‘Devils that suck the blood of the Malayan People’: The Case for Post-Revisionist Analysis of Counter-insurgency Violence

Introduction

In June 1952 the Malayan Communist Party’s Freedom News delivered its verdict on British imperialism, with which it had been in conflict since June 1948: ‘devils that suck the blood of the Malayan people, giving low wages while extracting vast profits’. These were vampire colonialists, whose tactics were worse than the Japanese, who had killed tens of thousands of Chinese in massacres that had started in Singapore in February 1942, and then spread through the Malayan peninsula. ¹ The British, declared Freedom News, had burned villagers’ huts, abused farmers, raped women, and, with increasing ferocity and desperation, corralled the masses into ‘concentration camps’. ²

If we strip away the emotional aspects, and the attempt to frame the British as the new Japanese, we are forced to recognize that much of what Freedom News stated was fact. Villagers’ huts were burned in 1948–49, some women were raped, ³ and forest fringe masses corralled, until after 1952 more than 500,000 out of Malaya’s rising population of 5–6 million (1948 and 1960) were in resettlements or ‘New Villages’. ⁴ The majority of those resettled were from the 38% of the population who were Chinese, a percentage slightly outstripped by the largest single category of Malays. Though most New Villages would survive and even thrive, their beginnings were fraught. A further 600,000 rural workers (including Malays and Indian rubber tappers) had been ‘concentrated’, that is, moved short distances to cluster them defensively. In addition, by February 1953 just under 40,000 people had been detained for more than 28 days, and by 1955 more than 31,000 had been ‘repatriated’ to a China many scarcely remembered, and some had never seen. ⁵

All of this seems to directly support the ‘revisionist’ trend in European and American literature on counter-insurgency of the past decade: a trend that downplays the centrality of ‘winning hearts and minds’ questions how far ‘minimum force’ was an aim, and emphasizes widespread, calculated use of violence, coercion, and intimidation against civilians. ⁶ Malaya has been overtly cited in support of this revisionist case. At the extreme, MCP leader Chin Peng has bundled together the above characteristics in support of his argument that ‘the British were the real terrorists’. They, he has argued, were the ones who executed prisoners (around 200 captured insurgents), retried people who were initially not convicted, and allowed decapitation of corpses for identification purposes until 1952. ⁷

By contrast, Anthony Short has argued that shootings, more typically of one or two people who ran away in a remote area – were more a result of uncertainty and nerves in contested areas than of policy; this despite labelling the 1948 to early 1949 period one of an ‘approach to counter terror’. ⁸ Even Huw Bennett, in a ‘revisionist’ attempt to enumerate killings of civilians for the early months of the conflict, came up with figures that suggest a more nuanced reading. Trawling the situation reports from the period, he identified 77 cases of people being killed ‘escaping’ up to April 1949. ⁹ Tellingly, there are a small number of accounts of up to 11 people (and at Batang Kali 24) being killed subsequent to searches of ‘kampongs’ (the Malay term for villages). Yet of that 77 a significant percentage would have been genuine escape attempts by insurgents and Min Yuen (civilian supporters of insurgents), whether armed or not. In addition, such killings sharply taper off from 1949 onwards. It is clear, therefore, that we need to look in more detail at why such ‘extreme’ levels of violence happened, and at why and how they and their underlying logic changed over time. ¹⁰

Before analysing how and why such high levels of force were possible, what varieties they took, and how colonial authorities attempted to temper, limit, and replace them, we also need to define ‘excessive’. Extreme or excessive counter-insurgency ‘violence’ and coercion is defined, for the purposes of this article, as the use of high to lethal levels of physical force against non-combatants (civilians, detainees, prisoners, and corpses).

This definition leaves open the question of whether some highly coercive actions that do not cause direct and immediate physical harm to individuals, such as population displacement, can be embraced by the above definition. My argument is that it depends on the nature of such actions, and to illustrate this I deploy, in adapted form, the term ‘structural violence’. This term has, up to this point, been very sparsely used in counter-insurgency literature. Instead, it is more usually applied in fields such as human rights, development studies, ‘peace medicine’, or ethnography, to describe ‘structural’ factors that embed social discrimination to the extent of seriously reducing the quality of life of particular groups (genders, ethnicities, social classes, communities), despite that harm not being attributable to particular intents or individuals. For instance, a policy might produce higher infant mortality amongst deprived groups. ¹¹

Those literatures are, however, too expansive to be of much utility in counter-insurgency analysis, embracing as they do vast swathes of government policy. Here I adapt the term for specific use in counter-insurgency, by defining it as the application of policies tightly bound up with counter-insurgency, in a way which impairs human life to the extent that fundamental needs are systematically denied or seriously deteriorated. ‘Counter-insurgency structural violence’ must do both lasting and significant harm to the people affected. We might consider it as a form of delayed-effect or slow-burn violence. Hence, for instance, structural counter-insurgency violence might cause, intensify, or embed damaging poverty levels by dramatically lessening access to jobs, land, and other resources. In so far as forced population moves left people with equal or superior resources and life chances, however (at least after initial disruption), it would make little sense to label such displacement as ‘structural violence’.

Malaya provides promising ground for testing types and levels of such force, and concepts such as counter-insurgency structural violence. This is because of its place in the shifting historiography on wars of decolonization, and in writing on the practice of counter-insurgency (COIN). It has passed from being a supposed beacon of good practice, to playing a part in post-2004 revisionism of Anglo-American historiography. This pattern can be simplified into three phases:

The post-2004 historiographical trend has been towards enumeration of abuses and towards seeing a hard repressive core as at the heart of post-war British COIN successes, however well supported by political and ‘hearts and minds’ tactics. The extension of our knowledge of the level of coercion in Malaya has played a key part in this revisionist turn, and this article will continue to add detail as to the scope and nature of physical violence employed against Malaya’s non-combatants. ²² But it will also argue that we now need to take full and detailed cognizance of the differences in the nature, logic, and timing of such violence. ‘Post-revisionist’ analysis is thus not about arguing there was more, or less, excessive or exemplary violence, but rather about giving greater attention to the variations in causes and types of violence, to the lifecycles of these types, to the role of violence limitation, and to the development of new concepts and approaches.

As this article will amply demonstrate, blanket stereotypes that claim to explain all policies over all periods can be little more than propaganda, or at best naïve over-simplifications. Different policies and phases need distinctive treatment. ²³ Even categorizing events such as the massacre of 24 Chinese villagers at Batang Kali in December 1948 can be problematical. Was this an atypical ‘excess’, one of a few such excesses, or a manifestation of more ‘systematic’ policies?

The rest of this article provides a case study in differentiating types of extreme violence, while beginning to analyse their causes and government responses at the time and after. It focuses on just a few contrasting types of high-level ‘violence’ against non-combatants. These are of course highly selective. Additional types – such as ‘violent’ interrogation and the beheading of corpses for identification reasons – have been left out, in favour of greater depth of analysis of a few types. Those additional types would require entire articles in themselves, for instance to explain how decapitation of enemy corpses to remove their heads from deep jungle for identification purposes, and display of corpses and their images for additional propaganda purposes, were normalized for long stretches of the campaign. ²⁴

The types of violence examined are instead those that best allow the illustration of three issues. First, what sorts of orders and contextual factors tended to encourage or permit extreme violence, notably in killings of groups of civilians. Second, the way the type and level of violence towards non-combatants varied over time, as exemplified in the changing way in which civilians in liminal areas were treated, and the radically different types of population displacement that were used. Third, the potential and difficulties in attempting to employ the category of ‘structural violence’.

The article is therefore a demonstration of a post-revisionist approach by detailed analysis of just a few types of extreme violence and coercion, rather than an enumeration of all the types of violence to be found in this one campaign.

The rest of the article divides as follows: Mass Killings of Civilians; Counter-terror; and Population Displacement.

Mass Killing of Civilians

The most extreme form of violence was the killing of groups of several or more civilians simultaneously or in very close succession, which happened an absolute minimum of two or three times. ²⁵ Hence the communists’ Freedom News of 15 January 1949 reported on ‘Japanese Brutal Acts Again Staged in Kuala Kubu’ (Freedom News 1, 15 January 1949). The report covered the activities of a 14-man patrol from G Company, 2nd Battalion Scots Guards, at Sungei Remok Estate, in Batang Kali district, on 11 and 12 December 1948. The nearest significant town was Kuala Kubu, around 45 miles north of Kuala Lumpur, hence the headline.

Prior to the Freedom News article, the local Straits Times had reported one Chinese killed ‘in the Batang Kali area’ on 11 December, and a further 25 on Sunday 12 December, subsequent to ammunition being found in their huts, and while making a concerted attempt to escape. The Singapore Free Press had more starkly declared that ‘A patrol shot dead 26 terrorists’. ²⁶ A telegram to the Colonial Office on 13 December and a military ‘Sitrep’ report of 17 December persisted in labelling the dead ‘bandits’. ²⁷

The numbers were later amended to a total of 24 Chinese shot over the two days. But complaints followed. By 22 December Thomas (‘Tom’) Menzies, European owner of the rubber estate, had claimed the victims were ordinary workers. Newspapers and the Chinese consul-general expressed concern, sparking off an inquiry. In January 1949 the Straits Times was still concerned that unarmed men should have been shot. ²⁸

Attorney-General Sir William Foster-Sutton took two main actions. He visited the site, talking to patrol commander 22-year-old Lance Sergeant Charles Douglas, his fellow NCO, and some of the local police who had accompanied the patrol. In 1970 he would tell the BBC Radio 4’s ‘World at One’ programme that he had found sten gun shells in burned huts, and had concluded that a ‘bona fide’ mistake had been made. ²⁹ He also ordered an inquiry, though without sworn statements being taken. The man chosen to investigate the actions of Scots Guardsmen, however, was Scottish-trained Federal Counsel John V.M. Shields, who had been attached to the Argyll and Sutherland Highlanders in the war. ³⁰

According to statements given by several patrol members after the story resurfaced in the People newspaper on 1 February 1970, in 1948 they pre-agreed a story that the prisoners ran away. According to National Service Guardsman William Cootes his inquisitor – having expressed doubt that prisoners would have split into the separate groups they died in – told him that ‘he hoped we got away with it’. ³¹ Neither Foster Sutton nor the ‘inquiry’ sought information from the villagers – other than one ‘Chong Hung’ – arguing they would have been too frightened to speak, and if they spoke, likely to lie. ³² Over 11 to 12 December Chong had told the patrol that bandits did visit for food but everyone else was too afraid to talk. ³³ The failure to question villagers was inexcusable, since government was aware of accusations of ‘the alleged bashing of people’. ³⁴

The government subsequently announced an inquiry had confirmed the Guards’ version that the men had made a run for it, and continued running after warnings. ³⁵ On Monday 3 January 1949 it issued a fuller statement, which suggested the workers had raised a shout as a pre-agreed signal, run, and received further warnings before being shot. General-Officer-Commanding for Malaya, Major-General Boucher, confirmed in a 5 January press conference that his men would only shoot when absolutely necessary, but were expected to shoot to kill when they did so. ³⁶ In short, the authorities struggled to convince a press sceptical as to how 24 unarmed men, rubber tappers ranging from teenagers to old men, could justifiably be shot dead, leaving no survivors.

By contrast, the communist’s Freedom News of 15 January 1949 described the patrol arriving around 3 p.m. on 11 December to surround ‘three coolie houses’ and order 70 workers out while the premises were searched. Finding nothing, shots were fired in the air to frighten them, and the soldiers continued intimidating them as ‘bandits’. Next day, the 12th, the soldiers burned the coolie lines and possessions, drove off families (one male was taken away too), and machine gunned all but one of the rest. ³⁷

The accusation that this was like Japanese wartime massacres – where entire villages of men, women, and children had been massacred in cold blood – was histrionic, as was the image of machine guns being used, as opposed to rifles. Putting that aside, Freedom News came closer to the truth than the government-endorsed version in two respects. It touched on issues central to understanding why this patrol went so catastrophically wrong. First, it revealed the mock executions and intimidation of 11 December that are a clue to the patrol’s mindset and the final killings. Based on pre-patrol briefing, on spotting two insurgents en route, and on information received from one (intimidated) worker, the soldiers decided the workers were supplying insurgents, and probably included in their number part-time ‘bandits’. In turn, the patrol was part of a bigger jigsaw of action by several units against the 8th Regiment Malayan People’s Anti British Army (MPABA). Overconfidence that workers were withholding additional information may help to explain the patrol’s attempts to extract that by threatening executions. A single killing on 11 December and mock executions raised the stakes. ³⁸ If they could not provide hard evidence of worker complicity, the troops’ behaviour would lay them open to accusation of serious abuses. The intimidation that Freedom News highlighted also increased the chance of captives running in response to any troop behaviour that might seem threatening.

The second way Freedom News came closer to the truth was a negative one: it was free of key elements of the government version which subsequent statements from patrol members suggest were fabricated. Several of these were suggested as early as Monday 13 December by patrol leader Lance Sergeant Charles Douglas, in an interview with a Straits Times correspondent. ³⁹ They included 13 sten gun rounds being found (in 1970 Douglas would tell a journalist ammunition was found later, and the Attorney-General would also claim they were found afterwards). ⁴⁰ They also included workers running following a shout from one of their number, the soldiers shouting a warning, and the prisoners ‘running into the guns’ of soldiers positioned roundabout in guard positions. Later accounts would revise this to a ‘chase’ and further warnings before the final shootings. ⁴¹ It was this latter combination that facilitated the final government statement that the firing was justified under existing orders, given that the prisoners were warned, and would have made good their escape had they not been shot. This approach shows just how easily soldiers might frame killings as justified, by suggesting good reason to believe people were armed (or even planting weapons and ammunition on bodies), and claiming people were obviously making a calculated attempt to ‘escape’. ⁴²

In 2012 a United Kingdom High Court judge, while stopping short of establishing why the soldiers started shooting, accepted that one Chinese had been shot the day before, that there had been mock executions, and that one further Chinese had told soldiers that bandits had visited and one more been trucked off in a poor state, before 23 Chinese were shot on 12 December. ⁴³ This confirmed key elements of the Freedom News version, while maintaining it was not possible to know exactly how the final killings were triggered, and that insufficient could be learned from any inquiry to justify ordering the British government to open one. That latter was critical since the claimants – four relatives of those killed – were asking for judicial review of British refusal to reopen an inquiry. ⁴⁴ The case went on appeal to the Court of Appeal in September 2013, and to the Supreme Court in November 2015. It failed at the latter on a technicality. The court ruled that individuals could only invoke the European Convention on Human Rights for events that had occurred up to 10 years before 1966. ⁴⁵

Regardless of what any court decides on ‘Batang Kali’, British reports of 1948–49 covered up serious abuse, and prevented a proper inquiry at the time. As a result, the ‘trigger’ for the final killings will probably remain forever disputed: calculated massacre on the orders of one or more of the two patrol Sergeants, or accidental chain reaction on the day? What we can be sure of is that that most, if not all, of the workers were not running when firing started, and that some of the patrol subsequently regarded the killings as murder. ⁴⁶

The question then becomes: Was this an aberration that tells us little about the campaign in general, or just the most violent result of policies that made the killing of civilians more likely? The context for the killings is clear. The patrol, accompanied by local guides, had been sent to a jungle fringe plantation area which they viewed as disputed territory. They had been briefed that there might be insurgents around, and had brief contact with two uniformed guerrillas en route. They had also been relayed orders that emphasized the danger of not shooting after uniformed bandits had been seen in any area, and that they could not be prosecuted if they reasonably believed that people were armed, with armed individuals, or consorting with insurgents.

Thus geared up, the Scots Guards seem to have prejudged that villagers were helping insurgents and included part-time ‘bandits’ amongst their number, and resorted to extreme pressure in their quest to extract information. Whatever the spark for the final killings, the events leading to them were thus not untypical of the early Emergency. Small patrols, sometimes led by non-commissioned officers and operating in disputed territory close to jungle and hill, worked through interpreters to try and force information from frightened villagers, whose huts were sometimes burned afterwards. The uncertainties ensured that villagers who ran, whether from guilt or fear and whether armed or not, carried a high risk of being shot.

Counter-terror: Episodic Failure, Systematic Violence, or ‘Excesses’?

It could be argued that the events of Batang Kali on 11–12 December 1948 were exceptional or even unique, the extremity of violence a result of an unlikely confluence of several distinct factors that tended to increase the likelihood of civilians being killed. There are only a handful of incidents when a dozen or more insurgent kills were claimed, and most are verifiable. Another way of looking at this incidence, however, is as a particular manifestation of contextual characteristics, which repeated elsewhere to facilitate multiple shootings of civilians in ones, twos, and small groups, and a general air of ‘counter-terror’ against isolated rural Chinese in 1948 to early 1949.

First of all, the main ‘battlefront’ was not just around but in villages and surrounding rural workplaces. Malaya’s topography features a main range of mountains and jungle dissecting it as a spine from the border with Thailand almost to Singapore in the south, with a relatively narrow coastal plain either side. Between the two-thirds of the country covered by jungle, and the coastal plains with their towns, mines, and road networks, was a jungle fringe zone. This was the most contested area, and the most contested part of the population consisted of rural Chinese farmers, often ‘squatters’ who held land there on one-year licenses or illegally. This squatter community had swelled during the Japanese Occupation, when it provided support to anti-Japanese guerrillas widely seen as heroes (and supported by British Special Operations Executive men from 1943 to 1945). These squatters, and to a lesser extent plantation and mine workers, were in many places imbricated with the insurgents. For a minority, sons might be in the jungle, and uncles and children (in a ‘Little Devils’ Corps’) suppliers in the Min Yuen (Masses Organization).

The rural workers’ quarters and kampongs (villages) were not so much a nomansland as an everymansland in 1948–49. The British, MCP, and anti-communist groups and leaders, all aspired to influence and control. Security forces knew that while they might patrol an area one day, the next day insurgents might visit relatives there. The proximity of villages to forest and mountain on one side and to the well-developed infrastructure of the rubber- and tin-producing west coast on the other, factored in high permeability to both government and insurgent forces. The soldiers’ job was to support the civil power under police direction, employing both the normal law and Emergency Regulations (ERs). In 1948 this meant constant patrols, and attempts to ‘screen’ villages in support of local police. This might involve fruitless questioning, or else lining up villagers in the dark in the glare of jeep headlights, from which a hooded informer – a ‘ghosthead’ or kuai tao – pointed out alleged insurgent collaborators. Or informers in a van might knock as suspect villagers passed. ⁴⁷

The ‘everymansland’, liminal nature of the contested forest frontier, combined with minimal intelligence from pro-communist or geographically vulnerable and frightened villagers, encouraged high levels of coercion against Chinese squatters. There is another major incident that illustrates the more extreme implications of this liminal, everymansland status of many rural settlements, combined with the early dearth of intelligence.

In June to July 1948 insurgents briefly took a small number of small towns and villages, most notably holding the remote northern town of Gua Musang, in Kelantan, for a couple of weeks. These headline-grabbing operations included the seizing of mining town Batu Arang, in Selangor about 20 miles north of Kuala Lumpur. The communists occupied it for a day on 12 July 1948.

Security forces decided that responsible insurgents and supporters must be broken. At the end of the month police, and men from 26 Regiment Field Artillery, quick-marched into the area inland of the town, clearing this and what they described as two ‘camps’ with armouries. One soldier subsequently described this as like a rabbit shoot, with no shots returned despite a few pistols being found. A military report claimed 27 were killed; later on Chinese would claim many if not most were civilians. At least 47 were taken prisoner. The 26 Field Regiment War Diary insisted the area had included two bases – possibly rural communities where insurgents had received support – and that one man was shot, pistol in hand. But the fact that the Colonial Office demanded to be told more, and that in excess of 20 were killed for only one claimed to be armed and no return fire, spoke volumes, as did the fact that ‘the entire area of about 10 square miles’ was razed by the army and police. Soldiers seem to have been primed to clear an everymans area that provided mixed civilian habitation and bandit support, as a result killing anything they imagined threatening, unless it promptly and unambiguously surrendered. The result seems to have been a mix of genuine captures and probable civilian casualties. ⁴⁸

Batu Arang thus echoes the scale and logic of Batang Kali, even though the nature of the two operations was very different. In both cases troops were primed to go into an everymansland, and to expect local villagers to be ‘bandit’ supporters intermixed with actual bandits. Batang Kali and Batu Arang thus demonstrate the extremes respectively of what could happen when a patrol attempted to use force sufficient to get cooperation from frightened villagers, and when military and police launched a large operation against an ‘everymans’ area that hosted both civilians and insurgents.

The more normal pattern was for patrols to screen settlements and make a few arrests, or return with no contact, or for one, two, or at most a few people to be shot ‘escaping’ or running towards jungle or plantation. Between June 1948 and April 1949 at least 77 were shot ‘escaping’, 37 captured in the attempt of whom several were wounded, 7 wounded escaping, and 4 killed in accidents. A proportion of ‘escapes’ were insurgents or armed supporters, with multiple reports of weapons and documents found. ⁴⁹ Some of those shot running will have been unarmed but carrying supplies or communist documents. Some will have been shot in error. In cases where death was suspicious coroners sometimes made friendly reports, and it would have been easy to claim weapons were found. ⁵⁰ Nevertheless, with the majority of patrols and screenings ending peacefully, the picture is not one of calculated or exemplary counter-terror killing, but of the authorities enabling quick and decisive action, and encouraging the understanding that transgressions would not be closely scrutinized. The early months thus established a pattern of many rural areas being contested, liminal, intelligence-sparse space where troops might easily assume all villagers were potential informants, that everyone who did not talk was culpable, and therefore that pressure might yield results. In such a context, a proportion of those killed ‘running away’ were civilians. ⁵¹ Similar if not much higher levels of killings of people ‘running away’ seem to have characterized the early phase of the Kenyan emergency (1952–53), and much of the Vietnam War. ⁵²

The excess deaths and counter-terror in rural areas were not, however, merely contextual. Security force guidance contributed a structural underpinning for such violence. On 14 August 1948 Headquarters Malaya Command provided guidance to North and Central subdistricts and the Malay Regiment. This addressed when troops could use deadly force. The formal instructions insisted troops could only open fire after giving a warning, and only if they had good reason to believe people were armed, or with armed people. They also limited use of force against potential escapees (unless armed or consorting with armed insurgents) to such non-lethal force as necessary to ensure arrest. But a covering letter to HQ Malaya Command’s ‘Notes’ also warned there would be uncertainty. It emphasized that ‘An incident occurred recently of troops withholding fire in a case of doubt and thereby guilty men escaped, later these men turned and fired on the troops.’ Recipients were instructed that in cases of doubt concerning arms they had a responsibility to fire. The instructions themselves also made it clear that, provided they obeyed the instructions, soldiers would be safe from prosecution even if people in the vicinity were accidentally shot. In short, the instructions attempted to encourage aggressive action and a willingness to make snap judgements about who was consorting with bandits, while straining to remain technically within ER limitations. Given the document’s internal tensions, the message that may have stood out most for soldiers was that, provided they could claim they were following these orders, they would not be prosecuted. ⁵³

These instructions still could not cover Batang Kali, where the patrol knew the workers were unarmed by 12 December having searched their premises on the 11th, even if they help to explain it. But they may also help to explain excess deaths elsewhere. The key issue is that several parts of the ‘Notes for the Guidance of Commanding Officers’ combined with overall strategy and geographical context to produce more killings. The guidance to shoot if in doubt and that escapees could compromise patrols, in the context of a strategy of applying ‘pressure’ to insurgents and civilian supporters in late 1948, helped frame a rural atmosphere of ‘counter-terror’. The proximity of the most disputed settlements to lallang (tall grass), rubber plantations, or jungle and mountain reduced the time for judgement and increased fear that escapees could quickly direct insurgents against a patrol.

The government’s response to the Batang Kali incident, with an ‘inquiry’ quickly backing the troops’ account, initially reinforced the danger of ‘counter-terror’. This reluctance to scrutinize actions properly was in turn part of a deliberate signalling that forces would be backed up if they took decisive action, even when things went wrong. The first government response was thus to allow the Attorney-General to make cursory inquiries. In December 1948 the High Commissioner, Sir Henry Gurney, reflected that ‘the Police and the Army are breaking the law almost every day’, and that though he could issue new Emergency Regulations (ERs), if this became too strong a flow it might undermine the law, and in extremis it would then be better to have martial law. This was the tension in Britain’s approach, not to use martial law, nor simply the rule of law, but through ERs to form a parallel quasi-legal world, regulated but with significant security force discretion and thin safeguards and, to the end of 1948, only limited attempts to signal that excesses would be reined in. ⁵⁴

The government was, nevertheless, keen to bring army actions under greater civilian control. After the police burned the small village of Kachau on 2 November 1948, the government made it clear that only actual buildings which harboured insurgents could be burned, and only on civilian authority. After Batang Kali, it issued ER27A in January 1949. This was a dual-edged response. On the one hand, it legalized shooting to kill if necessary to prevent escape, and was made retrospective, thus covering what the troops claimed happened at Batang Kali. On the other hand, it surrounded this with safeguards, insisting on a loud warning, and on the potential escapee being given opportunity to react to that warning. Official reports monitoring Chinese attitudes argued that ER27A had a reassuring effect. The government thus trod a delicate path between backing up security forces’ previous actions, and trying to moderate future behaviour. ⁵⁵

A number of conclusions arise from this period of June 1948 to early 1949. First, mutual permeability of contested settled areas, combined with poor intelligence and tortuously ambivalent ‘Notes’ for officers, framed rural villages as potential killing fields, where nervous villagers might run, and soldiers were primed to shoot runners. As more remote settlements were resettled, especially from 1950, these incidents would decline.

Second, the covering note we have on Malaya Command’s August 1948 guidance suggests that reading ERs and formal policies does not give the full picture. There is an element of institutionalized violence here, in the sense of government emphasis and omissions creating the environment for extreme violence. Unsurprisingly, the rural population felt itself crushed between two terrors. The ‘counter-terror’, however, would soon take on new forms. In 1949 it would be extended to large-scale legalized destruction of, or removal from, homes, and the mass deportation of squatters from ‘bad’ or ‘black’ areas. This raises the question of how far these latter two actions constituted a new, very different, type of ‘counter-insurgency structural violence’.

It was the combination of the above security force actions, together with intense patrols and screening, which led the High Commissioner to believe the policy of constant pressure was working. On 19 December 1948 Sir Henry Gurney telegraphed the Secretary of State for the Colonial Office about the Chinese – including the commercial leadership he was courting: ‘They are as you know notoriously inclined to lean towards whichever side frightens them more and at the moment this seems to be the Government.’

Population Displacement as ‘Structural Counter-insurgency Violence’?

From October to November 1948 the policy of exerting ‘pressure’ saw police and military units attempting to extract cooperation from unwilling or terrified rural Chinese, whose villages were being used to support anti-British forces (from 1949 renamed the Malayan National Liberation Army, MNLA). ⁵⁶ The government’s growing belief that it could not reverse the interpenetration of some areas would ultimately push Gurney to a policy of systematic, state-sanctioned removal of their inhabitants.

At first, however, pressure on the most isolated villagers was sporadic and ad hoc. We have already seen how early attempts to operationalize ‘pressure’ on villages meant screenings (with most people immediately released) and sweeps. We have also seen how for some units it turned into more kinetic impact.

Another way that the security forces tried to operationalize ‘pressure’ (and reverse the intelligence deficit) was by threatening villagers with repercussions for not providing information, including loss of homes. Many rural Chinese plots were occupied on one-year Temporary Occupation licenses, or ‘illegally’ on private or state property, or on ‘Malay’ or forest-reserved land. From August to November 1948 a couple thousand such villagers were resettled from a few particularly ‘bad’ areas, with indifferent results. In the largest case many of those resettled soon started to drift away.

With resettlement marginal, and deportation of less than 500 Chinese by the end of 1948, security forces sought other ways of exerting non-lethal but still ‘violent’ pressure. This came to a head at Kachau on 2 November 1948. It lay about 20 to 25 miles southeast of the capital, Kuala. The surrounding area had played host to an anti-Japanese force in 1942–45, and now boasted an insurgent group who drifted in and out of nearby forest. Army units kept up patrols, occasionally killing a few ‘bandits’. But the flow of information from the population remained feeble into 1949.

The Officer-Coordinating-Police District (OCPD) Kajang warned Kachau’s villagers that if nearby attacks continued, along with no information, he would burn their houses. Following attacks on a mine on 30–31 October and Dominion Estate early on 2 November 1948 he duly took a Police Jungle Squad, ordered the inhabitants out, and within half an hour was burning around 60 houses, rendering over 400 people homeless, with no planning for their subsequent fates. ⁵⁷

Officially the government stood firm. The police report claimed the villagers had had two hours to remove property, that the effect was good, and that other villagers had then moved voluntarily. Privately, Gurney had had enough of such ad hoc action. It was clear to him (as he expressed it in December) that security forces were ‘breaking the law almost every day’. His policy was both to reign in action beyond the law, and simultaneously to reel out the law to embrace and organize more draconian measures. He made it clear to Colonel Gray (Commissioner of Police) in November ‘that we cannot tolerate wholesale burning of Chinese villages without provision for resettling the inhabitants and giving them some opportunity for employment other than with the bandits’, and prepared a raft of new measures. ⁵⁸

These measures made manifest a policy which had started to coalesce as early as September 1948. Government had then called for aggressive patrolling to keep ‘bandit’ gangs moving and breaking up, combined with ‘pressure’ on civilian areas supporting them in the form of sweeps, screenings, and burning property thought to have been used by or in support of bandits. In this period ‘population displacement’ (by directly destroying some people’s properties, and by others taking the hint to move from the most insurgent-dominated areas) was as prominent, if not more so, as resettlement. In addition, the reaction of individual OCPDs varied, as did units, with the Scots Guards at Batang Kali on 11–12 December being the most extreme manifestation. Gurney’s response to this unstable situation culminated between November 1948 and early 1949, with new directives that fell into four main parts.

First, he attempted to regularize security force action as opposed to allowing ad hoc initiatives that meant that the SF ‘were breaking the law almost every day’. ⁵⁹ A slew of new ERs followed. ER18A and B of 13 November 1948 made it legal for an OCPD to order any property used by, or in support of, bandits to be seized, or if not amenable to being held, to be destroyed. ⁶⁰ Likewise, to cover shootings of people ‘escaping’ (as opposed to only those consorting with armed insurgents), January 1949’s ER24A allowed force to be used up to deadly force provided a clear, shouted warning was issued first, and the potential escapee given an opportunity to respond. The ultimate new weapon versus ‘bad areas’, however, was ER 17D of 10 January 1949. ER17D was to transform the way ‘structural violence’ (in the form of destroying select villagers and squatters’ homes and farms) was executed.

Second, then, ER17D was devised to address the most challenging rural settlements, offering a systematic answer as to what to do with ‘black’ squatter areas where the government did not feel able to disentangle innocent civilians. Selected areas were to be given a soft eliminationist approach: their populations totally removed from the area and potentially the country too. This compares with the previously much more limited powers to deport individuals under ER17C. ⁶¹ ER17D allowed for the collective detention of all the residents of any area that was deemed to have been assisting the bandits or withholding information on them. The logic was not just punitive, but also deeply practical. The government believed itself unable to separate out insurgent and civilian in the worst areas. Removing them to camps and then, for some, to China was intended to sterilize these areas, to help restore confidence to their neighbours, to act as warning to other areas, and to convince Malays concessions were merited to those Chinese ‘squatters’ who would then remain. In early 1949 Gurney hoped to deport up to 2000 rural Chinese per month for 12 months, making a total of around 24,000 men, women, and children. ⁶²

Third, Gurney attempted to add safeguards and bureaucratization for the most drastic types of security force action. It would be easy to caricature the new ERs as simply punitive, but they were also intended to limit abuse and increase government control over security forces. In reality, for instance, individual patrols were already burning huts on their own judgement. ER18B of November 1948 did regularize this, but it also required every destruction to be reported to the Menteri Besar (the first Minister in each Malay State), who could order compensation if property seized or destroyed did not properly fall under the regulation. In addition, Gurney told Gray in November 1948 that wholesale burning could not be tolerated until arrangements were made for reception or resettlement of any affected villagers. For ER24A, meanwhile, the ‘safeguard’ was the need for a warning ‘in a loud voice’ and a ‘reasonable chance’ to respond. The knowledge that Batang Kali had caused media scrutiny possibly also alerted patrols to the danger that the press would cause trouble over future excesses. As for ER17D, using the word ‘safeguard’ might seem a bit strong, as one of its effects was to remove the safeguard of review before deportation. Nevertheless, 1949 was to be very different from 1948, the latter of which saw ad hoc actions with operationalization from below, by individual military and police officers. Policy in 1949 envisaged formalized 17D operations following a considered police file being drawn up against each village, with the evacuation integrated with the Social Welfare department, detention camps, and army and police forces. ⁶³ This was to be state counter-terror bureaucratized, magnified in scale, and strictly controlled from above.

The fourth area that Gurney stressed was the need for more positive action, in the form of resettlement of rural Chinese in roughly the area they were in. Gurney had formed a committee to investigate the squatter issue by December 1948, and in January 1949 it recommended settlement of most rural Chinese in roughly the area they were in. It would take most of 1949 for the Malay States to accept this wholeheartedly, and commit to providing land. So while Gurney wanted this more constructive policy to predominate, the reality in 1949 (Malay opposition, and constrained central government finances) would be that ER17D remained more important.

From January to December 1949 16 ER17D operations saw over 6300 villagers sent to detention camps, some for screening and release, and others for deportation. In each case, it was as if Kachau had been bureaucratized. The first village detained under ER17D was surrounded pre-dawn, dogs shot, goods enumerated, and receipts issued, and residents screened by Social Welfare to identify health issues and remove the vulnerable, before most were shifted via transit sheds and truck and rail to a detention camp. This became the rough pattern for the next 15 such operations. ⁶⁴

Sergio Agamben’s idea of sovereignty as the ability to create ‘states of exception’ is relevant here, ⁶⁵ the critical issue being the ambivalence of citizenship in the Federation of Malaya. Most Malays were de jure citizens, but before September 1952 most Chinese were not. Before 1930 most of the latter had been temporary economic migrants, but the Sino–Japanese war and then Chinese civil war meant that by 1948 most were locally rooted. Most Chinese children would either scarcely remember China, or have been born locally. In June 1948 discussions in the Colonial Office it was recognized that deportation would remove people who had, de facto, made their home in Malaya, but was felt expedient to stick to the principle that non-citizens could be ‘repatriated’. ⁶⁶ Is such denial of habituated domicile ‘structural violence’? It certainly looks like cruel treatment, even if the British authorities rationalized it as the removal of aliens who had acted illegally, and so as something any sovereign state was entitled to do. It also caused real hardship to many early deportees, for many of whom compensation payments for losses also proved woefully insufficient.

The authorities did not achieve Gurney’s initial ‘24,000 in a year’ target for deportations, or ‘repatriations’ as the government called them. Nevertheless, the clearing of 16 ‘bad’ areas in 1949 did result in a few mass deportations, which were accompanied by a still larger number of ER17C deportations of individuals from other areas. By March 1953 more than 29,000 Chinese had been sent to China, with peaks of 10,262 in 1949 and 8719 in 1951. Deportations began to decline sharply from 1953, by when most resettlement was complete, and insurgent fortunes had gone into steep decline. Even then, from 1954 (when there were 915 repatriations) the British introduced a ‘fast-track’ deportation procedure, so they could tell villagers within 52–60 days of arrest that a person was on a boat, and point out that anyone else who helped the communists could enjoy the same hospitality. ⁶⁷

In short, ER17D acted as an aide to regaining control of rural areas in 1949, before resettlement took over this role on a far larger scale. Together with ER17C (individual repatriation orders) it acted as a safety valve afterwards. You did not need Kenya-style ‘dilution’ of prisoners with those already ‘turned’, for instance, when many of the harder cases could simply be deported. ⁶⁸ Of the total 31,249 deportations up to March 1953, they can be broken down as follows: 29,287 Chinese, 1786 Indians, 12 Ceylonese, and 164 Indonesians. ⁶⁹

Hence 1948 and 1949 look very different. In 1948 we have episodic killing of civilians, albeit with a ‘systemic’ element of army guidance and state ambivalence towards investigating excesses. The key locus was the rural everymanslands between the plains and the main range. The geodemographic conditions there – tight mixing of insurgents and supporters with villagers – also fuelled non-lethal structural counter-terror measures such as the burning of huts.

In 1949, this ‘counter-terror’ was systematized and bureaucratized and excess killings and uncoordinated burnings declined sharply. No more large examples have come to light for this latter period so far, and reports of individual shootings decline. But the new, more ‘structural’, non-lethal counter-terror intensified with the big 17D operations, despite the insurgents reducing activity in order to reorganize. This sort of major qualitative change makes it important for us to use our terms with precision, and to make sharply delineated observations about the changing nature of state violence, counter-terror, and pressure over different periods.

This point can be further pressed home by summarizing the major changes from 1950. While 1949 was the year of 17D operations, 1950–51 saw just one 17D operation alongside 18 months of breakneck resettlement. As of 10 March 1950 just 18,500 had been resettled, despite the acceptance of the Squatter Committee Report in January 1949, with its call for reintegrating Chinese ‘squatters’ by settling and protecting. By the end of 1951 there were, by contrast, 461,000 people in more than 353 resettlements (later rebranded ‘New Villages’).

The transformation was down to a number of factors. The MNLA, having reorganized in 1949, increased activity late in the year, causing a crisis in British and Malay confidence. This facilitated Britain’s appointment of Lt-General Briggs as the first Director of Operations (DOO) to coordinate all forces under the High Commissioner. He arrived in time to take advantage of the Korean War boom (meaning Malaya’s rubber and tin commanded high prices and generated higher wages and more government revenue), and the state’s willingness, given the crisis, to allow central coordination of rapid resettlement.

OPEN IN VIEWER

Table 1.

Progress of resettlement, 1950–52.

Resettlement (cumulative totals)

1950 – 10 March 18,500 (the estimate then being that 300,000 more need resettling)

1950 – end 117,198

1951 – end 401,698 in 353 New Villages

1952 – end 461,822 in 509 New Villages

Source: CO1022/29 Resettlement. ⁷⁰

From June 1950 onwards Briggs put into place an integrated plan at the core of which was rapid resettlement, intended to protect villagers, to separate them from insurgents, and ultimately to reintegrate them into normal society. By late 1951 some of these resettlements would be raising part-time Home Guards, and the administration was ready to put renewed emphasis into their as-yet primitive ‘aftercare’. From 1952 that brought increasing numbers of schools, community centres, visiting nurses and medics, and elected committees. Vegetable production, after an initial slump, also started a strong resurgence.

The communist view of these ‘resettlement areas’ was that they were little more than ‘concentration camps’. The British view was that the resettlement areas were embryonic communities (hence their rebranding as ‘New Villages’ in 1952), which it was eager to become organic entities with everything from scout groups and committees to badminton and football league matches. ⁷¹

Both representations have merit. Initially the British planned lines of houses, behind barbed wire, and some resettlements starting as little more than tropical slums. A few failed altogether. At the peak of operations in any area its New Villages might experience extended curfews, strict gate searches, central cooking of rice or reduced ‘operational rations’, sudden arrests of hard-core suppliers, and then gradual compromising and turning of their softer replacements. From 1952, if a village was deemed to be withholding information restrictions might also become punitive, albeit for a short, sharp spell usually of a few days to a week. At its extreme this might see all but one shop closed for up to a week-long curfew, and everyone kept in while questionnaires were distributed and house visits made (Operation Questionnaire). ⁷² Yet at the same moment one village was subjected to intensified control as part of intelligence–army–police operations to generate ‘live intelligence’ for ambushes, another would experience relaxed controls, increased facilities, or from 1953 might gain ‘white area’ status and a lifting of most Emergency controls.

In other words, the structuring of operations around resettlement involved a rich blend of coercive and rewarding strategies, of militarization and yet also gradual integration into normal civil structures. Rigid geodemographic control (inter-related control of space and people) did not sit alone. ⁷³

So did ‘population control’ constitute a form of counter-insurgency ‘structural violence’? The answer depends on which aspect of population control, and which definition of structural violence, is used.

For 17D operations of 1949, and large-scale deportations of Malayan-domiciled Chinese, the degree of attack on people’s livelihood and ability to continue in Malaya constituted an intense and in that sense ‘violent’ attack on rural Chinese, which undoubtedly caused many intense losses and suffering. It was also valued for its ability to frighten if not terrorize surrounding villagers and villages, with ‘voluntary’ moves from surrounding areas sometimes a welcomed knock-on effect. Finally, ER17D involved physical destruction of homes, and (despite variable compensation) deportation removed people from their means of support. The period of intense bureaucratized counter-terror does seem to constitute ‘structural violence’.

For resettlement, however, though it was strongly coercive, it is problematical to label it ‘structural violence’: if we insist on a definition that implies the necessity of physical hurt, or uncompensated destruction, or sustained reduction of means of livelihood. That resettlement was accompanied by use of barbed wire, exit points, and more police is undeniable. The means of livelihood also suffered a temporary setback in the first 18 months or more of resettlement, as people had to rebuild homes and lives. In terms of effect, however, one set of land – often held tenuously and with few facilities – was replaced with another, ultimately with better facilities, and more often with more secure tenure. Slowly, villagers also started to take up land title for their plots.

In addition, the ultimate aim of resettlement was to inoculate the village from violence, which was to be displaced first to the zone between village and forest fringe (with information extracted from the village used to arrange ambushes), and then into the forest itself. Resettlement, furthermore, was intended to contrast the bureaucratized ‘violence’ of action under ER17D, and to reduce the danger of bodily harm inherent in the counter-terror of 1948. The first stage often involved establishing a police post, which might have to fight off increased insurgent attacks, and which could then provide cover and a measure of security for resettlement. Once the fence was up and houses erected, the army could patrol the area. Once a Home Guard was effective, police presence could gradually be reduced. Thus there is a tension between, on the one hand, the clearly coercive nature and massive scale of resettlement and severe temporary dislocation, and yet its lack of bodily violence, and compensation for losses due to any loss of property. ⁷⁴

Even increased gate and food controls were both restrictive and protective, since only a villager who could argue they could not smuggle out essentials could risk refusing insurgent demands. ‘Punitive’ measures also underwent a metamorphosis. Collective fines were used occasionally in 1950–52. From 1952 they tended to be associated with visitations by General Sir Gerald Templer, the new High Commissioner, and with Operation Questionnaire. These operations were refined to emphasize not being a ‘punishment village’ as a continuing state, but rather a sharp phase of one to two weeks, in order to try and extract information and scare insurgent supporters. A village would be warned that information was being withheld by it despite high activity around it, and both asked what help it needed and warned of punishment if information didn’t follow. Upon failure the curfew would increase, rations go down, questionnaires appear (providing avenues for information from frightened villagers), and after the latter were opened arrests be made and the village rewarded according to information given. These operations were of dubious validity, and in part bluff since even the pretence they worked could affect insurgents. But they make the point that the highest level of control, mixed with more than a touch of psychological fear, could be non-violent by nature and intended to prevent physical violence.

In post-1950 resettlement, therefore, we see both high levels of coercive control and yet also detailed development, with the vast majority of villages increasing their numbers into the 1970s. ⁷⁵ Whether population displacement constituted a form of ‘violence’ therefore seems not to have been inherent in the activity itself, but rather to be determined by the particular approach to it, and in this case, the particular phase.

Conclusion

A ‘post-revisionist’ analysis of counter-insurgency builds on, rather than rejects, the post-2004 work of French, Bennett, Anderson, Hughes, Hack and others, that has documented the calculated and sometimes exemplary British use of coercion and intimidation of civilian populations, and episodic and systematic extreme violence.

It makes the analysis of coercive measures more forensic, in an attempt to identify their precise nature, and their changes and also limitation over time. In Malaya, for instance, ‘population displacement’ cannot be reduced to a single stereotype, whether of ‘structural counter-insurgency violence’ or of ‘winning hearts and minds’. It ran through a first counter-terror phase (1948) that emphasized ‘pressure’ to cooperate or move through sweeps, burnings, and aggressive patrolling, a second, bureaucratized counter-terror phase (1949) that included 16 mass clearances and largescale deportation, through to a third phase (1950 following) that emphasized resettlement.

This article thus confirms how types and causes of counter-insurgency violence against non-combatants were distinctive, and demand in-depth analysis each on its own terms. Take three of the forms of state action towards civilians that have been analysed here.

First, there is the killing of civilians in groups of 1 to 24 between 1948 to early 1949. To take just the most extreme example, the killing of 24 Chinese villagers at Batang Kali must be related to the geographical struggle to dominate forest fringe, everymansland, liminal rural settlements in 1948, where neither side could exclude the other. This cross-fertilized with a severe intelligence shortage, with a strategy of exerting ‘pressure’ on insurgents and their supporters, and with August 1948 instructions priming troops to fire if in doubt. The combination of geodemographic context and instructions and government omissions framed a context where recurring excess killing was likely in 1948 to early 1949, though such killing was neither an aim nor a desire.

Second, there is the more controlled, bureaucratized approach to ‘black’ areas adopted in 1949. A raft of new ERs were brought in between November 1948 and January 1949, with the express aim of bureaucratizing the more ‘violent’ actions such as burning of huts and expulsions of villagers. In part, they were intended to limit the excesses and mistakes of the earlier period. ⁷⁶ Two 1949 court cases probably had the same signalling effect, resulting in convictions for culpable homicide of a child, and for extorting a confession using violence. ⁷⁷ This article has nevertheless argued that ER17D and mass deportation constituted a new, bureaucratized ‘counter-insurgency structural violence’ towards selected settlements and individuals, which centred around population displacement. Security forces violently destroyed people’s homes with no replacement, and dramatically removed people from their long-established country of residence.

Third, there is mass resettlement, which largely displaced ER17D as the main type of action from 1950, and could not be described as ‘violent’ in the same way. Despite some communities being removed at gunpoint (others by prior agreement), it substituted a new set of land and facilities for previous resources that were often limited and insecure. Though temporarily traumatic, resettlement generally did establish viable and durable communities, and if anything the trend over time was towards these having better means of livelihood than what they replaced. Already by 1954 some of the resulting ‘New Villages’ had better facilities and access to markets, towns, infrastructure, healthcare, education, or jobs than their originating squatter area had possessed, despite high levels of controls.

So even within the broader category of ‘population displacement’ there are massive differences in approach, and a real debate to be had about which policies constituted ‘structural violence’. Each of the very different approaches to population displacement involved different types and extents of ‘violence’ and of alternatives to and limitations of violence.

In summary, broad generalizations about the ‘excessively violent’ nature of counter-insurgency in Malaya, and elsewhere too, cannot be allowed to waltz over the major differences in types of violence and non-violence that characterized constantly shifting cocktails of counter-insurgency policies. Instead, we need more in-depth analyses that trace the lifecycle of particular types of violence, and of violence limitation.