• 1 Bayart (1993), Messiant (1998), Herbst (2000), Vidal (2002), Chabal (2002), Cooper (2002), Kyle (20 (...)
• 2 New sources include unpublished Junta de Exportação de Algodão reports from the Arquivo Histórico U (...)

1Through a geographic and relational reinterpretation of the start of armed nationalist struggle in Angola, this article helps to critique and move beyond common interpretations of Angola (and Africa more generally) as characterized by long-standing socio-spatial divisions between enclaves of coastal creole elite and the impoverished hinterlands that they purportedly dominate through combinations of force, patronage, and/or neglect.1 Drawing on a range of important sources that have hitherto been unused or under-used, the paper makes a range of contributions to the agrarian history of Angola, African geography, spatiality in theories of African political economy, relational geography, and to histories of nationalism, gender, colonial concessions, churches, cotton, and rural protest in Africa.2

• 3 The designation «Baixa de Cassange» is not an indigenous term, and is used throughout the text to r (...)

2Contrary to most prevalent histories of Angola — academic, official, and colonial — Angola’s armed nationalist struggle appears to have been first launched in Kassanje, preceding the famed prison attack in Luanda on February 4^th, 1961. The key Kassanje revolt in northern Malanje Province that began more than a month earlier is conventionally (mis)portrayed as a colonial massacre of farmers protesting particular economic abuses by a foreign company operating a concession in a purported enclave of the «Cassange» lowlands («baixa») where it had the state-delegated exclusive right of purchasing of farmers’ cotton.3 Although there were attempts to enforce certain enclavist provisions, in practice these and other activities involved much more dynamic and layered geographies of connection. The so-called «cotton revolt» was actually also a polysemic — and often explicitly nationalist — mobilization forged through multiple connections spanning urban and rural areas, Malanje and Luanda, and Angola and the newly independent Congo. The revolt happened at Malanje’s crossroads of an underground Luanda-Malanje political network of churches, contracted laborers and administrative personnel that intersected with Congo-based provincial political mobilization organized through trans-border social ties. These combined nationalist networks articulated primarily but not exclusively with discontented peasants who faced joint state-corporate attempts to use intensified labor, control, and risks to overcome resistance and stagnating cotton production.

• 4 Note, I have tried to use post-independence Bantu spelling, substituting K for Q, and hard C’s.

4It was during fieldwork — when an elderly soba (chief) offhandedly mentioned to me, between bites of fresh sweet cassava while we took a break from weeding his intercropped rows in western Malanje that he too had been arrested after the 1961 revolt — that I began to really appreciate that the dynamics of the revolt were not restricted, as commonly perceived, to the Baixa lowlands in the east. Later, when I heard on the nightly news of the provincial radio station that the author of a recent book about the revolt would be speaking at Malanje’s agricultural college at Quéssua (Keswa) near the old American Methodist mission and school, I set out early and headed to the amphitheater at Keswa built by Cubans and reconstructed by Chinese.4 This book — the first participant’s extended account of the revolt — delicately but insistently showed the distinctly nationalist (and not only anti-colonial) character of the mobilization and revolt, and I have tried to carefully assess, situate and build upon this otherwise yet unappreciated work.

• 5 As a controversial topic, the revolt has also attracted more study with some new insights but still (...)
• 6 This contrasts with Weberian notions of determination through the interaction of abstractly discret (...)

5Since much of the revolt’s basic aspects and chronology have been described by other authors (who sometimes draw selectively on each other), I will not rehearse them in detail here (see MARCUM 1969; PÉLISSIER 1978; FREUDENTHAL 1999; KAMABAYA 2007).⁵ Instead, I will here provide a very brief summary in order to be able to then use a discussion of further details to point out the varying ways that different people have interpreted the revolt, and present my own new, geographical interpretation. I am specifically drawing on and developing a critical geographical conceptual approach that emphasizes that social relations are always also spatial relations, mutually produced through multiple temporal and spatial scales.6

• 7 Cotonang abbreviates Companhia Geral de Algodões de Angola. The main shareholders were the Banco Na (...)

6In the Baixa de Cassange in the eastern part of the Province of Malanje, some 500 km east of the capital Luanda, black alluvial soils fed by north-flowing tributaries of the Congo River were used by African farmers under Portuguese colonial compulsion to grow cotton for sale since the 1930s, as well as to pay obligatory colonial taxes. The Portuguese state allocated concessions with the rights to recruit labor and promote, purchase, and process cotton in the area to a company (the Angolan General Cotton Company, known by its Portuguese acronym, Cotonang) held by about a dozen mostly Portuguese and Belgian interests.7

• 8 The villages were Ganga Muchica (Nganga Muxika), Quivota (Kivota), Mulundo (Mulundu), and Gombe (Ng (...)
• 9 Monteiro, op. cit.
• 10 Martins, M. (1961) «Apontamento 14: Os incidents da Baixa de Cassange», January 30, p. 9; Informaçã (...)

7The main period of active revolt ran from January of 1961 through March, though since December people had been resisting, slowing down their work, «killing white animals, refusing to pay taxes and sow cotton» (FREUDENTHAL 1999: 283). January was the beginning of the cotton planting season, and farmers and non-farmers also refused administrative orders. The first incidents began in January as officials returned from Christmas break. On January 4^th there was unrest evident around four villages where discontent with state sepoys — «cipaios» — had simmered for years, just a few kilometers north of the administrative and commercial cotton post Milando.8 Villagers tried to tie up a local interpreter for the state, but his accompanying cipaio intervened, and the two African functionaries returned to the administrative post confirming rumors frightening the state staff.9 Villagers continued organizing amongst themselves, and then on the 12^th, after villagers again refused cipaios’ orders about taxes and cotton and threatened to kill them, that night a new mestiço Cotonang agent was killed when crossing into one of the resisting areas at night.10

• 11 There were some Angolans listed in the 4^th Companhia de Caçadores Especiais, though the timing of t (...)
• 12 Martins, M. (1961), «Apontamento 34: Incidentes da Baixa de Cassange (II)», March 7, p. 15.
• 13 Cann (2012: 515, 2015: 64) states unequivocally «For the mutineers there were 243 deaths», citing E (...)
• 14 Cann (2012: 509) also notes «The officer would fly with a box of grenades between his feet and, whe (...)

9Violent Portuguese military repression of the revolt lasted from early February through March at least, and consisted of two units from the Portuguese army, as well as air force surveillance and bombing raids.11 After a high-level meeting was held in Luanda on January 18^th, and further monitoring and negotiations were held, an army unit was sent from Luanda to Malanje on February 4^th. Another army unit had left from the provincial capital, and encountered armed resistance, with bombing raids reported on February 5^th. On February 8^th they reported that some 10,000 armed Angolans in the revolt had gathered near Cunda dia Base.12 The precise total number of people killed during the revolt is disputed and appears to have been deliberately obfuscated by the Portuguese at the time with a low partial figure of 243, which has sometimes been uncritically repeated.13 The actual rough number of Angolans killed is very likely in the thousands, but difficult to know particularly because of the indeterminate number killed by Portuguese bombing raids (reportedly of approximately 17 villages), with various estimates of fatalities ranging up to a perhaps improbable 30,000.¹⁴ Freudenthal (1999: 267) summarizes figures of the toll of «7,795 injured, 4,876 prisoners, [and] 5,524 Angolans shot down (men, women and children)» (based on «Angolan testimonies» and citing a PIDE process).

• 15 See AHU, AHM reports, and «Relatório Perintrel, n.º 22/64 da SCCIA - Malange (Angola)» (PT/AHD/MU/G (...)
• 16 As João César Correia later recounted, «The people that fled to the Congo, the sobas, Soba Marimba, (...)

10Thousands more people fled across the border to the Congo or were arrested or injured by the time the area had been re-conquered by the colonial government, at which point subsequent major connected nationalist armed revolts had also occurred in Luanda (February 4^th) and in the coffee areas to the north and west of Malanje (March 15^th onwards). The legacy of the revolt continued somewhat across Malanje in the years immediately after the revolt in the form of dozens of minor political incidents and threats.15 And the nationalist armed struggle that began in Cassange would continue in the broader region as some 10,000 refugees from the Baixa settled around the nearby Panzi District in the Congo and formed «a revolutionary support base for UPA [Union of Angolan Peoples] military operations back into Kasanje» (MARCUM 1969: 126).16

• 18 See perhaps Conceição Neto (1995) «Primeiro, primeiro, foi Janeiro: Guerra e "Memória"», Correio da (...)
• 19 Notable, for example, is Santos’ (2005) scholarship on the growth of the city of Malanje.

13Post-colonial states emerging out of liberation movements sometimes grapple with harnessing, directing or limiting the potent significance of liberation rhetoric, histories, and claims. However, as the war years recede in Angola and public discussions increase, exclusive claims to nationalist origins have weakened in favor of recognizing various contributions made to the anti-colonial struggle, including by Kassanje peasants.18 In addition, the emphasis on colonial and foreign brutality has also given way to more of an appreciation of the complex social transformations in Malanje.¹⁹ The rural concerns and family backgrounds of leading nationalist state officials have also gained publicity, particularly as the state grapples with trying to use oil revenues to boost agricultural development in order to reduce poverty and growing urban concentrations and to diversify the economy. In this respect, a nationalist recasting of the Kassanje revolt bolsters claims by current rural residents of the Baixa de Cassange to a portion of the vast oil wealth, not to mention constituting an implicit critique of inequality.

• 20 Reported in Jornal de Angola.

14For several years, plans for spectacular new projects in some areas of the revolt — including a model village, social centers, infrastructure, etc. — have been discussed, building on the yet unfulfilled promises made by Angola’s first president on a visit to the area in August 1979.²⁰ Such pronouncements of plans and projects serve as opportunities for publicity about purported commitment to rural development, but also are an ambivalent encounter in which people make implicit (and sometimes explicit) critiques of the government’s lack of fulfillment of obligations and prior promises, and even outright immoral enrichment.

• 21 Even he, though, limits their grievances to «precarious life conditions». E.g. (2013) «Sobrevivente (...)

15For example, in 2013 a historian was cited by the more independent Radio Ecclesia, asserting continuity between the peasants’ struggle and current struggles for social justice and redistribution: «It is necessary that we continue to go for social justice, love for one’s neighbor, know how to share the riches of the country [...]».²¹ Similarly, a spokesman for National Association for the Baixa de Kassanje stated that people there were requesting solutions to many problems, including «potable water, schools, health services, roads, bridges, rehabilitation of infrastructures supporting cotton production that have been inactive for half a century». He noted specifically:

• 22 ANGOP (2007) «04 de Janeiro de 1961 catapultou os nacionalistas para o movimento libertador», Jan 3

16And the secretary of information for the main opposition party, UNITA, was summarized as saying «that the commemoration was a homage to all who fought against colonialism and for national independence», and he went on to say, in what could be read as an implicit critique of government inaction, «We now have to have capacity to give dignity to the survivors of this act. In the last few years people have made [this] manifest, asking for greater dignity for the survivors and inhabitants of that region».²²

• 23 UPA is União das Populações de Angola, a group amongst others that emerged in northern Angola, Luan (...)
• 24 He writes «nationalist leaders were entirely taken by surprise when urban guerrillas began action i (...)

19Many academic analyses have uncritically repeated the narrow economic or Portuguese propaganda-influenced interpretations of the Baixa de Cassange and the revolt. Moorman’s (2008) otherwise excellent social history, for example, says that Malanje cotton producers «rebelled against the system of forced cotton production» (81) but says nothing of their nationalist aims. A popular tourist guide to Angola notes that the revolt was by «workers» who «protested violently against appalling working conditions and forced labour» (295). The pastor Malcom McVeigh (1961), seeking to garner sympathy for purportedly innocent and humble peasants, explicitly denied political overtones: «It is a mistake to consider these [MPLA and UPA] foreign or Communist movements [...]. It was a spontaneous demonstration against abusive practices by the Cottonag [sic] Company [...]».²³ For many years one of the best textbooks on Angolan history was that of Wheeler and Pélissier (1971), which notes in passing the possibility of political, religious, and Congolese influence, but the book generally portrayed the revolt simply as «an active of defiance against the system of obligatory cotton cultivation for which Cotonang, a monopolist company, had the concession in the region east of Malanje […]. This was a rebellion of poverty; it failed because the rebels lacked arms and leaders and support from neighbouring tribes». Newitt (1981: 228-229) cites Pélisser (1978) to dismiss the revolt as a «peasant jacquerie» that was done by «local groups», was «almost totally unorganized», and «took the nationalist leaders as well as the authorities unawares».²⁴ One of the most esteemed detailed analyses of the conditions leading up to 1961, Christine Messiant’s 1983 doctoral thesis likewise inaccurately describes,

• 26 Birmingham (1992: 35) mentions only rural «anti-colonial despair». He moreover erroneously puts Kas (...)

20And Marcum’s (1969) otherwise brilliant and extensive chronicle calls the uprising «spontaneous and localized» (126), while Cann’s (2015: 65) recent but problematic account also calls it «spontaneous and localized».²⁶

• 27 Ferguson (2005: 381), for example, notes, «The highly simplified sketch I present here has its uses (...)

21The nuances of contemporary Angolan debates discussed in the previous section about the meaning of the revolt in relation to current projects are not sufficiently appreciated in most academic and policy interpretations of Angola’s political economy. Perhaps this is partly because of a lack of sensitive on-the-ground study in Angola outside of Luanda.²⁷

• 28 See footnote 4.

22Perhaps more important though is the problematic framework used to understand political economy in Angola, which is then often depicted as a particularly extreme or clear example of dynamics that are characteristic of Africa more broadly.28 The essential dynamic is usually perceived as one of a neopatrimonial-rentier-gatekeeper state, in which a regime of cosmopolitan state, party, and business elites in the coastal capital cities profit by controlling trade and the proceeds of allocating valuable enclaves to foreign companies, and retain power by using such wealth to buy off political support through selective patronage, repression where needed, and otherwise neglect of the mass of low-productivity farmers. In this sense, the delimited foreign cotton concession covering the Baixa de Cassange, would be understood as a prototypical enclave. An indicative account for Angola that builds on many of the above approaches and specifically cites the Kassanje cotton area is given by Soares de Oliveira (2013: 168):

• 30 For example, soba Bumba reportedly stated, «All this "confusion" is not ours. We don’t want anythin (...)

29Sobas and leaders of the revolt were reported by colonial officials to have expressed in unequivocal terms that they were in touch with «Pessa», that is, P.S.A., or the Parti Solidaire Africain, a Congolese political party that had boomed in the southern Congolese districts bordering on Malanje Province during the run-up to Congo’s independence in June 1960. People in the Baixa had PSA calendars with the distinctive serpent and tree emblem, and reported receiving orders from PSA and Kasavubu via an intermediary Maria.³⁰

• 31 See also Lamal (1954).
• 32 See also Cann (2015: 54).
• 33 Boone (2013: 44) is exceptionally right to mention this work.

30The role of PSA however has gone relatively un-examined in accounts of Kassanje, despite a remarkable 1967 book by Herbert Weiss who used extensive travels and close cooperation to describe in detail the party’s rise. The PSA was buoyed by popular discontent with rural Belgian colonial administration and, after riots in 1959 in the capital Léopoldville (Kinshasa), the forcible re-location of politicized urban youth back to their villages.31 Weiss is at pains to point out that the PSA, led by the leftist Antoine Gizenga and Maoist Pierre Mulele (who would go one to lead widespread rural rebellions in 1964), was at first somewhat elitist and tentative, but the level of popular discontent exceeded the ability of the party leaders to conservatively channel it — «the leaders’ hands were forced by local occurrences» (229). In manners very similar to the subsequent Cassange revolt, rural people of the Cuango (Kwango) and Cuilo (Kwilu) districts in the neighboring Congo embraced PSA and refused to obey Belgian orders, pay taxes, perform obligatory labor, show respect, and so on.32 Weiss presciently recognized that such «rural radicalism» «clearly runs counter to the received view of the process of decolonization in Asia and Africa» (293), but such insight has thereafter been largely unappreciated by academics.33 However, some recent revisionist work on Cameroon, for example, draws together newly available archives and has illustrated the extensive deliberately cosmopolitan geographies of nationalist rural mobilization (TERRETTA 2013). To understand the ways by which the PSA, Kasavubu, Lumumba, and other aspects of Congolese independence came to have such an importance in Malanje requires examining the region’s deep historical and geographical connections through kinship, political rule, and trade.

• 34 See for example some of Neto’s articles in Angola Norte: «30 dias em Malange», 1948/09/25, «A juven (...)

31It is worth briefly foreshadowing that Congo-Malanje relations were also shaped by nationalists’ networks that stretched both through Malanje to the Congo and from Luanda to the Congo and back down to Malanje. Key here was the figure of Malanje-born Andre Rosário Neto (discussed more in the next sub-section). Rosário Neto was a civil servant in finance and part-time journalist who studied in a church seminary in the Baixa de Cassange, and then cultivated relationships with nationalists in Luanda during seminary school there, and subsequently joined the Union of People of Angola (UPA) in the Congo in June 1960, where he produced and translated from French into Portuguese and Kimbundu radio programs and printed newspapers and propaganda to reach Malanje and nearby areas (e.g. Figure 3).³⁴

• 35 On the border formation, see Vellut (2006).
• 36 See esp. Lamal (1965), and also Thissen (1965); Redinha (1971b), see also Bourgeois (1985).
• 37 The eighth Queen, Kamwangu, was invested in 1958, replacing her cousin, Kanadi, who reportedly was (...)

33Central to the cross-border aspect of the Kassanje revolt is the role of a motley group of people called Suku. A critical examination of their reification helps illuminate the complex socio-spatial regional geographies that both resulted from and facilitated decades of movement and relations, and eventually enabled the cross-border anti-colonial mobilization. One group of Suku resided predominantly north of the Angola-Congo border, and another straddled the national border.35 Both groups traced their lineage back to the old Seventeenth Century Congo Kingdom to the west (the «Musucu» were mentioned in a February 1535 letter by King Alfonso of the Congo to the Pope). The southern Suku came to be governed by a queen (called the Ngudi a Nkama), following the military victory in the 18^th Century over Lunda groups to the east by the sister of the king of the northern Suku.³⁶ The subsequent descendent 8^th Queen of the southern Suku, newly crowned in 1958 for being less fearful, is repeatedly mentioned in the 1961 colonial reports as identified by Angolans as a key agitator of the Kassanje revolt.37

• 38 There was also perhaps an elision of Ngudi a, Nguri a, Muria, and Maria (one report by a long-time (...)
• 39 Such dynamics also shed light on the real limitations of the approach of Kopytoff, who conducted hi (...)

34In various colonial reports on the revolt, Angolans were said to have associated Nguri a Nkama with «Maria», a name she was reportedly baptized with at the old Catholic Mussuco mission at Luremo after returning from newly independent Congo.38 Angolans near the initial unrest at Milando as early as January 13^th mention Maria in Kasule Kwenda (some 70 km to their south east), suggesting that by then the Ngudi a Nkama had been there, possibly traveling through Kulashingo (Kulaxingu), across the sparsely inhabited forested hills and then to the east of the Kwango River. Indeed, at the Kwango town, near Kasule Kwenda, by February 5^th and 6^th there was mention of an agitator from the Congo, and Angolans at Kwango also refused to plant cotton, while some 600 Angolans thereabouts had by then abandoned their work at the nearby burgeoning diamond operations on the Kwango River. Understanding the gendered roles of the Suku in trans-border trans-ethnic nationalist mobilization is thus key.³⁹

• 40 See also Vicente «A situação actual da zona da "Tribu Holo"», do sector de Cunda Ria Baza e dos Pos (...)
• 41 See also Vellut (1972, 2006); Vieira-Martinez (2006); Atkins (1955).
• 42 Carvalho (1890: 118), Valle and Castro (1908), Diniz (1918), Cancela (1922b: 126), JMGIC (1948), Vi (...)
• 43 Carvalho, J. (1961) «Situação na Baixa de Cassange», Letter to Administrador-Delgado, February 17 ( (...)
• 44 The history of ethnogenesis in Angola still remains to be told, but a key part has to do with the e (...)

35The important role of Suku also must be understood in relation to the complex shifting ethnic makeup of the region. The major ethnic group singled out by the Portuguese and Cotonang as agitators of the Kassanje revolt was actually the Holo, who trace their lineage not to the Congo kingdom to the east, but rather to the Kimbundu-speaking groups to the south.40 Yet despite persistent colonial efforts at ethnic codification and mapping in order to facilitate forced labor, both Suku and Holo, however, have generatively combined, merged and differentiated various linguistic, cultural and socio-political aspects from broader Kimbundu- and Kikongo-speaking social groups.⁴¹ There are, moreover, other related smaller groups, such as Kadi, Paka, Koxi, as well as larger bordering Ngola, Jinga, Xinji, and Lunda. Over a dozen different maps since the late Nineteenth Century show these groups in overlapping and/or shifting areas or zones, including both neighboring villages and ethnically mixed villages.42 And colonial reports on the revolt clearly mention the adherence also to the revolt of the most prestigious chiefdoms of Jingas, Bondos and Bangalas.43 This stands in stark contrast to now crude-looking portrayals by some commentators, scholars or analysts that map the liberation movement parties MPLA and UPA/FNLA onto discrete, fixed ethnic blocs of «the Kimbundu» or «the Bakongo».44

• 45 On March 11^th the New York Times reported, «A Ministry statement said the "inter-tribal" fights bro (...)

36Part of the explanation why some of scholars cited above did not recognize such support and connections between different ethnic groups, and emphasized the lack of support and connections, was because that was how colonial Portuguese propaganda deliberately sought to obfuscate the revolt in the tense context of the Cold War, United Nations hearings and national independence movements. In particular, Angola’s Governor General wrote to the Overseas Minister’s Office on March 9^th that the Kassanje revolt «could be presented as a fight between tribes for external influence in which the Army intervened to defend the injured and peaceful», and Lisbon swiftly drafted a press statement to such effect, which was then repeated in major international newspapers.45

• 46 These processes are largely excluded from the static structural-functionalist approach that Kopytof (...)
• 47 See Lamal (1965: 70). Holo were able to break off from the southern Kassanje formations in the late (...)
• 48 See Pélissier (1986) on the military campaigns. See von Oppen (1993), and Cancela (1923), and Osóri (...)
• 49 Also worth noting is that careful research by Cameroonian scholars has shown that there was actuall (...)

37In contrast, at work behind the actual shifting, layered mix of relations between people that facilitated trans-national political mobilization were a series of longer-standing social, economic, cultural and political processes.⁴⁶ The area’s dynamics took shape in particular as a key slaving hinterland from the Seventeenth through the Nineteenth Centuries, using control over the Kwango River, and with the Holo also controlling salt pans on the Lui River.⁴⁷ However, at the end of the Nineteenth Century, as the Portuguese sought to establish «effective occupation» following the Berlin Conference, it was with the Suku queen Ngudi a Nkama that state-sanctioned missionaries established relations (THISSEN 1960). There followed booming rubber trade until the market collapsed around 1909 and then large-scale (but by no means complete) migration and displacement related to colonial military campaigns of occupation that sought to establish political-administrative control in order to allow a census of the population that would facilitate collection of taxes and mobilization of labor for infrastructure and construction projects and agricultural production, which are discussed more in the next sub-section.⁴⁸ What is key in all this is that rather than migration constituting an «exit option» to some other comparable abstract unit of land (a la Kopytoff), migration instead entailed a shifting relationship to both fixed resources and sets of social relations, and this would eventually facilitate anti-colonial organizing and the Kassanje revolt.49

• 50 Around Mussuco in 1915-1923, there was hunger due to a disease outbreak amongst livestock (COSTA 19 (...)
• 51 The spelling varies: kifumbe, quifumbe, quifombo, kifumbi, etc. Thissen (1960: 170) interestingly d (...)

38The turbulence of colonial conquest also was interpreted metaphysically with specific ritual practices that would be resuscitated during the 1961 revolt.⁵⁰ The Kifumbi figure, for example, is reported in some Nineteenth Century accounts with ritual sacrifice of human hearts, but its meaning shifted closely in relation to the forced labor associated with colonial spatial-economic transformation.⁵¹ Kifumbi haunted the rural paths hidden in tall grasses and would capture people to use the oils of brains, in varying accounts, for the engines of the labor-intensive industries to which many people were forcibly rounded up and sent without return: railroads, sugar factories, diamond machinery, or plantation equipment (ESTEVES 2008). The Kifumbi illustrate metaphysically some of the long-term and extensive transformations and responses that shaped the region and lead to the particular form of the revolt, rather than a simple reflex to the economic conditions of cotton work.

• 52 Anon (1903), Anon (1910), Sarmento (1916), Gago (1921), Osório Junior (1921), Milheiros (1972).
• 53 See Diniz (1929) Alberto (1928), Bebiano, Jose Bacelar (1926) «Observações geológicas e apontamento (...)

39In contrast to the notion that the Baixa was simply delegated to management by a foreign company, there was increasingly intensive government presence during and after decades of Portuguese military-administrative occupation, and an associated history of protest against such presence, even in the far northern reaches along the Congo border. Government interest had included, for example, designating a Holo and Jinga district as early as the 1850s, national border demarcation in the 1890s, and then opening of military posts at Mussuco in 1898 and Tembo Aluma in 1903, followed by years of conquest campaigns.52 Then came non-military activities — for example, taxation, labor recruitment, a study expedition in 1925, plans for a train line, inventories of peasant holdings, phytosanitary measures against locusts, etc.⁵³

• 54 Palma, op. cit. See also DSEEG (1967), and Martins 1961, op. cit., p. 3, who also notes some people (...)
• 55 See fn 102, Pélissier (1986), and various AHM and AHA reports; Vaz 1921, op. cit.; (1916) «Atentado (...)

40This growing government presence operated in increasingly close collaboration with Cotonang — detailed in section 5 below — to such an extent that the many of the practices of revolt were explicitly addressed not simply at cotton company practices but also at administrative measures. For example, although a compulsory national census (with stiff fines for non-compliance) was supposed to be carried out in January of 1961 (preparatory surveys had already covered the country), at that time the Queen Ngudi a Nkama sought to substitute for this authoritative colonial practice of censusing by ordering people to not obey, and to instead gather and be counted for her (presumably for propaganda and enlistment in the revolt).⁵⁴ This continued decades of revolts, passive resistance and migration against censusing for tax and labor ever since military occupation in the early 1900s-1920s.⁵⁵ It was only fitting that the 1960 Christmas Eve front page edition of the Malanje colonial elite’s newspaper featured both the modernist questions «How Many Were We? How Many Are We?» and, a smaller type-faced article headline titled simply «Communism».

• 56 Vicente’s (1959) internal Cotonang report notes that the area around Tembo Aluma saw 270 farmers fl (...)
• 57 See coverage in the Malanje newspaper Angola Norte: (1960) «Os acontecimentos no Congo», July 16, p (...)
• 58 Teixeira, op. cit.

41Similar processes were likewise unfolding unevenly across the border in the Belgian Congo (DENIS 1962). By the 1930s these state-corporate exactions in parts of the Congo region had reached such an extent that they provoked a rebellion in areas with Pende, where some 500 people were killed by the Belgian army, and many others brought stories of the events as they subsequently crossed over to Angola where relatives lived (VANDERSTAETEN 2001). These processes also lead to migration from border areas of the Congo to urban centers such as Kinshasa and Matadi, as well as colonial projects and plantations, and sometimes to Angola also, where some labor contracts were more favorable, but there was also movement of farmers from the Angola to the Congo also.⁵⁶ With the increasing commercialization of the countrysides (see sections below), the northern border town of Tembo Aluma became a commercial center in the 1950s for people from both Angola and Congo, with Portuguese merchants also trading goods from Kinshasa (DENIS 1962). These sorts of linkages would be critical to facilitating the PSA political mobilization when area youth then living in Kinshasa were forcibly returned home in 1959. The independence of the Congo was widely publicized and experienced in Malanje (and in the front pages of Luanda newspapers), not least given the significant numbers of white settlers and traders that fled into Angola, including some 500 to Malanje.⁵⁷ Tembo Aluma would be one of the earliest major areas of the Kassanje revolt, and Angolans in the area reportedly talked of «expelling the whites from their lands as was done to the whites in the Belgian Congo».58

• 59 There is also mention of an apparently different «António Mariano» joining the Methodist church aro (...)
• 60 See also «Informação 216», February 27, p. 4 (AHU), which specifically describes «calcinhas» at Kun (...)
• 61 Lamal (1965: 60) reports that the Kizamba sector «between Kwango-Lukunga» had 2,742 Baholo and 6.86 (...)

43A reportedly key organizer of the Kassanje revolt, António Mariano, remains largely unknown, but he apparently made trips back and forth from Malanje to the Congo. He was born in the suburbs of Malanje, just across from the neighborhood that had been transformed by the growth of the cotton industry with massive new Cotonang warehouses, processing facilities, and segregated white housing complexes. While Kamabaya (2007: 79) describes Mariano as a «Catholic catechist», there are also references to an António Diogo Mariano who worked for many years for the Protestant Methodist church in the Baixa de Cassange, first in the south at Kambundi Katembo in the 1940s and then in the heart of the cotton lands at Kela in the 1950s.59 Either way, Mariano both saw the deplorable conditions and organized clandestinely with other church members, with peasants, and sympathetic traditional authorities, in addition to underground nationalist political activists in Malanje and the Congo, fusing syncretically Christian, African, and nationalist iconography in the form of holy water, incense, and incantations to Maria and Lumumba.60 Working as driver in the cotton zone for an important Holo soba named Kizamba (in an apparently mixed Suku and Holo area where a Pastor Morton had installed a mission in 1929), he also drove his Chevrolet up on through to Kinshasa several times, and reportedly brought back UPA literature and nationalist ideas.61

• 62 Kamabaya (2007: 106-107).
• 63 «Informação 696», March 29, 1961, p. 2 (AHU).
• 64 A range of other nationalist pamphlets, which made a point of emphasizing agrarian issues, had been (...)

44Another activist born in Malanje, João Cesar Correia («Makwiza»), had studied in the mid-1940s at the American Methodist college at Keswa, studied in Benguela, and, after meeting clandestinely in Luanda with inspectors from the International Labor Organization in 1959, also made the trek through the Baixa to the Congo in 1960, meeting along the way with Mariano and sobas, and returning in 1961 as a military commander of the revolt.⁶² Another activist, João Bernardo Domingos («Kyoza»), a cousin of Rosário Neto, spent more time Luanda, helping found the short-lived Movement for the National Independence of Angola around 1957. In order to avoid arrest by the secret police, he fled around 1960 to the Congo by passing through the clandestine political network of the Movement for National Liberation operating with the assistance of sympathetic Methodist priests in Keswa in Malanje. Shortly after the revolt Keswa would be called a center of subversion, having allegedly taught in geography lessons that «Angola is one of the independent African countries […]»63 Kyoza brought with him copies of a «Manifesto de Kasanji» for Rosário Neto. This Manifesto, authored by a ranged of nationalist figures in Malanje and Luanda, analyzed conditions in the Baixa, denounced forced labor, and argued for the total and complete independence of Angola, as in the Congo.⁶⁴

• 65 1934 legislation provided for 172 cipaios in Malanje (Diploma Legislativo 603). The number was repo (...)
• 66 See e.g. Decreto of September 13, 1906.

46A key driver of Malanje-Luanda connections was the expansionary dynamic, in which taxation in cash and kind both was an objective to fund the state and a means for compelling labor for business, projects, services and a broader «civilizing mission». To help enforce tax collection and administrative rule generally, by 1936 Malanje alone was apportioned 313 cipaios across some several dozen administrative posts.65 Taxation was a key subject of a great deal of attentive and varied colonial thought, design and action (often in response to the difficulties and contestation of implementing taxes) — by the early 1930s, there were already well over a 100 different legal measures issued on domestic taxation (not including issues of domestic and international trade tariffs and duties) (RAMOS 1970). Given the experience of 19^th Century tax revolts in Angola, the new 20^th Century legislation explicitly addressed in its start in 1906 the need for «smooth and progressive» «gradual extension» of taxation in order to «avoid resistance».66

• 67 Ibid., Portaria 375 of 1913, Portaria 11066 of 1959, Decreto 22793 of 1933, Decreto 23417 of 1934, (...)

47State functionaries had a personal material incentive to boost tax collection, since they received percentages of the revenue as personal income bonuses, as well as a professional incentive, since the local government received 20% of the total, a practice that continued until 1961. This principle of «co-participation» also included small but significant percentages of tax revenue paid to participating sobas.67

• 68 Although the great historical importance of Angolan laborers has been widely recognized, and numero (...)

50Contracted labor was a key aspect of the Malanje economy, and connected people from diverse areas of the province, rural and urban, and connected people from Malanje with other places and peoples. Surprisingly, the extensive social relations of widespread contracting of labor throughout Angola have received little detailed scholarly analysis.68

• 69 One could compare the number of contractees given the population in the respective areas (particula (...)

51Recruitment from the central highlands going to the northern coffee lands was the most important pattern of labor contracting, and that which is often mentioned. However, no in-depth study of that regime has been conducted, as far as I know. Moreover, no study has examined the significance of the patterns of the massive amount of other labor contracting. In 1958, for example, about two-thirds of the total contract labor in Angola was from areas outside the southern highlands of Huambo, Benguela, and Bie that fed the coffee plantations in the north.⁶⁹ Earlier statistics show from 1948 show Malanje supplying a quarter of the total contractees

• 71 The monopsony on recruiting labor in Kassanje was established with the Cotton Intendency, by Diplom (...)
• 72 See Ferreira (1948: 4-5).
• 73 Informação 262, op. cit., p. 5.
• 74 See Freudenthal (2012); compare also Vos (2015: 115), and the work of Shana Melnysyn on the Bailund (...)

53The pool of possible contract labor in Malanje was reduced by Cotonang’s monopsony since 1930 on recruiting labor in the cotton areas, stirring controversy, or at least that was the sentiment felt in internal Cotonang reports.⁷¹ There was competition over labor, colored though with variations of elitist and/or white nationalist business’ critiques of foreign businesses. Exposés and reports at the time mention the extensive use of contracted labor, and there was an elaboration of various administrative bureaucracies for managing this labor. The state Agência da Curadoria monitored labor. Licenses were given to labor contractors. Wages and conditions were regulated. Identity cards for taxation and control of movement were designed, distributed, monitored and enforced. But the administration of these contracts also appears to have been rife with patronage, as fees were paid by contractors to government officials regulating contracting.⁷² So, it was not simply the harsh economic exactions that were felt and resentment, but also the loss of freedom of movement and autonomy, as well as the arbitrary depredations by the expanding bureaucracy. Some revolters put their passbooks, annual tax forms, and other cards on spikes.73 People did nonetheless try to use such forced labor movement to their own ends of sharing information and political conscientization.74 Repeated encounters with other laborers and residents from across Malanje and Angola helped forge a sense of a broader collective identity and struggle.

• 75 Alves (1935), Ross (1925), de Matos (1944), CGC (1939), Santos (1927), Johnson (1929: 207).

55Much of the early forced and contract labor went to constructing the Luanda-Malanje railway in the early 1900s, which was revamped in the 1930s, as well as a huge system of roads in the 1910s and 1920s, which were also repaired and maintained through ongoing manual labor in the 1930s to the 1950s.⁷⁵ This investment of labor in turn allowed Portuguese administrative occupation and movement, and enabled the censusing, taxing, and mobilization of even more labor for other construction projects and for agricultural production. The rail, for example, was key to bringing heavy supplies and rations from Luanda for military expeditions to occupy Malanje and the diamond-rich Lunda Province in the North East from the 1900s through the 1920s (MACHADO 1913; PÉLISSIER 1986). Again, when the revolt broke out in Malanje in 1961, the railroad was used to quickly bring army units to Malanje.

• 76 Ferreira (1927: 29-30); Alves (1935: 19 & Appendices). See also the 1918 map of roads and planned r (...)

56Since the start of de Matos’ reign, a staggering 25,000 kilometers were reportedly constructed from 1912-1923. By 1927, Malanje Province itself was reported to have 2,200 km of motorable roads, and by 1935 also a «considerable extension» of paths, some 118 bridges, and 10 pontoons.76 Various road boards and brigades continued work through the 1930s and 1940s, increasing after World War II by the public investment of the colonial development promotion plans.

• 77 See Portaria 375 Art. 26 §5 and §22, Art. 58§3, and Art 78, and the 1933 Overseas Administrative Re (...)
• 78 See Decreto 235, February 2, 1923.
• 79 Castro (1974). Such work also entailed possibly thousands of manual rock breaking sites; by 1958 th (...)

57Equally if not more important than the main formal railroad and highways was the extensive quotidian forced labor supervised by local colonial administrators in the construction and maintenance of 18,000 km of dirt roads and 29,000 km of paths. Forced labor for road construction was legally enshrined in a series of general administrative laws throughout the first decades of the Twentieth Century. Although absent from national statistics, local labor was used by local administrators for local roads whose construction and maintenance such functionaries were responsible as part of their overall administrative duties.77 The burden of forced local road work increasingly fell disproportionately on women who remained in their villages, since men were away as migrant labor to pay their taxes, from which women were exempt after 1923.78 By 1973, Malanje was reported to have 4217 km of picadas, in addition to the 2200 km of state paved roads and 249 km district roads.79

• 80 See Cooper (1996) and Higgs (2012).

58Because the huge contributions of such forced labor were not quantified with a monetary value and thereby included in official colonial financial statistics, this effectively gigantic «payment» of taxes through labor has been grossly ignored by analysts emphasizing the «gatekeeping» aspects of states in Africa and the limited administrative reach. Indeed, the actual extent — quantified or not — of forced labor was often deliberately hidden or obfuscated because of the ongoing international publicity, heated debate, and scandals surround the heinous and often illegal and/or informal abuses and conditions.⁸⁰ Simply because such forced labor was not given a financial value by colonial administrations in their accounting books does not mean that such forced labor did not occur. Indeed it is impossible to understand the subsequent shift to mechanization and the contemporary emphasis on mechanization in construction and agriculture in Angola without really documenting and appreciating the extensive painful histories of such pervasive but unquantified forced labor.

• 81 See also Cooper (1996: 28-31).

59The decades of labor abuses incurred in building and maintaining the roads generated widespread resentment. Some such abuses in Malanje were documented in a published report transmitted in June 1925 to the League of Nations’ Slavery Commission about the continuing labor abuses in Portuguese Africa, in turn contributing to the League’s passage in September 1926 of the Slavery Convention.⁸¹ The Report and Convention came after international outrage several decades earlier about forced labor from Angola being used on cocoa plantations in São Tomé — and which in turn spurred some minor changes as well as Angola’s new labor code a few years later. Likewise, the Norwegian anthropologist Amandous Johnson had also traveled to Malanje around the same time, and documented similarly outrageous abuses of women laborers constructing roads leading east from Malanje city towards the Baixa:

• 82 3,703 at Cambunze, 8,647 at Cacuso, 1,870 at Matete, 1,451 at Aldeia Formosa, 399 at Cacolo, 1,943 (...)

60An end result of such stark and widely resented and remembered abuses however was greater fixed capital in the form of roads that facilitated circulation of forced labor to plantations and other works. Transport facilitated such contracted labor, and hence also a growing commercial economy, which entailed the significant movement within Malanje and to other provinces, as illustrated by the traffic figures for the smaller train stations in the year 1960. From small towns that were not even municipal capitals, thousands of people rode the train.82

• 83 The routes were outlined in the Anuário Estatístico.
• 84 See «Informação n.º 145/61-G.U.» 06/02/61, AHM; Anuário Estatístico 1957, p. 215; Cleaveland (2008) (...)

61Passenger traffic on trains was gradually facing competition by growing networks of bus services. The state allocated particular private companies with specific routes.⁸³ In 1957, for example, there were 45 of these regular bus and trucking routes throughout nearly 10,000 km of the country. By the late 1950s transport from Luanda eastward through Malanje picked up, particularly for the heavy machinery needed to power, process and support the increasingly mechanized diamond extraction that was spreading west from the dwindling initial alluvial finds towards the Kwango River on Baixa’s eastern edge, where hundreds of diamond workers would join the revolt.⁸⁴

• 85 See Diploma Legislativo 859, BO 47 of 1936 (this also modified Diploma Legislativo 242 of 1930, whi (...)
• 86 See also Cann (2015: 55).

62Beyond the railroad terminus, the road network in the Baixa de Kassanje was improved partly through financing from cotton sales. Under a 1936 law, 2.5% of the total amount that Cotonang paid in purchasing cotton from farmers was allocated to a special government fund for purposes of road building and conservation. So, while this could be interpreted as colonial delegation to a private concessionaire of state functions of tax collection, in reality that was only a relatively small task, as the funds were still managed and spent by state authorities, who had to find and mobilize labor for the road works. The situation was one of increasing state-private collaboration, rather than any one-off delegation. The result, after ten years, was an extensive network of 865 km of roads in the cotton zones, which more than doubled in the next ten years to 1711 kms.⁸⁵ By 1961 Kassanje peasants would protest these state exactions of labor for road works — not just cotton prices — and aim to turn the road network to their advantage militarily by setting up roadblocks and ambushes, destroying bridges, and refusing to obey state officials’ orders.86

• 87 For further details and references on this section, see deGrassi (2015).
• 88 1911 Regulamento, Art.110; see also Portarias 377 and 378 of 1912, Decreto 1224 of 1914; Portaria 1 (...)

64In an important decades-long process beginning in the early Twentieth Century and variously called «villagization», «concentration», «regrouping», and «rural reordering» — which deplorably has yet to be studied thoroughly — thousands of rural villages throughout Angola were ordered by colonial administrators to be concentrated from dispersed locations near water sources and moved into dense linear settlements along ever more flat and straight roads.87 The legal foundations for this were laid as early as 1911 in laws specifying reduced hut taxes for people living closer to municipal capitals, regional roads, and military posts, soon followed by orders for local administrators to force Angolans to live in «concentrated» settlements of no less than 10 structures, or face stiff fines. Chiefs were also given a financial incentive to concentrate villages since their stipends were based on village size.88

• 89 See also the colonial urban development plans for Duque de Bragança (Kalandula) and Malanje.
• 90 See (1948) Angola Norte, May 29; (1958) Angola Norte, Dec 6
• 91 Cabral and Fonseca (1956); Vicente (2011: 55); IICT and INEP (1988); Cabral (1980). The presence of (...)

69During his year in Malanje in 1945, for example, Agostinho Neto, as he walked up the city’s main road, passed by visible buildings embodying this new order.⁸⁹ He passed the bustling trading stores, the car and agriculture suppliers, the end of the train line, with its hulking Cotonang warehouses, and, as he reaching his work at the medical office, he would have seen the construction work on Cotonang’s new segregated residential neighborhood nearby.⁹⁰ It was in the burgeoning forced-labor cotton-plantation hub city of Malanje that he is reported to have experienced life-altering racial discrimination when he was refused entrance by a low-tier hotel for whites only. Ten years later, Amílcar Cabral, too, after meeting with Neto in Lisbon, would walk these streets past Cotonang on his way up to do research tests at silos quickly filling with plantation settlers’ maize crop from that season’s first harvest. From Cabral’s impressions and the statistics from his agronomic research in Malanje and elsewhere in Angola, he would several years later in 1960 pen his polemic The facts about Portugal’s African Colonies, decrying humiliating segregation, land expropriation, and forced labor in cotton and coffee in Angola, noting how some of such agricultural wealth ended up in ownership of «blocks of flats in Lisbon».⁹¹

• 92 Banco de Angola 1958.

70In addition to being invested in real estate and construction, the profits from the expanding contract-labor based plantation economy also often circulated through usurious trading stores that sold imported goods on credit at high interest rates to farmers (SANTOS 2005). From 1954 to 1958 the number of registered trading stores in Malanje Province increased by a quarter from 362 to 448.92

• 93 In the 1950s, the Methodist church in Malanje also had a presence in the southern Congo, including (...)
• 94 See Polanah (1970), Ferreira (1942).
• 95 Anuário Estatístico.

71Catholic and Protestant Church networks linking Malanje and Luanda, and perhaps Congo, and numerous small towns were also key in the facilitating the revolt, through both education and connecting people from disparate areas.93 The Catholic churches in Malanje expanded from the city in 1890 to the Baixa in 1900 (Luremo/Mussuco) and 1913 (Bangalas), with another two in the mid-1930s, and four more by 1960, some with close relations with Cotonang.94 These often involved schools, and a range of indigenous catechists — by 1959, across 13 parishes in Malanje there was a total of nearly 250 staff.⁹⁵ In the Methodist church also, there were a range of different commissions, activities, and forums, involving the direct engagement of thousands of people across the Baixa. Rotating positions through the church exposed catechists to diverse and challenging circumstances and communities, and helped forge connections, foster debates and circulate ideas.

• 96 See Kamabaya (2007), and also «Minutes of the Annual Conferences of the Methodist Episcopal Church» (...)
• 97 Agostinho Neto, the MPLA leader and Angola’s first president, also spent time in Piri in the planta (...)
• 98 See also the seminarians between Keswa and Cacuso later arrested, in «Informação 691», April 1, 196 (...)

72One particularly important figure was the Reverend Job Baltazar Diogo, who studied at the old American Methodist college in Keswa established in 1887, but also served in the historic rocky town of Pungo Andongo amidst proliferating European plantations.96 During the coffee boom years, he served in the rebellious Dembos coffee district, which saw massive land alienation for European plantations.⁹⁷ Upon his return to Keswa, he worked on translating the bible into Kimbundu (working on II Kings, Ecclesiasticies, Genesis, and Exodus in 1959-1960), served on a range of church commissions (youth, music, memory, and ritual), and coordinated a clandestine cell of the National Liberation Movement at Keswa, as well as joining the UPA. Reverend Diogo’s son, Bornito de Sousa, is now Minister of Territorial Administration, and presided over commemorations of the Kassanje revolt, together with Boaventura Cardoso. Cardoso was then governor of Malanje, and a former Minister of Culture, and was the son of the popular nurse, dentist and underground political activist Sebastião Cardoso, who was also active in Malanje at the time of the Kassanje revolt. Other major nationalist figures that were active in the church included Cónego Manuel Joaquim Mendes das Neves, who was born in the coffee areas west of Malanje, and was in contact with Diogo and a wide range of other nationalist and church figures, including Father Joaquim Pinto de Andrade. It was under Father Pinto de Andrade that Rosário Neto had studied at the Seminary in Luanda (KAMABAYA 2003, 2007).98

• 99 See Marcum (1969: 47), who says Cardoso was also known as «Canguia». Marcum reports that Kanguya’s (...)
• 100 Marcum (1969). On Black Moses in Quela (Kela) in 1924, see dos Santos, cited by Pélissier (1978).
• 101 See Martins, M. (1961) «Apontamento 14: Os Incidentes da Baixa de Cassange», January 30, p. 14, AHU (...)
• 102 Arcaute and Asúa (1961: 49) mention reports around northern Malanje of a «Casonsola Mission».

73So, rather than view nationalist politics as restricted to a cosmopolitan coastal creole elite, with separate parochial peasant protests in exploited interior enclaves, the political mobilization in the larger Malanje region was part of a larger extensive history and geography of explicitly nationalist protest that connected issues of land and labor exploitation to demands on the colonial government and the international community for equal rights and independence. As illustrated above in relation to the Congo connections, people protested both in relation to experiences in the past and familiarity with other places, as well as their immediate concerns. Rosário Neto, for example, drew on the experiences of his grandfather Joaquim «Kanguya» Filipe Cardoso, who was jailed for trying to petition the League of Nations in 1927 for Angolan independence (as described for Jamaica by Marcus Garvey). Kanguya wrote as part of a broader «Nativist Movement» throughout parts of northern Angola since 1917 at least, which included various attacks and threatened bombings of the railway, with August 15^th appointed as Angola’s Independence Day.⁹⁹ Earlier, in the 1910s and 1920s, the famed writer António de Assis Júnior penned a diatribe against land expropriation in the coffee areas, and others such as Quimamuenho (Kimamwenyu) worked tirelessly defending peasants against dubious settler expropriation of their land (FERREIRA 1989; CORRADO 2008; ASSIS JÚNIOR 1917). And in the 1920s and 1930s, there were related syncretic Garveyist, Kimbanguist, Kasonzola and Jehova’s Witness Watchtower/Kitawala activities ideas circulating around the region.100 Likewise, Marcum (1969: 47) notes «a movement known as Moïse Noir [Black Moses] spread through the countryside. According to its message, American Negroes would come to liberate the African from European oppression». Six years prior to the 1961 revolt, there were reports around Tembo Aluma, Milando and Ginga posts of renewed religious «Kasonzola» agitation similar to that underpinning the revolt, with killing of white livestock in expectation of a savior «who would descend the Cuango and bring arms, gunpowder and money so that the indigenous would have means to wage war to liberate themselves from whites».101 And again, in June 1960, the time of Congo’s Independence, there were still further reports of Kasonzola in Malanje.102

• 103 See fn 64, and AHM reports: da Silva, Major José Inácio (1915) «Inspecção 9.ª Companhia Indígena de (...)
• 104 See Neto, J. (1960) «Algumas considerações acerca da Zona Algodoeira da Baixa do Cassange», Delegaç (...)

74The Portuguese colonial regime responded to such unrest with a mix of increasing administrative surveillance and policing, particularly through increasing the numbers of the several hundred cipaios then spread across the administrative posts of Malanje.103 However, there were also token efforts at providing social measures in some of the key and contentious areas, either directly or through church or charitable organizations, in the form of a commission on cotton prices, corporate and labor laws (weak and often unenforced) and social development schemes and projects (such as, water, education/training, health, housing, etc) (e.g. CUNHA 1953; CRCCAU 1961).¹⁰⁴ So, rather than view the Kassanje revolt as spastic reaction against dropping prices in a purported delegation by the state to a privately run enclave, instead the revolt grew out of broader and longer-term expansion of contested negotiations between different social groups and state and corporate agents.

• 105 On the impact of the rains on production, see figures in Figure 1. Price indexes are in the Banco d (...)
• 106 In particular, the weed imperata cylindrica, known locally as senu or seno, and pests such as lygus (...)
• 107 Cotonang (1960b); Dias dos Santos (1961: 13); JEA (1960: 4).
• 108 See Portaria 6919 of 1949.
• 109 Dias dos Santos (1961: 4).

76In addition to the important roles of connections through to the Congo and Luanda, the precise timing of the revolt should also be understood in relation to the particular conjuncture of circumstances in Malanje. The prior year had seen heavy rains destroy crops, the head tax rate rose 15% in 1959, and costs of living and particularly food prices had risen in Malanje Province since 1957.¹⁰⁵ The late 1950s saw declining numbers of people growing cotton due to mounting difficulties with intensive cotton production, including increased pest and weed problems, concentration of farms distant from peoples’ villages, and heavy investments in lowland drainage and soil conservation and erosion works.106 State and Cotonang reports had complained for years about the lack of labor due to natives’ primitive and lazy nature, and also problems of inadequate food production in the area. In response, and contradicting notions of a private enclave, in these projects, state officials from the Cotton Export Board (Junta de Exportação de Algodão) worked closely with Cotonang employees (there were roughly 400 Cotonang agents by the late 1950s) to concentrate tens of thousands of farmers in lowland areas where they could be monitored and forced to work the heavy dark soils that were fertile but also susceptible to flooding. Moreover, local state administrators received bonuses based on cotton production levels and not only helped enforce these measures but also illegally prevented many Angolans from leaving to seek work elsewhere.107 Under a 1949 law, everyone in the cotton zone was required to show proof of cultivation through a special green-colored cotton passbook («caderneta»), which showed how much they had produced and sold, where, when, and with how much seed.108 In a stinging report written shortly after the revolt, the Director in Angola of the state Cotton Export Board could not have been clearer in blaming local state agents rather than just Cotonang employees: «the cause that contributes most to the poor condition of people in the cotton zones» was «too intense intervention by the administrative authorities».109 The scholars mentioned above got it wrong in emphasizing delegation to a private company enclave and economic protest against only poor prices — state intervention in the cotton concession areas was prevalent, intensive, and insidious, illustrated powerfully by the issue of field and settlement concentration.

• 110 See also João Vunje in Freudenthal (2012: 18), and fn 128.

Source: Dias dos Santos (1960: 8)110

• 111 Legislation set tax collection for the start of January (see, e.g., Portaria 3817 of October 22, 19 (...)
• 112 See (1958) «Cultura do algodão no distrito de Malange», Angola Norte, March 5.

78There are two rainy seasons for growing crops, one from roughly late September to December and one from late January to April. Cotton was cultivated from December to May. The difficult male labor of preparing fields in December falls precisely at the end of the major hungry season, in which the stored food reserves have been exhausted, the crops planted in September and October have yet to yield. In that context, farmers are required to perform heavy labor, and also to spend a significant proportion of the money earned in cotton sales in July and August on taxes.111 Another potential conflict had to do with the strict state-appointed time for sowing in January and February, which could detract from time for food harvesting.112 At the end of the cotton season, some cotton weeding and harvesting would be done in the lean season long after important short-cycle food from the first growing season had been harvested (such as maize, peanuts, beans, tomatoes, squash, cabbage, and sweet potatoes, but not yam and cassava). Closely related to temporal conflicts between food and cotton production were spatial issues of designated cotton plots, distances from villages to such plots, and relocation of villages to concentrated cotton plots.

• 113 The total land reserved in such concentrations was 60,000 ha. See Cotonang 1960; Gouveia (1956).
• 114 Neves (1954 : 12-15). Farmers had also to arduously carry cotton loads to the set markets whose loc (...)

80By 1960, most of the people growing cotton for Cotonang in the Baixa, some 31,000 people, were farming at 400 «concentrations» in 21 different «sectors» of the Baixa, with each sector’s lands parceled into reserves based on presumed optimal aptitude for growing cotton.113 Concentration was said to facilitate technical productivity, and control over production, workers, and sales, all deemed necessary to address the accumulating problems of the preceding intensive years 1933-1943, namely increasing pests, soil fertility, disease, low production, and fleeing labor.114 As the commercial cotton agro-industrial economy boomed across the region, such changes in production and settlement occurred together with changing movements of people, goods, money and ideas.

• 115 Decreto 35844 of 1946, Art.3; See also Isaacman (1996: 133-134).
• 116 Decreto 35844 of 1946, §9c.

81Concentration in the Baixa in particular had been pursued as production increased and the colonial state’s Cotton Export Board was created in 1938. A 1946 law tasked the Cotton Export Board with promoting concentration, which together with crop-rotation was promoted as a solution to scarcity and fleeing of labor, as well as lack of food crops. The state Cotton Export Board would delimit which areas within the overall concession zone were suitable and allowed for cotton growing, and Cotonang was obliged follow the Board in promoting concentration accordingly.115 Efforts to concentrate production on «appropriate» soils were further bolstered after an extensive soils survey was done in mid-1954 by a state agronomist with the help of state and Cotonang employees. Colonial prejudice prevented administrators from addressing how concentration to enforce food-crop production would not resolve nutritional problems, and in fact could exacerbate them by increasing traveling times, land clearance and weeding labor, and susceptibility to flooding. For example, in the published official legislation, the Colonial Minister dismissed food scarcity in cotton zones as due to «the naturally lazy and fatalist psychology of bantu people».116

• 117 See Salazar (1968: v1 p.153; v2 p.65).
• 118 Palma, op. cit.; Informação 216.

82A fairly exhaustive household nutrition survey across the Baixa by the President of Angola’s Provincial Nutrition Commission had found that «The separation between the villages and the fields for cultivation creates a modus vivendi on the part of the populations that results in a low nutritional level» (NUNES 1960: 399). However, the report did not criticize the problem of the state designating distant cotton areas, but instead proposed increasing the ongoing concentration of people near designated cotton plots. Such dislocation was simultaneously meaningful and material, as people were uprooted from their homes where they had ritually treated, or planted beneath milolo trees, their children’s placentas so as to ensure their children’s health and intelligence, which were in evident decline.117 Reported instructions during the revolt spoke to the political importance of these specifically domestic aspects of grievances, requiring revolters to «be done with the grass houses» of the sort built as temporary encampments in the cotton concentrations, and consequently avoid the irritating cramped space and changed gendered division of labor in such temporary encampments: «[do] not have utensils and items in the bedroom but rather in the kitchen, another house, or outside», and «only women can cook».118

• 119 See Cotonang (1956) Relatório e Contas.
• 120 Dias dos Santos (1961: 5).
• 121 There major different zones are between Chromic Luvisols (LVx) and Ferralic and Haplic Arenosols (A (...)
• 122 See Angola Norte, November 30, 1957.
• 123 Cotonang (various years) Relatório e Contas.
• 124 Neto, J. (1961) «Apontamento 41: Considerações sobre a cultura algodoeira da Baixa de Cassange», Ma (...)

84The heavy rains in 1960 marked the return of a grievance about forced cultivation in lowlands that had — after similar March 1956 flooding in lowlands (planted after a dry January and February) — prompted hundreds of farmers in Kambo to stop cultivation and/or flee, with the government and Cotonang in turn tightening enforcement and concentration.119 The Cotton Export Board had actually proposed in 1948 that certain zones be dropped (Tembo Aluma, Kwango, and Luremo), but this was rejected by the Malanje Governor.120 These areas of sandy soils were also given the lowest aptitude for cotton by the 1954 soil survey (Gouveia 1956).121 Continued high production was desired to recoup the heavy costs of various new deseeding machines and a massive new 18,000-conto cotton seed oil factory opened in late 1957 in Malanje.122 Soil conservation works were implemented on nearly 30,000 ha of land (cotton and food) in 1957. «Stabilization», said Cotonang in 1957, was the «object of our utmost attention», and again soil conservation works and «fixation to the land» — in tight collaboration with the state — was emphasized in 1958-60.123 These measures to control erosion and concentrate cotton production in flood-prone lowlands helped boost production to an all-time high in 1958, but at the cost of vulnerability to rains in the next two years. The largest drop in production after the 1960 rains was in Kambo (dropping 43%). A report after the revolt noted that farmers did not receive assistance designated by law during the 1960 flooding, and «it was on the margins of the Lui River, which forced producers to do two or three plantings each year, where there were the first incidents [of the revolt]».124

• 125 Antonio Bernardo Cabrita, Report, January 30, 1961, p. 2 (AHU).

85The mixing of material and meaningful, of land, water, and religion, was evident in the way that adhesion to the revolt was reportedly secured through a sort of baptismal pledge in which drops of water were put on people’s heads, then a token payment made, gun shots fired, and a cross drawn on the ground.125 As the pledge holy water dripped over their heads like the rivers spreading across Angola that they knew so well, they would try to use the river water that had last year flooded their fields to their advantage this time, planning destruction of bridges and pontoons in order to impede state troops and reprisals.

• 126 As a rhizome, the sharp imperata re-grows if any root material is left over, requiring prolonged fa (...)
• 127 MacDonald (2004). For example, in 1936, all uprooting and burning was supposed to occur from 20-30 (...)
• 128 Salazar (1968: 86). See also interview with Elisa Kapunga, recounted in Paredes (2015: 216). Senu w (...)

86In addition to susceptibility to flooding, the designation of lowland cotton cultivation areas and the strict planting schedule also entailed heavy labor demands for weeding, another source of grievances. There was thus a combination of structural grievances about long-term hunger produced as women spent increasing time on weeding imperata in the new concentrations) and proximate grievances that December 1960 when men were preparing fields. Being forced to work on heavy lowlands and/or areas infested by the rhizomatic weed imperata could make the work much more arduous than official calculations.126 Conducting participant observation research in which I farmed and conversed with farmers in their fields helped give me a closer understanding of imperata — known locally in Kimbundu as «senu» — and prompted me to be more attentive to imperata in the historical records. Gouveia’s (1956) soil survey of the Baixa mentions imperata 34 times. Imperata increased the heavy labor by men in the initial preparation to clear the weed’s dense root mats, as well as by women in rounds of weeding the imperata shoots that will sprout from any small fragment of the weed’s root mat. Illustrative of its importance, a known saying at time held that «Seno land is cotton land» (SALAZAR 1958: vol. 1, p. 32). Colonial and Cotonang administrators may have helped spread imperata by requiring annual burning of cotton plant residues because fires stimulate imperata growth. Administrators required annual burnings of cotton plant stalks in order to reduce pest accumulation in the monocropped fields, and such burnings also required strict adherence to state established planting calendars so as to prevent damage to any still unharvested cotton — it was such a strict calendar that also increased the susceptibility to flooding of early planted fields.127 It is hardly any coincidence that instructions for revolters in the Baixa de Kassanje were to use the tough stems of imperata to braid belts and bracelets that would help ward off bullets (in conjunction with holy water and incantations of Maria).128

• 129 Perreira, A. (1961) «Informação», Cotonang-Sector de Cunda Ria Base, February 2 (AHU). Also later s (...)
• 130 Palma, op. cit. This quote could perhaps also refer to encounters with colonial administrators, but (...)
• 131 Martins, «Apontamento», op. cit, p. 41. See also Karaje (2006).

88We also see revolters reject this distinctly patriarchal sort of Portuguese authority. Whereas an initial administrator trying to calm revolters referred to them patronizingly as «my children», the revolt included proliferating shouts and chants debasing such authority by proclaiming the Portuguese state was actually just «bird shit» (Maniputo, tuge ia gingila).129 Revolt reports seemed taken aback by, and critical of, the unusual assertiveness and lack of obedience by revolters, who are described as threatening, offending, and disrespecting. Relatedly, and perhaps also pertaining to domestic conflicts over different labor contributions to obligatory cotton quotas, revolt prescriptions included an emboldened roles for women in violent encounters: «If a man hits a woman, she will do the same to him and will say reconciliation and peace accomplished».130 Moreover, the sexual assaults by colonial administrators and African functionaries that were used as punishment or abuse of discretionary authority were also inverted as some revolters threatened administrators’ wives in Kunda dya Baze.131

• 132 The posts were Kwango and Luremo (in the Circunscrição of Kamaxilo), Kambo, Brito Godins, and Tembo (...)

90Similarly, Cotonang’s actual concession did not begin as a discrete unit, but rather entailed a transformation from the 1926 legislation that allowed sparse ginning factories with rights to the surrounding areas in a maximum 25 km radius (later 60 km) (see Figure 12 above). As Cotonang cotton production expanded whilst global markets declined during the 1930s, the state emphasized the need for closer regulation of cotton quality, production, and labor. The Portuguese state wanted to encourage investment, and yet balance concerns about unfavorable monopolistic practices and also national versus foreign capture of economic opportunities (CLARENCE-SMITH 1985). From this early period onward, there was increasingly close collaboration between Cotonang and the Portuguese administrative apparatus in the Baixa de Cassange. The multiple fragmented concessions were re-organized in 1940 from scattered factory-radius zones to concessions that matched administrative districts. Then, by 1947, these were united into what was subsequently known as a singular concession consisting of 12 administrative post areas that covered most of the Baixa de Cassange.¹³² Cotonang consistently emphasized what it referred to as its «collaboration with the authorities», and dozens of detailed pieces of legislation spelled out detailed roles for the state in shaping and participating in Cotonang’s activities.

• 133 Freudenthal (1999: 253) cites Marcum (1969: 124-126) as suggesting cotton prices declined, however (...)
• 134 Of note is that Ernesto Vilhena, through the Banco Burnay, held large financial stakes and influent (...)

91In other words then, the revolt was about the intense and extensive presence of the state in conjunction with the company in the Baixa de Cassange, and the particular spatial transformations in production, rather than simply low or declining prices paid by an abusive concessionaire. Indeed the actual official nominal producer prices paid to farmers for cotton do not appear to have been declining, though perhaps these official prices were not being paid by Cotonang agents since the real world and colonial prices had been declining (see Figure 16 below).¹³³ Moreover, rather than a confined enclave, Cotonang had a regional and international presence — in the shareholders that had direct interests in infrastructure, mining, commerce, etc in the region, in the actual facilities in the cities of Malanje and Luanda, and in the network of cotton business experts in neighboring Congo and elsewhere in Africa and the world.134

• 135 Administrators and Chefes de Circunscrições would receive 0.02 Angolars for each kilogram of unproc (...)
• 136 Neto, «Apontamento 41», op. cit., and «Algumas considerações acerca da Zona Algodoeira da Baixa do (...)
• 137 Letter, Alberto M. Quadros, February 6, 1961, p. 2 (AHU); Martins, «Apontamento 34», op. cit., p. 2 (...)

93What was key was that state agents had personal financial incentives to boost production, which they did by various means, because by law they received salary bonuses based on levels of cotton production.¹³⁵ So ostentatious were some state agents’ conspicuously consumption — evident in ostentatious new houses and cruising cars — based on these cotton bonuses (which could effectively double their salaries) that they were criticized by military commanders called in to quell the revolt.136 And the benefits to state agents working in the land of «white gold» had long been recognized. The Malanje Governor noted in 1935 for example that state agents’ «greater remuneration obtained in the area of [cotton] intendancy [...] influences the interest with which the authorities follow the march of agricultural operations, and it benefits production» (26). The bonuses prompted greater numbers of functionaries with better qualifications to want to work there, putting strains on other areas, becoming «so flagrant that it could collapse the whole system [...]» In early January, 1961, villagers responded that they would not send people for public works, and «for a long time they had come work for the State, but now they would not do it».137

• 138 Ferreira (1948); Georges (1955, 1960).

94Cotonang was also required by law to collaborate with the JEA and to fulfill a number of social obligations and labor protection requirements. Informally, Cotonang actively and intensely cultivated high-level political relationships in Lisbon and Luanda (particularly given the emphasis on national industries), but also with ground-level administrators in the cotton growing areas. In the city of Malanje, Cotonang was close with the commercial associations, with the provincial newspaper, and emphasized its paternal relations in «the Cotonang family».¹³⁸

• 139 Explicit post-World War II theorizations of «growth poles» was done most notably in France by the e (...)

96Understanding the extensive connections that gave rise to the Kassanje revolt is also important to understanding the subsequent late-colonial spatial development programs that have in turn shaped contemporary reconstruction. Many of the plans for Angola’s post-2002 reconstruction of infrastructure were outlined in a massive multi-volume 2025 Long Term Development Strategy, which often drew on comparisons with the late colonial era conditions and regional development plans (cf. POWER 2012). These contemporary projects draw partly on Angolan colonial regional development planning which in turn drew on various currents of thinking, some linked through Brazil to classic French theories, as well as the emerging experience with hydroelectric irrigation projects across Africa. ^{But, in particular, it was after the 1961 revolts in Kassanje}, in urban Luanda, and on the coffee plantations in the north, that the spatial experiences of and plans for Angola were laid out in great detail.139

• 140 The study was published as Marques (1964), though some of the basic parameters of planning and fina (...)

97Regional economic development inventories and plans were compiled in 1962 in a significant two-volume study by the influential Portuguese economist and planner Walter Marques, then working in the office of Angola’s top army commander. Marques subsequently became Angola’s Provincial Secretary for Finance and proceeded to push through funding for the planned projects.¹⁴⁰ In a systematic fashion his study inventoried the plans for regional poles and linkages in Angolan development. The study and its approaches were key parts of executing a broader counter-insurgency strategy that was geared around a multi-pronged approach implemented primarily by small local mobile units.

• 141 See Pinto and dos Santos (1961), Pinto (1959), dos Santos (1960). Perroux’s student, Jaques Boudevi (...)
• 142 The key points, according to Cann’s (1997) account (rather apologetic), that the Portuguese came to (...)
• 143 Tembo Aluma and Kambo were swiftly reorganized administratively. See Portaria 12195 of April 18, 19 (...)

98Development poles were part of a broader colonial state spatial program of regional development for counter-insurgency.¹⁴¹ This program — directly influenced by British and French colonial experience — had as one key plank «social development», as formally outlined in the guiding document, «The Army in Subversive War».¹⁴² In addition to psycho-social action involving culture and propaganda, the social plank emphasized four dimensions: (1) education, (2) sanitary assistance, (3) economic development in agriculture and cattle husbandry, and (4) local infrastructure improvements. Administrative presence throughout the country was significantly increased through increased numbers of posts, and more numerous smaller administrative districts (see Figure 17).143

• 144 See Rebocho Vaz (1993).

100The importance of the overlap between military planning and post-1961 regional economic development is also illustrated by the fact that the governor of Angola for the key period 1966-1972 was Lieutenant Coronel Camilo Rebocho Vaz, who was promoted to that position after being the on-the-ground commander of the repression of the 1961 revolts in the Baixa de Cassange and northern coffee areas (and subsequently being the regional administrator of the rebellious Uíje northern coffee district from 1961 to 1966).¹⁴⁴

• 145 See also Cooper (2014).

104Simply because some parts of the whole complex revolt did not conform to what were later forms of Angolan nationalism does not mean that the revolt did not have nationalist aspects, that it was «only anti-colonial».145 It was a mix, but there were direct links with what many people see as the main forms of modern nationalism, and there were aspects of the revolt that were unequivocally clear about political demands for independence for a new supra-ethnic entity in the form of an Angolan nation-state. The above analysis has traced the variety of processes that conjoined to produce the revolt, and consequently the revolt was not always entirely coherent, it was sometimes partial and contradictory, and much remains occluded and yet to be researched.

• 146 J. M. Pereira de Carvalho, Letter #543, February 27, 1961 (AHU). According to Marcum, citing Neto ( (...)
• 147 See Apontamento 14, op cit, p. See Vicente’s report 16th of January «Considerações sôbre a situação (...)
• 148 J. M. Pereira de Carvalho, «Situação na Baixa de Cassange», Letter #250, January 27, 1951 (AHU).
• 149 «Informação 216», op. cit.
• 150 J. M. Pereira de Carvalho, Letter #543, February 27, 1961 (AHU).
• 151 «Informação 144», February 6, 1961 (AHU).

105Nonetheless, from the colonial reports we can see that several key forms of the modern nation-state were invoked during the revolt (in addition to the notions directly linked to UPA and the MPLA), including a new president, capital, government, symbology, and common identity. The new president would be called Kulashingo (which was the old title of the King of the Bangalas).146 There are repeated reports that the revolters were asserting that their own government held power, not the Portuguese.147 Sometimes, contradictorily, this was conflated or merged with the Congo, as in already having «their government headed by Kasavububu [sic]».148 Soba Kivota was reported to have said that he and his community «wouldn’t pay taxes or follow orders because they already had "their State", which was "Pessa"».149 There was mention of a new capital being at Kasule Kwenda.150 The Portuguese flag at Marimba was cut down. Some prominent national symbolism distinct from colonial Portugal seems to have been largely that of PSA, but there are also vague references in the colonial reports to banners and flags. As one report noted, amidst a crowd of 600 armed people in the border town Tembo Aluma at the end of January, «one of the indigenous exhibited, as if it was a flag, a 1960 calendar edited by PSA».151

• 152 Maia (1994).
• 153 Palma, op. cit.; Salazar (1968).

106Since there are hardly any Kimbundu statements transcribed in known colonial reports on the revolt, it is hard to tell to what extent the revolters used the language of «nation» or «nação», which has loose translation in Kimbundu as people («muntu»), or people of the land («akwa’xi»).152 What does come across clearly though is that the revolters thought of themselves as Angolans, as having a broad new common non-local identity going beyond older ethnic identifications and distinct from the Portuguese. Maria «had been concerned about the suffering of Angolans», and was going to «save the people of Angola». The revolters «were all brothers», it had «become necessary that all Angolans initiate themselves in their religion and adhesion to the liberation movement», and Angola and Portugal would be separated forever.153 The revolt managed to mobilize a significant portion of population, tens of thousands of people, across a large area about half the size of Portugal, who were not asking for reform, not asking for return to old, but rather stating that they were their own entity, if albeit a still unclear one, and one prevented by sheer overwhelming Portuguese firepower.