The following people are to be thanked, without implication of responsibility, for their helpful comments on an earlier draft of this paper: David Booth, W. Morris Jones, Bruce Matthews, Eric Meyer, V. Ramakrishnan and, above all, Jonathan Spencer.

¹ Most of those who were active in the recent JVP campaign, including the entire leadership, are dead. The JVP will never be able to tell its own story of the events of the late 1980s. The account provided here relies on several other sources. A specialist on Sri Lankan politics, I was resident in Colombo for three months in mid-1989 at the peak of the JVP's campaign and of the armed forces' counter-offensive. It would be difficult, and in many cases unwise, to mention by name all the people who provided useful information. A special debt is acknowledged to: the first-class journalism of Gunasekera, Rohan of The Island newspaper (Colombo);Google Scholar and to the written accounts of the JVP produced by Chandraprema, C. A. (Sri Lanka: The Years of Terror. The JVP Insurrection 1987–1989, Colombo, 1991)Google Scholar and Gunaratna, Rohan, (Sri Lanka: A Lost Revolution? The inside Story of the JVP, Kandy, 1990).Google Scholar Gunaratna's work was produced very rapidly, is heavily descriptive, and depends almost exclusively on military intelligence sources. While useful, and not in fact an anti-JVP tract, it cannot always be assumed to be fully reliable. Chandraprema is much more analytical and insightful, and had a more balanced portfolio of sources of information. It is, however, possible that his history as one of the most active and long-standing political and ideological opponents of the JVP has coloured his interpretation.

² It seems likely that ‘only’ two or three thousand JVPers or suspects were killed in the repression of the 1971 insurgency; six thousand is the maximum likely figure. By contrast, forty thousand is a likely order of magnitude for the recent killings.

³ In most cases I refer in this paper to ‘Tamil separatists’ or ‘Tamil guerillas’. Largely as a result of internecine conflict, one particular organization, the Liberation Tigers of Tamil Eelam (the LTTE, or ‘Tamil Tigers’) emerged as dominant, and was the only organization to fight against the ‘Indian Peacekeeping Force’ when the latter occupied the Tamil areas between August 1987 and March 1990. Where appropriate, explicit reference is made to the LTTE.

⁴ The full story is, naturally, a little more complex than this. The ideological and organizational heartland of Tamil separatism is the Jaffna peninsula, which has a relatively dense population and depends more heavily than other Tamil areas on non-agricultural sources of income. In addition, by 1991 the Tamil Tigers had developed a considerable capacity to wage conventional warfare against the Sri Lankan armed forces. They had graduated from guerilla status to an army territorial responsibilities.

⁵ For a good review of the literature, see Goldstone, J. A., ‘Theories of Revolution: The Third Generation,’ World Politics 32, 3 (1980), pp. 425–53.CrossRefGoogle Scholar

⁶ Ibid.

⁷ This latter perspective is epitomized in the substantial debate, triggered by the publication of Eric Wolf's book (Peasant Wars in the Twentieth Century, London, 1971), over the question of which class or stratum among the peasantry was intrinsically the most revolutionary. The impact of emerging capitalism on different peasant strata was taken by Wolf, and many other participants in the debate, to be an explanatory variable of central importance.Google Scholar

⁸ An especially good example of the fruitfulness of the structuralist perspective is J. A. Goldstone's detailed analysis of the causes of the breakdown of the English state in the seventeenth century (‘State Breakdown in the English Revolution: A New Synthesis,’ American Journal of Sociology 92, 2 (1986), pp. 257–322).Google Scholar

⁹ Skocpol, Theda, the most prominent of the structuralist theorists (States and Social Revolutions; A Comparative Analysis of France, Russia and China, Cambridge, 1979)CrossRefGoogle Scholar, has co-authored recent work on revolutions in the Third World which focuses on the issue of constructing revolutionary coalitions, gives considerable prominence to the entrepreneurial activities of professional revolutionary organizations, and thus implicitly adopts much of the agenda of the agency perspective: Goodwin, J. and Skocpol, T., ‘Explaining Revolutions in the Contemporary Third World’, Politics and Society 17, 4 (1989).CrossRefGoogle Scholar

¹⁰ The most eloquent statement of this position is to be found in Chandraprema, Sri Lanka: The Years of Terror, esp. chs 3, 4, 7, 8, 14, and 15.Google Scholar One must, however, bear in mind Chandraprema's hostility to the JVP (see footnote 1). See also Leelananda, S., ‘The Rise of the JVP—A Sociological Perspective (3)’, Lanka Guardian (Colombo) 12, 18 (1990).Google Scholar

¹¹ Gates, J. M., ‘Toward a History of Revolution,’ Comparative Studies in Society and History 28, 3 (1986).CrossRefGoogle Scholar The JVP provides an illustration of this process in the way which the mid-1980s they constructed a ‘patriotic’ ideological position—and attempted to construct a corresponding ‘patriotic’ coalition—modelled closely on the practices of the Vietcong in the 1960s and 1970s.

¹² Skocpol, States and Social Revolutions, p. 41.Google Scholar

¹³ Kowalewski, D. (‘Periphery Revolutions in World-System Perspective, 1821–1985,’ Comparative Political Studies 24, 1, 1991) reviews some of this literature and presents data which indicate that there is a statistical association between instability at the ‘core’ and revolution on the ‘periphery’.CrossRefGoogle Scholar

¹⁴ See, for example, Halliday, F. and Molyneux, M., The Ethiopian Revolution (London, 1981), ch. 1.Google Scholar

¹⁵ This point applies particularly to the (early stages of ) the revolutions in North Korea, Cuba, Vietnam, Cambodia, Mozambique, Guinea Bissau, Angola, and Ethiopia.Google Scholar

¹⁶ The naionalist idiom may continue to be as important in mobilizing mass support for revolutionary movements as it has been over recent decades. However, nationalism is increasingly being interpreted in terms of local differences.Google Scholar

¹⁷ This was related to Wijeweera's personal experiences as a student in the Soviet Union. He came into conflict with the authorities, and failed to complete his course.Google Scholar

¹⁸ Insofar as the JVP did receive any external support in the 1980s, this appears to have been confined to limited amounts of cash from Iraq (Chandraprema, Sri Lanka: The Years of Terror, pp. 59–60).Google Scholar

¹⁹ Indian support for the Tamil separatists reflected in part Delhi's persistent concerns about separatist sentiment among the Tamils of Tamil Nadu. To have failed to become involved would have risked the creation of strong, independent linkages between Tamils on both sides of the Palk Straits.Google Scholar

²⁰ Evidence for most of the points in this list may be found in Moore, M., The State and Peasant Politics in Sri Lanka (Cambridge, 1985);CrossRefGoogle Scholar one might note in particular the evidence there on p. 135 that, at the end of the 1970s, about one-third of the active labour force was employed in the state sector. Sri Lanka's superb record of high levels of education and literacy and low levels of mortality is widely known; see, for example, Isenman, P., ‘Basic Needs: The Case of Sri Lanka’, World Development 8, 8 (1980).CrossRefGoogle Scholar A few relevant comparative statistics which are easily available are as follows. They are given in the form of comparisons between Sri Lanka and the average of thirty-six other ‘low income economies’ (excluding China and India) at various points in the 1980s: (a) number of persons per square kilometre—239: 32; (b) percent of Gross Domestic Production derived from agriculture—27: 36; (c) infant mortality rate—36: 112; (d) population per physician—7,460: 17,350; and (e) percent of school-age children enrolled in school—103: 70. See World Bank, World Development Report 1987 (Washington, 1987), pp. 202, 206, 258, 260 and 262.Google Scholar

²¹ For evidence to support the interpretation made here see especially Moore, , The State and Peasant Politics;Google Scholar and Moore, M., ‘Sri Lanka: The Crisis of the Social Democratic State’, in Mitra, S. (ed.), The Post-Colonial State in Asia (New York and London, 1990).Google ScholarManor, J. (The Expedient Utopian: Bandaranaike and Ceylon, Cambridge, 1989) provides evidence on how far the apparent leader of the Sinhalese Buddhist electoral uprising in 1956, S. W. R. D. Bandaranaike, was in fact responding as best he could to ‘grassroots’ (middle class) forces which he did not control.Google Scholar

²² See Jupp, J., Sri Lanka: Third World Democracy (London, 1978).Google Scholar

²³ This pattern is not completely linear. For example, changes in voting arrangements introduced in the late 1950s reduced the degree of physical intimidation of (manily lower caste) voters (Manor, The Expedient Utopian, p. 321).Google Scholar

²⁴ This episode is very rarely mentioned in interpretations of recent Sri Lankan history. I am grateful to Jagath Seneratne for providing information on this from his current research.Google Scholar

²⁵ Most of the training was conducted in India, but some of the smaller groups had cadres trained by Palestinians and others in the Middle East. See interview with the former military commander of the Eelam Revolutionary Organization in Island International (Colombo), 15 03 1989, p. 6.Google Scholar

²⁶ Details and evidence on the events summarized in this and succeeding paragraphs may be found in Manor, J. (ed.), Sri Lanka in Change and Crisis (London and Sydney, 1984).Google Scholar A number of reports from international human rights organizations also document the situation. See for example Sieghart, P., Sri Lanka: A Mounting Tragedy of Errors. Report of a Mission to Sri Lanka in January 1984 on behalf of the International Commission of Jurists and its British Section, JUSTICE, International Commission of Jurists (1984).Google Scholar

²⁷ Because this incident evokes such strong emotions it seems necessary to say that it was triggered by some combination of: (a) political incitement on the Sinhalese side; (b) a genuine Sinhalese reaction against the killing of thirteen soldiers by Tamil militants; (c) the deliberate attempts by the Tamil militians to provoke such reactions; and (d) the clear unwillingness of the police and the armed forces to take any action to prevent the anti-Tamil violence.Google Scholar

²⁸ These were monks belonging to various militant Sinhala Buddhist organizations.Google Scholar

²⁹ For more details on the various dimensions of this ‘political decay’ see Moore, M., ‘Economic Liberalization versus Political Pluralism in Sri Lanka?,’ Modern Asian Studies 24, 1 (1990).CrossRefGoogle Scholar

³⁰ Almost the entire leadership survived the defeat of JVP's first insurrection in 1971. Most leaders left the movement, and several have provided their own interpretations of what happened. Good sources, for both analytic and descriptive purposes, are: Alexander, P., ‘Shared Fantasies and Elite Politics: The Sri Lankan “Insurrection” of 1971’, Mankind 13, 2 (1981);Google ScholarJiggins, J., Caste and Family in the Politics of the Sinhalese, Cambridge (1979);Google Scholar and Keerawella, G. B., ‘The Janatha Vikmuthi Peramuna and the 1971 Uprising’, Social Science Review (Colombo) 2 (1980).Google Scholar For the more recent period, see Alles, A. C., ‘Wijeweera's Changing Political Ideology’, Island International, 21 02 1990, p. 9;Google ScholarChandraprema, Sri Lanka: The Years of Terror; Gunaratna, Sri Lanka: A Lost Revolution;Google ScholarLeelananda, S., ‘The Rise of the JVP—A Sociological Perspective’, Lanka Guardian 12, 15 (1989);Google ScholarLeelananda, S., ‘The Rise of the JVP— Sociological Perspective (2)’, Lanka Guardian 12, 16 (1989);Google ScholarLeelananda, S., ‘The Rise of the JVP–A Sociological Perspective (3)’, Lanka Guardian 12, 18 (1989);Google ScholarLeelananda, S., ‘JVP Learning from Vietnam?’, Lanka Guardian 13, 19 (1990);Google ScholarMatthews, B., ‘Sinhala Cultural and Buddhist Patriotic Organizations in Contemporary Sri Lanka’, Pacific Affairs 61, 4 (1988–1989);CrossRefGoogle Scholar and Matthews, B., ‘The Janatha Vimukthi Peramuna and the Politics of the Underground in Sri Lanka’, The Round Table 312 (1989).Google Scholar

³¹ Except for a small but influential Christian minority, virtually all Sinhalese are Buddhists.Google Scholar

³² The JVP leader, Rohana Wijeweera, ‘stood out as the only individual with a background of foreign travel and at least a smattering of English. He was thus able to overawe the others by exuding an aura of learning on account of his access to literature in English and his course work in Marxism while in Russia” (Chandraprema, Sri Lanka: The Years of Terror, p. 21).Google Scholar

³³ Alexander, ‘Shared Fantasies and Elite Politics,’ Chandraprema, Sri Lanka: The Years of Terror, ch. 5.Google Scholar

³⁴ Wijeweera himself was in prison at the time of the insurrection; the general looseness of the JVP organization is reflected in the fact that a number of other leaders of the movements did not participate in the coup.Google Scholar

³⁵ These castes comprise about one-third of the Sinhalese population, and one quarter of the total population (Jiggins, Caste and Family, p. 35).Google Scholar

³⁶ Mrs Bandaranaike was a member of the so-called ‘Kandyan aristocracy,’ the pinnacle of the ‘traditional’ social hierarchy.Google Scholar

³⁷ Chandraprema, Sri Lanka: The Years of Terror, p. 98;Google ScholarGunaratna, Sri Lanka: A Lost Revolution, p. 144.Google Scholar

³⁸ This measure was undertaken by a Presidential Commission which assumed retrospective powers—powers which were declared ultra vires by the courts but reinstated through constitutional changes made possible by the UNP held 80% of Parliamentary seats.Google Scholar

³⁹ The SLFP was founded by Mrs Bandaranaike's assassinated husband, S. W. R. D. Bandaranaike, and run as a personal fief (Manor, The Expedient Utopian). She succeeded to the leadership soon after his assassination, and continued the tradition.Google Scholar

⁴⁰ This candidate, Hector Kobbekaduwa, received much of his organizational support from the Communist Party and from Mrs Bandaranaike's son-in-law, the actor and aspirant politician, Vijaya Kumaranatunge.Google Scholar

⁴¹ Most of the previous leadership had given evidence against Wijeweera to the Criminal Justice Commission investigating the insurgency (Gunaratna, Sri Lanka: A Lost Revolution, p. 111).Google Scholar

⁴² Chandraprema, Sri Lanka: The Years of Terror, chs 2, 4, 9 and 33.Google Scholar

⁴³ ‘Official’ in the sense that the Communist Party was recognized by Moscow and the Lanka Sama Samaj Party (LSSP) was a longstanding member of the Trotskyite Fourth International. The electoral support of the ‘old left’ had been in almost continual decline since the late 1940s.Google Scholar

⁴⁴ Since organized thuggery tended to enjoy the protection if not the active involvement of UNP politicians at local level, one might expect persons with known UNP connections to be especially active in the attacks on Tamils. It seems clear that the attacks were not organized or stimulated by the UNP leadership as a collective, although powerful individual leaders were heavily involved.Google Scholar

⁴⁵ The Tamil movement was also claimed to be motivated by drugs and the drug trade. The phrase ‘narco-terrorism’ became widely used in the official media.Google Scholar

⁴⁶ There is no evidence that the JVP, as a party, participated in organizing the attacks on Tamils. Individuals associated with the party were almost certainly involved (Chandraprema, Sri Lanka: The Years of Terror, p. 60), but this was probably true of every significant Sinhalese political party at the time.Google Scholar

⁴⁷ The decision to re-embark on the road of armed struggle was made in early 1984 (ibid., ch. 11). The first armed confrontation between the JVP and the police after the proscription appears to have come in later 1984.

⁴⁸ Gunaratna, Sri Lanka: A Lost Revolution, pp. 145–52 and 156–61. This persecution was partly at least in revenge for the fact that the police had been the JVP's main target in 1971. It seems unlikely that it was authorized by the President, and may not have been known to him, although many UNP members were implicated. It is clear that the state apparatuses had no clear and consistent attitude to the JVP; different elements pursued contradictory policies. Assaults on actual or suspected JVPers by the police were to prove a useful source of new and committed recruits for the movement (ibid., pp. 199 and 203–5).

⁴⁹ The JVP is also known to have recruited from among those personnel dismmissed from the armed forces in the early 1980s for insubordination—often for atrocities against Tamils.Google Scholar

⁵⁰ President Jayawardene has claimed—inter alia, in an interview with the present author in Colombo on 6 August 1989—that he had been receiving regular warnings from the police that the JVP was using its ‘parlimentarism’ as a cover for plans to take power by violent means. Jayawardene admitted that, because it set the JVP on the revolutionary road, the banning of the party may have been a mistake. See also Gunaratna, Sri Lanka: A Lost Revolution, pp. 187–8.Google Scholar

⁵¹ See for example Steinhoff, P. G. (‘Hijackers, Bombers and Bank Robbers: Managerial Style in the Japanese Red Army’, Journal of Asian Studies 48, 4, 1989) on the organizational transition in the Japanese Red Army.CrossRefGoogle Scholar

⁵² Gunasekera, R., ‘Taking the Reins of a Fighting Army’, Island International, 31 08 1988.Google Scholar

⁵³ The motives for Indian intervention were complex and do not centrally concern us here. They include: disquiet about the political backlash in Tamil South India should the Government of India stand by while the Sri Lankan army over-ran Jaffna (almost certainly with considerable brutality); longer-term worries about the apparent insensitivity of the Sri Lankan government to India's objective of excluding all other great powers from the region; and a certain element of ‘expansionism’ in the Indian armed forces.Google Scholar

⁵⁴ The police and the armed forces proved reluctant to fire on rioting mobs of fellow Sinhalese. The government may have been saved by its foresight in quickly transferring to Colombo a detachment of the police commands, the Special Task Force, who had become hardened by their brutal campaign against Tamil militants in the Eastern Province. On 29 July they fired on aud dispersed a march into the centre of Colombo by a large mob which had gathered in the poorer southern suburbs of the city.Google Scholar

⁵⁵ ‘It is clear that the JVP in the post-accord period modelled its tactics on the Vietnamese experience. They identified an external aggressor (India), saw the UNP government as a puppet regime of that power (like the Diem regime in South Vietnam), claimed that cliques (Jayawardena–Dissanayake–Thondaman)—(Premadasa–Ranjan Wijeratne) were at work within these regimes, brought in women, students and monks as a “frontline” in street battles, emphassised the “patriotic” nature of the opposition wooed “patriotic intellectuals”’…. (Leelananda, ‘JVP Learning from Vietnam’, p. 18). The fact that President Jayawardene was known to have non-Sinhalese ancestors[!] and had in his younger days been nick-named ‘Yankee Dicky’ for pro-Americanism all added to the plausibility of the basic scenario, as did the relative prominence in the government of Minister S. Thondaman, the leader of the ‘Indian Tamil’ community of estate workers, who was of Indian origin and maintained close contacts there.Google Scholar

⁵⁶ Matthews, ‘Sinhala Cultural and Buddhist Patriotic Organization’.Google Scholar

⁵⁷ Chandraprema, Sri Lanka: The Years of Terror, ch. 3Google Scholar; Moore, M., ‘What did the Southern Province Results Reveal?’ Lanka Guardian II, 6, 7, and 8 (1988).Google Scholar

⁵⁸ Chandraprema, Sri Lanka: The Years of Terror, chs 20 and 27.Google Scholar

⁵⁹ The most significant victim, Vijaya Kumaranatunga, was not from either of these categories. A famous actor, Kumaranatunga had married Mrs Bandaranaike's politically active and radical daughter, Chandrika. They had recently established their own party, the SLMP, in response to the situation in the SLFP described above in the main text. Largely because of Kumaranatunga's popular appeal—boosted by a brief imprisonment in 1982 on false charges as part of Jayawardene's harassment of the SLFP—the SLMP had become the dominent element in the coalition which it formed with the two main Marxist parties, the LSSP and the Communist Party. Aided in part by his humble and ‘low’ caste origins, Kumaranatunga was the most potent direct competitor with the JVP for the votes of the poor. It was for this reason that he was assassinated.Google Scholar

⁶⁰ The JVP effectively prevented large numbers of SLFP supporters from voting owing to its strength in areas where the SLFP had traditionally enjoyed a great deal of electoral support. This was a deliberate strategy: the replacement of the of the UNP regime by an elected SLFP government would have destroyed the JVP's own claim to power. In addition, in late 1988 the JVP had begun to slaughter large numbers of SLFP cadres and thus severely weakened the party. The SLFP was less well armed than the UNP or the main Marxist parties, and thus unusually vulnerable. Some of the murders of SLFPers were carried out by their political opponents in the UNP and, in a few cases at least, within their own party. Political violence was endemic at this time.Google Scholar

⁶¹ Moore, The State and Peasant Politics, ch. 6.Google Scholar

⁶² Sri Lankan employment statistics are not very accurate. They suggest that the high levels of unemployment did fall somewhat in the late 1970s and the early 1980s, at a time when the economy was growing fast. By the mid-1980s the situation had worsened again.Google Scholar

⁶³ Exactly the same strategy was applied in the Tamil political sphere. Jayawardene toyed with the democratic, bourgeois party which took most Tamil seats at the 1977 general election, the Tamil United Liberation Front, but ultimately offered them nothing, and thus contributed substantially to the undermining of their credibility and their replacement by the armed militant groups, especially the LTTE. For some participants's evidence on this see Wilson, A. J., The Break-Up of the Sri Lanka: The Sinhalese-Tamil Conflict (London, 1988), ch. 6.Google Scholar

⁶⁴ Moore, M., ‘The 1982 Elections and the New Gaullist–Bonapartist State in Sri Lanka’, in Manor, J. (ed.), Sri Lanka in Change and Crisis (London and Sydney, 1984), pp. 70–1.Google Scholar

⁶⁵ There is no space here to present and discuss in detail alternative explanations of the rise of the JVP. Predictably, the (partial) liberalization of the economy introduced by the UNP government elected in 1977 is widely blamed for most of the political mess of the succeeding decade. I have elsewhere tried to examine carefully all possible interpretations of this general case, and conclude that it does not appear to have much substance (Moore, ‘Economic Liberalisation’).Google Scholar

⁶⁶ See references in footnote 10.Google Scholar

⁶⁷ Leelananda, ‘The Rise of the JVP–A Sociological Perspective (2)’, p. 6;Google Scholar and Gunaratna, Sri Lanka: A Lost Revolution, p. 78.Google ScholarChandraprema (Sri Lanka: The Years of Terror, p. 27) suggests that Wijeweera's involvement was peripheral and fleeting.Google Scholar

⁶⁸ This may have been a deliberate choice on Wijeweera's part in response to the fact that in 1971 the JVP cadres in the main urban areas had been almost completely inactive. With the urban areas and thus the core of the state apparatus intact, it had been that much easier for the armed forces to ‘mop up’ the JVP's rural strongholds one after another.Google Scholar

⁶⁹ The JVP membership had always been almost exclusively Sinhalese Buddhist, and a degree of Sinhalese Buddhist nationalism (or chauvinism) had always been implicit or explicit in its programme and discourse. However, in the early 1980s the party adopted a fairly orthodox cosmopolitan Marxist stance in relation to the Tamil issue, and indeed expressed some limited support for the right of Tamils to national self-determination. When this position was rejected in favour of an explicitly chauvinist position in December 1983, the Secretary General left the party, taking with him a substantial fraction of the membership. The JVP's changing—and opportunistic—attitudes to the ethnic issues and to India are explained at some length in Chandraprema, Sri Lanka: The Years of Terror, chs 13–16.Google Scholar

⁷⁰ ‘Undoubtedly the most powerful arm of the JVP, second only to the military wing, happened to be its student wing. It is from the student's wing that a considerable percentage of recruitments were made for higher and important positions of the JVP. It must also be clear that 90% of the JVP hierarchy happened to be either university graduates, drop outs or teachers’ (Gunaratna, Sri Lanka: A Lost Revolution, p. 45).Google Scholar

⁷¹ Chandraprema, C. A., Putschism, Ethnic Chauvinism and Social Revolution; Reflections on the Sri Lankan Experience (Colombo, Independent Students Union, 1989), pp. 28–9.Google Scholar

⁷² Chandraprema, Sri Lanka: The Years of Terror, ch. 12.Google Scholar

⁷³ Relatedly, the capture of state power has always appeared as the main JVP objective, an end in its own right rather than a means. This was especially evident in 1971, when the ‘patriotic’ objective lacked real potency. There are plausible accounts of how, once they attained local power in 1971, rather than creating a ‘revolutionary administration’, JVP Local leaders elevated themselves to roles in the existing state apparatus.Google Scholar

⁷⁴ Chandraprema, Putschism, Ethnic Chauvinism, p. 29.Google Scholar

⁷⁵ Chandraprema, Sri Lanka: The Years of Terror, ch. 25;Google ScholarMatthews, ‘The Janatha Vimukthi Peramuna’, p. 432.Google Scholar

⁷⁶ One of the first contract killings organized by the JVP was that of Daya Pathirana, the leader of the Independent Students Union, in 1986. See Chandraprema, Sri Lanka: The Years of Terror, ch. 23.Google Scholar

⁷⁷ See report in Island International, 21 September 1988.Google Scholar

⁷⁸ For example, a high school student was killed by police firing in the town of Badulla on 13 September 1988, allegedly because the police had exhausted their supply of rubber bullets (Island International, 21 September 1988).Google Scholar

⁷⁹ Chandraprema, Sri Lanka: The Years of Terror, p. 54 and passim. In some cases, large numbers of cadres left at one time. This, normally reflected some dispute at leadership level: ‘By 1979 bickering within the party had emerged and a group headed by Nandana Marasinghe resigned from the party. After the JVP's decision not to participate in or support the general strike of 1980, a large group, among whom were several middle class young intellectuals who had been attracted to the JVP's “new image” after 1977, left along with H. N. Fernando and the Ceylon Teachers’ Union. Another group headed by “Captain” Kularatne left the party in 1981 over differences with the party leadership. After the Presidential election fiasco in 1982, the JVP lost about half its “strength”. In April 1983, Vaas Thilakaratne and Mahinda Pathirana were removed from the party. In 1983 December, a large group among whom were Lionel Bopage, left the party because of the change of policy on the ethnic issue.’ (ibid, p. 54).

⁸⁰ ‘Though we paid lip service to “the leading role of the proletariat” in the revolution, our emphasis has always been on the provincial sector in organising students and unemployed youth…. They will be the first to desert the party in a situation of repression.’ This quotation is from the resignation letter of Lionel Bopage, the JVP's General Secretary, in 1983 (ibid., p. 113).

⁸¹ The JVP named its military wing the DJV [Deshapremi Janatha Viyaparaya—the People's Patriotic Organization]. The name DJV first appeared in 1987 in the context of the reactions against the Indo-Sri Lanka Peace Accord, although the military wing was established in mid-1986 (ibid., p. 65). The JVP position was that the DJV was a separate organization of patriots, not part of the JVP (See the interview with Rohana Wijeweera in Sunday Times (Colombo), 13 11 1988, pp. 9–10). While in 1987–89 the JVP was engaged in various kinds of dialogue with the major political parties, especially the SLFP and the UNP, in search of political advantage, it was convenient for all sides to attribute violence and murder to the DJV and thus absolve the JVP. The DJV was headed by a member of the JVP's Politburo (Alles, ‘Wijeweera's Changing Political Ideology’, p. 9).

⁸² Gunaratna (Sri Lanka: A Lost Revolution, p. 279) claims that the JVP put more effort into rescuing from captivity their political cadres than their military cadres. It is not clear whether, on balance, this reliance on ‘auxiliaries’ made the JVP more or less vulnerable to counter-intelligence. It seems likely that, in the first two years or so of the recent campaign, when the state's security and intelligence services were weak and heavily penetrated by JVP sympathizers, use of auxiliaries presented few problems. However, once the security forces began their crackdown on the movement in August 1989, the whole movement became vulnerable to betrayal under pressure. The top leadership appears to have been traced and eliminated through the classic counter-insurgency technique of first catching the ‘small fry’ and forcing them to inform on the source of their information and instructions, leading eventually to the top.Google Scholar

⁸³ Alles, ‘Wijeweera's Changing Political Ideology’; Chandraprema, Sri Lanka: The Years of Terror, chs 21, 23 and 31;Google ScholarGunaratna, Sri Lanka: A Lost Revolution, 221.Google Scholar

⁸⁴ The leadership appears to have remained physically dispersed, in many cases taking cover by leading the lives of relatively affluent members of the middle class. Wijeweera spent his last months living with his family in the guise of a retired plantation owner in a large house in a rural area. For more general information on the JVP's organization, see Chandraprema, Sri Lanka: The Years of Terror; ch. 2;Google Scholar and Gunaratna, Sri Lanka: A Lost Revolution, pp. 22, 37–48, 213–20 and 324–5.Google Scholar

⁸⁵ On the importance of the these linkages, see for example Chandraprema, Sri Lanka: The Years of Terror, ch. 22. Jonathan Spencer has provided me with some illustrations of this from his own experiences.Google Scholar

⁸⁶ Chandraprema, Sri Lanka: The Years of Terror; p. 200;Google ScholarGunasekera, R., ‘JVP Tactics’, Island International, 23 09 1987.Google Scholar

⁸⁷ Generally speaking, the various orders of Buddhist monks in Sri Lanka are drawn from and serve different caste groups. It seems that the monks which actively supported the JVP tended to represent those low castes which were especially active JVP supporters. The priesthood was to suffer grievously in the long run. Even before this time the priesthood was, in clear contradiction to what are often described as Buddhist ideals, widely involved in national electoral politics. The JVP, however, raised the stakes considerably by going to the lengths of murdering those monks who stood in their way. When the counter-terror got underway, little compunction was shown about killing monks believed to have JVP connections. In addition, independent monks who tried to protest about the brutality of what was happening or demand the release of detainees were silenced by threats and, at least in one or two cases of recalcitrance, by the bullet. Gunaratne suggests that in 1989 about a hundred Buddhist monks were killed by anti-JVP forces (Sri Lanka: A Lost Revolution, p. 307).Google Scholar

⁸⁸ Chandraprema, Sri Lanka: The Years of Terror; chs 30 and 46, and p. 299.Google Scholar There is some evidence to suggest that the JVP succeeded in penetrating almost to the top of the Joint Operations Command (See Lanka Guardian, 1 February 1990, p. 7).Google Scholar

⁸⁹ de Silva, M., ‘The “Phoney Peace” is Over—Protracted War Begins’, Lanka Guardian, 12, 5 (1989), p. 4.Google Scholar R. Premadasa, a streetwise politician of very humble social origin, had occupied the Prime Minister's office under Jayawardene's Presidency, but had been excluded from real power. The UNP was forced to nominate him for the Presidency when Jayawardene's term expired because Premadasa's populist image, his humble social origins, his evident independence of the previous administration and his known opposition to the Indo-Sri Lanka Accord provided the only real prospect for the party to obtain sufficient popular support for them to ‘win’ the election without indulging in an unacceptably blatant degree of ballot rigging. Premadasa always adopted a sympathetic approach to the JVP, refused to accept publicly that they were engaging in violence, and attempted to reach a political agreement with them. He adhered to this line while and after the armed forces and the vigilantes they supported were massacring JVP suspects and destroying the organization.

⁹⁰ Chandraprema, Sri Lanka: The Years of Terror; pp. 125–32, 193, 198–9, and 232. This may have been partly the result of a self-fulfilling expectation in the minds of the security forces. It is known that, in classic style, many of the military operations against the JVP generated fierce reaction and added support for the insurgents. The military may have targeted low caste villages. It seems clear that the few cases of generalized punitive actions against whole communities of JVP suspects were directed against low caste settlements. The clearest and worst case is that of the massacre of the villagers of Menikhinna, near Kandy (ibid., p. 296; Gunaratna, Sri Lanka: A Lost Revolution, pp. 296–7).

⁹¹ Chandraprema, Sri Lanka: The Years of Terror, ch. 19. In Sri Lanka, public reference to caste is generally held to be in bad taste. One euphemistic device is to refer to caste by the first letter of their name. The Karava caste, traditionally associated with fishing, are thus described as the ‘Ks’. Since Che Guevara was a popular international cult figure in 1971, the JVP was sometimes referred to as the ‘K Guevaras’. The reason for the prevalence of members of the Karava caste among the leadership at that time relates, at least in a proximate sense, as much to factors of locality as to those of caste: Wijeweera recruited his deputies very heavily from those who had attended the same distinguished secondary school as himself, Dharmasoka College, which is located in a Karava area, Ambalangoda.Google Scholar

⁹² Ibid., ch. 17. In its recent campaign the JVP was noticeably more active than in 1971 in the suburbs of Colombo and in Kandy, the second largest town. This greater urban presence enabled the JVP to make more active use of one of its most effective propaganda techniques: the simultanous pasting up over one night of a large number of posters containing the same message. This helped create an image of ubiquity. The messages themselves were not purely symbolic: this was one of the main ways in which the movement communicated with the public.

⁹³ It is likely that many were true (Gunaratna, Sri Lanka: A Lost Revolution, p. 268).Google Scholar

⁹⁴ The JVP never achieved that level of discipline and willingness to self-sacrifice in favour of a cause that characterized the Tamil Tigers. Unlike many JVP activists, LTTE cadres were full-time soldiers. They excited widespread admiration and respect for their total devotion to their cause. The most visible evidence was the cyanide capsule that every LTTE cadre wore around his (or her) neck, and was expected to take to avoid capture. In most cases the cyanide was taken (Gunasekera, R., ‘Cyanide Cult: Yogi Gives the Lowdown’, The Island, 6 08 1989, pp. 11–12)Google Scholar. In late 1987 a group of thirteen Tigers in Indian custody simultaneously and publicly took cyanide to avoid being transported to Colombo and handed over to the Sri Lankan authorities. In other respects conspicuous discipline and self abnegation was demanded of Tiger cadres. The ideologies and symbols that underpinned such devotion to the cause of Tamil nationalism were deeply racist in the fullest sense of that term, and reflected a long history of perceived discrimination and, more recently, bloodshed. It was partly the lack of any comparable recent Sinhalese collective national trauma which frustrated attempts to create a comparable degree of commitment among the JVP's cadres. Wijeweera had always tried to insist that his cadres abstain from liquor and tobacco (Gunaratna, Sri Lanka: A Lost Revolution, p. 79), but such discipline never became a prominent feature of the movement.Google Scholar

⁹⁵ On the prevalence of local particularism in poor countries in general, see Dunn, J., ‘The Politics of Representation and Good Government in Post-Colonial Africa’, in Chabal, P. (ed.), Political Domination in Africa: Reflections on the Limits of Power (Cambridge, 1986), pp. 167–8.Google Scholar Especially in the more rural Tamil areas of Sri Lanka, recruitment into the various competing militant organizations was significantly influenced by local connections of individual leaders. See Taraki, , ‘Wrong Strategies of TNA and Raw Moves’, Island International, 20 12 1989, p. 6;Google Scholar and Taraki, , ‘The Struggle for the Wanni’, Island International, 10 01 1990, p. 6.Google Scholar

⁹⁶ For an account of these recurrent themes in Sri Lankan, especially Sinhalese, political culture, see Moore, M., ‘The Ideological History of the Sri Lankan Peasantry’, Modern Asian Studies 23, 1 (1989).CrossRefGoogle Scholar

⁹⁷ See the interview with Wijeweera, Rohana in Sunday Times, 13 11 1988, pp. 9–10.Google Scholar This implied at least a tactical modification of the statism which had previously dominated the JVP's comments on economic organization (see above in main text).

⁹⁸ It is for example noticeable that the JVP made relatively little in their propaganda of the real and widespread hardships suffered by large numbers of poor people as a result of the major cuts made after 1978 in the coverage and value of the hitherto-extensive food subsidy scheme. For details of these cuts and their consequences see Sahn, D., ‘Changes in the Living Standards of the Poor in Sri Lanka during a Period of Macroeconomic Restructuring’, World Development 15, 6 (1987).CrossRefGoogle Scholar

⁹⁹ One of the reasons why low caste status was a logistically-viable basis for JVP recruitment was that in rural Sri Lanka different castes tend to live in separate villages or localities. In the preparatory stages, JVP activity could proceed relatively unhindered in a supportive micro-environment of low caste settlements. The situation was however reversed once the armed forces began actively to pursue the JVP: these ‘supportive micro-environments’ became prime targets.Google Scholar

¹⁰⁰ A list of the elaborate network of institutions existing in Jaffna under the rule of the Tamil Tigers in early 1987 is given in Odchimar, M., ‘Tigers Make Planes and Bombs, Run Farms’, Sri Lanka News (Colombo), 4 02 1987, p. 1.Google Scholar

¹⁰¹ During 1988 the JVP did in fact establish a substantial degree of territorial control in the districts of Matara and Hambantota, and managed to exclude administrative activity of which they did not approve. They did not, however, attempt totally to disable or exclude the state. An impression that they were attempting to do so was relatively widespread, presumably in large part (a) because of a priori expectations about the characteristic behaviour of contemporary revolutionaries and (b) because it was expected that the JVP would emulate the tactics of the Tamil separatists. This latter impression was reinforced by the fact that, during 1988, the JVP purchased lessons in the technology of directional (Claymore) mines from a ‘retired’ Tamil separatist movement (PLOTE—the Peoples Liberation Organisation of Tamil Eelam) (Chandraprema, Sri Lanka: The Years of Terror, p. 266Google Scholar; and Gunaratna, Sri Lanka: A Lost Revolution, p. 133Google Scholar) and were very quickly using it to good effect against the armed forces in the Southern districts. The use of landmines had a direct military rationale: in the densely populated rural areas of Southern Sri Lanka, the armed forces could rarely use either helicopters or cross-country vehicles to move troops during operations. They were heavily dependent on transport on surfaced roads, and thus vulnerable to directional mines located along the roadsides. The adoption of landmine technology did not imply any particular commitment to a policy of excluding the state apparatus from particular localities; it was the armed forces that were the targets.

¹⁰² Sri Lanka had always lacked an effective political intelligence system; this was one of the reasons for the early successes of the JVP in 1971, and for the inability of the government to tackle the Tamil militant movements in the late 1970s and early 1980s when they were still small, clandestine organizations. The UNP apparatus performed political intelligence functions, but not very effectively, especially after the JVP began to attack it. The establishment of a political intelligence system was a major function of the Israeli intelligence advisers brought into Sri Lanka, under the aegis of the American Embassy, in the mid-1980s. Muslim opposition to their presence was so strong that they eventually had to leave without having accomplished much. A Counter Subversive Division [CSD] was established in late 1986 to counter the JVP. Its head, Superintendent Perera, was assassinated in mid-1987. The head of the Special Investigations Division, and former head of CSD, was shot in June 1989 while investigating sensitive allegations against senior politicians and military personnel, apparently including possible JVP links. The unreliability of even the CSD is demonstrated in the fact that one particularly zealous anti-JVP Police Superintendent, in whose hands a lawyer suspected of JVP connections had died, had secretly held the man in his own official quarters and not informed the local unit of the CSD for fear that they would leak the information (Island International, 21 August 1988).Google Scholar

¹⁰³ In 1988 the JVP began a fairly successful campaign to block the distribution in many areas of newspapers printed by the government-owned Lake House group.Google Scholar

¹⁰⁴ See in particular the account in footnote 59 of the murder of Vijaya Kumaranatunga.Google Scholar

¹⁰⁵ Foreign aid continued to flow in relatively abundantly throughout this period, and was indirectly of major political significance by permitting harsh economic policy decisions to be delayed until the end of 1989, when the JVP was already defeated. There was, however, no direct military support for the regime, mainly because no other power had sufficient interest in doing this that it was prepared to incur the displeasure of the Government of India, which had clearly reaffirmed that Sri Lanka was to be regarded as within India's geo-strategic domain.Google Scholar

¹⁰⁶ At the same time the heads of all the Buddhist orders issued a joint statement calling on the government to dissolve Parliament and hold elections under a caretaker government. This was one of the demands of the entire opposition, including the JVP.Google Scholar

¹⁰⁷ There was no sign in Sri Lanka of a right wing, business-backed movement to crush the JVP forcibly and use the opportunity to install a military-backed government of an openly capitalist and repressive nature.Google Scholar

¹⁰⁸ For example, the attack of 1 November 1988 on the Pannala Camp of the National Auxiliary Force, which is about thirty miles from the centre of Colombo, is estimated to have lasted about two hours. Three members of the Force had deserted with their weapons the previous night. See Sunday Times, 6 and 22 November 1988. The last attack of this nature was on the Thimbirigasyaya Police Field Force Station in central Colombo on 2 August 1989. The attackers escaped, and it was later revealed that some officers who should have been on duty there were absent at the time.Google Scholar

¹⁰⁹ Chandraprema, Sri Lanka: The Years of Terror, pp. 278–9.Google Scholar

¹¹⁰ Silva, M. de, ‘The “Phoney Peace” is Over—Protracted War Begins’, Lanka Guardian, 12, 5 (1989), p. 3.Google Scholar

¹¹¹ In the sense of being tied to a variety of political parties and vulnerable to control by the party in power.Google Scholar

¹¹² Ibid., p. 3; Leelananda, ‘The Rise of the JVP—A Sociological Perspective’, p. 9.

¹¹³ The JVP's concern was that involvement would expose their cadres to the police and threaten the future of the organization (Gunaratna, Sri Lanka: A Lost Revolution, p. 17).Google Scholar

¹¹⁴ The JVP was able to make deep inroads into the workforce of the Sri Lanka Transport Board in part because of major differences of strategy between senior politicians involved in the dispute.Google Scholar

¹¹⁵ The government faced a major credit crisis, and found the large international banks, by then very wary of any Third World involvements, unwilling to help. Salvation came from three smaller international banks with a substantial stake in Sri Lanka.Google Scholar

¹¹⁶ Other examples of successful involvement in trades union disputes include, for example, a pay grievance of the universities' minor employees in June 1989. The JVP insisted that the unions take a firm stand; a pay rise resulted. Sunday Times, 11 06 1989.Google Scholar For more details of the JVP's relations with trades unions, see Chandraprema, Sri Lanka: The Years of Terror, ch. 38Google Scholar; and Gunaratna, Sri Lanka: A Lost Revolution, pp. 51–5Google Scholar).

¹¹⁷ Wijeweera was found of quoting Danton's favourite maxim: ‘Audacity, Audacity, and once more Audacity’ (Leelananda, ‘JVP Learning from Vietnam’, p. 6).Google Scholar

¹¹⁸ Intelligence operations had originally been the responsibility of the police. They failed completely.The army assumed responsibility for intelligence and propaganda in late 1988, and was far more effective.Google Scholar

¹¹⁹ For details of the way in which the armed forces and vigilante groups eliminated the JVP see Chandraprema, Sri Lanka: The Years of Terror, chs 45–50;Google Scholar and Gunaratna, Sri Lanka: A Lost Revolution, pp. 274, 285–6 and 318–42.Google Scholar Success depended in particular in getting the armed forces out of barracks and into the streets and villages, operating in small groups, and equipped in particular with handguns and civilian vans.

¹²⁰ It failed partly because the armed forces were willing to fire on and disperse crowds of civilian demonstrators. Many if not most of the demonstrators had been forced onto the streets by the JVP. Several hundred were killed over two days.Google Scholar

¹²¹ Ibid., ch. 45.

¹²² The gross brutality of the methods of murder, torture and mutilation and display of corpses employed by both the JVP and their opponents is something that requires mention but no elaboration. It closely parallels the gory nature of much JVP propaganda. The story of the creation of anti-JVP ‘vigilantes’ is of more analytic interest. Most of the leading participants remain alive, and more details may emerge eventually. For present purposes it is adequate to mention a few key points. These units appear to have been created mainly through combining the equipment, information and resources available to the armed forces with the commitment, selfabnegation and bloodlust of those who had lost close family members to the JVP. In the initial stages at least, the physical and social space needed to establish vigilante movements when the JVP was felt to be everywhere was provided through the failure of the JVP, this time as in 1971, to generate support among the relatively small but well-organized and influential Sinhalese Christian communities. The JVP was identified as a Sinhalese Buddhist movement. The anti-JVP vigilante groups were largely created in the predominantly Sinhalese Catholic areas north of Colombo.Google Scholar

¹²³ Up to this point only one current JVP Politburo member had been captured (and killed) (Gunaratna, Sri Lanka: A Lost Revolution, p. 341).Google Scholar

¹²⁴ Ibid., p. 269.

¹²⁵ Kowalewski, ‘Periphery Revolutions’.Google Scholar

¹²⁶ The vast majority of deaths were those of actual or suspected JVPers.Google Scholar

¹²⁷ One could say much the same of the other contemporary revolutionary movement which bears so much similarity to the JVP—the Shining Path (Sendero Luminoso) movement in Peru. Both movements have used indigenism to build up a support base distinctively different from that enjoyed by established left or communist movements with international connections, and assaulted the established left as traitors in the service of foreign powers. See Davis, R. B., ‘Sendero Luminoso and Peru's Struggle for Survival’, Military Review 70, 1 (1990);Google Scholar and Palmer, D. S., ‘Rebellion in Rural Peru: The Origins and Evolution of Sendero Luminoso’, Comparative Politics 18, 2 (1986).CrossRefGoogle Scholar