1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Sonderdruck aus
Bernadette Descharmes / Eric Anton Heuser /Caroline Krüger / Thomas Loy (eds.)
Varieties of friendship
Interdisciplinary perspectiveson social relationships
V&R unipress
ISBN 978-3-89971-787-7
ISBN 978-3-86234-108-5 (E-Book)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Contents
Ronald G. AschPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Bernadette Descharmes / Eric A. Heuser / Caroline Krüger /Thomas LoyPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Heather DevereIntroduction. The Resurrection Of Political Friendship: MakingConnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Semantics and Conceptions
Julian P. HaseldineFriends or amici? Amicitia and monastic letter-writing in the twelfthcentury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Christian Kühner“Quand je retournai, je trouvai toutes les cabales de la cour chang�es”:Friendship under the Conditions of Seventeenth-Century Court Society . 59
Tanja ZeebMoralist Concepts of Friendship: An Interplay of Stability andDynamism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Simon Meier„Nur im Gespräch können Freunde einander finden“ – ZurReaktualisierung des Freundschaftsdiskurses in der modernenKonversationstheorie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Danny KaplanChemistry and Alchemy : Narrative Building Blocks of Friendship andNationalism in Israeli Culture . . . . . . . . . . . . . . . . . . . . . . . . 119
Practices of Friendship
Agnes Brandt / Eric Anton HeuserFriendship and socio-cultural context. Experiences from New Zealandand Java / Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Sebastian Kühn“We have not fail’d to remember you on all occasions & to drinkconstantly your health” – Drinking Rituals and the Social Model ofTriads in Early Modern Scholarly Friendship . . . . . . . . . . . . . . . . 175
Laura Polexe„In alter Freundschaft, Dein Baron“ – Die Freundschaft zwischen denSozialdemokraten Karl Kautsky und Pavel Aksel’rod . . . . . . . . . . . 191
Vincent LeuschnerPlaying the field – dealing with interpersonal relationships in thepolitical arena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Judith GurrPolitics, Emotions, and Rationality. The Thatcher-Reagan and theBlair-Brown Relationship . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Patronage and Corruption
Iskra Gencheva-MikamiSelf at Play : Friendship and Patronage under Female Rulers . . . . . . . 239
Ronald G. AschFreundschaft und Patronage zwischen alteuropäischer Tradition undModerne: Frühneuzeitliche Fragestellungen und Befunde . . . . . . . . . 265
Niels GrüneFreundschaft, Privatheit und Korruption. Zur Disqualifizierung sozialerNähe im Kräftefeld frühmoderner Staatlichkeit . . . . . . . . . . . . . . . 287
Contents6
Dietmar NeutatzStalin als Freund? Überlegungen zu den persönlichen Nahbeziehungenim Politbüro der 1930er und 1940er Jahre . . . . . . . . . . . . . . . . . 309
Limits and Transgressions of Friendship
Ana Claudia MarquesIntrigas and Quest�es. Blood Revenge and Social Network inPernambuco, Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
Tilo GrätzSocial-Anthropological Perspectives on Friendship in Africa . . . . . . . 355
Bernadette Descharmesphilia und prodosia – Grenzen der Freundschaft in den Texten derattischen Tragödie oder Ein Versuch über den Verrat . . . . . . . . . . . 377
Contents 7
Limits and Transgressions of Friendship
Ana Claudia Marques
Intrigas and Questões. Blood Revenge and Social Network inPernambuco, Brazil
This paper focuses on brigas de fam�lia (lit. family fights) in the sert¼o (back-lands) of the Brazilian state of Pernambuco, commonly known as blood feuds. Byavoiding a description of a vengeance system as a whole and describing how thesestruggles are experienced by their participants and in the communities in whichthey take place, this paper aims to show that, while kinship relations are veryimportant in these processes, other kinds of relationships, such as friendship andpersonal loyalty, also play an important and necessary role.
I first distinguish the native categories of quest¼o (lit. issue) and intriga (lit.intrigue) in order to underscore two moments or fundamental states of rela-tionships, aggression and adjustment respectively, between the sides that takepart in them. Subsequently, I analyse the different local meanings of family andfriendship. This analysis reveals that these conflicts follow unpredictable courses.Particular relationships made, recognized, and presumed by those involved arethe only nexus among different episodes of the fights.
The elucidation of these native concepts interferes with the approach to feudingprocesses proposed in this article. It also problematizes the adoption of purely ego-centred or socio-centred perspectives for the analysis of pertinent social units ofprivate and public lives in the social field here in focus.
One of the persistent and fundamental tensions in anthropology, the in-dividual/society opposition, unfolds from different perspectives in methodologies,concepts, and theories. Every easy attempt to achieve a synthesis between thesetwo determining poles, if not of social life, of its conceptions and descriptions,seems destined to failure.1 Friendship, godparenthood, patronage, and clientshiphave been analysed by the anthropological literature on “complex societies,”particularly from the 1960s on, as forms of dyadic relations or coalitions centredon individuals. These forms of relations or coalitions are seen as filling the gaps
1 Eduardo Viveiros De Castro : Society. In: Encyclopedia of Social and Culture of An-thropology. Alan Barnard/ Jonathan Spencer. London, New York: Routledge, 1996, 514 –522.
left by formal institutions (considered to be less complete in terms of bureaucraticdevelopment), in social contexts to which they are thus added. The aim of thispaper is to discuss the limits of intelligibility imposed by purely sociocentric oregocentric analyses of vengeance relations I have observed in the sert¼o of con-temporary Pernambuco.
I will first put forth some general observations about blood feuds in the sert¼oof Pernambuco in the past and present time, and describe in detail a few episodesof one of the family quarrels I witnessed during my fieldwork. I will then discusstwo native categories, intriga and quest¼o, in order to describe the dynamicsparticular to local hostility relations and its cohesive and divisive effects on theentire social field, beyond the individuals or groups immediately involved. Sub-sequently, I will analyse the notion of family in its different local meanings. Theforms of alignment and opposition in vengeance relations will provide insightinto the regulation of municipal political life, also organized in accordance with afamily principle. Family, vengeance, and politics: for each of these subjects a shiftin focus will be proposed, one that moves from a systemic, functional, and to-talizing perspective to a more, say, transversal point of view, in which one observesthe partial and concrete realization of their functioning. This shift will afford us abetter glimpse of the importance of relations of friendship, patronage, and alli-ances as agents of collectivities, in this case formulated in terms of family.
Family Feuds
In the most inland regions of the Brazilian Northeast, the so-called sert¼o, acertain means of conflict regulation persists despite being in contradiction withwhat is provided for by current Brazilian legislation. In these regions bloodrevenge emerges as a possibility and an effective alternative response to anattempt against life personally suffered by the avenger or by a relative or a personclose to him. There are registers of “family feuds” that date back as far as thecolonial period and whose trajectories are contemporary with the settlement andfoundation of certain communities, producing long-lasting effects on the termsof their internal and external social and political relations.2 In the currentconfiguration of illegal acts, crimes of revenge are not uncommon in socialnetworks connected to drug trafficking and other forms of banditry, and signs ofthese forms of retaliation occasionally arise in the municipal political life of
2 Billy Jaynes Chandler : Os Feitosas e os Sert�es dos Inhamuns. Fortaleza: UniversidadeFederal do Cear�, Civilizażo Brasileira, 1980; Lu�s Aguiar da Costa Pinto : As Lutas deFam�lia no Brasil. (Era Colonial). In: Revista do Arquivo Municipal, Ano VIII (vol.LXXXVIII), 1949, 7 – 125; Henry Koster : Viagens ao Nordeste do Brasil. S¼o Paulo: Com-panhia Editora Nacional, 1942 [1817].
Ana Claudia Marques338
smaller communities. There are a number of studies on such objects which,however, have yet to systematically focus on the problem of revenge. The fol-lowing remarks regarding conflict and hostility relations between family groupsin the sert¼o cannot, therefore, be exposed in comparative perspective in theBrazilian context. The intention of the specific focus I take to conduct theanalysis of such conflicts, looking at how the actors themselves and theircommunities experience them instead of approaching them from the totalizingangle of a vengeance or feuding system, is to avoid separating the conflict fromthe set of relations that compose it and in which it takes part. At the core of thisproposal is the idea that personal relationships, friendship, and loyalty are su-perimposed on kinship, giving meanings to and defining courses for the proc-esses of blood feud in the region here explored.
In 1999 I started my fieldwork in Pernambuco, carrying the unconsciousassumption that “family feuds” consisted of two opposing family groups.Gradually, I came to realize the mistake in that implicit assumption. It was soonclear to me that there were not two groups or groups at all, if we attribute to suchnotion a discrete character. In my first days in the field, members of one of the“families” that had been taking part in a conflict of this sort for years revealedthat kinship ties connected them to both sides of the feud, towards which theyremained neutral. The more I gained insight into quarrels and family relations,two aspects of social life which shed light on each another, the more this in-formation became significant to me. Due to the distribution of the members of acertain “family,” the Santanas, my research has focused primarily on threemunicipalities in the countryside of Pernambuco – Monte Verde, Monsanto, andJord�nia – as well as the state capital – Recife – located on the coast.3 This spatialarrangement has proven to be a meaningful fact for my research.
The quest¼o between the Santanas and the Gouveias in Jord�nia began in1990, due to a quarrel between some young men over the alleged theft of a pair ofboots. Reports about this and many successive episodes disagree on a series offacts, including the real reason for discord, which, for many, involved the illegaltrade of cannabis. These disagreements reveal the way in which the actors of thequest¼o and their public position themselves in relation to each other and to theconflict as a whole. Nonetheless, only the information necessary for the in-telligibility of the chain of events will be considered here, and as briefly aspossible.
The original owner of the boots was the son of a woman of the Santana family
3 As part of my field interlocutors are under circumstances of illegality, and to the benefit of the(often fragile) social relations in force in these locations, I have chosen to use pseudonyms(given names and surnames) for people, cities, towns, settlements, and municipalities, but notfor the states and their capitals.
Blood Revenge and Social Network in Pernambuco, Brazil 339
who held a municipal public office and was linked to the dominant factionholding that family name. The young man accused of the theft was the adoptedson of a woman who lived in the lands belonging to a woman of the Gouveiafamily. The alleged recipient of the boots was a young policeman of the sameGouveia family. The accused, in reprisal for the public accusation against him,shot the young Santana dead in a bar. From this moment on the young Santana’sbrothers, supported by many of their cousins, adopted an attitude that suggestedretaliation and self-defence and transformed their mother’s house into a sort offortified headquarters. And thus it remained for several months. On a certainevening, rain and a serenade interrupted the daily routine in Jord�nia when oneof these cousins left his aunt’s house with the excuse of buying cigarettes and raninto a friend of his opponents. Worried about his cousin’s unexpected delay, asecond cousin, son of the town’s mayor, left in search of him and came across thetwo men in the midst of an argument. Things were stirred up by his questioning,and certainly also by the additional unbalance his arrival had caused. Previouslyconcealed guns were pulled and a shootout between the first two opponentsensued. Wounded, they were both taken to the municipal hospital. There, themayor’s son, who was accompanying them, attempted to force the prioritizationof his cousin’s transfer, in the only available ambulance, to the regional hospital– a better equipped centre located in another municipality some dozens ofkilometres away. The doctor stated that the definition of the procedures to betaken and health-care prioritization were at his discretion. This disagreementstirred unrest once again and other people present also stepped into the argu-ment. In the heat of the moment, the mayor’s son made a verbal and physicalassault upon one of these people, a lawyer who belonged to the Garcia “family,”also very prestigious and politically influential in some of the municipalities ofthe sert¼o. People in Jord�nia knew that his father, a man who was known to bebrabo (raging), would not forgive that insult. Rumours of vengeance reached thenewspapers in the state capital, Recife.
At this critical point, some effort was made to have the mayor and the lawyer’sfather reach an agreement. This was attempted by traditional means, in whichmen linked to both parties by kinship and friendship try to establish peacebetween them.4 The mayor, however, would not give in to the demand, certainlyhumiliating for someone in his position, to send his son away from town, in akind of customary exile. The widespread suspicions were confirmed in the mosttruculent manner. A couple of days after this meeting there was an attempt on the
4 The means of peacemaking in this region of the sert¼o, the usual modes of agreement, and theconditions for mediation have been analysed by Jorge Mattar Villlela : ViolÞncia e Mediażode VinganÅa de Sangue no Sert¼o de Pernambuco, Nordeste do Brasil. In: Conflitos, Pol�tica eRelaÅ�es Pessoais. Ana Claudia Marques (ed.). Fortaleza, Campinas: Universidade Federaldo Cear�/Funcap/CNPq-Pronex/Pontes Editores 2007.
Ana Claudia Marques340
life of the mayor and his son, in which the former was killed and the latterremained unharmed. The men inside the car that ambushed them were identi-fied, and from then on a series of theories began to emerge concerning who wasresponsible for organizing the homicide.
It is said that the crime was committed by hitmen, as often happens in suchcases, even when associated with revenge.5 However, this does not precludeassumptions about the identity of the responsible party, which will be madebased on the accumulated knowledge about the status of the relationships inwhich the victims are inserted and on the networks the killers themselves be-longed to. The first hitman was the paternal cousin of the mayor, who washimself politically affiliated with the Santanas faction by maternal kinship. Twoother hitmen belonged to the district of the deputy mayor, who at that momentwas pleading for the emancipation of his district. The fact that the mayor, andnot his son, was the privileged target aroused suspicions about whether politicalmotivations had – in a proportion never sufficiently distinguished and neverconfirmed – accompanied the personal and family motives that connected thisepisode to the original event of the narrative, the theft of the pair of boots. Themayor’s homicide has always been considered one of the incidents of the quest¼obetween the Santanas and the Gouveias, despite the insufficient clarity con-cerning its more specific motivations or authorship.
In the ten following years there were several sequels to this story, which I havedescribed elsewhere.6 These included the murder of members of the Santanas,the Gouveias, and other families. Among the murdered were another mayorfrom the Santana family, several cousins of the first Santana victim, and relativesof his enemies, as well as men whose relation of friendship or subordination tothese actors are inferred and disputed. In each episode responsibility is placedon actors who were not necessarily implicated in the former episodes. Themotivations for retaliation change depending on the specific relationships thevictims had, though the quest¼o in its totality is also a determinant for estab-lishing targets and defining avengers. Indeed, there is no linear relation of causeand effect among all of the episodes. None of the murdered mayors, for instance,
5 According to Wadley (Lethal Treachery and the Imbalance of Power in Warfare and Feuding.In: Journal of Anthropological Research 59 (4), 2003. 531 – 554), the use of concealed am-bushes or sneak attacks as a murder method is a recurring element in revenge systems. Theauthor also proposes a functional and intercultural analysis of this method and of betrayals ofpeacemaking pacts. Invoking third parties as murderers seems to be less widespread, possiblydue to more positive rules concerning duties of revenge. For the role of bandits in vendettaprocesses in Calabria, see Jan Brogger : Conflict Resolution and the Role of the Bandit inPeasant Society. In: Anthropological Quarterly 41, 1968, 229 – 241.
6 Ana Claudia Marques : Intrigas e Quest�es. VinganÅa de Fam�lia e Tramas Sociais no Sert¼ode Pernambuco. Rio de Janeiro: Relume Dumar�/ Ufrj, Nfflcleo de Antropologia da Pol�tica,2002.
Blood Revenge and Social Network in Pernambuco, Brazil 341
were directly involved in the revenge. Some of my informants consider that theirbecoming a target was related to the expectation or suspicion that their politicalor personal influence and alleged control over the performance of the judicialand police apparatus could lead to their own relatives being favoured in thequest¼o. Others claim that personal and political rivals took advantage of thequest¼o to disperse suspicions about their participation and their real motiva-tions. At the opposite end of the prestige chain, even hitmen invoke privatereasons for their participation in the murders, declaring themselves to have beeneither personal enemies of their victims or motivated by friendship. Theirsupport of the main participants of a quest¼o is justified with claims that thequest¼o is also theirs. The epithet hitmen is avoided by those who are consideredso, since they argue that they act out of friendship and not in exchange formaterial compensation.7
Family and Personal Relationships
In the sert¼o of Pernambuco, “family” represents a pertinent agent in variouscircumstances of social interaction, being stressed, for instance, in relations ofsociability, associations of productive life, alliances, and political followings.Even though there is no space here for an exhaustive elaboration of the problem,it is necessary to distinguish the scope and the circumstantial character of localnotions of “family.”
The term designates, for instance, individual households that are often alsoproductive units, especially in rural areas. In a much wider but never explicitlystated sense, however, “family” includes all cognate descendants of the same
7 In a society endowed with a central bureaucratic apparatus, those that commit a crime forrevenge unavoidably become outlaws. Even so, in the sert¼o, these actors might not be ex-cluded from the networks of community relations, especially if they are able to avoid gettinginvolved in more reprehensible illegal practices such as robberies and drug trafficking. Unableto move around freely in public places and to engage in formal productive activities, theirarmed power often becomes a socially legitimate, though not legal, way of life. Armed supportmay be and indeed is often used as a way to compensate for the protection offered by relativesand friends. Barreira discusses the contexts of other states in Northeastern Brazil and analysesthe moral ambiguity in which the material authors of contract killings, who do not forgoassertions of honour, get involved (see C�sar Barreira : Crimes por Encomenda. ViolÞncia ePistolagem no Cen�rio Brasileiro. Rio de Janeiro: Relume Dumar�/ Ufrj, Nfflcleo de An-tropologia da Pol�tica, 1998.). With regard to the terms of the relationship between outlawsand the rest of the community, many of Brogger’s (ibid.) insights about the role of bandits invendetta processes in mid-twentieth-century Calabria seem pertinent here. The BrazilianNortheastern sert¼o has also established a type of rural banditry, denominated cangaÅo, withits own characteristics but also many overlapping points with the revenge system (cf. Villela,O Povo em Armas).
Ana Claudia Marques342
ancestral couple, and corresponds to a genealogy. This sense is basically appliedto the families of pioneers who, about half a dozen generations ago, foundedtheir cattle raising farms in the sert¼o.8 In spite of the fragmentation caused bypartible inheritance and property transfer, its unity has not necessarily beenlost. Land cultivated by different family groups is aggregated and constitutes alocal unit, a neighbourhood (ribeira, s�tio, or fazenda in local terms). Accordingto the residential composition of these neighbourhoods, such localities are as-sociated with one or several family names. Reciprocally, genealogical sectionsare also singled out by their association with specific locations. The Santanas, forexample, never constitute a unit for whatever effect. However, the Santanas of theImbuzeiro farm or the Santanas of Jord�nia, designations on which there is areasonable degree of consensus, may function as units in different circum-stances. Under these conditions, they are also designated as a “family.” Localitytakes priority over consanguinity in this third sense.
Social relations, solidarity, and mutual help relations develop more intenselywithin neighbourhoods. However, neighbourhoods also create opportunities fordisagreements of various levels of importance. Though consanguinity tiesstrongly inform solidarity alignments with multiple purposes, they are not ex-clusive or sufficient here. The internal divisions of kinship, always (even locally)present, are no less relevant in the orientation of these alignments. Preferentialrelationships are established among neighbours, whether they are relatives ornot. Such relations may be horizontal or vertical, implying either a condition ofequality in status among actors or a relation of subordination or dependence ofsome sort. Even within a set of descendants of the same ancestor, one or morefamily members may, for various reasons, take on a position of leadership,centralizing decisions that will affect the lives of their relatives, friends, andneighbours.
Some generations ago, when parcels of land were transferred much morerarely than they are nowadays, these leaders intervened in decisions concerningthe acceptance or repudiation of new inhabitants who had arrived as new in-dependent land owners or as dwellers of lands that did not belong to them. Thus,the connections among the members of the population in these places are not,and have probably never been, exclusively blood-related. At the same time,however, these people have always been somehow subjected to the control ex-erted by locally dominant families. Yet these local units, though marked by a
8 Some of the descendants of these large families devote themselves to mapping family trees thatusually have the male ancestors as starting points. Their wives’ names are not always knownand the information regarding their side of the family is usually vaguer. Some information onthe genealogical continuity of possible Portuguese families holding the same surname may beincluded, though also vaguely. The most significant piece of information on the genealogicalformation of a “family” from the sert¼o is the name of the male ancestor, founder of the farm.
Blood Revenge and Social Network in Pernambuco, Brazil 343
family character, do not imply (and have probably never implied) an inner socialdifferentiation of the kind generally attributed to a family as a coherent unit froman outside viewpoint.
Two further meanings of “family,” one moral and the other politico-factional,would deserve a more careful description that exceeds the limits of this paper,and will therefore only be briefly outlined here. The known reputation of aleader, of a homem manso (gentle man) or a homem brabo (raging man), forinstance, tends to be extended to his domestic group or to the domestic groupsunder his influence. Thus, to blood (sangue) and place, certain moral qualitiesare usually added, a reputation or fame (fama) consolidated in the sphere of thesocial relations experienced by the actors who take part in these units internallyor externally. Under this moral qualification, the group also gains a certain socialunity, as opposed to other subgroups of relatives. The contexts of conflictualrelations resulting from municipal politics and family vengeance are particularlyrelevant in this sense. The most expressive leader of the Imbuzeiro farm, MajorClemente Santana, received the reputation of homem manso as he was not animpetuous man but someone who was prone to easing tensions arising bothinside and outside of his residential and political haven. The same “Santanas ofImbuzeiro” acknowledged the descendants of a certain sister of Major Clementeas their cousins. These, however, came to be distinguished, following the line oftheir paternal ancestors, as gente braba (raging people), keen on discord andimpetuous acts of retaliation which displeased many of those who were closer tothe major. This reputation, together with the status of inferior paternal origins,produced, during more than one generation, divisive effects on several spheresof sociability, though they were ineffective in some and positively pertinent inothers.
Much less circumscribed is the notion of “family” when it refers to a wholepolitical faction, to which a significant number of relatives – although never itstotality – and members of other families simultaneously belong. Despite thehomogeneity and cohesion that the attribution of a family name to a factionmight suggest, it should be highlighted that such alignments do not necessarilyfollow kinship ties. Major Clemente was also a municipal political leader whohad never run for an elective post but whose support was vital for his main allyMajor Jo¼o Gouveia, who, curiously enough, started to centre the opposition tothe faction in power, represented by the Santana family, on himself. Jo¼o Gouveiaand Clemente Santana were neighbouring farmers who shared common socialcontexts – they went to the same church, the same Sunday fair, and faced thesame local problems – and developed a friendship capable of rising above moresocially remote consanguinity ties that Clemente had with relatives who do-minated the other political faction to the extent that they trusted it with theirfamily name. In the municipality of Jord�nia, the two opposing factions of local
Ana Claudia Marques344
politics have been denominated by Santana and Gouveia since 1913, regardlessof the national political parties with which their candidates have been affiliatedover the years. Therefore, members of a family may back the faction opposingthe one carrying their surname (and they indeed do so).
My effort to summarize these meanings of family should not exclude anobservation related to these many senses – domestic, virtual-genealogic, local,moral, and political. In order to make their descriptions more straightforward, Idifferentiated them in terms of more or less inclusive spans. They should not,however, be seen as concentric. Even though the set of relatives does offer,theoretically, the most plausible options of alignment for any purpose, it is interms of concrete relationships that the allies (partners, partisans, and friends)are effectively recruited.
Although everyday intimacy provides many opportunities for differ-entiations within this environment, it usually seems homogeneous from a for-eigner’s perspective. This is due to the fact that from an outside viewpoint – andthe farther away the observer of the object is, the stronger such an impressionwill be – the families, in any of the described senses, are presented as homo-geneous and consistent units, and the inaccuracy of these characteristics is ofteneven reinforced by the family members when they interact with outsiders. Onedoes not have to suppose, though, that they are thus lying or concealing thetruth, for the inclusion and exclusion criteria are, as we have seen, variable.Nevertheless, one should not lose sight of the instrumental use of such varia-bility. At any rate, preferential relationships lead to divisions between the rela-tives which, although not pertinent to an outside interlocutor, are nonethelessvery real and are compounded further by status and moral distinctions.
At this point it is necessary to make some preliminary considerations on therelations of friendship and dependency in the sert¼o. In general, friendship is apersonal form of relation, chiefly recognized among men, who use the term torefer to a preferential affective tie, a disposition of mutual support for in-determinate purposes. Friendship is a form of bonding between non-relatives. Italso stresses pre-existing kinship ties (as does godparenthood, generated inreligious rites) and confers upon relationships a moral egalitarianism thateventually contradicts existing inequalities of class or prestige. Therefore, theterm is applied verbally whenever one intends to suggest and affirm both the tieand the horizontality of the relation between two agents.
Not less common in the sert¼o are the vertical relations of cooperation, whichimply personal dependency and subordination. These two forms, however, arenot irreconcilable in social rhetoric. In the sert¼o, when someone says that he isdoing something for someone else out of friendship, he intends to assert thesolidary, voluntary, and disinterested character of the act. It also shows the non-submission of the subject, who is in this context capable of such an act of
Blood Revenge and Social Network in Pernambuco, Brazil 345
liberality despite the undeniable social distance between the collaborator andthe beneficiary. If this means a moral credit on the part of the collaborator, healso holds himself responsible for the act, which would not be the case if he actedout of simple obedience and subjection.
The multifunctionality of family as an informal association that acts withinformal structures – which, as Wolf has noted, persists in “complex societies” –certainly applies to this social universe, although some qualifications are nec-essary.9 For openers, the postulated low cost and high efficiency are debatablewhen one is dealing with potentially lethal hostility relations such as those beingstudied here. The parallel, interstitial, and supplementary character of family (orpersonal) relationships has also already been challenged by a realization of theimportance of these relationships in the constitution of formal structures andtheir operability.10 As it seems, the exclusion of this type of agency from thesocial analysis is perhaps more related to the (institutional and totalizing)perspective one adopts, as well as to the philosophical and political convictionsof the social scientist, occasionally inserted in “simple,” “complex,” and “semi-complex” notions.11 In the sert¼o of Pernambuco, “family” is regarded as apertinent agent in various circumstances of social interaction, as previouslyseen. In none of them, however, do “families” take on the character of corporategroups. A series of divisions is introduced at the heart of kinship relations, andheterogeneous additions are similarly indispensible to the always precariousand ad hoc constitution of groups or functional units on any pretext.
For those who followed a certain sociological literature in the 1960s and 70s,the concepts of dyadic contract,12 quasi-groups,13 or dyadic non-corporategroups14 could be approximated here. The objections I consider necessary fortheir application help to express conceptions which, in my view, can be used tomore accurately describe family groups from the sert¼o. The most important ofthese is undoubtedly the fact that these concepts are forged with individuals in
9 Eric Wolf : Kinship, Friendship, and Patron-Client Relations in Complex Societies. In: TheSocial Anthropology of Complex Societies. Michael Banton (ed.). London: Tavistock Pu-blications, 1966, 1 – 22.
10 Carl H. Landé: Introduction: The Dyadic Basis of Clientelism. In: Friends, Followers, andFactions. Stefen W. Schmidt/ Laura Guasti/ Carl Landé/ James Scott (eds.). Berkeley, LosAngeles, London: University of California Press, 1977, XIII-XXXVII.
11 Nigel Rapport/ Joanna Overing : Society. In: Social and Cultural Anthropology. The KeyConcepts. London, New York: Routledge, 2000, 333 – 343.
12 George M. Foster : The Dyadic Contract: A Model for the Social Structure of a MexicanPeasant Village. In: American Anthropologist 63, 1961, 1173 – 1192.
13 Adrian Mayer : Quasi-Groups in the Study of Complex Societies. In: The Social An-thropology of Complex Societies. Michael Banton (ed.). London: Tavistock Publications,1966, 97 – 122.
14 Landé, Introduction.
Ana Claudia Marques346
mind, interconnected as dyads, either directly or indirectly. Although a bilateraldescent system does not structurally generate corporate groups, we can stillinquire about the possible constitution of non-ego-centred collectivities thatoccasionally apply to the notion of “family” in its multivocality.15 Entering into adialogue with authors such as Freeman,16 Firth,17 and Leach,18 among others, andin light of her own ethnography among the Vezo of Madagascar, Astuti19 dis-tinguishes, but at the same time also highlights, relations of continuity anddiscontinuity between cognatic descent, cognatic kinship, and unilineal descent.In the same way, we may consider that the definition of a collectivity as a cognaticgroup or ego-centred network is dependent upon the perspective of the analysis,which varies according to the actors, the different contexts and moments in thecycle of life in which they are inserted, and based on which the conception isconstituted.
Revenge
During my fieldwork, I collected information about various blood feuds, most ofwhich were already virtually closed, and some still going on. In either closed oropen cases, however, the families involved considered themselves to be quar-relling with each other. Intriga (quarrel) is a category that describes a state ofrelations which includes avoidance and estrangement behaviours. In its weakestsense, it describes only the situation of a personal quarrel of some sort. None-
15 Freeman designates kindred-based action groups as those whose members are “recruitedwholly or predominantly from an individual’s cognates and organized for a specific purpose(such as blood-vengeance or swearing of oaths of compurgation), which may, for a period,become solidary groups” (Freeman, On the Concept of Kindred, 203). Such groups wouldalso recruit non-consanguineous individuals, with affinity, friendship, and godparentshipties. The problem faced by the author remains related to the constitution of operationalgroups in bilateral descent systems, and its solution matters to the distinction between thecategorical character of cognatic descent groups, which under these conditions include allrelatives who share common ancestors either via matrilineal or patrilineal descent, and thepragmatism of the constitution of groups, which would be done based on the individuals.
16 J. D. Freeman : On the Concept of the Kindred. In: The Bobbs-Merrill Reprint Series in SocialSciences. The Journal of the Royal Anthropological Institute of Great Britain and Ireland 91,1961, 192 – 220.
17 Raymond Firth : Bilateral Descent Groups: an Operational Viewpoint. In: Studies inKinship and Marriage. Isaac Schapera (ed.). London: Royal Anthropological OccasionalPaper No. 16, 1963.
18 Edmund Leach : On Certain Unconsidered Aspects of Double Descent Systems. In: Man 214,1962, 130 – 134.
19 Rita Astuti : Kindred and Descent Groups: New Perspectives from Madagascar. In: Culturesof Relatedness: New Approaches to the Study of Kinship. Janet Carsten (ed.). Cambridge:Cambridge University Press, 2000, 90 – 103; Anthropological Quarterly 41, 1968, 229 – 240.
Blood Revenge and Social Network in Pernambuco, Brazil 347
theless, when a conflict unravels to the point of an imminent or effective bloodshedding, intriga concerns not only the individual, but the more strict nucleus ofsolidarity in which its participants are included – generally speaking, themembers of their “family.” The scope of meanings for family varies according tothe status of individuals and their families in the community, even though itcould be anticipated that the members of their household, their parents, andsiblings will be inevitably involved.
It is important to point out that intriga defines the terms of relations betweenthose who take part in the opposite sides of the conflict through a series ofavoidance and estrangement marks, as well as public performances of retaliationand confrontation threats. As these threats materialize (or convince the audienceof their intention of being materialized), a quest¼o is established. Thus, in bloodfeuds, intriga and quest¼o correspond to two moments of the same cycle. Theformer corresponds to a state of latency in hostility relations, while the latterconcerns the triggering of reciprocal aggressions against the lives of thoseconsidered to be implicated in the conflict. The limits between the two momentsof the cycle are not clearly discernable and can only be applied a posteriori.Nevertheless, this cyclical composition helps us understand the sense in whichthe condition of conflict becomes perennial in the relations among “families.”“Intriga never ends,” say the sert¼o people. Once a quest¼o de sangue (bloodfeud) is established, even if a state of peace is reached, members of the familiesinvolved will maintain for generations, even if subtly, reservations about the oldrivals of their relatives. They know or assume that if relatives, friends, or com-mon allies attempt an approach, any friction might again stir up the flame of theold quest¼o.
According to Boehm, “it is this idea of a well controlled but potentially in-terminable conflict that defines the nature of feuding”20 in relation to vendetta, towar, and to raids, and in opposition to the definitions by Peters21 and Black-Michaud.22 The eternity of feuding postulated by both definitions, would resultfrom the adoption of a native Bedouin formulation as a category of analysis, usedby the former and endorsed by the latter, which would prevent the under-standing of the containing and codifying function of homicidal violence insocieties marked by all those forms of conflict. For Boehm, peacemaking and,consequently, the end of feuding as a possibility, defines it, though it goes against
20 Christopher Boehm : Blood Revenge. The Enactment and Management of Conflict inMontenegro and Other Tribal Societies. Philadelphia: University of Pennsylvania Press, 1984,206.
21 Emrys Peters : Some Structural Aspects of Feud among the Camel-Herding Bedouin ofCyrenaica. In: Africa 37, 1967, 261 – 282.
22 Jacob Black-Michaud : Cohesive Force. Feud in the Mediterranean and the Middle East.New York: St. Martin’s Press, 1975.
Ana Claudia Marques348
the native meanings conferred on this form of hostility relation. As regards thesert¼o of Pernambuco in relation to the definitions suggested by these authors,one may point out the existence of vendettas and feuds in the region, though notwars. The social function of peacemaking, in the sense conferred by Boehm,would be jeopardized here. This small controversy is one of the reasons why Ibelieve extreme caution is needed when one applies necessarily generic conceptsto particular cases, as in the case of feuding. Combined, the categories of intrigaand quest¼o in the sert¼o of Pernambuco allow us to understand persistencemodes of social divides which are valid for a series of effects and which can onlybe discerned through ethnographical study. The perpetuity of an intriga shouldnot be taken, in my view, as an ideology, for it expresses a highly pertinentdivisive potential which includes – but also transcends – the individual quest�es,guiding means of establishing alliances and oppositions in the entire social field.Both of these native categories help us understand conflict relations not as anaspect of a somewhat stagnant domain of social life, but rather as an aspect thatconstitutes it.
In 1987, an incident cut short an old friendship that united two neighbouringfamilies living on the same street in the small town of Monte Verde.23 A youngman belonging to one of these families was clumsily guiding a half-tamed calfdown a slope when he suddenly slipped and fell, inciting laughter in a child whobelonged to his neighbours’ family. Irritated, the young man threatened to hit thechild, but an uncle, slightly inebriated according to witnesses, came to the child’sdefence. The disagreement between the two men attracted attention from otherfamily members and, in the midst of the unrest, whose details vary widelydepending on the narrator,24 the child’s grandfather – the progenitor of one ofthe neighbouring families – stepped into the scene. Coming from his errands inthe rural area, he held a knife, with which he mortally wounded his old friend, theprogenitor of the opposing family, who had also entered the fight. The incidentput an end to the cordial relations cultivated for so many years among themembers of both families and established an intriga between them. The mur-
23 Monte Verde is the administrative seat of a municipality located slightly over 350 km awayfrom the state capital, Recife. Its total population is close to 16,000 inhabitants, and ap-proximately one quarter of them live in the town. Though located in the sert¼o mesoregionand Pajeffl microregion, Monte Verde stands out from the other municipalities in the sert¼ofor its high-altitude tropical climate, which results from its mountainous location. Such mildclimate and rainfall regime has enabled an agricultural and horticultural exploration to takeplace, favouring the supply of local markets, in which organic livestock farming was the mostprominent productive activity. In conjunction with this form of exploration, a smallholdingregime prevails in this municipality, as opposed to the large properties that are typical inneighbouring pastoral areas.
24 Marques, Intrigas e Quest�es, 81 ff.
Blood Revenge and Social Network in Pernambuco, Brazil 349
derer went into hiding, and his married children, the child’s mother, the unclewho had gone to the child’s defence, and a third brother – my informant – had toleave their urban residences to live in the rural area, taking their families withthem. Even so, they were still wary of the tacit or announced threat of revenge,echoed by the other neighbours, and avoided being seen in public on the streetsof Monte Verde. The vengeance, a crime committed against one of the mur-derer’s brothers, took place eight days later. No other aggressive episodes hadtaken place since. The first murderer was eventually arrested. His son – who alsoparticipated in the fight – moved (retirou-se) to the South and his other two sonsgradually resumed their normal community life. The rival family did not think itnecessary to protect themselves as the other family had done. They certainlybelieved to be in an advantageous position, probably due to their status, but alsoto the greater number of males in their family as compared to that of the in-trigados (intrigued family). A few years later, the avenger was tried in freedomand found not guilty. The friendship between the old neighbours was neverrestored and no one expects it to be re-established in the future. Among oldfriends, mutual respect is expressed by deliberate demonstrations of estrange-ment and, at the same time, good intentions.
This case describes a rather simplified vengeance process, and yet it contains afew ambiguities. Though the retaliations ceased, their protagonists perpetuatedthe conflict as an intriga,25 which is applicable to and helps to define all of thosewho got involved in it: the direct descendants of the first victim and those of themurderer. Existing friendship relations did not prevent the fight or the homicidefrom happening. But they might have influenced the revenge. My informant statedthat he considers a son of the former neighbouring family as his own brother.Neither of these two men had participated in the quest¼o directly. This also sug-gests that the fact that he was friends with the people from the street (tinha muitaamizade na rua, meaning that these people were from the urban as opposed to therural area of the municipality), where he is acknowledged to be a cabra direito(righteous man), has guaranteed respect to his physical integrity despite his un-cle’s questionable reputation – as a man prone to drinking, women, and jokes –which could have victimized him. Such an old friendship also discursively covers anot so large but significant difference in status between the two conflicting fam-ilies. The target of the revenge, as well as the ceasing of the retaliations, seems tohave been decided in favour of the lowest social cost incurred.
When the process of revenge hits members of dominant “families” that al-ready oppose one another at other levels of social life, i. e. the process of selectingtargets and setting in motion a net of active solidarity, defining those who will
25 Under these circumstances, the defining criteria of vendetta or feuding defended by Black-Michaud apply only with certain difficulties. See above.
Ana Claudia Marques350
take up the actions of defence and attack becomes a very complex task. Usuallythe families of the victims and aggressors are cognatically related and might evencarry the same surname. This is due to the fact that in this region cognatism iscombined with a matrimonial regime through which family alliances are con-solidated, enormously increasing the intersection between kindred.26 Although Ihave not systematized the degree of kinship between those involved in thevindictive acts, the available data converge and allow me to state that the relativesinvolved in the retaliations are usually closely related to the victims (especially asparents, children, and siblings); though theoretically and potentially, the duty ofrevenge (as well as the selection of targets) falls on all male relatives of a victim.On the other hand, cases in which more distant relatives have actively taken partin the conflict have shown that it is not only in the case of a lack of male relativesthat effective friendship relations are necessary for recruiting avengers.27
As Land�28 points out, the act of revenge propels the creation of groups, whoseformation will depend on the relations of solidarity previously established, butwhich cannot include members who have obligations to the opposite side. As adhoc groups they will have to face this problem of recruitment after each vin-dictive act, introducing and excluding new and former participants. This dy-namics produces unpredicted results concerning its course, extension, and
26 In the neighbouring state of Para�ba, Lewin (Linda Lewin : Politics and Parentela in Para�ba:A Case Study of Family-Based Oligarchy in Brazil. Princeton: Princeton University Press,1987) analysed the reproduction of political oligarchies during the Brazilian Empire and OldRepublic and established correlations between endogamous and exogamous emphases fromthe consanguineous and local viewpoints and specific political regimes. Matrimonial regimeis an essential ingredient to the analysis carried out by CaÇedo (Let�cia Bicalho Cañedo : LaProduction G�n�alogique et les Modes de Transmission d’ un Capital Politique Familial dansles Minas Gerais Br�silien. In: Gen�ses, Juin, 1998, 4 – 28.) with oligarchic families of MinasGerais of how local political factions have combined and reproduced themselves in the localand national political spheres for centuries (see also Cid Rebelo Horta : Fam�lias Gover-namentais em Minas Gerais. In: II Semin�rio de Estudos Mineiros. Belo Horizonte: Uni-versidade Federal de Minas Gerais, 1956, 45 – 91). Also in the sert¼o of Pernambuco, in thenot so distant past, the consolidation of strict nuclei in municipal politics took place throughalliances reinforced by marriage ties. In the context of politics, marriage certainly reinforcesexternal boundaries and internal consistency characteristic of corporate groups in the coreof social networks deprived of such structural definitions. From this functional perspective,marriage would work as feuds in Turkish villages, according to Stirling’s analysis (PaulStirling : A Death and a Youth Club: Feuding in a Turkish Village. In: AnthropologicalQuarterly 33 (1), 1969, 51 – 75.), in which conflicts of this nature are the only situation underwhich agnatic kinship ties are activated by exclusion of other types of relation. For thecohesive role of feuds in segmentary systems, see Black-Michaud (op. cit.).
27 A very similar tendency was observed by Wilson in nineteenth-century Corsica, where,according to his descriptions, a combination of endogamy and cognatism also prevailed.Nevertheless, the absence of male relatives is the main reason found by the author to justifythe engagement of more distant relatives in feuds.
28 Landé, Introduction, XXXII-XXXIII.
Blood Revenge and Social Network in Pernambuco, Brazil 351
duration, and has a divisive impact of great intensity that affects the whole life ofa community, including those who do not take part directly in these processes.Moreover, there is a high degree of uncertainty concerning those responsible foracts of vengeance and the motivations at stake in a quest¼o.
These conditions of uncertainty and unpredictability of actors and theirmotivations should not be understood as deviations in the system, but as its verymeans of operation. A quest¼o de fam�lia (family quest¼o), such as the one thatopposes the Gouveias and the Santanas, may undoubtedly reach epic pro-portions. What would distinguish the previously narrated case, in which thesuccession of homicides was interrupted after the second act, from this one?Black-Michaud’s terminological distinction between vendetta and feuding doesnot perfectly fit the case, as vengeance always seems to involve collective and notmerely individual relations (the defining element of a vendetta), a characteristicthat might be able to justify the native idea of infinitude.29 The collectivities thattake part in the different episodes of a quest¼o, however, are not social unitspreviously defined for this effect or any other, despite being encompassed infamily names. If we added up, for instance, the set of victims and targets ac-counted for by both sides in this specific conflict, we would still not be able tomake up two recognizable units in reciprocal opposition. This is different fromthe distinctive trace found in many feuding systems, in which there is a balancein the amount of murders committed on each side. In this case there is not evenconsensus on the victims who should or should not be counted as part of thequest¼o, nor on who the participants are or which groups they belong to. Thevarious subsets composed of kinship and friendship ties are involved in differentacts and moments of hostility and for different reasons. Despite their internaldistinctions, these subsets still constitute organized totalizations based on anopposition, itself totalizing, defined by the quest¼o between two “families.”
Due to this specific dynamics, it would not be precise to equate the cohesiveeffects of vengeance relations in such a cognatic system either to the dyadic basesof non-corporate groups or social networks or to a normative collectivistprinciple that supposedly constitutes corporate groups. On the contrary, it couldbe said that these two modes of articulation are equally pertinent, each dealingwith different aspects of the same phenomenon, combined even in their in-consistencies. Thus, a dynamics of this kind that combines both modes of ar-ticulation is important in blood feuds but is not reduced to it, since it alsocomprises an essential operator of local politics as it is lived in the sert¼o. That is
29 Boehm, Blood Revenge, 199.
Ana Claudia Marques352
why the semantic slip between acts and relations of family revenge and acts ofpolitical life that mark the historical conflicts in this region is not casual.30
These two principles of recruitment and organization are irreducible to eachother and are crucial for resituating the relations of friendship, godparenthood,and patronage in this scenario. Anthropology classically conceives such relationsin terms of “coalitions centred on individuals”, following Wolf ’s formulation,regardless of their rather instrumental and emotional value. In a way, these rela-tions would fill in the gaps left by institutional links and breaches generatedbetween social units of solidarity defined by kinship, and between distinct sociallayers from the economic and political point of view, in which they ambivalentlyintroduce moral, symbolic, solidary, and affective continuities capable of blurringand even contradicting status inequalities. In this sense, informal ties would es-tablish essential connections, through the interindividual pathway, between nor-mative and institutional social divisions of different types, which define the socialfield but do not possess an autonomous status in relation to them.
The ethnographic approach to family feuds31 and political disputes32 in thesert¼o of Pernambuco shows that the significant relationships in both fields ofconflict are not reducible to organizing principles centred on individuals. Per-haps we should extend this statement to include collective and institutionalorganizing principles, whose normative character may inadvertently tend to getmixed up with an ontology of the social.33 Moreover, relations centred on in-dividuals introduce functional devices without which it would not be possible tounderstand the dynamics among collective agents. It seems that the problemshould be put in a different way.
The fact that two political factions and two sides of a quest¼o are both des-
30 Cf. Ana Claudia Marques : Pol�tica e Quest¼o de Fam�lia. In: Revista de Antropologia 45 (2),2002, 417 – 442.
31 Marques, Intrigas e Quest�es.32 Marques, Pol�tica e Quest¼o de Fam�lia; Jorge Mattar Villela/ Ana Claudia Marques :
Municipal Elections. Favor, Vote and Credit in the Pernambucan Sert¼o of Brazil. In: TheLatin Americanist 49 (2), 2006, 25 – 63; Moacir G. S. Palmeira : Voto: racionalidade ousignificado. In: Revista Brasileira de CiÞncias Sociais 20 (7), 1992, 26 – 30; Id./ Ead.: Pol�tica,facÅ�es e voto. In: Antropologia, Voto e Representażo Pol�tica. Moacir Palmeira / MarcioGoldman (eds.). Rio de Janeiro: Contra-Capa, 1996, 41 – 56. Moacir G. S. Palmeira/ MariaBeatriz A. de. Heredia : Le Temps de la politique. In: Etudes Rurales 131 – 32, 1997, 73 – 87;Id./ Ead.: Pol�tica Amb�gua. In: O Mal Brasileira. Patricia Birman/ Regina Novaes/ SamiraCrespo. (eds.). Rio de Janeiro: Editora da Universidade Estadual do Rio de Janeiro, 1997,159 – 184; Id./ Ead.: O voto como ades¼o. In: O Sufr�gio Universal. Let�cia Cañedo. (ed.). Riode Janeiro: Estażo Liberdade, 2005, 323 – 345.
33 The irreducibility of the social fact to its rules is not a recent issue in anthropology, and muchink has been spilled over the problem since Malinowski. This is not enough, however, for ourdiscipline to extricate itself from an inadvertent transit between the analytical model and theset of institutional or informal rules.
Blood Revenge and Social Network in Pernambuco, Brazil 353
ignated by family names does not suppress the internal segmentation of theparties (and therefore should not be understood as transcending its compo-nents). It is a fact both in the sense of given and constructed. The ideal cognatic“family,” thought of as being absolutely inclusive, is never realized. The sert¼ofamilies have, however, a virtual mode of existence,34 since it extends con-ceptually to the pragmatism of the association forms. Under family orientation,concrete ties cannot be considered independent and do not derive from the freecourse of individual encounters. The infinity of nexuses that precedes an in-dividual prevents him from being considered a social subject. This is true notonly from a chronological point of view, due to the segment in which each personis engendered, but also from a logical perspective, due to the relations that thesebelongings condition. The different modes of association actualized in politicsand in blood feuds, in all their contingency, depend on these nexuses. This,however, does not translate into any predictable coherence for these formations.
The considerations here presented suggest that relations of personal solidarityand loyalty constitutive of friendship (as well as patronage and godparenthood)are not quite individual nor quite collective. They are both, depending on theperspective from which the matter is approached. These relationships establish theconditions of realization or formation of social units that operate in the fields ofvengeance, politics, or others. In this sense, their pertinence to social analysis goesbeyond the status of mere addenda to formal institutions. These bonds become adevice or instrument through which the concept of “family” itself, in its multiplesenses, is realized as such. More than an ideology or a chimera, families in thesert¼o are real, realized through these relations.
Acknowledgements
Preliminary versions of this paper have been presented to and discussed withstudents at the Hybris Study and Research Centre, coordinated by Jorge MattarVillela and myself. I greatly appreciate the remarks and suggestions made at thatoccasion. I would also like to thank the participants of the Friends, Patrons,Followers conference for their stimulating comments. I am especially grateful toAnna Catarina Morawska Vianna for her inestimable contribution during thepreparation of the first English version of this article, which went far beyond thetranslation work.
34 Jorge Mattar Villela : O Povo em Armas. ViolÞncia e Pol�tica no Sert¼o de Pernambuco. Riode Janeiro: Relume Dumar�/NuAP, 2004, 27. Id.: Pol�tica e EleiÅ�es no Sert¼o de Pernam-buco. O Povo em Armas. Fortaleza, Campinas: Universidade Federal do Cear�/Funcap/CNPq-Pronex/Pontes Editores, 2008, 158.
Ana Claudia Marques354
