia, in
Nicaragua,duringthe late 1920s, MichaelSchroederfound that
478
September 2003 I Vol. 1/No. 3
they "hadlong genealogies,and were deeply institutionalisedat
the locallevel.... [They]emergedfrom the contingentintersection of ethnic, village-level,regional,and national-levelpolitical
struggles.... [T]he violence expressedmany ongoing struggles
within Segoviansociety,a micro-universeof conflict-riddenrelations, developedover time, among and betweenfamilies,households, parties,communities,patronsand clients,and variouslayers of the state. In this light, perhapsthe most strikingthing
about this violence is its utterlyhomegrown,local character."31
Similardynamicsemergedlater on, during the Sandinistaand
Contracivil wars.Policemenin Quilalf,Nicaragua,werebasically the "armedfollowingof the Talaveraclan,whoseturf thiswas,"
Paul Bermanreports,addingthat clan politicswas "anembodiment of every ruralNicaraguanevent that never did get adequatelyreportedto the outsideworld in the yearsfollowingthe
Sandinistarevolution."32
A studyof a northernSpanishtown found that the maincleavage in its centralneighborhoodbeganin the early1930s as a dispute betweentwo doctorscompetingfor the title of officialtown
doctor,which entaileda lucrativestate-guaranteed
practice.Many
familiesbecameengagedon the side of one doctor or the other:
the politicalturmoilof the end of the Republic
"Simultaneously,
a
added wider politicaldimensionto what was in essencea dispute basedon localissues.The tug-of-waris often describedtoday
in termsof the liberal-conservative
issuesof the time, but most
informantsagreethat the basicissueswerelocal and personal."33
Clan rivalriesin Chinesevillagesshapedpeasantdecisionsabout
whetherto side with or againstthe Communistsduringthe civil
warthere.PeterSeybolt'sanalysisof the ChineseCivilWarduring
the Japanese occupation uncovers a similar disjunction between
centerand periphery:"Manyof the battlesfoughtamongChinese
had little to do with collaborationor resistance.They werestruggles for power and economic spoils that pit centralauthorities
againstlocalauthorities;local authoritiesagainsteach other,bandits againstmerchantsand landlords,secretsocietiesagainstbandits, GuomindangmembersagainstCommunists,and so on."34
Duringthe Colombian Violencia,the "eliminationof membersof
the oppositionfromparticularhamlets... appearsto haveobeyed
the logic of personalfeuds,partisandifferences,and intermuniciA reportby the envoyof the ConservativeGovernor
pal rivalries."
revealed"asordid,corof Antioquiain the town of Cafiasgordas
rupt, divided, and violent society riven by factionalism,family
feuds,localanimosities,personaljealousies,vindictiveness,greed,
conflicts between haves and have-nots, and struggles over
power."35The mass killings that took place in Indonesia in
1965-1966 were ostensibly articulatedaround the communism/anticommunismcleavage,yet a sustainedexaminationof
regional massacresunearthedall kinds of local conflicts. For
instance,in the southernSumatraprovinceof Lampung,the violencewascausedby a conflictbetweenlocalMuslimsandJavanese
transmigrantsettlers.In some areasof Timor, the victims were
Protestants,while in othersthey were followersof local cults;in
Lombokthey were Balineseand Chinese.The killingsin Central
and East Java were caused by hostility between local Muslim
cultural-religious
groupsknownas abangan;in BalitheywereassoOn
ciatedwith long-standingrivalriesbetweenpatronagegroups.36
a visit to the Lebanesecountryside,the travelwriter William
Dalrymplewas surprisedto discoverthat a bloody raidby Samir
of
Geagea's(Christian)Phalangistmilitiaagainstthe headquarters
the (likewiseChristian)Maradamilitialed by Tony Franjiehwas
only ostensiblya struggleabout political issues (the Phalangists
preferringLebanon'spartitionand the Franjiehswishingto keep
it whole):"Infact [it] had its true rootsin somethingmoreprimitive still: a century-oldblood feud between Bsharre,Geagea's
home town, and Ehden and Zgharta,the Franjiehstrongholds
forty miles to the west."Dalrymplereachesthe conclusionthat
"thestory of the raidwas remarkable,and revealedmore clearly
than anything the medievalfeudal realitybehind the civilized
twentieth-centuryveneerof Lebanesepolitics."37When told by
the milithe armyto makean exampleof the local "subversives,"
tia leaderin the Guatemalanhamlet of Emol Centralchose his
victims from Kotoh, "Emol Central'straditionalrivals."38
The
1983 massacreof journalistsby the inhabitantsof Uchuraccay,
Peru,led to an extensiveinvestigationthat eventuallytracedthe
massacreto the animositybetweenhighlandersand lowlanders;
the lowlandswereeasierfor SenderoLuminosorebelsto penetrate
becausetheyweregeographically
moreaccessible.Once, however,
Sendero became associatedwith the lowland communities, it
sparkedthe enmityof the highlandones-an enmitythatanthropologistshad alreadytracedto a long traditionof rivalrybetween
The Liberiancivilwardurhighlandand lowlandcommunities.39
the
tens
of
local
1990s triggered
ing
cleavages:
It is saidthatin someareasthe warin the south-east
reopenedold
feudsdatingbackto the 1930s.Certainly
it militarized
the factional
beenthe stuffof localpolitics,and
disputeswhichhad previously
whichlinkedlocalstruggles
to nationalinterests.
Asthewaritselfgave
riseto localvendettas,
orasolderantagonisms
weresettledbyforceat
a timeof war,thereemerged
a micro-politics
of warin whichcertain
territories
suffered
morethanothersat particular
Theareas
moments.
worstaffectedwerethosewhichweredevastated
as local
repeatedly
rivalslaunched
see-saw
raidsandcounter-attacks
oneanother.40
against
The reason that Toposa tribesmenacceptedweapons from the
Sudanesegovernmentto fight againsttheir formerDinka insurgent comradesin southernSudanis to be found in old disputes
and cattle thieving among the two groups.41Most recentlyin
Congo, "analystsdistinguishbetweenthe big war,the main conflict betweenthe Congolesegovernmentand the rebelarmiestrying to topple it, and the many smallerwars being waged deep
insideCongo'sjungles."As one analystput it: "Thenationallevel
and the local level aretwo differentthings in Congo."42
All in all, the salience of local cleavagesis ubiquitous in
ground-leveldescriptionsof civil war and holds for societiesthat
are sharplypolarizedin terms of class,43religion,44and ethnicity.45It would not be an exaggeration to say that references to the
disjunctionbetweencenter and peripheryare presentin almost
everydescriptiveaccount.46
This disjunctionis consistentwith the observationthat civil
warsare"weltersof complexstruggles"47
ratherthan simplebinaconflicts
a
ry
neatlyarrayedalong singleissue dimension.In this
sense, civil wars can be understoodas processesthat providea
medium for a variety of grievancesto be realizedwithin the
greaterconflict, particularlythrough violence. As Colin Lucas
notes about the counterrevolutionin southernFrance,the revolutionarystruggleprovideda languagefor other conflicts of a
social,communal,or personalnature.48
An understanding
of civilwardynamicsas substantiallyshaped
local
is
also
by
fullyconsistentwith recurringsuggestions
cleavages
that mastercleavagesoften fail to accountfor the natureof the
conflictand its violence49and that violenceis eitherunrelatedor
incompletelyrelatedto the dominantdiscourseof the war;50that
civil warsareimperfectand fluid aggregationsof multiple,more
or less overlapping,smaller,diverse,and localizedcivil wars,51
entailingByzantinecomplexity52and splinteringauthorityinto
"thousandsof fragmentsand micro-powersof localcharacter."53
This evidencejibeswith the anthropologicalinsightthat local
politics is not just (or primarily)the local reflectionof national
politics. In his analysisof local politics in Sri Lanka,Jonathan
Spencershows that "villagersdid not simplyhave politicsthrust
upon them; ratherthey appropriatedpoliticsand used them for
their own purposes."He adds that "peoplewere not necessarily
enemiesbecausethey were in differentparties;more often they
had ended up in differentpartiesbecausethey were enemies."
Hence, he points out, "atleast part of the apparentideological
and sociologicalincoherenceof politicalpartyallegiance"can be
tracedto the factthatpoliticsprovidesa meansof expressinglocal
conflicts:
It is possibleto seea greatpartof villagepoliticsas littlemorethan
thedressing
of partypolitiup of domesticdisputesin thetrappings
cal competition,
of troublewhich
exploitingthe publicexpectation
accompanies
partypoliticsin orderto settleprivatescoresin the
idiomof publicaffairs.Partypoliticsareestablished
so firmlyin Sri
Lanka,in partbecauseof theirelectiveaffinitywiththosedividedor
whichotherwiselackan everydayidiomin
dividingcommunities
which to characterize
theirown disunity:politicsprovidejust such an
idiom.54
While local cleavagesare by no means the only mechanism
producingallegianceand violence, they appearto have substantial impacton the distributionof allegiancesas well as the content, direction,and intensityof violence.This evidencelendssupport to the view that both the distributionof allegiancesacross
the populationand the violencethattakesplaceareoften (though
not always)a function of preexistinglocal rivalrieswhose connection to the cleavagethat informsthe civil war is tenuousand
loose-even when conflicts are framedin the discursiveterminology of the mastercleavage.Of course,evidencecan only be
anecdotalsince,for obviousreasons,we lacksystematicstudiesof
the dynamicsof civilwarsat the locallevel,as well as measuresof
local cleavages.55
Leavingasidethe often questionablequalityof
data on civil wars,it is worth noting that the
(macro)
aggregate
availableevidenceis particularlystrikingand deservesattention
sincemacro-levelstudieshaveconsistentlyoverlookedand misinterpretedthesedynamics.Althoughit is impossibleto ascertainat
this point the relativeweight of local cleavageswithin and across
wars,it is necessaryto acknowledgethe significanceof this phenomenon;this should sparka researchprogramleadingto a rigorousempiricalstatementaboutits prevalence.One obviouspath
www.apsanet.org 479
Violence"
Articles I The Ontologyof "Political
is to incorporatethese insights explicitlyinto deductivemodels
whose predictionscan then be independentlyand systematically
testedwith fine-graineddata.56
Although ubiquitous in the descriptive literature, these
dynamicshave been overlookedby macro-levelstudies of civil
wars, both descriptiveand theoretical-with very few exceptions.57Instead,most accountsinfer local and individualidentities and actions directlyfrom the war'smaster cleavage.Local
cleavagesareneglectedfor a numberof reasons.Firstis a division
of laborseparatingthe tasksof collectingevidenceat the micro
level and interpretingmacro-dynamics;second is an epistemic
preferencefor the universalover the particular,and the easily
codableovermessyevidence;thirdis the ambiguityof local-level
dynamics,which in some ways parallelsthe distinctionbetween
"objective" structures and "subjective" actions;58 fourth is the
fact that local cleavagesare typicallyarticulatedin the language
of the war's master cleavage,often instrumentally.To give a
recentexample,local factionsin Afghanistanaccusedone another of being Talibanor al-Qaedaso as to have rivalsbombed by
the U.S. Air Force.59As a result, naive observersand participants, including the principals, tend to miscode local cleavages.60Overall, academic studies often share with "official"
historiographiesthe tendency to erase troubling internal divisions-"class fissures,acts of treachery,or peasantinitiativesthat
were independentof elite control"-and to smooth over "the
past'sjaggededges."61
who areattunedto the grassroots
At the sametime, researchers
micro-oriented
historians) report
(anthropologists,journalists,
these dynamicsbut fail to theorizethem. A startingpoint in the
directionof theorizingis to sketcha few broaddistinctions.Local
cleavagesmaybe preexistingor warinduced;theymayalignneatly with centralcleavagesor subvertthem; and they may be consistentovertime or more fluid and random.
With preexistinglocal cleavages,war activatesthe fault lines.
When prewarlocal cleavageshave alreadybeen politicizedand
graftedonto the nationalstructureof cleavages,their autonomy
and visibilityqua local cleavagesis diminished;even then, however,the mastercleavagemay not erasethem.To understandviolence,one has to takeinto accountlocalcleavages,as suggestedby
the followingdescriptionof EastTennesseeduringthe American
Civil War:
extensive
Thepolicyof granting
powersto nativeUnionistsandmakaimedatrestorof EastTennessee
in
the
them
occupation
ing
partners
that
But
as
as
a
policy,coming loyalgovernment quickly possible.
seriousrisks.It
harshFederal
binedwithincreasingly
policies,carried
on secesforUnioniststo takerevenge
furtheropportunities
provided
violence
ratherthanconstrained,
sionists,andit encouraged,
partisan
Unionistshadtheirownagenda,anagendathatdidnot
anddisorder.
created
aims,andthisdifference
frequently
alwaysmeshwithFederal
fortheUnioncommand.62
complications
In the most extremecases,local cleavagesmay lose all autonomy
and turn into mere local manifestationsof the centralcleavage.
Conversely,a centralcleavagemaybranchout into localcleavages
cleavageof Liberalsand Conservativesspawnedresidentialsegrepatterns.63
gation and intermarriage
Often, local cleavagesare preexistingwithout being grafted
onto the mastercleavage-which increasestheir visibility.Thus,
the conflict between Royalists and Parliamentariansin
Leicestershireduring the English Civil War was also a conflict
betweenthe Hastingsand the Grey familiesthat "wentback to
personalfeudsof farlongerstandingthanthe CivilWar,in factto
theirrivalryfor the controlof the countrysincethe mid-sixteenth
century.For these two families,the Rebellionwas, at one level,
simply a furtherstage in the long drawn-outbattle for local
dominion."64The Protestant-Catholic
violence that erupted in
southeasternFranceduringthe FrenchRevolutionwas not simply
religious;it pitted againsteach other particularfamilieswith a
trackrecordof pastfeuding:the Lanteirisagainstthe Labastinein
Chamborigaud,the Bossieragainstthe Roux in Vauvert,and the
Rousselagainstthe Devaulxin Bagnols.65Likewise,"familyand
faction dictated the course of the IRA split in units all over
Ireland"duringthe civil war. "Once again, it was the Brennans
against the Barrettsin Clare, the Hanniganites against the
Manahanitesin east Limerick,and the Sweeneysversus the
The
O'Donnellsin Donegalas all the old feudswerereignited."66
clash in Colombia"frequentlygrew out of
Liberal-Conservative
long-standingfamilyfeuds. LiberalUrregos,for instance,joined
Franco,while theirlong-timeenemies,the Cossiosand Montoya
Montoyasfrom Caicedo, made up the ranksof the police and
Conservative contrachusma [bands] in nearby towns."67
Journalistsoften encountersimilarpatterns:the war betweenthe
pro-IraqiKurdjashmilitiaand Kurdishrebelswas also a conflict
on the otherside
betweenthe Sourchiand the Barzanifamilies;68
of the border,in easternTurkey,the warbetweenethnicKurdsand
the Turkishstate in the village of Ugrakwas also between the
Gucluand the TangunerandTekinfamilies,both Kurdish.69
Warmay generatenew local cleavagesbecausepowershifts at
the local level upset delicate arrangements.After Shining Path
rebelsappointednew villageleaders,"theguerrillacolumnwould
leave,without realizingthat it had left behind a hornet'snest of
contradictionsthat could not be resolved.Even if in these cases
no overtrebelliontook place,the impositionof the new authorities generatedinitialresentmentsand the firstpeasantalliesof the
armedforces,'informers'(soplones)in the senderistaterminology."70In the centralPeruvianvalleyof Canipaco,the population
enjoyeda "kindof honeymoon"with ShiningPathuntil a dispute
eruptedbetweentwo communitiesoverthe distributionof lands
previouslyusurpedby haciendas:
of armedShiningPathcadreson thesideof oneof
Theparticipation
in a massiveconfrontation
the communities
againsta confederation
of rivalcommunitiesprovokeda rupturewith the latter,who decided
in thescuffleto
cadrestheyhadcaptured
to turnovertwosenderista
the authoritiesin Huancayo.This actionprovokedShiningPath
in theexecutionof thirteenpeasantleadwhichculminated
reprisals,
ers.The victimswerekidnappedfrom theircommunitiesand assassinatedin the centralplazaof ChongosAlto.71
that remain active even after the central cleavage has died. This
One of the most potent cleavages produced by civil wars is gen-
seems to have been the case in Colombia,wherethe ideological
erational:rebels(but alsoincumbents)often recruityoung people
480
September 2003 I Vol. 1/No. 3
who then proceedto represstheir village'selders.The war may
also lower the cost of opportunisticbehavior,triggeringtens of
local cleavages.
When local cleavagessubvertcentralones, factionalconflicts
emerge within supposedly unified political camps. McCoy
describes how two factions in Western Visayas, Philippines,
becamesplit ratherevenlybetweenthe resistanceand collaboration regimesduring the Japaneseoccupation.However,during
the war,membersof the same politicalfactionon oppositesides
cooperatedclosely with each other, while membersof opposite
factions,within the resistanceand the Japanese-sponsored
government, respectively, fought bitterly against each other.72
Similarly,CarlosRafaelCabarrusshowsthat in some of the rural
communitieshe studiedin El Salvador,kin-basedconflictscaused
importantdivisionswithin politicalfactions.73
An exclusive focus on cleavages(both local and nonlocal)
would failto accountforvariationin levelsof victimization.Local
cleavagesmay be compatibleboth with an escalationof violence,
as competingfactionstry to gain advantage,and with moderation, as they have the meansto strikelocal deals,may anticipate
futurecooperation,and can resortto effectivein-grouppolicing
in orderto preventdecentralizedescalation.74
Accountingforviolence requiresthat local dynamicsbe embeddedin an analysisof
war dynamics,especiallythe logic of territorialcontrol.75
In sum, examininglocal cleavagesopens up fascinatingempirical possibilities for exploring the various paths, trajectories,
modalities,and combinationsof centraland local cleavages,as
well as theirconsequences.Researchon clientelism,76
networks,77
and local factionalism78
constitutesan obvious resourcein this
respect.
Theoretical Implications
It may be possibleto overlookdynamicsat the microlevel if the
goal is to attain a historicalinterpretationof the conflict at the
macrolevel and the longueduree.The fact that much violencein
Missouriduringthe AmericanCivilWarwas relatedto localconflicts ratherthan the issue of slavery79undercutsthe broadlines
of standardmacro-levelinterpretationsof the AmericanCivil
Waronly in part-while also causinga loss of descriptiveaccuracy. However,analysisof the dynamicsof civil war (how and why
people join or defect, how violence takes place, et cetera) is
impossiblein the absenceof close attention to local dynamics.
Such attentionis also necessaryfor achievinga closerfit between
macro-and micro-leveltheory80and interpretingcross-national
findingsaboutkey variables,such as the onset, duration,and terminationof civil wars.For instance,one of the most robustpredictorsof civil waronset, per capitagrossdomesticproduct,may
poor, nonmoderncapturein part the effect of local cleavages;81
ized states have failed to penetratetheir peripheryeffectively,
which would have reducedthe salienceof local cleavages82
and
thus createdopportunitiesfor rebelsto tap into them.
Severaltheoreticalimplicationsfollow from an understanding
of civilwarsinformedby the dynamicsof localcleavages.Identity
labelsshould be handledwith caution:actorsin civil war cannot
be treated as if they were unitary. Labels coined at the center may
be misleadingwhen generalizeddown to the local level; hence,
motivationscannot be derivedfrom identities at the top. The
interchangeabilityof individualsthat underliesthe concept of
groupconflict and violenceis variableratherthan constant.The
locus of agencyis as likely to be at the bottom as at the top, so
civilianscannot be treatedas passive,manipulated,or invisible
actors;indeed,they often manipulatecentralactorsto settletheir
own conflicts.
The analyticalprimacypresentlyenjoyedby mastercleavages
implies that local dynamicsare perceivedas a mere (and rather
irrelevant)localmanifestationof the centralcleavage-automatic
and unproblematicaftereffectsof actionsand decisionslocatedat
higherlevels.In this perspective,local actorscan only be replicas
of centralactors,and theirstudy is justifiedsolelyon groundsof
local historyor antiquarianinterest.It followsthat it is unproblematicto generalizedirectlyfromthe centerto the locallevel;in
other words, actors (e.g., Serbs)are unitary,and motives (e.g.,
ethnic domination)hold for all individualmembersand actions,
including violence. Thus, we speak of actors such as Shias,
Albanians,or workersfollowingdescriptionsof civil wars along
the "modular"
themesof religion,ethnicity,or class.These labels
are not neutral;they typicallyimply a theoryof causation.Civil
wars (and their violence) are assumedto be directlycaused by
religious,ethnic, or classcleavages.
However,the disjunctionbetweencentraland local cleavages
challengesthe validityof such labels.Althoughmastercleavages
inform and motivate local dynamicsto a varying degree, the
observeddisjunctionbetween the two raisescritical questions
about the dynamicsof civil war and its violence. Likewise,the
pronouncedtendencyto infer motivationsdirectlyfrom identities at the center is undermined.Violence in an ethnic or class
warmaynot be ethnicor classviolence.Forinstance,Stoll shows
how the first Ixil Indianswho collaboratedwith the rebels in
Guatemala"werenot impoverishedseasonalplantationlaborers,
as [rebel]strategistsseem to have expected.Instead,they were
prominent men from San Juan Cotzal, relativelywell-situated
merchantsand laborcontractors,who wishedto enlist the guerrillasin the bitterpoliticalfeudsof theirtown."Conversely,their
local enemies "whohad disgracedthemselvesin office and were
being defeatedin electionscould now denouncetheiropponents
to the army."83
The concept of group conflictor groupviolence (and, hence,
ethnic conflict and ethnic violence, and so on) entailsthe total
of individuals,either as participantsand perinterchangeability
petratorsor as targets."Groupconflict"makessenseonly if group
membersarefully substitutablefor eachother.84If targetsof violence are selected along lines that go beyond group attributes,
then the violence cannot be describedas simply ethnic, classbased,et cetera.One indicationthat this may be the case is the
highly intimatenatureof interaction,particularlyas expressedin
violence:
TheEastTyroneBrigade[oftheIRA]werenot an armybuta band,
a companyof latter-daywoodkernes,of ordinaryfarmworkers,
tractordrivers,the unemployed,
the oddschool-teacher,
mechanics,
inheritorsof the dispossession,who gatheredtogetherto killparticular knownenemieslike EdwardGibson,ThomasJamesonand Harry
Henry.The IRA were not waginga war but a sporadicassassination
www.apsanet.org 481
Articles I The Ontologyof "Political
Violence"
campaign in the tiny rural communities of Tyrone to attack the
enemy in theirmidst [emphasismine].85
Though class informed politics in revolutionary America, there is
a consensus among historians that class tensions cannot explain
the extensive variations in levels of internecine violence in
Virginia and the Carolinas.86The same appears to have been true
in Nicaragua: "There were poor peasants who ran to tell the
Guard when they saw the Sandinistas, and there were members of
wealthy urban families who deserted the guerrillas and told the
authorities everything they knew about their former comrades."87
In some areas of predominantly Croatian rural Herzegovina,
much violence during the 1990s was an outgrowth of local
vendettas.88 The violence between the neighboring villages of
Coagh and Ardboe, in Northern Ireland, which cost the lives of
30 men in the space of three years in the late 1980s and early
1990s (for a combined population of just over a thousand people), was not simply violence between the Catholic Irish
Republican Army and the Protestant Ulster Volunteer Force, but
also a "bitter vendetta" and the "freshest cycle of a blood feud"
that pitted these particular two villages against each other. In
other words, the nature of the violence in this area cannot be
understood by simple reference to the religious cleavage in
Northern Ireland but requires knowledge about the local cleavage
between Coagh and Ardboe.89
Likewise for individuals. Often, the master cleavage establishes
a baseline that determines what the relevant groups are. However,
the assumption of noninterchangeability of individuals is violated
with the introduction of a secondary selection criterion based on
individual characteristics unrelated to group identity. Motives
vary, but grudge and loot appear to prevail. Intergroup victimization spurred by looting among neighbors is common.90 Because
the class cleavage defined the relevant group identities in
Republican Barcelona during the Spanish Civil War, concierges,
maids, and other domestic personnel in well-to-do neighborhoods could victimize the upper- and middle-class families residing in the buildings where they worked.91 Yet, as a resident of
Barcelona told me, concierges often handpicked their individual
victims based on their own grudges that went beyond class.
Individualized selection may take place even under the extreme
circumstances of ethnic cleansing and genocide. A former prisoner
of the notorious Omarska camp in Bosnia describes violence
inflicted by Serb guardson Muslim inmates. One day, a Serb guard
came in at night and insulted a prisoner who, as a judge, had fined
him for a traffic offense in the late 1970s! In another instance,
from my village, "disappeared"
SakibPervanic,a thirty-two-year-old
becauseof an old grudgeagainsthis father.Sakib'sfather,Mustafa,
had had businessdeals with Rade Gruban-but over the yearsthey
had failedto settlesome businessdebts.Radeowneda coupleof small
groceryshops also selling home appliances.One of the shops was in
my village.The businesswas going well and he decidedto expandit
throughbulk salesof cement, but he did not havethe necessarystorage space.Mustafalet him use a partof his basementfor this purpose,
but they could not agree on the amount of the rent. As a result,
Mustafarefusedto payRadefor some applianceshe had purchasedon
credit.Radenow wantedrevenge-but Mustafawas in the Trnopolje
camp.It savedhim, but not his son.92
482
September 2003
1 Vol. 1/No. 3
After the Kosovowar ended, a journalistreportedthat "Captain
Kevin Lamberttold me of an Albanianwoman who accuseda
Serbof kidnappingher duringthe war.CaptainLambert'stroops
arrestedthe man, but upon investigating,they discoveredthat the
woman'sfamily had been trying to coercehim to sell them his
apartment.Was this a case of falselyaccusingthe Serbto get his
home? With no proof, the U.S. Army decided it was."93Jan
Gross'sobservationabout the violence that eruptedin western
Poland during the Soviet occupation of 1939 captures this
aspectparticularlywell:
private-grudge
an explosionof combinedethYet,muchastheviolencerepresented
andnationalist
struckby its
nic,religious,
conflict,I amnevertheless
kneweach
Moreoftenthannot,victimsandexecutioners
intimacy.
couldstillname
otherpersonally.
Evenafterseveralyears,survivors
to getevenforpernames.Definitively,
peopletookthisopportunity
sonalinjuriesof the past [emphasismine].9
Becauseof the prevailingemphasison the top at the expenseof
the bottom, thereis a pronouncedtendencyto locatethe agency
of violencein the former;hencethe propensityto portraythe violence of civil wars as being externallyimposed on unsuspecting
and, therefore,innocent civilians.95In this view, civilians are
objectsratherthan subjectsof the violence.Guatemalanpeasants
tend to describethe civil war as "somethingruralcommunities
werecaughtin but not of theirmaking."96
Referringto the experienceof a GreekvillageduringGreece'scivil war,an anthropologist points out: "The villagerswere, as always,the victims of
strugglesof othersratherthan the activeelementof the struggle
itself."97
This perspectiveis succinctlyexpressedin varioussayings
about the proverbialants caught between fighting elephantsor
buffalo. Indeed, much of the contemporaryhuman-rightsdiscourse entails this assumption, which is also echoed in instru-
mentalisttheoriesof ethnicconflict,whereindividualsaremanipulated by politicians in pursuit of political power. When not seen
as victims, individualssimply vanish.They are aggregatedinto
groups ("the Serbs," "the people") whose actions are otherdirected.The termpuppet,used to describethe collaboratorarmy
duringthe Japaneseoccupationof China and similarsituations
indicatesthe prevalenceof an "instigator"theoryof
elsewhere,98
violent conflict. This theory is not necessarilyinaccurate,especially when the focus is just on the visible portion of violence;
however,it underplaysor downrightdenies that there are also
"instigatees,"whose participationis essentialto transformanimosity into violence.9
Many detaileddescriptionsof violencesuggestthe presenceof
considerablelocal input and initiativein the productionof violence. Ratherthan being imposed upon communitiesby outsiders, this evidence suggests, violence often (but not always)
growsfromwithin communitiesevenwhen it is executedby outsiders;it is, in otherwords,often intimate.The followinganalysis by a Sinn Fein councilorin Coalisland,NorthernIreland,suggests that the "religious"cleavagein this area,though activated
along the lines of the conflict'smastercleavage,overlappedwith
a (local) conflict between two subsetsof people in Coagh and
Coalisland-distinct
from other local conflicts
between
Protestantand CatholicgroupsacrossNorthernIreland:
The UDR [Ulster Defense Regiment] from Coagh came into
Coalisland,which is a ninety-nine per cent nationalisttown, and
patrolledaroundthe town. They would stop schoolchildrenon their
way to school,get them to turnout theirschoolbags,or stop cars....
They would search and read anything, letters, private documents
from your solicitor,even if it was obviousthat therewas no security
forceconnection.The UDR man could readeveryone of those documents,he could even count the money in your pocket, eventhough
he wasyournext-doorneighbour.
The only qualificationhe neededwas
thathe wasa memberof the UDR. It led to greattension .... It made
people feel low and it engenderedtotal hostilitytowardsthe Loyalist
community and gave the impressionthat this is a Catholic versus
Protestantwar.But it had nothingto do with religion;it wasthesimple
arming of one sectionof the communityagainstthe otherwhilstyou
deprivethat othersectionof any meansof defendingthemselves
[emphasis mine].100
Descriptions of police, army, or guerrilla sweeps, arrests,or assassinations reveal that violence in civil wars often entails the participation of community members, who either act as suppliers of
information or (less often) participate in more direct ways. The
reliance of political actors on local information is typically conveyed by the widespread use of blacklists, as suggested by the following report from Colombia:
sufferedat the handsof the nationalistsduringthe SpanishCivil
War,summarizesit best: "It wasn'tFrancowho harmedus, but
people from here-the village."109
Local participationis compatiblewith all sorts of motives,
rangingfrom the most ideologicalto the most opportunistic.
Evidencesuggeststhat a key motiveis settlingprivatescoresunrelated to the war'smastercleavage.Many acts of violencethat on
the surface(and to outsiders)appearto be generatedby exclusivelypoliticalmotivationsoften turnout, on closerexamination,
to be "causednot by politicsbut by personalhatreds,vendettas,
and envy."110
Thucydidesarguesthat personallymotivatedcrime
maskedby politicalpretextis one of the essentialfeaturesof civil
war,11while Machiavellidescribesa situationwhere politically
motivatedriotsoffera pretextfor privateviolence.l2 Tocqueville
makesa similarobservationwhen he arguesthat "privateinterest,
which always plays the greatest part in political passions, is ...
skillfullyconcealedbeneaththe veil of public interest."113In her
studyof Guatemala,KayWarrenfindsa "deepermessage"hidden
in the local and privateunderpinningsof a murderthat seems
politicaland impersonal."4The anthropologistwho assertsthat
Greek villagerswere "alwaysthe victims of strugglesof others
ratherthan the activeelement of the struggleitself" lists, a few
pageslaterin her book, a host of privatemotivesbehindthe vioAt leasteightpeasantswerekilledin the northernvillageof SanRoque lence of the GreekCivil War;for
example,"oneman joined the
in what the police said they suspectedwas a right-wingparamilitary Communistswith
the expressintentionof killinga rivalinheritor
attack.Gunmenkilledfourmembersof a familyat a gasstation,then
of his father'."115
stormedinto the homes of four farmworkersand opened fire after
The storiesof Aristogitonand Harmodiuson the one hand,
checkingtheiridentitiesagainsta list they carried,the policesaid.The
and
PavlikMorozovon the other, are particularlysuggestivein
areais also a frequentstagefor leftistrebelattacks.'0
this respect. Thucydides tells the story of Aristogiton and
In his postwar trial, Lieutenant General Takeo Ito, a Japanese Harmodius,two Athenianscelebratedfor havingkilled the diccommander in Papua New Guinea, told the judges that "the lists
tator Hipparchus: "In fact the bold action undertaken by
for executions were compiled in this way. Information would be
Aristogitonand Harmodiuswas due to a love affair.I shall deal
to
a
soldier
a
native
that
some
a
was
with this in some detail,and show thatAtheniansthemselvesare
given
Japanese
by
person
spy
and had contacted Australian soldiers."'02When Federal forces
no better than other people at producingaccurateinformation
invaded central Arkansas in 1863, a delegation of Unionists from
about their own dictatorsand the factsof their own history."It
Pine Bluff went to meet them and escort them to their town. On
turns out that Hipparchus, without success, approached
Harmodius,"a most beautifulyoung man in the flower of his
arriving in Pine Bluff, the troops proceeded to ransack the homes
of Rebel sympathizers; as one resident noted, "They knew every youth [who] was loved and possessed by Aristogiton."
ones name & where they lived."103After the Whites captured a
HarmodiusrebukedHipparchus's
advancesand told Aristogiton,
"who,being in love as he was, was greatlyupset and was afraid
city during the Russian Civil War, "it was enough for someone to
that Hipparchus,with all his power,might take Harmodiusby
point a finger" for a person to die.104The list of victims in the
Colombian town of Buritica was routinely submitted in advance
force. He thereforebeganat once, so far as he could in his posito the parish priest for approval.'05After he was denounced and
tion, to plot to overthrowthe dictatorship."Eventually,after a
arrested, during the Biafran Civil War, a man recalled: "I should
complicated sequence of events, Harmodius and Aristogiton
not return to Uyo, for my people were after my blood."106 Almost
assassinatedHipparchus.As Thucydidesconcludes:"Inthis way
case
of
indiscriminate
in
violence
Guatemala
the conspiracyof Harmodiusand Aristogitonoriginatedin the
every
apparently
described in detail by Robert Carmack and his associates turns
wounded feeling of a lover."116
PavlikMorozovwas the Soviet
out to have entailed some measure of local input: name lists used
who
informed
on
his
kulak father and was killed by his
boy
in army massacres were composed with information from local
uncles in revengein September 1932. Pavlik became famous
when the Sovietregimepromotedhim as the upstandingyoung
people, "orders to kill . . . had a local origin," and people were
killed after the intervention of old enemies.107Local Serbs particPioneerwho, in a situationof conflictingfamilyand stateloyalin
the
massacre
of
40
about
ethnic
in
Albanians
the
ties, nobly put the interestsof the state first.The writerMaxim
ipated
village
of Slovinje in Kosovo (April 15 and 16, 1999); according to a witGorky cited PavlikMorozov as an exampleof Soviet heroism,
ness, "When the army came, our own Serbs put on masks and
and for decades Pavlikwas treatedas the patron saint of the
joined in the butchery. They knew who to single out. They knew
Pioneers and eulogized in public monuments, meetings, and
who had money."108
A Basque peasantwoman, whose family inspirationalchildren'sbooks. Anticommunists,however,cited
www.apsanet.org 483
Articles | The Ontologyof "Political
Violence"
his case as indicative of the moral decay of totalitarianism,
whereby ideological control undermined and destroyed even
family bonds. But a carefulinvestigationuncovereda different
motivationbehindPavlik'saction:his father,the chairmanof the
local rural soviet, had abandoned his wife and children and
moved in with a youngerwoman from the same village. Pavlik
either denounced his fatherout of personalresentment(as the
eldestchild, at 13 or 14, he had to takecareof his family)or was
prompted by his mother out of revenge,or by a cousin who
wanted to become chairmanof the ruralsoviet.ll7
Forall its manifestimportance,this aspectof violenceremains
hidden to most observers,who, when not dismissingall violence
as "criminal,"tend to code it automaticallyas "political"(ethnic,
religious,partisan,et cetera).Indeed,the violenceof civil warsis
described and classified as "politicalviolence." Most macro
studies disregardthe privatecontent of "politicalviolence"and
miscode individual cases. However, identifying the locus of
agencyis highly consequentialfrom a theoreticalpoint of view.
The intersticesof politicaland privateviolenceprovideconsiderablespacefor manipulation-a fact noted by participantsand
observersalike.Forexample,the Frenchtroopssent by Napoleon
to suppressthe rebellionin Calabrianoticedin 1807 thatthe local
peoplewerehijackingtheirwar.The local volunteerswho joined
the Civic Guardshad a "tendencyto pursuelocalvendettasquite
apartfrom the war effort.There is much evidencethat the desire
to settlea long-standingfeudwith a localrivalfamilywas a strong
impetusfor joining the Civic Guards.On severaloccasionslocal
town dwellers asked the French to allow them to execute
Calabrianprisonerswho happenedto be membersof a rivalfamily or from a rivaltown."1l8This certainlyechoesrecentdevelopmentsin contemporaryAfghanistanand Iraq.
Although in some instances political actors willingly under-
write local factionsin every respect,in other instancesthey are
manipulatedby such factionsand led to act in ways they would
have otherwisepreferredto avoid. Local actors sometimessucceed in getting centralactorsto directtheir violenceagainstprivateenemiesby describingthem in the idiom of the mastercleavage. Sheila Fitzpatrick and Robert Gellately's comparative
overview of denunciation in modern European dictatorships
emphasizesexactlythis point:
state'sexceptional
Becauseof the totalitarian
willingnessto receive
fromitscitizensandto actuponthem,thatstate'sfordenunciations
citimidablepowerswerein effectput at the disposalof individual
zens.Ifyouhavea privateenemy,whynotdenouncehimto thepolice
Thenthe Gestapoor the NKVDwouldtake
as a Jewor Trotskyite?
him awayto a concentration
camp,and yourproblemwouldbe
wasextremely
denunciation
solved.... This kindof manipulative
commonin bothsocieties.Classenemiesweredenouncedin Stalin's
whocovetedtheirapartments;
SovietUnionby neighbors
Jewswere
the
same
for
in
Nazi
denounced
purpose,and
Germany
byneighbors
well as conflicts about local political power, led to violence
because "they tried to resolve them using their political
In a Guatemalantown, "asguerrillasentered local
groups."121
socialrelations,neighborswho felt they had beenwrongedin the
distributionof land were presentedwith new ways to settle
scores."122Sometimes, the process entails more complicated
chains of principalsand agents, as in the following description
from Punjab,India:
withinthevillagesare
factionalandfamilyanimosities
Undoubtedly
of new
thedevelopment
exploitedby thestateas a wayof hindering
in maritaldisIn its fightagainstterrorism
loyalties.
policeinterfered
andhencecomputesandlanddisputesin thevillages,supporting,
would
be
False
one
by one
complaints
registered
promising, party.
bythestate,to theeffectthattheoppopartyto a dispute,supported
nent had linkswith terrorists.
The individualnatureof the many
overlandbetweenandwithinfamilies. . . [was]eclipsedby
quarrels
thewidespread
useof suchquarrels
by thepolice.Disputesspiraled
of state,usedallsuchconoutof controlasthepolice,asinstruments
flicts to advancetheir missionagainstterrorism.Incidentswere
framework.
Policeofficers
and convertedinto a terrorist
processed
In thistheyweregivenprowouldthenclaimthe resulting
rewards.
In themidst
officersandrarelyheldaccountable.
tectionbysuperior
of situationssuch as these,innocentswith no connectionsto militan-
in desperate
trouble.123
cy foundthemselves
The realizationthat agentsoften manipulatetheirprincipalsproduces paradoxicalstatements,as when Ralph Thaxton reports
that in occupied China "Yang'spuppet regime exertedits own
interestoverthat of its Japanesemasters."24
The interactionof the politicaland the privatepoints to a crucial puzzlethat can be succinctlyexpressedin Lenin'sfamousformulation:Ktokovo?Who is takingwhom in hand?Who manipulateswhom?Arecentralactorsusinglocalones, or is it the other
way around?In a book about his mother'sexecutionduringthe
GreekCivil War,Nicholas Gage sets up this puzzleas his main
theme:
As I drovetowardthecentralsquare,I kepthearingoverthesoundof
a
thecar'senginea phrasethatmysisterandmyfatherhadrepeated
was the villagerswho
hundredtimes: "Tinfaganei horiani"--"It
like Katis
devouredher."To my family,the Communistguerrillas
on ourvillagebywar,like
actof God,unleashed
wereanimpersonal
for my
whomtheyheldresponsible
a plague.It wasour neighbors
secretsto the security
mother'sdeath;the villagerswho whispered
I had
policeandtestifiedagainstherat thetrial.Thiswassomething
to resolve:perhapsthe villagersreallyweremoreculpablefor her
deaththanthe menwhopassedthesentenceandfiredthe bullets.I
if something
aboutmymotherincitedthepeopleof Liato
wondered
hadonly
lamb.Or perhapsthevillagers
offerherup likea sacrificial
whoexploitedtheirmoralweakbeenmanipulated
by theguerrillas,
wantedmy
and fears,becausethe guerrillas
nesses,pettyjealousies
Whatwastherealreasonshe
motherkilledforsomepoliticalpurpose.
was executed?125
with similarsuccess.119
Both duringthe Japaneseoccupationof the Philippinesand during the Huk rebellion,local authoritiestook advantageof the
situation"tosettleold quarrelsfromprewardaysby accusingenemies of being antigovernment without showing any proof."120In
El Salvador,waterand land disputesamong peasantfamilies,as
484
September 2003 | Vol. 1/No. 3
Interaction
Both the relativestrengthof centralvis-a-vislocal dynamicsand
the locusof agencyareperenniallypuzzling.The questionis nicely
formulatedby Howell:"Whatone needs to know is the manner
in which the local issues, local perceptions,and local problems
shapedand informedthe nationalperspective... and conversely
how that sense of generality,which is so integrala part of the
nationalperspective,was transferredand perhapstranslatedback
into the frameworkand languageof local politics."126
I havealreadydiscussedthe propensityof macro-levelaccounts
to completelyoverlooklocaldynamics;this papermakesclearthat
it would be equallymisguidedto deprivethe local and private
sphereof agency.Indeed,the evidenceadducedso farwouldappear
to underminethe Schmittianthesisin favorof the Hobbesianone,
supportinga view of civilwaras a processso utterlydecentralized
and uncontrolledas to be almostanomic,pointless,and random.
Arewe then to reducecivilwarsinto simpleaggregations
of private
feudsandlocalconflicts-much as Homerdid in describingwaras
an aggregationof duels?127
Are civil warsnothing but "feudswrit
To paraphrasea well-knowndictum, are all civil war
large?"128
politicslocal?The answeris negative.
Among the researcherswho stressthe importanceof private
and localconflicts,some strikea correctcautionarynote by arguing that while these conflictsinvolvelocal individualsand communities, their originsare external.BenjaminPaul and William
Demarest'sdetaileddescriptionof the operationof a deathsquad
in a small town of Guatemalashowshow a groupof individuals
was vestedby the armywith exceptionalpower,which they used
in pursuitof vengeance,local power,"money,liquor,and sex."
They conclude:
needed no outside energyto continue, though it was of course
The processof interactionis capexploitedby outsideagents."135
tured at the individuallevel by the practiceof denunciation.
Fitzpatrickobservesthatwhile it "canbe seen in 'top down'terms
as a state controlmechanismand a meansof monitoringpublic
opinion ... thereis also a possible'bottom up' interpretationof
the function of denunciation:if the state used this practiceto
control its citizens, individualcitizens could also use it for the
This is also nicely conpurposeof manipulatingthe state."136
veyed in a letter from occupiedGreece,in 1944: "Jason,son of
P.," this letter goes, served the Italians on his island so well that
Cobb also capturesthis interthey "carriedout all his desires."137
actionwhen he describesinstancesof violenceduringthe French
Revolutionas situations"wherethere was no frontierbetween
privatevengeanceand collectivevengeance,"which was exercised
by people who put their "privateviolence to public use."138
Violence in Congo-Brazzaville
is portrayedas a situationwhere
"therewas no distinctionmade betweena privatesphereand a
a point echoedby a studyof Nicaragua,where
publicsphere,"139
the motivesof violence"wereapparentlypersonalas well as political."140
The murderof Afonso Goncalvesin September1999 in
EastTimor was "aspersonalas it was political";Goncalveswas
killed not only for the pro-independenceviews he held, but also
for a family feud relatedto a niece who eloped, againstfamily
resistance,with a pro-Indonesiamilitiaman.A yearlater,during
the terrorthat engulfedEastTimor in the wake of the referenIt maybetemptingto blametheoutbreak
of violencein SanPedroon dum, membersof the militiaman's
family came to Goncalves's
socialdivisiveness
andthe settlingof old scores,but the temptation house and killed
him.'14 In CivilWarTennessee,participantsdid
shouldbe resisted.Religiouscompetitionand vigorouspolitical
not alwaysseparateviolencemotivatedby politicalends and vioinfightingwerefeaturesof SanPedrolife for decadesbefore1980 lence
originatingin personalgrievances.142
withoutproducing
violence.The samecanbe saidforinterpersonal
the extremepoliticizationof life undertotalitarParadoxically,
Theyarosein the pastandweresettledby meansshort
antagonisms.
ian
leads
to the extreme privatizationof politics. By
of murder.
regimes
Whatdisrupted
thepeacein SanPedrowasnot thepresto
turn
all
that is personalinto the political,totalitarians
enceof differences
anddivisions,
butthearmy's
of agents wanting
recruitment
the
exact
andspiesthathadtheeffectof exploiting
thesecleavages.'29
get
oppositeresult:they turn the politicalinto the private.JanGrossarguesthat the essenceof totalitarianism
was "the
It is rightthen to saythatthe decentralized
and localizednature institutionalizationof resentment."143
In his study of the Soviet
of the Republicanviolenceduringthe SpanishCivilWardoes not occupationof westernUkraineand westernBelorussiain 1939,
imply that it was an instanceof spontaneousand anarchicalvio- he finds that the new powerapparatuswas "motivatedby particlence by uncontrolledactors, as is usuallyassumedby histori- ular interests,like avengingpersonalwrongs, assuaginghunger,
ans,130or that violence in civil war is double-edged.13 These points
or satisfyinggreed"in a patternakin to the "privatization
of the
arewell takenas warningsagainstan interpretation
of privateand state."He describesthe violence there as a situationwhere "the
local conflictsthat overlooksthe politicalcontext in which they state was franchised,as it were, to local individuals,who used
occur.In most places,localconflictsand privategrudgesarepres- theirpowerto pursuetheirprivateinterestsand settlescores;the
ent without eruptinginto violence. State sanctionsand mecha- pursuitof privateinterestsbecamethe principalmethod of carnismsof socialcontrolpreventtranslationinto violenceand pro- ryingout officialdutiesand establishingauthority."He addsthat
vide ways of managingsocial tension.132Even in the context of "Sovietauthoritiesconductedsearchesand arrests. . . directlyin
civil war,such conflictsdo not alwaysresultin violence.33
response to denunciations by neighbors who had personal
It would seem obvious that both centraland local dynamics accountsto square .... [A]ccusations,denunciations,and permatter.As Howellwritesaboutthe EnglishCivilWar:"Atvarious sonalanimositiescouldleadto arrestat anymoment.Peoplewere
points throughoutthe century local and national politics had officiallyencouragedto bringaccusationsand denunciations....
intersectedin waysthat intensifiedthe natureof politicaldebate. [W]hoeverhad a grudgeagainstsomebodyelse, an old feud,who
Local grievancesbecame the medium through which many had anotheras a grainof salt in the eye-he had a stageto show
nationalconcernswere perceived,while the issuesand labelsof his skills,therewas a cockedear,willingto listen."144
JungChang
nationaldebatewereused to clothe the continuinglocal political locates the source of much violence perpetratedduring the
struggles."134
StanleyAschenbrennerdescribesthe Greek Civil Cultural Revolution in Mao's mobilization of envy and
War,in a Greekvillage,as "asequenceof actionand reactionthat resentment.In her familyhistory,she eloquentlyshowshow the
www.apsanet.org 485
Articles
I The Ontologyof "PoliticalViolence"
politicizationof privatelife ultimatelyleads to the privatization
of politics: "The Communistshad embarkedon a radicalreorganizationnot just of institutions,but of people'slives, especially the livesof thosewho had 'joinedthe revolution.'The ideawas
that everythingpersonalwas political;in fact, henceforthnothing was supposedto be regardedas 'personal'or private.Pettiness
was validatedby being labeled'political,'and meetings became
the forum by which the Communistschanneledall sortsof personal animosities."Changprovidesthe followingpersonalexample: "Mymotherwas also horrifiedto hearthat my grandmother
had beendenounced-by her own sister-in-law,Yu-lin'swife. She
had long felt put-uponby my grandmother,as she had to do the
hardwork aroundthe house, while my grandmotherranit as its
mistress.The Communistshad urgedeveryoneto speakup about
'oppressionand exploitation,'so Mrs.Yu-lin'sgrudgesweregiven
a politicalframework."145
This evidencesuggeststhatthe intimatecharacterthat "political
violence"often displaysis not necessarilythe reflectionof impersonal or abstractideologicalor identity-basedpolarizationand
hatred;it is alsothe surprisingresultof the interactionbetweenthe
politicaland privatespheres.
Cleavage and Alliance
To summarize,the interaction between supralocaland local
actors,and the privateand public spheres,is hinted at by various
works,but is left untheorized.Below,I outline the missingtheoreticalaccount.
Actorsat the centerare assumedto be linked with action on
the ground via the well-known mechanism of cleavage.This
impliesvariousunderlyingmicrofoundations,most notablycenor coorcommon preferences,147
tralizedorganization,146
fear,148
another
This
introduces
dination aroundfocal points.l49
paper
microfoundationlinkingcenterand periphery:alliance.The theoreticaladvantageof allianceis that it allowsfor multiplerather
than unitaryactors,agencylocatedin bothcenterand periphery
ratherthan only in either one, and a varietyof preferencesand
identitiesas opposedto a common and overarchingone. Alliance
entailsa transactionbetweensupralocaland local actors,whereby
the formersupply the latterwith externalmuscle, thus allowing
them to win decisivelocal advantage;in exchangethe formerrely
on local conflictsto recruitand motivatesupportersand obtain
local control, resources,and information'50 even when their
ideologicalagendais opposed to localism.'51Fromthis perspective, the selectivebenefitthat producescollectiveactionand supportis violence,whichoperatesherenot as an instrumentof coercion but as a resourceleadingto mobilization.152
Allianceis for local actorsa meansratherthan a goal, as conA greatdeal of action in
firmedby anthropologicalevidence.153
andlinkedto
decentralized
civil waris, therefore,simultaneously
which
can be both
includes
the wider conflict; this
violence,
in both the
resides
time.
at
same
Agency
politicaland private the
war
thus
be underCivil
may
privateand the political spheres.
actors'
the
collective
a
into
stood as transforming
joint process
for
local
the
actors'
and
local
for
advantage.
quest
quest
power
This view is an alternativeto the conventional dichotomy
betweenthe Schmittianand Hobbesianframes.Localand private
486
September 2003 | Vol. 1/No. 3
conflictsexplodeinto sustainedviolenceneitherbecausecivil war
is an instanceof Hobbesiananarchynor as a resultof the designs
and manipulationsof supralocalactors.What matters,instead,is
the interactionbetweenthe two.
The relevanceof this conceptualizationis twofold. First, it
allowsfor a theoreticalunderstandingof civil war that incorporatesthe puzzleof the disjunctionbetweencenterand periphery
and the relatedextensiveambiguity.Second, it turnsthe centerperipheryinterfaceinto a central issue and forces us to think
more preciselyabout the modalitieslinking distinct actorsand
motivations.This interpretationhas the addedadvantageof subsumingboth strategicactionsby politicalactorsand opportunistic actionsby local individuals.
We may,then, want to think of cleavageas a symbolicformation that simplifies,streamlines,and incorporatesa bewildering
varietyof localconflicts-a view compatiblewith the wayoutside
as a means
observers,like historians,relyon a "masternarrative"
of "emplotment,"to tell a straightcompellingstoryout of many
complexones.154Similarly,allianceallowsus to see civil warsas
concatenationsof multiple and often disparatelocal cleavages,
more or less loosely arrayedaroundthe mastercleavage.This is
consistentwith insights and interpretationsfrom a number of
Forexample,OlivierRoy interpretsthe Islamist/conresearchers.
servativecleavageof the 1992 civil war in Tajikistanin termsof
what he describesas the essentialfeatureof Tajikpolitics,namethat civil war'smasor localism.He disaggregates
ly mahalgeray,
ter cleavage(religion)into a numberof disparateconflictsalong
multipledimensions,such as region,profession,positionwithin
it is easierto disthe stateapparatus,and ethnicity.'55Predictably,
cern these dynamicsin recentcivil wars,which lack the sort of
modulardiscoursesprovidedby the Cold War.But the available
evidencesuggeststhe commonalityof these dynamics;perceived
differencesbetweenpost-Cold War conflicts and previouscivil
warsmay be attributablemore to the demiseof readilyavailable
conceptualcategoriescausedby the end of the Cold Warthan to
the fundamentallydifferentnatureof pre-ColdWarcivilwars.156
Likewise,the fact that ethnic or religiouslocal cleavagesaregenerally easier to discern by outside observersthan are factional
ones may also causea bias in reporting,coding, and interpreting
evidence.
Thucydideshints at the mechanismof alliancewhen he argues,
in his analysisof the civilwarin Corcyra,that "inpeacetimethere
would havebeenno excuseand no desirefor the callingof [external allies] in, but in time of war,when each partycould always
count upon an alliancewhich would do harm to its opponents
and at the sametime strengthenits own position,it becamea naturalthing for anyonewho wanteda changein governmentto call
At the sametime, externalintervention
in help from outside."157
is possible only when local factionsand individualsare willing
and able to call in outsiders.Determiningwhen this is the case,
and who allieswith whom, calls for a fine-grainedanalysisthat
takes into account both intracommunitydynamics and the
dynamicsof the civilwar.Forinstance,a recurringpatternis that
losersin local conflictsare more likely to move first and, hence,
be the first ones to call in outside forces.Localauthoritieswho
had been marginalizedby the governmentwere highly likely to
join the Renamo insurgencyin Mozambique;and in Sierra
Leone, "losersin a local land or chieftaincydisputemight sometimes side with the insurgentsto securerevenge.The beheading
of a ParamountChief, Gboney Fyle, in Bonthe District is
In this sense,civilwaris the ideal
thoughtto be one such case."158
revancheopportunityfor losersin local powerconflictsas well as
individualswho feel slightedand envious.It is hardto conveythis
betterthan a man who, afterthe Union ArmyenteredMadison
County in Alabama,announcedhis intentionto kill his localrival
and then "getsome of the Union soldiersand takeeverythingout
of [his rival's]house and burn the whole place up.... He has
been a big fellow for a long time, but now is my time to bring
him down."159
The dearthof systematicdatamakesit impossibleat this point
to recordand analyzethe modalitiesof interactionbetweencentral and local actors. Still, it is possible to put forwardtwo
hypothesesaboutthe relativeimportanceof alliancecomparedto
top-downmechanisms,such as centralizedorganizationor common preferenceswithin a civil war. First,the top-down mechanisms arelikely to do most of the "heavylifting"beforethe war,
duringits initialstages,or afterthe warhas ended.When the war
is underway,alliancemayprevailsincethe wartendsto fragment
geographicalspace,thus placinga premiumon localdynamics.160
Once a war has ended, the masternarrativeof cleavageprovides
a handyway to ex post facto simplify,streamline,and coverup
the war'sambiguitiesand contradictions-including the role of
alliance.161Sometimes, the invocation by local and individual
actorsof the mastersymbolor messagemay becomea self-fulfilling prophecyas local issues and identitiesget redefined,reconstructed,and projectedbackwardfollowingthe conflict'sconclusion. The recurrenceof the same alliancesover time and the
relianceon the samecentralsymbolsand messagesmay ultimately integrateand fusethe multitudeof localcleavagesinto the master cleavage-consistent with the observationthat warsarestateA secondhypothesiswould accountfor the
buildingprocesses.162
relativesalienceof allianceacrosscivil wars:the less powerfuland
centralizedthe political actorsfighting a war, the less able they
will be to impose control directlyand hence the more likely to
resortto local alliances.An implicationis that substantialthirdparty assistancemay make allianceless useful for at least one
party.
Conclusion
Civil war is a context that placesa premiumon the joint action
of local and supralocalactors,insidersand outsiders,individuals and organizations,civilians and armies: action (including
violence) resultsfrom their alliance in pursuit of their diverse
goals-whose main empiricalmanifestationis ambiguity.The
interpretiveframeelaboratedhere carriestwo majortheoretical
implicationsfor theories of civil wars and "politicalviolence."
First,and counter to Schmitt, "politicalviolence"is not always
necessarilypolitical;identitiesand actionscannot be reducedto
decisions taken by the belligerent organizations,to the discourses produced at the center, and to the ideologies derived
from the war'smastercleavage.So positingunitaryactors,inferring the dynamics of identity and action exclusivelyfrom the
mastercleavage,and framingcivil wars in binaryterms is misleading;instead,local cleavagesand intracommunitydynamics
must be incorporatedinto theories of civil war. Second, and
counterto Hobbes, civil warcannotbe reducedto a meremechanism that opens up the floodgatesto random and anarchical
privateviolence. Privateviolence is generallyconstrainedby the
modalities of alliance,which must be exploredsystematically.
Civil war fostersinteractionamong actorswith distinct identities and interests. It is the convergenceof local motives and
supralocalimperativesthat endows civil war with its particular
characterand leads to joint violence that straddlesthe divide
between the political and the private, the collective and the
individual.
References
Annis, Sheldon. 1988. Story from a peacefultown: San
Antonio Aguas Calientes.In Harvestof Violence:TheMaya
Indiansand the GuatemalanCrisis,ed. RobertM. Carmack.
Norman:Universityof OklahomaPress,155-73.
Arnson, Cynthia, and Robin Kirk. 1993. State of War:Political Violenceand Counterinsurgency
in Colombia.New York:
Human Rights Watch/Americas.
Aschenbrenner,Stanley.1987. The civil war from the perspectiveof a Messenianvillage. In Studiesin the Historyof
the GreekCivil War,1945-1949, eds. LarsBaerentzen,
John O. Iatrides,and Ole L. Smith. Copenhagen:Museum TusculanumPress, 105-25.
Ash, StephenV. 1995. Whenthe YankeesCame:Conflictand
Chaosin the OccupiedSouth, 1861-1865. Chapel Hill:
Universityof North CarolinaPress.
Bartolini,Stefano.2000. The PoliticalMobilizationof the EuropeanLeft, 1860-1980: The ClassCleavage.Cambridge:
CambridgeUniversityPress.
Bax, Mart. 2000. Warlords,priestsand the politics of ethnic
cleansing:A case study from ruralBosnia Hercegovina.
Ethnicand RacialStudies23:1, 16-36.
Remy. 1999a. The spreadof politicalvioBazenguissa-Ganga,
lence in Congo-Brazzaville.
AfricanAffairs98, 37-54.
. 1999b. Les ninja, les cobra et les Zoulou crevent
l'ecrana Brazzaville:Le role des medias et la construction
des identites de violence politique. CanadianJournalof
AfricanStudies33:2-3, 329-61.
Bearak,Barry.1999. Crisis in the Balkans:The atrocities;
Kosovo town'stale of betrayaland massacre.The New
YorkTimes,6 May, Al.
Beevor,Anthony. 2001. The SpanishCivil War.London:
Cassell.
Berdal,Mats, and David M. Malone, eds. 2000. Greedand
Grievance:EconomicAgendasin Civil Wars.Boulder,Colo.:
Lynne RiennerPublishers.
Berlow,Alan. 1998. Dead Season:A Storyof Murderand Revenge.New York:Vintage.
Berman,Paul. 1996. In searchof Ben Linder'skillers. The
New Yorker,
23 September,58-81.
Bernand,Andre. 1999. Guerreet violencedans la Greceantique. Paris:Hachette.
www.apsanet.org 487
Articles j The Ontologyof "Political
Violence"
Besteman,Catherine. 1996. Violent politics and the politics
of violence:The dissolutionof the Somali nation-state.
AmericanEthnologist23:3, 579-96.
Bobbio, Norberto. 1992. Guerracivile? TeoriaPolitica1-2,
297-307.
Brass,Paul R. 1997. Theftof an Idol: Textand Contextin the
Representation
of CollectiveViolence.Princeton:Princeton
UniversityPress.
Brovkin,VladimirN. 1994. Behindthe FrontLinesof the
Civil War:PoliticalPartiesand SocialMovementsin Russia,
1918-1922. Princeton:PrincetonUniversityPress.
Brubaker,Rogers,and David D. Laitin. 1998. Ethnic and
nationalistviolence. Annual Reviewof Sociology24,
243-52.
Cabarruis,Carlos Rafael. 1983. Genesisde una revolucidn:
Andlisisdel surgimiento
y desarrollode la organizacion
El
en
Salvador.
Mexico, D.E: Centro de Investicampesina
Estudios
Superioresen AntropologfaSocial.
gacionesy
Nationalism
and ethnicities:Lessons
2000.
Michel.
Cahen,
The Politicsof War,
In
Kills?
from Mozambique. Ethnicity
Peaceand Ethnicityin SubSaharanAfrica,eds. Einar
Braathen,Morten Boas, and GjermundSather. London:
MacMillan, 163-87.
Calder,BruceJ. 1984. The Impactof Intervention:The DominicanRepublicduringthe U.S. Occupationof
1916-1924. Austin: Universityof Texas Press.
Carmack,Robert M. 1988. The story of Santa Cruz Quiche.
In Harvestof Violence:The Maya Indiansand the
GuatemalanCrisis,ed. Robert M. Carmack.Norman: Universityof OklahomaPress,39-69.
Chang, Jung. 1992. Wild Swans:ThreeDaughtersof China.
New York:Doubleday.
Chivers,C. J. 2003. Feud between Kurdishclans createsits
own war. The New YorkTimes,24 February,A8.
Chwe, Michael Suk-Young.2001. RationalRitual:Culture,
Coordination,and CommonKnowledge.Princeton:Princeton UniversityPress.
Clastres,Pierre. 1999. Archeologiede la violence.Paris:Editions de l'Aube.
Cobb, Richard.1972. Reactionsto the FrenchRevolution.London: Oxford UniversityPress.
Collier, GeorgeAllen. 1987. Socialistsof RuralAndalusia:UnRevolutionaries
of the SecondRepublic.Stanacknowledged
ford: StanfordUniversityPress.
Collier, Paul, and Anke Hoeffler.2002. Greed and grievance
in civil war. Centre for the Study of AfricanEconomies
WorkingPaperSeries2002-01.
Crainz, Guido. 1995. Memoire des conflicts et inertie d'une
guerrecivile: L'2milieen 1945-1956. In La guerrecivile
entrehistoireet memoire,ed. Jean-ClementMartin. Nantes:
Ouest 1ditions, 121-7.
Cribb, Robert. 1990. Introduction:Problemsin the historiography of the killings in Indonesia.In The Indonesian
Killings,1965-1966. Studiesfrom Java and Bali, ed.
Robert Cribb. Monash University:Centre of Southeast
Asian Studies, 1-43.
488
September 2003 | Vol. 1/No. 3
Crow,JeffreyJ. 1985. Libertymen and loyalists:Disorder
and disaffectionin the North Carolinabackcountry.In An
Uncivil War:The SouthernBackcountry
duringthe American
Revolution(Perspectives
on the AmericanRevolution),eds.
Ronald Hoffman, Thad W. Tate, and PeterJ. Albert.
Charlottesville:UniversityPressof Virginia, 125-78.
Dale, Catherine.1997. The dynamicsand challengesof ethnic cleansing:The Georgia-Abkhazia
case. RefugeeSurvey
Quarterly16:3, 77-109.
Dalrymple,William. 1997. Fromthe Holy Mountain:A Journey amongthe Christiansof the Middle East. New York:
Henry Holt.
Davidson,Jamie S. Forthcoming.The politics of violence on
an Indonesianperiphery.SouthEastAsia Research.
Dean, Roger.2000. Rethinkingthe civil war in Sudan. Civil
Wars3:1, 71-91.
De Foxa, Agustin. 1993. Madrid de Cortea Checa.Barcelona:
Planeta.
Degregori,Carlos Ivan. 1998. Harvestingstorms:Peasant
rondasand the defeat of SenderoLuminosoin Ayacucho.
In Shiningand OtherPaths:Warand Societyin Peru,
1980-1995, ed. SteveJ. Stern. Durham:Duke University
Press, 128-57.
De Lupis, Ingrid Detter. 1987. The Law of War.Cambridge:
CambridgeUniversityPress.
Des Forges,Alison. 1999. LeaveNone to Tellthe Story:Genocide in Rwanda.New York:Human Rights Watch.
Du Boulay,Juliet. 1974. Portraitof a GreekMountain Village.
Oxford:Oxford UniversityPress.
Duyvesteyn,Isabelle.2000. Contemporarywar: Ethnic conflict, resourceconflict or something else? Civil Wars3:1,
92-116.
Elliott, David W. P. 2003. The VietnameseWar:Revolution
and Social Changein the MekongDelta, 1930-1975.
Armonk, N.Y.: M. E. Sharpe.
Ellis, Stephen. 1995. Liberia1989-1994: A study of ethnic
and spiritualviolence. AfricanAffairs94, 165-97.
1999. The Mask of Anarchy:The Destructionof
Liberiaand the ReligiousDimensionof an African Civil
War.New York:NYU Press.
Enzensberger,Hans Magnus. 1994. Civil Wars:FromL.A. to
Bosnia.New York:New Press.
Escott, Paul D., and JeffreyJ. Crow. 1986. The social order
and violent disorder:An analysisof North Carolinain the
Revolutionand Civil War. TheJournalof SouthernHistory
52:3, 373-402.
Essien,JeremiahMoses. 1987. In the Shadowof Death:Personal Recollections
of Eventsduringthe Nigerian Civil War.
Ibadan:Heinemann.
Everitt,Alan. 1997. The local community and the great rebellion. In The EnglishCivil Wars:LocalAspects,ed. R. C.
Richardson.Phoenix Mill: Sutton, 15-36.
Faivre,Maurice.1994. Un villagede Harkis.Paris:l'Harmattan.
Fawaz,LeilaTarazi.1994. An Occasionfor War:Civil Conflict
in Lebanonand Damascusin 1860. Berkeley:Universityof
CaliforniaPress.
Fearon,James D., and David D. Laitin. 1996. Explaininginterethniccooperation.AmericanPoliticalScienceReview
90:4, 715-35.
. 2003. Ethnicity,insurgency,and civil war. American
PoliticalScienceReview97:1, 75-90.
Fellman,Michael. 1989. InsideWar:The GuerrillaConflictin
Missouriduringthe AmericanCivil War.New York:Oxford
UniversityPress.
Fenoglio,Beppe.1973. La guerresur les collines.Paris:Gallimard.
Fiennes, Ranulph. 1975. WhereSoldiersFear to Tread.
London: Hodder and Stoughton.
Figes, Orlando. 1996. A PeoplesTragedy:The RussianRevolution, 1891-1924. New York:Viking.
Finley,Milton. 1994. TheMost Monstrousof Wars:The
NapoleonicGuerrillaWarin SouthernItaly, 1806-1811.
Columbia:Universityof South CarolinaPress.
Fisher,Noel C. 1997. Warat EveryDoor:PartisanPolitics
and GuerrillaViolencein East Tennessee,
1860-1869.
Chapel Hill: Universityof North CarolinaPress.
and SurFitzpatrick,Sheila. 1994. StalinsPeasants:Resistance
vival in the RussianVillageafter Collectivization.
New York:
Oxford UniversityPress.
Fitzpatrick,Sheila, and Robert Gellately.1997. Introduction
to the practicesof denunciationin modern European
Practices:Denunciationin Modern
history.In Accusatory
eds. Sheila Fitzpatrick
1789-1989,
EuropeanHistory,
and Robert Gellately.Chicago:Universityof Chicago
Press, 1-21.
Freeman,Simon. 1994. Learningthe fundamentallessons of
religiousconviction. The Scotsman,28 December, 14.
Freeman,SusanTax. 1979. The Pasiegos:Spaniardsin No
Mans Land. Chicago:Universityof Chicago Press.
Gage, Nicholas. 1984. Eleni. New York:BallantineBooks.
Geffray,Christian.1990. La causedes armesau Mozambique:
d'uneguerrecivile. Paris:Karthala.
Anthropologie
Gossman,Patricia.2000. India'ssecret armies.In Death
Murderwith Deniability,eds.
Squadsin GlobalPerspective:
Bruce B. Campbelland Arthur D. Brenner.New York:
St. Martin'sPress,261-86.
Gould, Roger. 1995. InsurgentIdentities:Class,Community,
and Protestin Parisfrom 1848 to the Commune.Chicago:
Universityof Chicago Press.
Gross,Jan T. 1988. Revolutionfrom Abroad:The Soviet Conquestof Poland'sWesternUkraineand WesternBelorussia.
Princeton:PrincetonUniversityPress.
. 2001. Neighbors:TheDestructionof theJewishCommunityin Jedwabne,Poland.Princeton:PrincetonUniversity
Press.
Groth, Siegfried.1995. Namibia-The Wallof Silence:The
Dark Days of the LiberationStruggle.Wuppertal:Peter
HammerVerlag.
Gunfight shattersKabul calm. 2002. BBC News,World Edition, 7 August. Availableat news.bbc.co.uk/2/hi/south
_asia/2178607.stm.Accessed30 May 2003.
Hamoumou, Mohand. 1993. Et ils sont devenusHarkis.Paris:
Fayard.
Hardin, Russell. 1995. Onefor All: The Logicof GroupConflict. Princeton:PrincetonUniversityPress.
Harding, Susan E 1984. RemakingIbieca:RuralLife in
AragonunderFranco.Chapel Hill: Universityof North
CarolinaPress.
Hart, Peter. 1997. The geographyof revolutionin Ireland
1917-1923. Past & Present155, 142-155.
. 1998. The I.R.A. and Its Enemies:Violenceand Communityin Cork,1916-1923. New York:ClarendonPress.
Henderson,James D. 1985. When ColombiaBled:A History
of the Violenciain Tolima.University:Universityof
AlabamaPress.
Henriksen,Thomas H. 1983. Revolutionand Counterrevolution: Mozambique's
Warof Independence,
1964-1974. Westport: GreenwoodPublishingGroup.
Herrington,StuartA. 1997. Stalkingthe Vietcong.InsideOperationPhoenix:A PersonalAccount.Novato, Calif.:
PresidioPress.
Hinton, William. 1984. Shenfan:The ContinuingRevolution
in a ChineseVillage.New York:Vintage Books.
Hoare, Marko. 2001. The partisansin Bosnia-Herzegovina
1941-1946. Unpublishedpaper,Universityof Cambridge,
U.K.
Hobsbawm,E. J. 2001. Bandits.London:Abacus.
Horowitz, Donald. 1985. Ethnic Groupsin Conflict.Berkeley:
Universityof CaliforniaPress.
Horton, Lynn. 1998. Peasantsin Arms:Warand Peacein the
Mountainsof Nicaragua,1979-1944. Athens: Ohio UniversityCenter for InternationalStudies.
Howell, Roger,Jr. 1997. Newcastleand the nation:The
seventeenth-centuryexperience.In The EnglishCivil Wars:
LocalAspects,ed. R. C. Richardson.Phoenix Mill, U.K.:
Sutton Publishing,309-29.
Hua, Linshan,and IsabelleThireau. 1996. Enquetesociologiquesur la Chine, 1911-1949. Paris:PressesUniversitairesde France.
Johnson, Michael. 2001. All HonourableMen: The Social Origins of Warin Lebanon.London: Centre for Lebanese
Studies and I. B. Tauris.
Jones, Gregg R. 1989. Red Revolution:Insidethe Philippine
GuerrillaMovement.Boulder,Colo.: WestviewPress.
Kakar,Sudhir.1996. The Colorsof Violence:CulturalIdentities,
Religion,and Conflict.Chicago:Universityof ChicagoPress.
Kaldor,Mary. 1999. New and Old Wars:OrganizedViolence
in a GlobalEra. Stanford:StanfordUniversityPress.
Kalyvas,StathisN. 1996. The Riseof ChristianDemocracyin
Europe.Ithaca,N.Y.: Cornell UniversityPress.
. 2001. "New"and "old"civil wars:A valid distinction? WorldPolitics54:1, 99-118.
. 2003. The logic of violence in civil war. Unpublished manuscript,Universityof Chicago.
Kaufman,StuartJ. 2001. ModernHatreds:The SymbolicPolitics of EthnicWar.Ithaca,N.Y.: Cornell UniversityPress.
Kedward,H. R. 1993. In Searchof the Maquis:RuralResistance in SouthernFrance,1942-1944. Oxford: Oxford
UniversityPress.
www.apsanet.org 489
Articles
I The Ontologyof "PoliticalViolence"
Keen, David. 1998. The economic functions of violence in
civil wars. Adelphi Paper320. London: Oxford University
Press,InternationalInstitute for StrategicStudies.
Kelly,RaymondC. 2000. WarlessSocietiesand the Originof
War.Ann Arbor:Universityof Michigan Press.
Kerkvliet,BenedictJ. 1977. The Huk Rebellion:A Studyof
PeasantRevoltin the Philippines.Berkeley:Universityof
CaliforniaPress.
Kheng, Cheah Boon. 1980. The social impact of the Japanese occupationof Malaya(1942-1945). In SoutheastAsia
underJapaneseOccupation,ed. Alfred W. McCoy. New
Haven:Yale UniversitySoutheastAsia Studies, 91-123.
Kriger,Norma J. 1992. ZimbabwesGuerrillaWar:Peasant
Voices.Cambridge:CambridgeUniversityPress.
Lacey,Marc. 2003. Letterfrom Africa:With all the little
wars, big peace is elusive. The New YorkTimes,9 April,
A4.
Lear,Elmer. 1961. The Japaneseoccupationof the Philippines, Leyte, 1941-1945. Data PaperNo. 42, Southeast
Asia Program,Departmentof Far EasternStudies. Ithaca,
N.Y.: Cornell UniversityPress.
LedesmaVera,Jose Luis. 2001. Espacios de poder, violencia
y revoluci6n:Una perspectivapolitica de la represi6n
en el Arag6n republicanodurante la guerracivil. In
El dificil caminoa la democracia,ed. Antonio Morales
Moya. Madrid: Sociedad Estatal EspaiiaNuevo Milenio,
249-68.
Lewis, Gwynne. 1978. The SecondVendde.Oxford: Oxford
UniversityPress.
Linn, Brian McAllister.1989. The U.S. Armyand Counterinsurgencyin the PhilippineWar,1899-1902. Chapel Hill:
Universityof North CarolinaPress.
Lipman,JonathanN. 1990. Ethnic violence in modern
China: Hans and Huis in Gansu, 1781-1929. In Violence
eds.
in China:Essaysin Cultureand Counterculture,
Harrell.
and
N.
Stevan
Jonathan Lipman
Albany:SUNY
65-86.
Press,
Lipset, SeymourM., and Stein Rokkan. 1967. PartySystems
and VoterAlignments.New York:Free Press.
Loizos, Peter. 1988. Intercommunalkilling in Cyprus.Man
23, 639-53.
Loyd, Anthony. 2001. My WarGoneBy, I Miss It So. New
York:Penguin.
Lucas, Colin. 1983. Themes in southernviolence after 9
Thermidor.In Beyondthe Terror:Essaysin FrenchRegional
and SocialHistory,1794-1815, eds. Gwynne Lewis and
Colin Lucas.Cambridge:CambridgeUniversityPress,
152-94.
Machiavelli,Niccolo. 1988. FlorentineHistories,trans. Laura
F. Banfieldand Harvey C. Mansfield,Jr. Princeton:
PrincetonUniversityPress.
Maier, Karl. 1995. A fragilepeace. AfricaReport40, 22-7.
Manrique,Nelson. 1998. The war for the centralSierra.In
Shiningand OtherPaths:Warand Societyin Peru,
1980-1995, ed. Steve J. Stern. Durham:Duke University
Press, 193-223.
490
September 2003 I Vol. 1/No. 3
Marks,Robert. 1984. RuralRevolutionin South China:Peasants and the Makingof Historyin Haifeng County,
1570-1930. Madison:Universityof Wisconsin Press.
Martin,Jean-Clement.1994. Rivoluzionefrancesee guerra
civile. In Guerrefratricide:Le guerrecivili in eta contemporanea,ed. GabrieleRanzato.Torino:Bollati Boringhieri,
27-85.
. 1995. Guerrecivile et modernite:Le cas de la revolution. In La guerrecivile entrehistoireet memoire,ed.
Jean-ClementMartin. Nantes: Ouest lditions, 57-64.
. 1998. Contre-revolution,
revolutionet nation en
France,1789-1799. Paris:Editions du Seuil.
. 2002. Dans la guerrecivile, tout est permis.
LHistoire267, 56-9.
Mazower,Mark. 1993. InsideHitlers Greece:The Experience
of the Occupation,1941-1944. New Haven:Yale University Press.
McCoy,AlfredW. 1980. "Politicsby other means":World
War II in the WesternVisayas,Philippines.In Southeast
Asia underJapaneseOccupation,ed. AlfredW. McCoy. New
Haven:YaleUniversitySoutheastAsia Studies, 191-245.
McCrady,Edward.1969. The Historyof South Carolinain
the Revolution,1780-1783. New York:Paladin.
McKenna,Thomas M. 1998. MuslimRulersand Rebels:
EverydayPoliticsand ArmedSeparatismin the Southern
Philippines.Berkeley:Universityof CaliforniaPress.
Mitter, Rana. 2000. TheManchurianMyth:Nationalism,Rein ModernChina. Berkeley:
sistance,and Collaboration
Universityof CaliforniaPress.
Moore, Jeanne. 1999. World briefing:Americas.Colombia:
ruralmassacre.The New YorkTimes,19 August, A10.
Mueller,John. 2000. The banalityof "ethnicwar."International Security25:1, 42-70.
Mydans, Seth. 1999. East Timor family'sterror:Trappedat
home by gunmen. The New YorkTimes,27 September,
Al, A6.
Nelson, Hank. 1980. TaimBilongPait The impact of the
Second WorldWar on PapuaNew Guinea. In Southeast
Asia underJapaneseOccupation,ed. AlfredW. McCoy. New
Haven:YaleUniversitySoutheastAsia Studies,246-66.
O'Leary,Brendan,and John McGarry.1993. The Politicsof
NorthernIreland.London:
Antagonism:Understanding
Athlone Press.
Outram, Quentin. 1997. "It'sterminaleither way":An analysis of armed conflict in Liberia,1989-1996. Reviewof
AfricanPoliticalEconomy24:73, 355-71.
Paul, BenjaminD., and William J. Demarest. 1988. The operation of a death squad in San Pedrola Laguna.In Harvest of Violence:TheMaya Indiansand the Guatemalan
Crisis,ed. Robert M. Carmack.Norman: Universityof
OklahomaPress, 119-54.
Payne, StanleyG. 1987. The FrancoRegime,1936-1975.
Madison:Universityof Wisconsin Press.
Pecaut, Daniel. 1996. Rfflexionssur la violence en Colombie. In De la violence,ed. FrancoiseHeritier.Paris:Odile
Jacob, 225-71.
Perkins,Lucian. 1999. Building trust in a Kosovo village. InternationalHeraldTribune,28 December.
in North
Perry,ElizabethJ. 1980. Rebelsand Revolutionaries
China, 1845-1945. Stanford:StanfordUniversityPress.
. 1984. Collectiveviolence in China, 1880-1980.
Theoryand Society13:3, 427-54.
Pervanic,Kemal. 1999. The KillingDays:My Journeythrough
the BosnianWar.London: Blake Publishing.
Peterson,Scott. 2000. Me againstMy Brother.At Warin
Somalia,Sudan, and Rwanda:A JournalistReportsfrom the
Battlefieldsof Africa.New York:Routledge.
Pettigrew,Joyce. 2000. Parentsand their childrenin situations of terror:Disappearancesand specialpolice activity
in Punjab.In Death Squad:TheAnthropology
of State
Terror,ed. JeffreyA. Sluka. Philadelphia:Universityof
PennsylvaniaPress,204-25.
Piattoni, Simona, ed. 2001. Clientelism,Interests,and DemocraticRepresentation:
The EuropeanExperiencein Historical
and Comparative
Perspective.
Cambridge:CambridgeUniPress.
versity
Pigou, Piers. 2001. The apartheidstate and violence:What
has the Truth and ReconciliationCommissionfound?
Politikon28:2, 207-33.
Posen, Barry.1993. The securitydilemma and ethnic conflict. Survival35:1, 27-47.
and InternalWar.CamPrice,JonathanJ. 2001. Thucydides
bridge:CambridgeUniversityPress.
Prunier,Gerard.1995. The RwandanCrisis:Historyof a
Genocide.New York:Columbia UniversityPress.
Ranzato,Gabriele.1994. Un evento antico e un nuovo
oggetto di riflessione.In Guerrefratricide:Le guerrecivili
in eta contemporanea,
ed. GabrieleRanzato.Torino:Bollati
Boringhieri.
Richards,Paul. 1996. Fightingfor the Rain Forest:War,Youth,
andResources
in SierraLeone.Portsmouth,N.H.: Heinemann.
Ricoeur,Paul. 1984. Timeand Narrative,trans. Kathleen
McLaughlinand David Pellauer.Chicago:Universityof
Chicago Press.
Roldan, Mary.2002. La Violenciain Antioquia,Colombia,
1946-1953. Durham:Duke UniversityPress.
Romero, Mauricio.2000. Changing identitiesand contested
settings:Regionalelites and the paramilitariesin Colombia. InternationalJournalof Politics,Culture,and Society
14:1, 51-69.
Rosenberg,Tina. 1991. Childrenof Cain: Violenceand the
Violentin Latin America.New York:Penguin.
Roy, Olivier. 1999. Etat et recompositionsidentitaires:
L'exempledu Tadjikistan.In Guerresciviles:Economiesde
la violence,dimensionsde la civilite,ed. Jean Hannoyer.
Paris:Karthala,221-34.
Salibi, KamalS. 1988. A Houseof Many Mansions:The HisLondon: Centre for Lebanese
toryof LebanonReconsidered.
Studies and I. B. Tauris.
Sambanis,Nicholas. 2002. A reviewof recent advancesand
future directionsin the literatureon civil war. Defenseand
PeaceEconomics13:2, 215-43.
Schmitt, Carl. 1976. The Conceptof the Political,trans.
George Schwab.New Brunswick,N.J.: RutgersUniversity
Press.
. 1992. La notion du politique: Theoriedu partisan.
Paris:Flammarion.
Schoppa, R. Keith. 2001. Patternsand dynamicsof elite collaborationin occupied ShaoxingCounty. In ChineseCollaborationwith Japan, 1932-1945: The Limitsof Accommodation,eds. David P. Barrettand LarryN. Shyu. Stanford:
StanfordUniversityPress, 156-79.
Schroeder,MichaelJ. 1996. Horse thieves to rebelsto dogs:
Politicalgang violence and the state in the western
Segovias,Nicaragua,in the time of Sandino, 1926-1934.
Journalof Latin AmericanStudies28:2, 383-434.
. 2000. "To induce a sense of terror":Caudillo politics and politicalviolence in northernNicaragua,
1926-1934 and 1981-1995. In Death Squadsin Global
Murderwith Deniability,eds. Bruce B.
Perspective:
and
ArthurD. Brenner.New York:St. Martin's
Campbell
27-56.
Press,
Seybolt, PeterJ. 2001. The war within a war:A case study
of a county on the North China Plain. In ChineseCollaborationwithJapan, 1932-1945: The Limitsof Accommodation, eds. David P. Barrettand LarryN. Shyu. Stanford:
StanfordUniversityPress,201-25.
Shy, John. 1976. A PeopleNumerousand Armed:Reflections
on the MilitaryStrugglefor AmericanIndependence.
New
York:Oxford UniversityPress.
Skinner,Jonathan.1995. La guerrecivile revolutionnaire:
Oubli ou heritage?L'exemplede la pressevauclusiennede
la IIe Republique.In La guerrecivile entrehistoireet memoire,ed. Jean-ClementMartin. Nantes: Ouest Editions,
143-53.
Spencer,Jonathan. 1990. A Sinhala Villagein a Timeof
Trouble:Politicsand Changein RuralSri Lanka.Delhi:
Oxford UniversityPress.
Starn, Orin. 1998. Villagersat arms:War and counterrevolution in the central-southAndes. In Shiningand Other
Paths:Warand Societyin Peru, 1980-1995, ed. SteveJ.
Stern. Durham:Duke UniversityPress,224-57.
Stoll, David. 1993. BetweenTwoArmies:In the Ixil Townsof
Guatemala.New York:Columbia UniversityPress.
Swedenburg,Ted. 1995. Memoriesof Revolt:The 1936-1939
Rebellionand the PalestinianNationalPast.Minneapolis:
Universityof MinnesotaPress.
Tambiah,StanleyJ. 1996. LevelingCrowds:Ethnonationalist
Conflictand CollectiveViolencein SouthAsia. Berkeley:
Universityof CaliforniaPress.
Thaxton, Ralph. 1997. Salt of the Earth:The PoliticalOrigins
of PeasantProtestand CommunistRevolutionin China.
Berkeley:Universityof CaliforniaPress.
War,trans.
Thucydides. 1972. Historyof the Peloponnesian
Rex Warner.London: Penguin.
Tilly, Charles. 1992. Coercion,Capitaland EuropeanStates
(Studiesin SocialDiscontinuity).Cambridge:Blackwell
Publishers.
www.apsanet.org 491
Articles | The Ontologyof "Political
Violence"
in America,ed. J. P.
Tocqueville,Alexisde. 1969. Democracy
trans.
New
Lawrence.
York:Harperand Row.
Mayer,
George
Tone, John Lawrence.1994. The Fatal Knot: The Guerrilla
Warin Navarreand the Defeat of Napoleonin Spain.
Chapel Hill: Universityof North CarolinaPress.
Toolis, Kevin. 1997. RebelHearts:Journeyswithin the IRAs
Soul. New York:St. Martin'sGriffin.
U.S. forces fire at second Afghan wedding. 2002. The Financial Times,25-6 May, 3.
U.S. is set to help the Afghan opposition, Rumsfeldsays.
2001. TheNew YorkTimes,19 October, B2.
VargasLlosa, Mario. 1998. Un barbarechez les civilises.Paris:
Gallimard.
Varshney,Ashutosh. 2001. Ethnic conflict and civil society:
India and beyond. WorldPolitics53:3, 362-98.
. 2003. Nationalism,ethnic conflict, and rationality.
on Politics1:1, 85-99.
Perspectives
Vick, Karl. 2002. In KurdishTurkey,a new enemy: Village
guards,empoweredduring war, turn guns on returnees.
The Washington
Post,31 October,A18.
2002.
A nation challenged:The commander;
Waldman,Amy.
link him to U.S. attacks. The New
rivals
warlord's
Afghan
YorkTimes,3 January,A15.
Warren,Kay B. 1998. IndigenousMovementsand TheirCritics: Pan-MayaActivismin Guatemala.Princeton:Princeton
UniversityPress.
Watanabe,John M. 1992. Maya Saintsand Souls in a
ChangingWorld.Austin: Universityof Texas Press.
Werth, Nicolas. 1998. Un etat contre son peuple:Violences,
repressions,terreursen Union Sovietique.In Le livre noir
eds. Stephane
du communisme:Crimes,terreur,repression,
Courtois et al. Paris:Robert Laffont.
White, Lynn T. III. 1989. Policiesof Chaos:The Organizational Causesof Violencein Chinas CulturalRevolution.
Princeton:PrincetonUniversityPress.
Wiesner,Louis A. 1988. Victimsand Survivors:DisplacedPersons and OtherWarVictimsin Viet-Nam,1954-1975. New
York:GreenwoodPublishingGroup.
Wou, Odoric Y. K. 1994. Mobilizingthe Masses:Building
Revolutionin Henan. Stanford:StanfordUniversityPress.
An
Yoon, Taek-lim. 1992. Koreans'Storiesabout Themselves:
Historyof HermitPond Villagein South
Ethnographic
Korea.Ph.D. diss., Universityof Minnesota.
Zimmerman,Matilde. 2000. Sandinista.Durham, N.C.:
Duke UniversityPress.
Zulaika,Joseba. 1988. BasqueViolence:Metaphorand Sacrament. Reno: Universityof Nevada Press.
Zur, Judith N. 1998. ViolentMemories:Mayan WarWidows
in Guatemala.Boulder,Colo.: WestviewPress.
Notes
1
2
3
4
5
492
Gunfight 2002.
Collier and Hoeffler 2002; Berdaland Malone 2000.
Price 2001, 29.
Mueller2000; Posen 1993.
Kaldor 1999; Keen 1998.
September 2003 I Vol. 1/No. 3
6 Schmitt 1976.
7 De Lupis 1987.
8 Ranzato 1994; Bobbio 1992; Payne 1987.
9 Varshney2003; Horowitz 1985.
10 Brubakerand Laitin 1998.
11 McCrady 1969, 139.
12 Fellman 1989, 85.
13 Wou 1994.
14 Hobsbawm2001, 18.
15 Mitter 2000; Chang 1992.
16 Roldan 2002, 132, 276.
17 Ibid., 230; Crainz 1995; Martin 1995.
18 For example,under the veneer of religion, the Algerian
civil war was reallyabout "secularand political"issues.
See Freeman1994, 14. The clashesamong Dayaks,
Malays,and Maduresein West Kalimantan,Indonesia,
were not about religion either;see Davidson (forthcoming). A recent popularargumentis that many civil wars
are about little more than looting-see Kaldor 1999,
Enzensberger1994, and (for a critique)Kalyvas2001.
19 Brass 1997.
20 McCoy 1980.
21 Lear 1961, 234.
22 Cobb 1972, 123.
23 Stoll 1993, 259.
24 Waldman2002, A15, and U.S. is set to help 2001, B2.
25 Howell 1997, 315.
26 Quoted in Shy 1976, 206.
27 Crow 1985, 162.
28 Fellman 1989, 90.
29 Gould 1995.
30 Kedward1993.
31 Schroeder1996, 424, 431.
32 Berman 1996, 65.
33 Freeman1979, 164.
34 Seybolt 2001, 202.
35 Roldan 2002, 251, 212.
36 Cribb 1990.
37 Dalrymple 1997, 253.
38 Zur 1998, 114.
39 VargasLlosa 1998.
40 Ellis 1999, 128-9.
41 Peterson2000.
42 Lacey2003, A4.
43 Stoll 1993; Gould 1995.
44 Dean 2000; Fawaz1994.
45 Richards1996; Hamoumou 1993; Gross 1988.
46 For similarevidence,see Chung Kunsik(in Yoon 2002);
Johnson 2001; Schoppa2001; Cahen 2000; Bax 2000;
Pettigrew2000; Romero2000; Schroeder2000; Bazenguissa-Ganga1999b; Hart 1997 and 1998; Horton
1998; McKenna 1998; Starn 1998; Besteman1996;
Figes 1996; Tambiah1996; Berlow 1998; Brovkin1994;
Stoll 1993; Kriger1992; Lipman 1990; Groth 1995;
Linn 1989; Jones 1989; White 1989; Collier 1987; Perry
1980 and 1984; Calder 1984; Hinton 1984; Marks
1984; Cabarruis1983; McCoy 1980; Fiennes 1975.
Harding 1984, 59.
Lucas 1983.
E.g., Roldan 2002; Dean 2000; Duyvesteyn2000.
E.g., Varshney2001; O'Learyand McGarry1993.
Roldan 2002; Beevor2001; Loyd 2001; Hoare 2001;
Dale 1997; Pecaut 1996; Fawaz1994; Schmitt 1992.
52 Johnson 2001.
53 LedesmaVera 2001, 258.
54 Spencer 1990, 12, 80, 184.
55 The EthnolinguisticFractionalization(ELF) Index obviously does not capturelocal cleavages.
56 Kalyvas2003.
57 E.g., Martin 1994 and 2002; Ranzato 1994.
58 It is possible to think of a person'senvy as an individual
manifestationof class struggle(e.g., Harding 1984),
or-the other way around-of a person'sparticipationin
abstractclass struggleas an individualalibi for the expressionof his or her subjectiveindividualenvy. Cribb
1990, 28, makes a somewhatsimilarclaim about the violence that took place in Indonesiain 1965-1966,
when he arguesthat killings motivatedby private
grudgesare political since they take place in a charged
atmospherewhere "verylittle was non-politicalin one
sense or other, and grudgesfell into that broaderpattern
of social polarization."Still, it is both valuableand possible to analyticallydisentanglethe two.
59 The governorof the provinceof Khost, in southern
Afghanistan,"saidhe was convinced that much of reported al-Qaedaactivitywas, in fact, tribal problems.
One tribe will try to eliminate its rivalsby calling them
al-Qaedaand getting the coalition to bomb them." U.S.
forces 2002, 3.
60 Parisianrevolutionariesfailed to graspthe complex dynamics of a civil war that eruptedin the FrenchSouth,
in 1790-1791, between the towns of Avignon and
Carpentras;this was a clash less about ideas and programs than about settling local and personalaccounts.
Yet Robespierreframedthe conflict along the lines of
the national cleavage.See Martin 1998; Skinner 1995.
61 Swedenburg1995, 21; Kedward1993, 160.
62 Fisher 1997, 143.
63 Henderson 1985.
64 Everitt 1997, 24.
65 Lewis 1978.
66 Hart 1998, 265-6.
67 Roldan 2002, 243.
68 Chivers2003.
69 Vick 2002.
70 Degregori 1998, 135.
71 Manrique 1998, 994, 7.
72 McCoy 1980.
73 Cabarrus1983, 189.
74 Fearonand Laitin 1996.
75 Kalyvas2003.
76 E.g., Piattoni 2001.
47
48
49
50
51
77 E.g., Gould 1995.
78 E.g., Aschenbrenner1987.
79 Fellman 1989.
80 Sambanis2002.
81 Fearonand Laitin 2003; Collier and Hoeffler 2002.
82 Lipset and Rokkan 1967.
83 Stoll 1993, 68, 76.
84 Kelly 2000; Loizos 1988.
85 Accordingto Toolis 1997, 81-2.
86 Escott and Crow 1986.
87 Zimmerman2000, 97.
88 Bax 2000.
89 Toolis 1997, 35.
90 E.g., Toolis 1997; Dale 1997.
91 De Foxa 1993.
92 Pervanic1999, 120; 156-7.
93 Perkins1999. Similarexamplescan be found in Rwanda.
Des Forges1999, 15, reportsa case where a Hutu family
was killed after being denouncedas being Tutsi by
neighbors"who coveted their wealth."Prunier1995,
184, 203, reportsthat Hutu militiamenused their power
in orderto "settleprivatequarrels";
"old privateaccounts
were settled in blood."After the genocide, Prunier1995,
358, points out, innocent Hutu villagers"weretargeted
by jealousneighbourswanting their property."
94 Gross 1988, 42.
95 E.g., Roldan 2002.
96 Warren1998, 93.
97 Du Boulay 1974, 237.
98 E.g., Thaxton 1997; Wou 1994; Henriksen 1983.
99 Kakar1996.
100 Quoted in Toolis 1997, 42.
101 Moore 1999, A10. Name lists are common in civil wars.
They have been used, amongother places,duringthe
Guerrillawar in Navarre(Tone 1994), the AmericanCivil
War (Ash 1995; Fellman1989), the RussianCivil War
(Werth1998), the SpanishCivil War (LedesmaVera2001),
Malaya(Kheng1980), Italy (Fenoglio1973), the Colombian Violencia(Roldan2002), Algeria(Faivre1994), Vietnam (Herrington1997; Wiesner1988), Angola(Maier
1995), Liberia(Outram1997; Ellis 1995), Guatemala
(Carmack1988; Stoll 1993; Pauland Demarest1988),
Punjab(Gossman2000), the Philippines(Berlow1998),
Bosnia(Pervanic1999), Colombia(Rosenberg1991; Arnson and Kirk 1993), SierraLeone (Richards1996), CongoBrazzaville
1999a). Rumorsthat name
(Bazenguissa-Ganga
lists have been compiledare alsoprevalent(Kaufman2001).
102 Quoted in Nelson 1980, 253.
103 Ash 1995, 127.
104 Brovkin 1994, 226.
105 Roldan 2002.
106 Essien 1987, 116.
107 Quoted in Carmack1988, 54; Annis 1988.
108 Bearak1999, A1.
109 Zulaika 1988, 21.
110 Harding 1984, 75.
www.apsanet.org 493
Articles I The Ontologyof "Political
Violence"
111 Thucydides 1972.
112 "Andmany citizens, to avenge privateinjuries,led them
to the houses of their enemies;for it was enough that a
single voice shout out in the midst of the multitude, 'to
so-and-so'shouse,' or that he who held the standardin
his hands turn towardit." Machiavelli1988, book 3,
paragraph15.
113 Tocqueville1969, 17.
114 Warren1998, 98.
115 Du Boulay 1974, 237.
116 Thucydides 1972, book 6, paragraph54-9.
117 Fitzpatrick1994.
118 Finley 1994, 73.
119 Fitzpatrickand Gellately 1997, 11.
120 Kerkvliet1977, 66-7.
121 Cabarruis1983, 189.
122 Stoll 1993, 116.
123 Pettigrew2000, 210-1.
124 Thaxton 1997, 275.
125 Gage 1984, 19.
126 Howell 1997, 309.
127 Bernand 1999.
128 Loizos 1988, 648.
129 Paul and Demarest 1988, 153. Roldan 2002, 286, writes
about Colombia that "in many instances,mid-twentiethcentury violence was not the spontaneousresult of inherent local partisanconflict but was ratherconsciously
spearheadedby selectivesectors of the regionalstate or
tacitly encouragedby local bosses to advanceinterests
that had little or nothing to do with ideologicaldifferences."The South AfricanTruth and Reconciliation
Commission made a similarpoint when it arguedthat
the apartheidstate pursueda policy "to manipulatesocial, ethnic and other divisionswith the intention of
mobilising one group againstanother"(quoted in Pigou
2001, 226). In Sri Lanka,Spencer 1990, 184, observes,
"if politics provide a necessarymedium for the working
out of local disputes and grievances,they do so by appeal to forces and powersoutside the local community."
130 LedesmaVera2001.
131 Warren1998. See also Seybolt 2001, 202. "Justas the
Japanesewere using Chinese to pursue their imperialist
interestsduring the war, many Chinese were using the
Japaneseto pursue their domestic interests."
132 Hua and Thireau 1996.
133 Watanabe1992, ix-x, found that in the small
Guatemalantown he studied (SantiagoChimaltenango),
personaland local disputesand animositiesabounded
but failed to produceviolence: "Evenduring the worst
months of the Guatemalanarmy'scounterinsurgency
campaignin 1982-1983, the town refusedto succumb
to the self-servingrecriminations,power-mongering,and
murderthat infected all its neighbors."
134 Howell 1997, 324.
494
September 2003 j Vol. 1/No. 3
135 Aschenbrenner1987, 116.
136 Fitzpatrick1994, 255.
137 Quoted in Mazower1993, xv.
138 Cobb 1972, 56, 90.
139 Bazenguissa-Ganga
1999a, 48-9.
140 Horton 1998, 290.
141 Mydans 1999, A6.
142 Fisher 1997.
143 Gross 2001, 4.
144 Gross 1988, 117-20.
145 Chang 1992, 134, 173.
146 Bartolini2000; Kalyvas1996.
147 Horowitz 1985; Lipset and Rokkan 1967.
148 Posen 1993.
149 Chwe 2001; Hardin 1995.
150 A stylizedexample:supposethat villagex is composedof
two factions,a and b. The rebels(usuallythe first movers)
show up (usuallyvia local brokers)and mobilizea; this
factionthen extractsresourcesfrom b, by relyingon rebel
might. Lateron, the armyshows up and chasesthe rebels;
b joins the armyand denouncesthe leadersof a.
151 The Vietnameseand Chinese Communistsconstitute a
clear examplein this respect(Elliott 2003; Hua and
Thireau 1996).
152 Local factions enforce internaldisciplinethrough norms
and effectivein-group policing.
153 Clastres1999.
154 Ricoeur 1984.
faction"included re155 Roy 1999. The "Islamo-democratic
and
ethnic
professional,
groups such as the
gional,
Gharmi (from the Karateginarea),the Pamiris(from the
Gorno-Badakhshan
area),and intellectualsfrom the
the "conservativefaction"was
whereas
area,
Pendjikent
from the Leninabadarea,
of
Leninabadis
composed
Koulabisfrom Koulab,Hissarisfrom Hissar,and ethnic
Ouzbeks. Salibi 1988 providesa similaranalysisof the
LebaneseCivil War.
156 Kalyvas2001.
157 Thucydides 1972, book 3, paragraph82.
158 Geffray 1990; Richards1996, 8.
159 Quoted in Ash 1995, 128.
160 Kalyvas2003.
161 Despite the nonethnicmotivesbehind many acts of violence (includingpervasiverobberyand the takeoverof
neighbors'apartments),ethnicitybecame"theprimary
categorywith which people on the ground narrateand
comprehendthe war'sviolence"(Dale 1997, 91). Religion
has been used opportunisticallyin Sudan,as a means of
justifyingactionsor assigningblame (Dean 2000); as
such, observesPeterson2000, 174-5, it "maybe window
dressing-a means of mobilizingtroops and cash for both
sides,"but the war has causeda deepeningof religious
sense "forthe populationshammeredby this conflict."
162 Tilly 1992.