Studying histories of violence in Uíge

Research for this paper was carried out as part of a comparative research project on rebel governance in Angola, Côte d’Ivoire, and South Sudan (Péclard et al. Reference Péclard, Santschi, Schubert, Lazaro, Moro and Zina2019). Given the dearth of archival material from that period of time, my argument is chiefly based on field research in the capital city of Luanda and the northern province of Uíge in late 2017, and further informed by continuous research in and on Angola since 2007. In Luanda, I mainly carried out interviews with UNITA cadres, former military commanders and party functionaries, as well as a number of civilians originally from the North (Uíge and Zaire provinces), who had experienced UNITA’s occupation and had relocated to Luanda during or after the war.Footnote ¹ In Uíge, I carried out a number of focus group interviews, as well as open-ended, narrative interviews with one, two, or three interviewees at a time in five of the province’s sixteen municipalities, namely: Uíge (the provincial capital, hereafter Uíge City or the city of Uíge), Quimbele, Sanza Pombo, Negage, and Songo. Most of these individual interviewees had, through their family, educational, or occupational status, been identified as local elites (teachers, pastors, administrators, for example) and were thus, to a certain extent, “targets” in the efforts to establish mechanisms of governance by the successive entities—Portuguese colonial, FNLA, MPLA, and UNITA—asserting authority over the areas. These interviews are complemented by secondary literature, mainly the many autobiographies written by UNITA cadres after the end of the war.

Access to UNITA cadres was surprisingly easy, thanks to the intercession of one former UNITA general I had interviewed during earlier fieldwork, and most were enthusiastic about telling their side of the story. Here, the key issue was rather to guard myself against “ethnographic seduction” (Robben Reference Robben1996), as most were genial hosts with charismatic personalities, considerable charm and, very often, a sense of self-deprecating humor. These were men who, despite their turbulent biographies, had not only survived the war, but come out of the conflict doing rather well for themselves, regardless of ending up on the losing side.Footnote ² Here, the statements of civilian interviewees provided a welcome counterbalance to the “politically correct” discourse of UNITA cadres, which very seldom conceded that there had been occasional, if any at all, downsides to the local population from UNITA’s presence. A second form of seduction also needs to be cautioned against, and that is that the fetishization of violence—the graphic nature and emotional weight of narratives of violence—might tempt the researcher into “believing that the meaning of the events unfolding is embodied by the violent acts themselves, rather than seeing those violent acts as embedded within a larger process of war that culminates rather than erupts in acts of profound brutality that produce lasting effects for entire communities” (Kovats-Bernat Reference Kovats-Bernat2006, 120). Accordingly, I will reproduce graphic accounts of the experience of violence only sparingly.

In Uíge, there were some evident constraints to doing field research as a foreigner, apart from the sheer remoteness of some municipalities and villages. Due its geographic location on the border with the Democratic Republic of the Congo (DRC), the SME—the Migration and Foreigners Service (Serviço de Migração e Estrangeiros)—were very closely controlling who entered and left the province. This translated into having to have my passport checked and stamped when crossing into the province from neighboring Bengo province, and to register with the provincial and local SME anywhere I stayed overnight. Though my visa had been sponsored by an Angolan university, given the overall distrust of Angolan authorities toward any form of independent research on potentially politically sensitive topics, my interviewees at various points advised me to downplay my role as researcher and emphasize the personal aspect of my visit.Footnote ³ Not quite “surreptitious ethnography” (Neto Reference Neto2019) then, but I certainly followed my interlocutors’ advice and tried to stay somewhat below the radar.

The sample of interviews from Uíge is thus necessarily small and somewhat improvised, though the recurring themes, complemented with interviews with “Northerners” now living in Luanda, allow weaving together an ethnographic micro-history of the Northern Front. It is, in that sense, clearly a narrative construct, also to protect the identity of interlocutors. For all interviews and focus groups, oral consent was sought and given. Due to the sensitive nature of the topic, however, only UNITA cadres are identified by their real name in the text; civilian interviewees are either assigned a pseudonym, or their statements collated into unified narratives, especially when drawn from focus group interviews.

UNITA’s state-building project in the South

Based on this new data, the article seeks to offer new insights into a largely underexplored geographical area of the Angolan civil war. Recent scholarship, on political control in the central highlands (Pearce Reference Pearce2015), and the role of Cold War ideologies (Pearce Reference Pearce2017) ethno-regional identities (Péclard Reference Péclard and Morier-Genoud2012; Martins Reference Martins2015), and repertoires and practices of stateness (Roque Reference Roque2024) in mobilizing adherents, has begun to paint a differentiated picture of UNITA’s “rebel” governance in the territories it controlled. All authors stress the importance of UNITA’s social project, listing the importance of education, agriculture, and “traditional” social structures alongside military logistics and the war economy. Some emphasize the coercive character of UNITA in creating the “illusion of a civilian normality” in Jamba (Beck Reference Beck2013, 116); others highlight political mobilization and the provision of state services to rehearse for the day the movement would take over the government of the country (Roque Reference Roque2024, 111–112); others stress the dimension of territorial control and “belonging to” either of the movements (Pearce Reference Pearce2015, 45–46). Overall, outside purely military history, the emphasis on UNITA’s efforts to build a “state within a state” in Jamba and in the southeast of the country, dominates the historiography of the conflict. The role of FNLA (Frente Nacional para a Libertação de Angola, National Front for the Liberation of Angola), the oldest of Angola’s three independence movements, is largely marginalized in postindependence history, mirroring both the MPLA’s rewriting of the history of the independence struggle (Messiant Reference Messiant1998; Pearce Reference Pearce2015; Schubert Reference Schubert2015) and the movement’s nigh-insignificance as a player in contemporary parliamentary politics.

Drawing up a micro-history of the “Northern Front” of the Angolan conflict allows us to expand and nuance this picture. In the historiography of the civil war, the North has long been, in the words of my interlocutors, the cassule (youngest child) of history, unfairly relegated to obscurity by the dominance of the coast for the social history of the MPLA (Luanda and Benguela, together with the Kwanza River valley towards Catete and Malange), and the Central Highlands (the provinces of Huambo and Bié), with its Ovimbundu population that, in the general view, constitutes the heartland and “natural” constituency of UNITA. A closer look at the history of nationalist mobilization in the North also calls into question history as written by the victors, that is, the historical inevitability of the MPLA’s dominance of Angolan politics until today.

As a corollary, the article contributes with new empirical evidence to discussions about “rebel governance,” where in the past decade, the consensus has been firmly established that rebel movements that exert control over territories and populations do so not only through the strength of arms, but by establishing “insurgent governance” and contributing to the production of social order (Mampilly Reference Mampilly2011; Kasfir Reference Kasfir, Arjona, Kasfir and Mampilly2015; Arjona Reference Arjona2017). Building on earlier collective work (Péclard et al. Reference Péclard, Santschi, Schubert, Lazaro, Moro and Zina2019), what emerges from this body of literature is the need to interrogate the legitimizing narratives deployed by armed groups to gain popular support, and to assess to what degree they engage in civilian rule and how this relates to other, preexisting sources of authority. Here, as we shall see, the legacy of earlier anti-colonial political mobilization by FNLA played a crucial role in how UNITA sought to construct its legitimacy.

Opening a second front in the North

Following the Carnation Revolution in Portugal in 1974, the Alvor Agreement between Portugal and the three liberation movements foresaw a period of transition, in which FNLA, MPLA, and UNITA would share power and prepare elections to form the government that would lead the country to independence. However, mutual distrust between the parties, as well as more or less open favoritism toward the MPLA by some of the left-leaning Portuguese military leaders, eroded what little consensus there had been at the signing of the agreement: “elections were foregone for military engagement between the three parties, each emboldened by Cold War resources and rhetoric” (Moorman Reference Moorman2008, 169). The three movements were soon pitted against each other, rushing to secure control over towns and populations.

As the MPLA secured control over Luanda, it proceeded to proclaim the independence of the Popular Republic of Angola on November 11, 1975. The same day, UNITA —together with FNLA—proclaimed the birth of the Democratic Republic of Angola in Huambo. UNITA and FNLA had the support of the West: FNLA chiefly from the United States, through CIA advisers to Zaire for the FNLA; UNITA from South Africa, which was pursuing its own strategy of destabilization and containment in its neighboring frontline states (Miller Reference Miller2013), an agenda couched in terms of anti-communism that also suited the United States. Both Zaire and South Africa sent army units to advance on Luanda. However, thanks to Cuban military support, the MPLA held on to Luanda, inflicting a decisive defeat on the FNLA and the Zaireans at the Battle of Kifangondo, some 40 km north of central Luanda, and halting the South African advance at Sumbe, 300 km south of the capital (Gonçalves Reference Gonçalves2017, 247–248). As a consequence of the aggression of the Apartheid regime, calls for compliance with the Alvor Agreement soon fell silent, and a wave of solidarity swept the continent, leading many countries to quickly recognize the MPLA as Angola’s legitimate government (Marcum Reference Marcum1978).

By early 1976, South Africa had withdrawn from Angola, leaving UNITA friendless, and enabling the MPLA to advance toward Huambo (Gleijeses Reference Gleijeses2010, 342). Routed, UNITA retreated in what was later recast in its own myth-making as a foundational “long march” to the remote southeast of the country, from where it gradually rebuilt itself as a movement, and where it established its new capital, Jamba (Sitte Reference Sitte1981; Höring Reference Höring2022; Roque Reference Roque2024).

Most studies of the Angolan conflict concentrate on the military evolution of the conflict between UNITA and the MPLA government from that point on, narrating how UNITA evolved from having very little external support during the independence struggle (Ribeiro Reference Ribeiro, Saunders, Fonseca and Dallywater2023) to gaining in strength and allies, and increasing its control over the countryside, with the government holding on to the cities. The FNLA remains marginal in these accounts, and is said to have ceased to exist as a military movement after its defeat, with its leader Holden Roberto ineffectively lobbying for support from Kinshasa, then from Paris. Whatever international—that is, Western—support existed for FNLA’s opposition to the MPLA government “naturally” shifted over to UNITA, according to the UNITA commanders I spoke with. UNITA’s leader Savimbi successfully positioned himself as master military strategist, despite the movement’s initial defeats, and despite initial reluctance by the United States to back them. Following the Clark Amendment in 1976, the USA could also not openly supply UNITA, so assistance came mainly from South Africa. Accepting South African support was controversial, even within UNITA, and the alliance between the “racist apartheid regime and the tribalist fantoches [stooges] of UNITA” proved a propaganda boon to the MPLA, both at home and abroad. In the words of my interlocutors, however, it was an alliance born out of necessity and common interest—the defense against communism in Southern Africa.Footnote ⁴

In the early 1980s, under P.W. Botha’s leadership, South Africa recommitted to UNITA (Davies and O’Meara Reference Davies and O’Meara1985). Secure in its military support from South Africa, the UNITA leadership decided to open a second front in the North (Figure 1). Having crossed the Kwanza River, initial advance was slow and gradual, and UNITA’s presence in the North was felt only sporadically from 1980 onwards.Footnote ⁵ Its units faced resistance from the Popular Armed Forces for the Liberation of Angola, i.e., The MPLA government’s army (FAPLA), including aerial bombardment when crossing the Kwanza River near Malanje (city), but they received logistics support from South Africa, in the form of airdrops of military supplies. The difficult terrain made progress difficult and led to a number of non-combat-related casualties.Footnote ⁶

Figure 1.

Map of Angola. Source: Author’s drawing, adapted from Open Street Maps.

Figure 1. Long description

A geographical map of Angola and its surrounding neighbors.

* External Borders: To the North is R o C (Republic of the Congo) and D R C (Democratic Republic of the Congo), with Kinshasa marked just across the border. To the East is Zambia, and to the South is Namibia. The Atlantic Ocean lies to the West.

* Northern Provinces: Cabinda is an exclave in the far Northwest. Moving East across the main territory are Zaire, Uíge (containing a red marker for C d F), and Lunda Norte.

* Central Provinces: Along the coast are Luanda and Bengo. Inland are Kwanza Norte, Malanje (with the city of Malanje marked), and Lunda Sul. The Kwanza River is depicted as a blue line flowing from the central highlands Northwest toward the coast between Luanda and Kwanza Sul.

* South-Central Provinces: Benguela and Kwanza Sul are on the coast. Inland are Huambo (with cities Huambo and Andulo), Bié (with the city Kuito), and the large eastern province of Moxico.

* Southern Provinces: Namibe and Huíla are in the Southwest. Cunene is in the South-Center. Kuando Kubango is in the Southeast, featuring the cities Cuito Cuanavale and Mavinga, and a prominent red label for Jamba in the far Southeast corner.

* Scale: A scale bar in the bottom right corner represents 200 km.

However, at the beginning of 1984, these dispersed UNITA units made contact with remaining FNLA pockets of resistance in the North. Just before his death in 1979, President Agostinho Neto had made a rapprochement with Zaire’s Mobutu, after which Mobutu ceased to support armed opposition to the MPLA government (at least for a while—later in the 1980s he started supporting UNITA again).Footnote ⁷ This was followed by a policy of clemency (política de clemência) towards FNLA, a period until 1984 during which many Angolan exiles returned from Kinshasa and some prominent FNLA leaders such as Hendrick Vaal Neto and Johnny Eduardo Pinnock joined the MPLA government. In the North, for example in Quimbele, some former FNLA cadres also became municipal administrators, as part of the MPLA government.

Despite the relative success of the policy of clemency, however, there remained “fierce pockets of resistance,” which UNITA encountered in 1984. Dr. Vicente Tembo, a former FNLA fighter who took part in these encounters, and who remains a UNITA cadre from 1985 on until today, emphasized:

The first military contact was made between FNLA’s remaining 3rd Battalion under Comandante Alberto Vicente and UNITA’s Bat. 517 under Major Ngongo, after which the two movements started collaborating in attacks. “UNITA had the military might and were evolving into a regular force, while FNLA knew the terrain and the population.”Footnote ⁹

Part of that evolution was that Brigadier (later General) António Dembo became commander of the Northern Front and reorganized the military strategy. Given UNITA’s military supremacy over FNLA, and thanks to Dembo’s parlaying—in which “much manioc, much banana” was eaten—most of the remaining FNLA groups were integrated into UNITA over the course of 1984–85, and their commanders were given appropriate ranks. Despite this “natural” alliance, some FNLA commanders refused to join either UNITA or the MPLA, and the “war of everyone against everyone” continued. Two of Dr. Tembo’s brothers, for example, were killed by the FNLA in retribution for his joining UNITA.

The war in those years took the form of hit-and-run ambushes on military units and convoys. The MPLA controlled the cities and towns, while UNITA roamed the countryside, establishing temporary bases in the bush under the cover of trees and living with, or off, the local rural population. Eventually a more permanent base was established outside Quimbele, in the regedoria of Marimba, by the banks of the River Cuilo. This Comando de Frente (Front Command [CdF] on the maps, see Figure 1, Figure 2) was to become “a Second Jamba,” a powerful base in the North from which UNITA could, in theory, advance more swiftly on to Luanda.Footnote ¹⁰ Yet while accounts of Jamba portray it as a “real” city with established structures, neighborhoods and services—including, famously, white-gloved traffic police directing almost nonexistent traffic at a major intersection (Roque Reference Roque2024, 133)—UNITA’s attempts to produce the semblance of stateness in the North in this first phase of the war were much more haphazard and patchy.

Figure 2.

Map of Northern Angola with major roads. Source: Author’s drawing.

Figure 2. Long description

A geographical map of Northern Angola. A legend in the top right corner identifies black lines as Roads, blue diamonds as Municipalities, and red circles as Provincial capitals.

In the Northwest, the enclave of Cabinda is visible. Moving South along the coast, the Congo River flows past Soyo. The Zaire province contains M’banza Kongo as its capital and N’zeto on the coast.

In the West-Central region, Luanda is marked as a coastal provincial capital, with Bengo province and its capital Caxito just inland.

In the Central region, Uíge province features the capital Uíge and municipalities including Maquela do Zombo, Damba, Quimbele, Songo, Negage, and Púri. The Cuilo river flows through this area.

To the South and East, Kwanza Norte province has N’dalatando as its capital, while Malanje province to the East includes the capital Malanje and the municipality of Cafunfo. The Kwanza river flows along the Southern border of Kwanza Norte and Malanje.

The D R C border is marked to the North and East, with the Cuango river forming part of the Eastern boundary. A scale bar at the bottom right indicates 100 km.

Mobilizing popular support?

According to former UNITA commanders, their implantation in Uíge was a largely harmonious process done with great respect for the local population, in line with its doctrine established in the South (Roque Reference Roque2024, 115). When entering a hamlet, the command would first approach the soba (the traditional authority) of the village and sit down with him; a goat was usually eaten, and the soba would be convinced of UNITA’s cause:

Each unit had a political commissar in charge of mobilizing the population, and distributing propaganda material from Jamba; “then we would form committees, village secretariats, LIMA [Liga da Mulher Angolana, Angolan Women’s League, i.e. the party’s female wing] and JURA [Juventude Revolucionária de Angola, Angolan Revolutionary Youth, the youth wing] chapters—we had to conquer the territory bit by bit and then join those bits together.”Footnote ¹²

UNITA’s message was that they were fighting to free Angola from the Cuban invasion, which, according to the ex-commanders, found much resonance with the population (Figure 3). As they told me, the population was able to see the difference between the presence of the Cubans at the side of the “alien” MPLA and the instrumental alliance between the South Africans and UNITA, who were fighting on behalf of the historically marginalized “real Angolans” from the interior. One possible explanation is that there was very little visible South African presence in the country (Pearce Reference Pearce2017, 23), barring the occasional military or medical instructor, whereas the Cubans were embedded within FAPLA units.

Figure 3.

A UNITA commander’s photo album. Photo by author, Oct 2017.

Another element that also turned the population against the MPLA government was its crackdown on internal dissidents after the May 27, 1977 “coup attempt” (Pawson Reference Pawson2014). While UNITA did also use corporal and capital punishment to enforce discipline among its soldiers and the population under its control (see below), one key difference was, according to most of my interviewees, that “UNITA, in contrast to the FAPLA, did not rape women.”

In addition, there appears, in the view of some mais velhos (elders) from Uíge, a profound ignorance of, and (deliberate) disrespect of, local mores and customs among not only the Cubans but also the FAPLA troops, who came from other parts of the country, and who thus were often seen as a foreign, occupying force.Footnote ¹³ This ties into UNITA’s legitimizing discourse as representing “real, interior Angola.” In the North, this discourse also latched onto preexisting anti-colonial political mobilization by FNLA, that is, the fight of “us Angolans” against the foreign occupier. UNITA tapped into this, managing to position itself as the “natural successors” of the FNLA, but as more inclusive, unlike the FNLA, who privileged the Bakongo elite, and who:

UNITA mobilized the population through rallies, at which attendance was compulsory, and where there would be much singing and dancing in addition to political speeches—often in local languages, which was certainly one of UNITA’s great advantages over the MPLA.

Interviewees recalled the songs they had to sing:

In more practical terms, UNITA had a logistics representative in Kinshasa, who would see to it that supplies would get to the border: ammunitions, uniforms, and food. The military command would then requisition youths and women from every village to come and help carry the materials. In mixed military-civilian columns, 500–600 people would march to the border and carry the material back to base on their heads. There was, however, like in the South, no money circulating (Roque Reference Roque2024, 149–50). UNITA would request in-kind contributions from the sobas, who could sometimes intercede if the amount requested was too much, and agricultural produce was used to feed the troops. They bartered for goods across the border in Zaïre—salt, cooking oil, soap, clothes. No stealing was allowed, and everything that was captured in combat had to be handed over to the commander; soldiers caught stealing would get a lashing and, in the case of repeat offenders, be executed.

Civilians were issued membership cards, and had to have a laissez-passer (guia de marcha) from UNITA to move from one area to the next, else they would be considered spies. UNITA also created a hospital at the Comando de Frente, where it would treat civilians. Each unit had a trained nurse who could provide first aid also to the population. Schooling was mandatory in all UNITA areas, and there were attempts to continue education on the fixed and moving bases, improvising classes with looted school material under the trees.

We can see here, from taxation to the issuing of documents, to the (attempted) provision of services such as health and schooling, how UNITA sought to espouse some of the formal and institutional mechanisms of statehood in this period, which would give support to a state-centric view on rebel governance. In this phase of the war in the North, however, UNITA did not chiefly aim for territorial control, but skillfully tailored its messaging to take advantage of existing grievances and a preexisting history of anti-colonial political mobilization by the FNLA.

A first captivity

Among the population, however, the experience was somewhat less harmonious. Many interviewees recalled this time as the first cativeiro (captivity), as a period of great uncertainty, remarking on some positive but mainly on the negative aspects of UNITA’s domination. In contrast to the government soldiers, UNITA troops were indeed known for their cleanliness and discipline—if rapes happened, these were usually punished swiftly and exemplarily, and UNITA normally respected both traditional and church authorities.

At the same time, there was great distrust and denunciations among the population, and anyone suspected of MPLA sympathies would be killed. As one participant in Sanza Pombo recalled, “one of my brothers, who was the soba here, was killed because he’d hid away an MPLA flag at home. This was enough [to kill him].”Footnote ¹⁵ Although a few youths joined the movement, “adherence” was thus rarely a matter of deep political conviction but rather an instrumental means of survival (cf. Pearce Reference Pearce2017, 15). And thus, the collaboration of the population appears in different light: Mamã Joaquina Pombo, who was born in Quimbele (Figure 4), recalls that there was:

no social contract, nothing; threats only. It was a great punishment. They used us as cheap labor force. If you were a girl, you had to be their wife. It was truly slavery. And when they noticed your father had been in the MPLA army … for them, o filho do peixe também é peixe [the child of the fish is also a fish]—there, they would take their measures!Footnote ¹⁶

Figure 4.

On the way to the focus group meeting, Quimbele. Photo by author, Oct. 2017.

The most dreaded of these was the kandambala, which civilian interviewees recalled with horror. It was administered for all sorts of infractions: if a soba could not muster the required number of porters, or people refused to serve as porters, if a girl resisted participating in the animation song and dance routines of JURA, or a boy disobeyed orders to join the army, they themselves or their family members would be pinned to the ground face down, and receive a set number of blows on the back, from 10 up to 120. These would be administered by two people simultaneously, so ten kandambalas would mean twenty blows. Several interviewees in their accounts mimicked the loud “thwack” sound the sticks or chicottes made, and underlined that many who had been punished that way were left with lasting damages, chiefly infertility. Holding people in water tanks, or burying them alive up to their necks were other forms of punishment, as was capital punishment, especially for treason and rape. This usually meant execution by shooting, though people recounted instances of public burning at the stake or by necklacing.

And while schooling was, in theory, mandatory in all UNITA territory as an essential part of their nation-building project (Roque Reference Roque2024, 142), education in that first phase of the 1980s—“bush schools” in the temporary bases—was, according to those who followed classes there, little more than a pastime for the kids. It was not very systematic, consisting mainly of political mobilization. Respondents recalled that UNITA had well-trained medical staff who sometimes performed life-saving operations in the bush under very basic circumstances, but that the supply of drugs was irregular: “Their health service was based on traditional therapy, a form of syrup (xarope), and the injection of ‘90’, a mixture of woods and tablets—6 paracetamol, 6 aspirins, 9 terciclina [tetracycline = antibiotics]. Also, pomades made from wax.”Footnote ¹⁷

UNITA captured young men and women to serve with the units, either as fighters or as porters and auxiliaries, which was an essential element of logistics in the liberated territories—though not, as implied in other cases, as voluntary contribution (Roque Reference Roque2024, 140). Elena, who was captured by UNITA from her family’s farm in a hamlet outside Sanza Pombo (Figure 5) in 1988 recalled:

They continued to capture youths. We knew that if you spent two years with them, they would start trusting that you wouldn’t try to escape. They thought we liked the life that they were offering us—ha! [laughs mirthlessly]. But you had to speak com jeito [skillfully], because there were some who would report you if you complained, then they would put you on the parada and kill you, shooting or burning you. This happened a lot.

Figure 5.

A view of Sanza Pombo. Photo by author, Oct. 2017.

Figure 5. Long description

In the foreground, a large, weathered warehouse with a corrugated metal roof dominates the right side. The building has cream-colored walls marked with graffiti and a blue metal door. Part of the roof structure is exposed with wooden beams. To the left of the warehouse, a tall papaya tree stands near a red metal fence. A person in a red shirt is visible in the mid-ground near a concrete wall. The background reveals a sprawling, dense settlement of small, reddish-brown houses with metal roofs, extending across a green, hilly landscape toward a distant line of trees. The sky is overcast with thick, grey and white clouds.

Elena’s sister Joana, captured in 1989 on her way to school explained that, “There was much mobilization, a proper evangelization. But I was a captive, it was not my free will.” Once the captives had proven their trustworthiness, they would be given more important tasks. Young women especially were deployed as spies in the MPLA-held cities—they would be given money to go to the beautician first, so that they would not be recognizable immediately as “coming from the bush.” They were tasked with charming key MPLA personnel to discover troop strengths and strategies; their targets were the battalion commander, the political commissar, the chief of personnel, and the cadres and men of the state security (intelligence) service.Footnote ¹⁸

Another element that is not usually in the records concerns trade. UNITA cadres emphasized the cross-border trade in agricultural produce with Zaïre, which was important for basic supplies for the population and the troops. According to interlocutors who spent time in captivity, there was also a source of “liquid mercury” near Quimbele—“it drops and slowly fills up a wine garrafão [flagon]; it was scary, one drop was so heavy, and it would become solid in your hand”—which was traded in exchange for arms, uniforms, and tinned rations (such liquid mercury might have been used as explosive). Much speculation surrounds this trade, and no UNITA commander officially confirmed it, though they “had heard rumors”. I was told that the mercury was traded with the South African Buffalo Battalion, or directly between General Dembo and the FAPLA commander of the northern zone, who would personally come in a helicopter and deliver the goods in exchange for the mercury.Footnote ¹⁹

A second cativeiro

With the signing of the Bicesse peace accords in 1991 under the auspices of the “Troika” of the US, USSR, and Portugal, UNITA started preparing for the first elections, and very swiftly opened delegation offices in cities and major towns across the country to win the population over. Despite UNITA’s ambivalent track record in how it controlled the population, many people from Uíge voted for them in the 1992 elections. Even Pastora Josefina, whose nieces Elena and Joana (above) were kidnapped by UNITA in 1988 and 1989, respectively, admitted, “We were evangelized by UNITA. I cannot lie, I voted for them. We were mobilized. They said: if you want change, vote for us. And we wanted change.”Footnote ²⁰ Official results give UNITA only 30 percent of the provincial vote, with the MPLA at 52 percent (IFES 1992), but, as Joana said, “Everyone knows that UNITA won—so the war starts again, with almost 80 percent of the territory in their possession.”Footnote ²¹

Joana’s is a contentious statement, though a sentiment shared by many. It is far from certain that UNITA could have won the elections, despite having been “promised” victory by its Western allies, as José Eduardo dos Santos and the MPLA, with the help of Brazilian PR firms and state media, skillfully rebranded themselves as the voice of reason and democratic maturity, while Savimbi campaigned in army fatigues, delivering revanchist speeches. However, while the MPLA and, subsequently, the international community always blamed UNITA for a return to war, it is now clear that elections were poorly managed and that both parties had secretly prepared for a resumption of hostilities.Footnote ²² The MPLA had already distributed weapons to the civilian population in Luanda “for self-defense,” which were deployed in the “Halloween Massacre” and the “Bloody Friday” of 1993 (Mabeko–Tali Reference Mabeko-Tali1995; Messiant Reference Messiant1995; Maier Reference Maier1996, 95; Schubert Reference Schubert2017, chap. 2). Yet the government had demonstrated better compliance with the creation of new, integrated army—the Angolan Armed Forces (Forças Armadas de Angola, FAA)—and thus had less of a presence in the provinces at the moment of the elections. When open hostilities resumed, UNITA made rapid advances, capturing Ndalatando (where it ransacked the National Bank branch to seize cash), and attacking Uíge City and Negage on November 20, 1992. UNITA also occupied the whole of the province of Zaire, establishing a provincial government in M’banza Kongo.

Most historiography of the second phase of the Angolan civil war focuses on the “war of the cities,” UNITA’s occupation of Huambo, Kuito, and Andulo in the central highlands, because of the terrible toll the two years of sieges and occupation (1993–94) took on the civilian population. There, UNITA’s attempts to govern the cities was largely unsuccessful. Based on in-depth research in the central highlands, Justin Pearce contends that “when UNITA took control of real towns, it failed to convince the townsfolk of its capabilities as an urban governing party,” because of its dictatorial control, mainly by way of killing; its “manners of the bush” (despite its attempts to produce urbanity in Jamba); and its failure to provide any services to the population—maintenance of peace and order, payment of salaries to civil servants, and the provision of food (Pearce Reference Pearce2015, 128–29). Similarly, Paula Cristina Roque characterizes UNITA’s rule over and from the highlands cities of Bailundo and Andulo as radically different from the state-building process it had undertaken in Jamba in the 1980s, and as “contested by the populace, politically driven by anger and without the underpinnings of national reform” (Roque Reference Roque2024, 113).

In the North, however, experiences were different. UNITA retained control of the whole of Uíge and half of Zaire until the reestablishment of government control (reposição da administração territorial) under the Government of National Union and Reconciliation in 1998. This was also the case between Uíge and Zaire, and more importantly, also within the province of Uíge from one municipality to the next. UNITA’s message was roughly the same as before, that is:

The “force of obedience,” though, had less to do with ideology and more with UNITA’s capacity to provide some sort of “normal life” for the people under its control (Hibou Reference Hibou2011). Rather than hiding away in semi-mobile bush bases, it now controlled towns and cities, facing the challenge of making things work there. That capacity was very much contingent on a municipality’s location—that is, available resources and insertion into more or less functioning markets—and the way local commanders exerted their authority. Here, many respondents gave credit to General Dembo, by then UNITA’s vice-president and supreme commander of the North, who “managed to temper this barbarous people,”Footnote ²⁴ and “who did not like to mistreat people from the North” (Dembo was from Bengo province, not from the central highlands like other UNITA cadres).Footnote ²⁵

Although most scholarship on Angola’s “war economy” in the 1990s rightly focuses on UNITA’s control of the diamond fields in the Lunda provinces and how that provided its army with money to buy weapons (Cilliers and Dietrich Reference Cilliers and Dietrich2000), we can see two economic circuits functioning—one, for strategic resources and military necessities, and one for the population’s day-to-day necessities, which was chiefly barter-based:

Incidentally, General Numa confirmed the ivory trade to me, though he insisted that this was in no way poached ivory, given that the parliament at Jamba had passed Angola’s first environmental protection law in the 1980s: “We found an elephant cemetery; that is the ivory that was sold. It was forbidden to shoot down elephants.”Footnote ²⁷ Mainly, however, it was General Dembo’s wife, Doutora (Dra.) Judith, who organized a flourishing trade with Zaïre. Uíge produced bombó (manioc flour), beans, groundnut, and coffee, which were exchanged with Zaïre for salt, soap and clothes; some salt and fish was traded from Soyo in neighboring Zaire province.

Joana had managed to escape captivity in 1991 only to be caught up again by the war in Uíge in 1992. Despite initial suspicions about her defection, she regained the trust of General Dembo and ended up working directly for Dra. Judith in the organization of the trade: “In Cangola, Sanza Pombo, but also Negage and Kwanza Norte, they would buy whatever the population produced, and exchange it for plates and glasses, salt … I would go and sell with the trucks of Dra. Judith on the border.”Footnote ²⁸

Nonetheless, it was mainly Negage (Figure 6), where General Dembo had his residence, that benefited from this trade, standing at the crossroads between the road to Zaïre, the city of Uíge, and the road towards Malange and the Lundas. Residents of Sanza Pombo and Quimbele, by contrast, recall a time of scarcity and violence. In Negage and Songo, the trading hub for commerce with Ambriz and Soyo, life was more or less normal: “we had freedom of movement, even if it was called the cativeiro,” said a former teacher recruited into UNITA’s administration in Songo.

Figure 6.

A main street in Negage. Photo by Author, Oct. 2017.

Schooling worked better in the more urban municipalities where UNITA could recuperate existing school structures and trained teaching staff. In most cases across Uíge, however, it could not pay regular salaries, only irregular subsidies, but teachers were expected to perform their duty: “Salaries? Don’t even think about remuneration! As a ‘patriotic gesture’ some student would give them a chicken or so.”Footnote ²⁹ Thus there was some functioning administration in Negage, Uíge City, and Songo, with UNITA raising its flag and issuing circulars, guias de marcha, membership cards, and even traffic fines (though no birth registers—this had to be done by the churches).

UNITA controlled access to Uíge by controlling the road from Ndalatando, though government forces eventually managed to penetrate into Zaire province. In 1996, the FAA entered Uíge province from the West, taking UNITA by surprise. The FAA then quickly advanced into Songo, taking the town against almost no resistance. Seventy-two hours later, the city of Uíge also fell back to the government, with the Candombe River as the demarcation line to Negage, which remained under UNITA control. Once the city changed hands, those found in possession of UNITA documents were again suspected of collaboration and treason and imprisoned, tortured, and sometimes executed. In October 1997, with the reestablishment of government administration, the MPLA finally entered Negage. Here, too, experiences with UNITA had been relatively positive, and the MPLA, by contrast, “had received the order that they were free, whatever you find is yours. I had a Hinda motorcycle that they immediately wanted to take. And you could no longer flee, otherwise the FAPLA would rape your wife or your sister in your presence.”Footnote ³⁰

Accordingly, civilian experiences of the second cativeiro remain widely divergent. As two gentlemen in Songo explained, “Experiences were different. In Quimbele or Mucaba the people lost everything. Bungo, Mucaba, they did not give one single vote to UNITA. The way they lived here was different—Songo, Ambuila, Bembe, all very good.”Footnote ³¹ That assessment, by the way, would appear to color attitudes up to the 2017 election results, where UNITA took 24 percent of the vote in Songo, 35.7 percent in Ambuíla, and 35.9 percent in Bembe, compared to 4.5 percent in Bungo, 7 percent in Mucaba, and 8.6 percent in Quimbele.Footnote ³²