¹ Third-world military spending dropped from a peak of $155,000 million in 1984 to $123,000 million in 1990 according to the United Nations Development Programme, Human Development Report, 1992 (New York, 1992), p. 85.Google Scholar However, this level of spending is still high — in contrast, the U.N. peace-keeping budget for 1992 was less that $3,000 million, as confirmed by Ghali, Boutros Boutros, An Agenda for Peace: Report of the Secretary-General to the Security Council (New York, 1992), p. 28.Google Scholar

² This point is made by Tilly, Charles, Coercion, Capital, and European States, AD 990–1992 (Cambridge, MA, 1992), p. 227.Google Scholar Most publications on the ending of the cold war have focused on the implications for either the United States and/or Europe — see, for example, Mearsheimer, John, ‘Back to the Future’, in International Security (Cambridge, MA), 15, 1, Summer 1990, pp. 5–56;Google ScholarChase, James, ‘The Consequences of the Peace: the new internationalism and American foreign policy (Oxford, 1992);Google Scholar and Gaddis, John Lewis, The United States and the End of the Cold War (Oxford, 1992).Google Scholar The book by Tilly, and that by Chomsky, Noam, Deterring Democracy (London, 1991), are exceptions in that they focus, at least in part, on the Third World.Google Scholar

³ A Mexican President was supposed to have declared, ‘Poor Mexico! So far from God, and so close to the United States!’. Skidmore, Thomas and Smith, Peter H., Modern Latin America (New York, 1992), p. 221.Google Scholar

⁴ Most studies of Angola have interpreted its ethno-cultural groups as primordial identities that endured without change from pre-colonial times. This kind of analysis results in the tautology that Angolans, like other Africans, are tribalistic because they are a tribal people. For an alternative view that examines the social construction of ethnic identity in historical perspective, see the introduction in Vail, Leroy (ed.), The Creation of Tribalism in Southern Africa (London, Berkeley, and Los Angeles, 1989), pp. 1–19.Google Scholar See also Mafeje, Archie, ‘The Ideology of “Tribalism”’, in The Journal of Modern African Studies (Cambridge), 9, 2, 08 1971, pp. 253–61,CrossRefGoogle Scholar and Adam, Heribert and Moodley, Kogila, The Opening of the Apartheid Mind: options for the new South Africa (Berkeley, Los Angeles, London, 1993), pp. 2–3.Google Scholar

⁵ In this respect, Angola is no different from most other nations in Africa and elsewhere that have begun as an incongruous composite of people, whose identity is then homogenised by deiberate state policy. Charles Tilly calls the latter state-led nationalism, and distinguishes it from state-seeking nationalism, in which a people first form a national identity and then demand a state. See ‘Past, Present, and Future Nationalisms’, in The New School Commentator (New York), 4, 2, 11 1992, p. 3. State-seeking nationalism has so far been fairly unimportant in Africa, although the recent formation of Eritrea is an exception.Google Scholar

⁶ The problem of how to unify disparate peoples is not unique to Angola. Eric Hobsbawm writes of Europe that ‘any nation of even middling size had to construct its unity on the basis of evident disparity’; see Nations and Nationalism Since 1780: programme, myth, reality (Cambridge, 1990), p. 91. Early in their history, European nations were no more united than many African ones today; only the passage of time and the propagation of nationalist myths by the state created such unity. Furthermore, unity is not guaranteed, as regionalist movements in the Basque country, Scotland, northern Italy, and other parts of Europe show.Google Scholar

⁷ Marcum, John, The Angolan Revolution, Vol. I, The Anatomy of an Explosion, 1950–1962 (Cambridge, MA, 1969), p. 5.Google Scholar

⁸ Although the offer resulted in concessions being offered to Jewish colonists, the provisions of the bill passed in June 1912 did not meet the conditions of the Zionist organisation. See ibid. pp. 3–4.

⁹ In contrast, armed resistance to the French in Côte d'Ivoire was ended by 1917. See Zolberg, Aristide R., One-Party Government in the Ivory Coast (Princeton, 1964), p. 22.CrossRefGoogle Scholar

¹⁰ Quoted in Bender, Gerald J., Angola under the Portuguese (Berkeley, 1978), p. 6.Google Scholar For a discussion of this kind of parasitism on an individual level, see Morrison, Toni, Playing in the Dark: whiteness and the literary imagination (Cambridge, MA, 1992), pp. 42–4, where she refers to the memoirs of William Dunbar, an eighteenth-century Scottish nobleman who professed that he did not feel truly free until he came to the New World and ruled over a plantation worked by slaves.Google Scholar

¹¹ Leys, Colin, Underdevelopment in Kenya: the political economy of neo-colonialism, 1964–1971 (London, 1975), p. 29. It should be noted that the white population in the country as a whole was somewhat higher than this.Google Scholar

¹² Humbaraci, Arslan, Algeria: a revolution that failed (New York, 1966), p. 4.Google Scholar

¹³ Bender, op. cit. p. 228. This author also reports on pp. 230–3 that 56 per cent of Portuguese emigrants (aged 7+) to Angola in the 1964–72 period had no formal education, while another 28 per cent had only one to four years of schooling. In 1968, the 1, 205 professionals in Angola represented less than 0·5 per cent of the total white population, most of whom were ‘petits blancs’, over two-thirds being engaged in commerce.

¹⁴ Presentation by Gerald Bender in Luanda, Angola, 24 September 1992.

¹⁵ Data from Marcum, op. cit. p. 5, and Bender, op. cit. p. 151. A further indication of the relative underdevelopment of Angola's educational system is revealed by the fact that 1967 was the first year for more than 25 when the money allocated for education exceeded that for white settlement. Even so, according to Bender, op. cit. p. 221, the amount – around $11 million – was less than the British had spent on education in Ghana in 1950.

¹⁶ Ibid. p. 221.

¹⁷ Vail, (ed.), op. cit. p. ix, makes the point that nationalism in much of Africa ‘had been a basically negative force, directed against colonialism, with little positive vision about the nature of the new society after colonialism's demise’.Google Scholar

¹⁸ Tilly, op. cit.

¹⁹ Scott, James, Domination and the Arts of Resistance: hidden transcripts (New Haven, CT, 1990), pp. 21–2. I would add, as an implicit corollary of this author's insight, that differences in colonial forms in Africa produced corresponding differences in nationalist reactions.Google Scholar

²⁰ I follow Eric Hobsbawn – who follows Gellner, Ernest, Nations and Nationalism (Oxford, 1983), – in defining nationalism as ‘primarily a principle which holds that the political and national unit should be congruent’. Proto-nationalisms are ‘feelings of collective belonging’ which could be used by states and nationalist movements to mobilise people, notably language, ethnicity, religion, and race. See Hobsbawn, op. cit. pp. 9 and 46.Google Scholar

²¹ According to Cabral, Amilcar, ‘Identity and Dignity in the National Liberation Struggle’, Lincoln University, Pennsylvania, 15 10 1972Google Scholar: ‘The masses in the rural areas and the larger section of the urban population… are untouched, or almost untouched, by the culture of the colonial power. This is partially the result of the…effectiveness of the cultural resistance of the people, who, subjected to political domination and colonial exploitation, find their own culture acts as a bulwark in preserving their identity’. Quoted in Bender, op. cit. p. 217.

²² See Zolberg, Aristide, ‘Angola’, in Zolberg, , Astri, Suhrke, and Aguayo, Sergio (eds.), Escape from Violence (New York, 1989), p. 76. The Angola case shows that the forces thought by modernisation theorists to promote national integration (capitalist development, social mobilisation, increased communication between regions) can actually increase feelings of hatred between sub-national groups. Certainly, labour migration from the Ovimbundu heartland of Huambo and Bié to the plantations of the north did much to crystallise the separate, anti-Luanda ethnic identity upon which Unita's mobilisation of support has been based.Google Scholar

²³ See Messiant's, Christine excellent analysis, ‘Social and Political Background to “Democratization” and the Peace Process in Angola’, in Eduardo Mondlane Foundation et al., Democratization in Angola (Amsterdam/Leiden, 1993), pp. 16–17.Google Scholar

²⁴ According to Wolfers, Michael and Bergerol, Jane, Angola in the Frontline (London, 1983), p. 109, the Ovimbundu provinces of Huambo and Bié were providing over 95,000 registered contract workers to various places outside the region in 1971Google Scholar. These two provinces provided the lion's share of the Unita/Savimbi votes in the 1992 elections, as may be seen from Maps 2–3.

²⁵ Cf. Brubaker, Rogers, Citizenship and Nationhood in France and Germany (Cambridge, MA, 1992).Google Scholar

²⁶ The class structure of pre-1975 Angola is taken from Sogge, David, Sustainable Peace: Angola's recovery (Harare, Southern African Research and Documentation Centre, 1992).Google Scholar

²⁷ For an insight into the politics of Angolan independence, as seen by Angolan students in the Portuguese capital during 1961, see the novel by Pepetela, , A Geraçāo da Utopia (Lisbon, 1992), pp. 11–118.Google Scholar

²⁸ In 1947, as High Inspector of Colonial Affairs, Captain Henrique Galvāo presented a report to the National Assembly that was highly critical of Portuguese policy in its African colonies, and in 1961 he hijacked a Portuguese luxury liner for ten days, as the first step of an unsuccessful plan to overthrow the régimes of Antonio de Oliveira Salazar and Francisco Franco in Portugal and Spain. His story has been translated into English as Santa Maria: my crusade for Portugal (New York, 1961); see pp. 57–71 for extracts from his 1947 report.Google Scholar

²⁹ Cf. Young, Crawford, Politics in the Congo (Princeton, 1965), pp. 307–43,CrossRefGoogle Scholar and O'Brien, Conor Cruise, To Katanga and Back: a U.N case history (London, 1962).Google Scholar

³⁰ For a memorable first-hand account of Luanda during the Portuguese exodus, see Kapuscińśki, Ryszard, Another Day of Life (New York, 1987).Google Scholar

³¹ Licklider, Roy, ‘How Civil Wars Have Ended Since 1945’, New School Proseminar on State Formation and Collective Action, New York, 10 11 1992, p. 5. Another source counts 125 wars and conflicts, leading to 40 million deaths, in the developing world from 1945 to 1990.Google Scholar See McNamara, Robert S., ‘The Post-Cold War World: implications for military expenditure in the developing countries’, in Proceedings of the World Bank Annual Conference on Development Economics (Washington, DC, 1991), p. 97.Google Scholar

³² Ibid. pp. 95 and 103.

³³ Sogge, op. cit. pp. 117–18.

³⁴ In another study, military expenditure was also shown as being four times larger than that for health and education; Mc Namara, loc. cit. p. 122.

³⁵ Sivard, Ruth Leger, World Military and Social Expenditures, 1991 (Washington, DC, 1991).Google Scholar

³⁶ According to Howard, Bertie, ‘Angolans Revel in First Free Vote’, in Africa News (Durham, NC), 37, 2, 10 1992, p. 7, the United States provided over $300 million in covert funding to Unita between 1985 and 1992.Google Scholar

³⁷ Sogge, op. cit. p. 27.

³⁸ For a critique of U.S. policy towards Angola, written by a former C.I.A. agent, see Stockwell, John, In Search of Enemies: a CIA story (New York, 1978).Google Scholar For recent analyses of U.S. policy towards Angola, see McFaul, Michael, ‘Rethinking the “Reagan Doctrine” in Angola’, in International Security (Cambridge, MA), 14, 3, Winter 1989/1990, pp. 99–135,Google Scholar and Tvedten, Inge, ‘U.S. Policy Towards Angola Since 1975’, in The Journal of Modern African Studies, 30, 1, 03 1992, pp. 31–52.CrossRefGoogle Scholar

³⁹ According to Wright, George, ‘The US and Angola’, in Z (Boston), 05–06 1992, p. 18, ‘The goal was to weaken the various regimes, thus forcing them to capitulate to Western political and economic formulas and policies, including “multiparty democracy”, and restructuring the economy to a free-market system and to an “open door” for foreign capital.’Google Scholar

⁴⁰ The World Bank estimated that in 1985 only 37 per Cent of the economically active population of Angola was to be found in the rural areas, down from 74 per cent in 1970. Sogge, op. cit. p. 121.

⁴¹ The withdrawal of Cuban teachers and health personnel in the early 1990s affected these sectors particularly badly.

⁴² Messiant, loc. cit. pp. 20–3.

⁴³ Ibid. pp. 25–6.

⁴⁴ For information on human rights abuses, see Watch, Africa, Angola: violations of the laws of war on both sides (New York, 1989),Google Scholar and Amnesty International, Angola: an appeal for prompt action to protect human rights (New York, 1992).Google Scholar In a poll of almost 4,000 people conducted in Angola in 1992, 78 per cent of those who claimed to have been the victims of violence said that Unita was responsible as against 18 per cent who blamed the M.P.L.A. Queiroz, Artur, ‘Larga Vantagem Para Dos Santos e MPLA’, in Sabado (Luanda), 5, 224, 25 09–1 10 1992, pp. 96–7.Google Scholar

⁴⁵ A revealing glimpse into the tightly controlled apparatus of Unita was revealed after its vicepresident, Jeremias Chitunda, had been killed in Luanda on 1 November 1992. According to the diaries found on his body, Chitunda had endured physical punishment on several occasions for disagreeing with Savimbi, often referred to as ‘the master’, ‘the father’, and ‘the oldest one’ (mais velho), and had once nearly been executed. Foreign Broadcast Information Service, Daily Report Supplement/Sub-Saharan Africa/Angola (Washington, DC), 29 01 1993.Google Scholar See also Jamba, Sousa, Patriotas (Lisbon, 1990), a novel about an Angolan exile living in Zambia who joins the rebels but eventually grows disillusioned with the cause.Google Scholar

⁴⁶ The U.N. Verification Mission in Angola had 350 military and 126 police observers, while an additional 400 arrived for the elections, in stark contrast to the more than 7,000 U.N. peacekeepers who were deployed in Cambodia in 1993.Google Scholar

⁴⁷ Messiant, loc. cit. p. 24.

⁴⁸ It was in order to avoid such a tricky proposition that in Cambodia the rival parties agreed in 1992 not to elect a President. Instead, the voters elected in May 1993 representatives to a constituent assembly, which was then responsible for writing a constitution and forming a government.

⁴⁹ For example, as in Nicaragua, the U.S. Administration authorised lavish expenditures on war, but gave little for peaceful reconstruction. It is also highly improbable, even if it were desirable, that the rich countries will create a trusteeship over Angola with the idea of restoring governability as suggested by Paul Johnson, ‘Colonialism's Back – and Not a Moment Too Soon’, in the New York Times Magazine, 18 April 1993, p. 44.Google Scholar

⁵⁰ Messiant, loc. cit. p. 29.

⁵¹ Bayer, Tom, Angola. Presidential and Legislative Elections, September 29–30, 1992. Report of the IFES Observation Mission (Washington, DC, 1993), pp. 6–7.Google Scholar

⁵² Wallerstein, Immanuel, ‘On the Interpretation of Nationalism in the Periphery: Marcum's Angola’, in Africa Today (Denver), 26, 4, 1979, p. 71.Google Scholar

⁵³ Hobsbawm distinguishes in op. cit. p. 22, between two concepts of the nation: (i) the revolutionary–democratic sees the state creating the nation, by granting common citizenship rights to a ‘people’, while (ii) the nationalist sees the nation creating the state: the prior existence of a community that is different from ‘foreigners’ requires the former to have its own state. The M.P.L.A. view of the nation is revolutionary–democratic, while Unita's is nationalist.

⁵⁴ The last census to be published was carried out in 1960, although the Portuguese held another in 1970. Since then only partial headcounts and estimates have been possible, and in the meantime, the population of Angola is thought to have doubled to around 10 to 12 million.

⁵⁵ Anderson, Benedict, Imagined Communities: reflections on the origin and spread of nationalism (London, 1991), pp. 163–4.Google Scholar

⁵⁶ The Angolan/Brazilian research team based its portrait of the national electorate on a partial census of 1983–4 and a government survey carried out between August and November 1991. See Queiroz, loc. cit. pp. 96–7.Google Scholar

⁵⁷ Ibid. The survey was said to show that the average M.P.L.A. voter tended to be Catholic, young, female, and relatively well-educated, whereas the average Unita voter tended to be Protestant, old, male, poorly educated, with Umbundu as a mother tongue. Some aspects of these profiles were confirmed by the election results.

⁵⁸ The U.N. Statement issued on 17 October 1992 said that ‘while there were certainly some irregularities in the electoral process, these appear to have been mainly due to human error and inexperience’, and that there was ‘no evidence of major, systematic or widespread fraud’. According to the Washington-based International Foundation for Electoral Systems, the elections ‘constituted a proper and effective application of the mechanism of elective choice of political representatives’. Bayer, op. cit. appendix F, p. 3.Google Scholar

⁵⁹ However, the returns from Cabinda, where the M.P.L.A. received a high percentage of votes, are not indicative of its support there. Only 16,000 of an estimated 84,000 eligible voters actually registered in that province (most of them government soldiers), presumably because of a secessionist movement's threats and its exhortations not to co-operate with the régime in Luanda.Google Scholar

⁶⁰ This may have been due in part to the desire of many Angolans to obtain proof of citizenship with which to bolster claims on the state. All those who registered to vote obtained an identification card with a photograph, a valuable item in a country in which migration flows put into question the nationality of many people.Google Scholar

⁶¹ The proportion of spoiled and blank ballots was considerably lower in the provinces of Benguela, Huambo, Luanda, Cabinda, and Kuando Kubango, where the majority of voters were undoubtedly exposed to considerable party political propaganda. Bayer, op. cit. p. 58.Google Scholar

⁶² Campbell, Horace, ‘Angolan women and the Electoral Process in Angola, 1992’, in Africa Development (Dakar), 18, 2, 1993, p. 58.Google Scholar

⁶³ This analysis of core regions of support for each party is taken from Gerald Bender, who is referred to by Marcum, John in his article, ‘Angola: war again’, in Current History (Philadelphia), 05 1993, p. 220.Google Scholar

⁶⁴ All calculations in this section were made by the author on the basis of final election results released by Angola's National Electoral Council. Maps 2 and 3 are based on the same source. I thank Tom Bayer of the International Foundation for Electoral Systems for supplying me with the relevant data.Google Scholar

⁶⁵ In the provinces of Luanda, Kwanza Norte, Melange, Benguela, Huambo, Bié, Kuando Kubango, and Bengo, the aggregate vote for Savimbi was 1,184,394 compared to 1,110,209 for dos Santos.Google Scholar

⁶⁶ Marcum, loc. cit. p. 220, understates the margin of the M.P.L.A.'s victory in the hinterland in suggesting that Unita gained 578,000 votes in these provinces. Although all other electoral results cited in his article match mine, this appears to be a mistake. My own figure is 317,395.

⁶⁷ Registered voters in the M.P.L.A.'s four core provinces (Luanda, Bengo, Kwanza Norte, and Malange) numbered 1,412,201, or 29·3 per cent of the national total, whereas Unita's four core provinces (Benguela, Huambo, Bié, and Kuando Kubango) listed 1,523,334 registered voters, 31·6 per cent of the total.Google Scholar

⁶⁸ If Luanda had been subtracted from the vote totals, the M.P.L.A. would have defeated Unita in the national assembly elections by a count of 1,592,832 to 1,206,677, but by only 1,420,767 to 1,407,555 in the presidential elections.

⁶⁹ In winning an average of 24 per cent of the two-party vote in Unita's four ‘home’ provinces, the M.P.L.A. would have had to receive some Ovimbundu votes. Nevertheless, in the absence of electoral data at the level of the polling place, detailed correlations between ethnicity and the vote are impossible to make. Such figures do exist – they were appropriately tabulated and entered into computers at the National Electoral Council in Luanda – but have not been published, probably to prevent them from being used as the basis of reprisals by either of the two armies. The available data still await a detailed regression analysis.Google Scholar

⁷⁰ Savimbi described the central highlands of Huambo, Bie, eastern Benguela, Kwanza Sul, and northern Huila provinces as an area peopled by ‘the majority in Angola’ that wants ‘the influence that it deserves’, that feels ‘Humiliated and rejected’, and that wishes ‘to be represented by our own representatives’. U.S. Foreign Broadcast Information Service, Southern Africa/Angola, ‘Savimbi Addresses Nation From Huambo’, 10 March 1993, p. 19.Google Scholar

⁷¹ The conflation of any party with the state is problematic for the process of democratisation, as recognised by Apter, David E., ‘Some Reflections on the Role of a Political Opposition in New Nations’, in Some Conceptual Approaches to the Study of Modernization (Englewood Cliffs, NJ, 1968), p. 83: ‘When the distinction between government and party breaks down, the representative government is at end, because embedded in the idea of democratic government is the concept that the party is a conveyer of the people's will through the institutions of government, but is not the repository of state power.’Google Scholar

⁷² In contrast, Biafran separatism in 1967 appeared viable to many of its supporters because of the existence of oil in Iboland. The extent to which geography can influence whether or not nationalist movements accept pre-existing national boundaries (centripetal), or are separatist (centrifugal), does not seem to have been dealt with in the relevant literature.Google Scholar

⁷³ See Kundera, Milan, The Book of Laughter and Forgetting (Harmondsworth, 1986), p. 197Google Scholar: ‘We are all prisoners of a rigid conception of what is important and what is not. We anxiously follow what we suppose to be important, while what we suppose to be unimportant wages guerrilla warfare behind our backs, transforming the world without our knowledge and eventually mounting a surprise attack on us.’ The surprising finding that national integration and homogenisation seems to have been taking place in Angola throughout the 16 years of internationalised civil war confirms the truth of this observation.

⁷⁴ Minter, William, a scholar at the Washington Office on Africa, feels that this figure is too high. Personal communication, September 1993.Google Scholar

⁷⁵ Sogge, agrees with such an analysis in op. cit. p. 141: ‘a basis for a national consciousness of Angolanidade has been built to no small degree. On that could be founded social consensus about a project of nation-building. If that holds, then norms of governance which embrace everyone, regardless of background or social standing, may meet no insuperable resistance.’Google Scholar

⁷⁶ Jamba, , op. cit. pp. 84–5. The author, who was educated in Zambia, wrote the novel in English, and hence my translation of the Portuguese version is unlikely to be the exact equivalent of the dialogue in the original English-language edition.Google Scholar

⁷⁷ Pepetela, , A Geraçāo da Utopia (Lisbon, 1992).Google Scholar

⁷⁸ Nationalist sentiments may not always be so benign. For example, in January 1993, angry crowds in Luanda attacked and killed a number of Zaïrians because of reports that troops from their country had helped Unita. Significantly, the victims of this violence were identified by nationality rather than ethnicity. From ‘62 Dead After Angolan Rampage of Rape and Murder of Zaïrians’, in The New York Times, 24 January 1993.Google Scholar

⁷⁹ The decline of nationalism in the contemporary era has been suggested by Tilly, loc. cit. p. 4, and by Hobsbawm, op. cit. pp. 181–3.Google Scholar

⁸⁰ See Cobbah, Josiah A. M., ‘Toward a Geography of Peace in Africa: redefining sub-state self-determination rights’, in Johnston, R. J. et al. (eds.), National Self-Determination and Political Geography (London, 1988). p. 83: ‘in order to cut down on ethnic strife this chapter has called for a reinterpretation of the concept of self-determination in a manner that will grant to geoethnic groups considerable control over their local affairs and ensure a fair share of state resources. To achieve this self-determination, a policy of consociation is recommended, one which will respect the existence and equality of geotechnic groupings, and strive to turn the natural cleavages between ethnic groupings into sources of conflict resolution rather than perceiving these cleavages as a threat to the state.’Google Scholar

⁸¹ See Hartlyn, Jonathan, The Politics of Coalition Rule in Colombia (Cambridge, 1988).Google Scholar

⁸² See, for example, Crocker, Chester, ‘Angola: can this outrageous spectacle be stopped?’, in The Washington Post, 13 10 1993, and the distorted piece by Margaret Calhoun, ‘A False Solution to Angola's Messy War’, in The Wall Street Journal, European edn., 16 June 1993.Google Scholar