Ethnic ‘Ferociousness’ in Colonial Wars

This article explores the myths and evidence surrounding extreme violence and the framing of ‘ethnic soldiers’ as loyal and indispensable Moluccan soldiers in the Dutch army in the Indonesian War of Independence, 1945–1949. In this article, we first inter-rogate the origins of this framing in the Dutch–Indonesian case and the type of sources underlying this perspective. Next, we present the results of our research, which combines a study of Dutch veterans’ ego documents and oral-history projects. Based on this analysis, we reconsider both the framing and the evidence, after which we conclude with some comparative observations on ‘ethnic soldiers’ and the sources and perspectives underlying the ambivalent, but increasingly critical, framing of these men. Our methodology includes the use of digital-humanities techniques.

followed the proclamation of Indonesian independence on 17 August 1945. The Dutch government would only officially transfer sovereignty over four years later, on 27 December 1949. The intervening years saw a hectic alternation of bloody warfare and protracted negotiations. In all, over 130,000 Dutch soldiers were shipped to the archipelago, only a few of whom had prior experience either in warfare or abroad, let alone in Indonesia. On arrival, many of these soldiers felt totally unprepared for the task allotted to them. It is therefore not surprising that memoirs of Dutch veterans contain countless passages expressing appreciation for the indigenous soldiers fighting on their side (Smeets and Steijlen 2006:38;Oostindie 2015:127-34).
This framing of 'ethnic soldiers' as loyal and indispensable but at the same time somehow exceptionally prone to violence is a recurring theme in narratives of colonial warfare. It is obviously also a very uncomfortable, Orientalist framing which tends to overlook the fact that indigenous troops were often ordered to fight in the most risky circumstances and were often encouraged by their white superiors to engage in extreme violence. While we reject an Orientalist framing, the reading of persistent stereotypes as quoted above did make us wonder about this framing and the reaction it evokes up to the present day.
So, what we attempt to do in this article is not so much to reconstruct what happened in [1945][1946][1947][1948][1949] but what was, and is, narrated about and by Moluccan soldiers. We first sketch in a bird's-eye view the history of Moluccans in the colonial army and their framing as ethnic soldiers. Next, we present the results of our research, which combines a study of Dutch veterans' ego documents with oral-history projects. Based on this analysis, we reconsider both the framing and the evidence, and conclude with some general observations on 'ethnic soldiers' and the sources and perspectives underlying the ambivalent, but often critical, framing of these men.

Indigenous Troops in Dutch Colonial Warfare
Dutch colonialism in Indonesia dates back to ca. 1600, when the Dutch East India Company (Vereenigde Oost-Indische Compagnie, or voc) established a monopoly on the spice trade from the Moluccas by brutal force, creating a colonial system on several Moluccan islands. This period also saw the first recruitment of Moluccan men to serve as soldiers for the voc. The colonial state only embarked on a consistent policy of territorial occupation in the early nineteenth century, through a series of minor and major wars that was more or less successfully concluded in the first decades of the twentieth century. Established in 1814, the Dutch colonial army (Koninklijk Nederlands-Indisch the Moluccan community in the Netherlands, the core of which was formed by demobilized knil soldiers sent over to the metropolis.6 The Japanese assault on the Dutch East Indies in 1942 demonstrated that the knil was no match for the Japanese army. After the colonial army's capitulation, its European and Indo-European personnel was interned and put to work in Indonesia and elsewhere in Southeast Asia, particularly Thailand and Japan; a small minority found rescue in Australia. While many of the native knil soldiers were soon released, the Japanese targeted the Ambonese as loyal to the Dutch and treated them more or less as the European knil staff-this meant continued internment and hard labour. The Japanese capitulation on 15 August 1945 and the proclamation of the Indonesian independence by Sukarno and Mohammed Hatta two days later set the stage for the final round of Dutch colonial history in Indonesia.7 The Republicans, understandably, perceived the arrival of Allied (British) troops in September as an overture to Dutch recolonization, and indeed the Dutch government prepared for re-instalment of its regime. Both the nascent Republican army and a wide array of independent armed groups prepared for the struggle against recolonization.
Mass violence flared up in these last months of 1945. In many parts of the archipelago, the Indonesian Revolution implied a reckoning not only with Dutch citizens but equally with communities that had had a long history of cooperation with the colonial regime. This chaotic and complex process, known as Berdaulat in Indonesian historiography and Bersiap in Dutch historiography and particularly in the memories of those in the Dutch East Indies, victimized native elites, pro-Dutch Indonesians, and specific ethnic groups, in particular the Europeans, Eurasians, Ambonese and-probably most vehemently-the Chinese.8 knil soldiers returning from internment on Javaincluding the newly formed Moluccan militia-acted in a provocative and 6 The cliché of 'loyal through the centuries' is the motto of Oom Ambon van het k.n.i.l. (Dames 1954). For later analyses, see Chauvel 1990:39-70, 197-210;Smeets and Steijlen 2006:29-35;and Steijlen 2015. 7 The actual final round was the handing over of Netherlands New Guinea to Indonesia in 1962Indonesia in /1963 via the United Nations. The 1949 transfer of sovereignty to Indonesia did not include Netherlands New Guinea. 8 The number of casualties is impossible to establish. Estimates of the European and Eurasian victims vary from some 3,500 to 20,000 (Frederick 2012:359-80). The concept of 'bersiap' derives from the word 'siap' , meaning 'get ready' , a much-used battle cry in these months to encourage Indonesians to defend the Revolution. The word gained currency in the Netherlands long after the war to denote the violence experienced by (Indo-)Europeans. The concept of 'Berdaulat' has a broader meaning, covering all intra-Indonesian violence in the early phase of the Indonesian Revolution.
trigger-happy manner, thus fuelling this bloody process, which lasted until early 1946. Moluccans were indeed both victims of the Bersiap period and perpetrators of counterrevolutionary violence.9 Due to their aspiration to reoccupy the colony, the Dutch began rebuilding the knil, starting in the eastern part of Indonesia. In eastern Indonesia, Moluccans in particular were recruited on a large scale.10 Although many of them were simply looking for a relatively well-paid job and may not have intentionally opted to stand against the Indonesian Revolution, their decision to serve with the knil was considered by friend and foe alike as confirmation of their pro-colonial allegiance as Belanda hitam (black Dutchmen). As part of the newly established Netherlands Indies Civil Administration (nica), they were described by the nationalists as 'andjing nica' (nica dogs).11 Tellingly, a dog is an animal that is considered haram (forbidden) by many Muslimsand just as tellingly, throughout the war, the Ambonese defiantly used 'andjing nica' as a proud moniker.12 It even became the name of the fifth battalion infantry of the knil, established in December 1945 in Bandung by the Dutch, Indo-Europeans, and Moluccans recently released from Japanese imprisonment (Matanasi 2007:115).
However, this does not present a complete picture. From the early days of the Indonesian Revolution, Ambonese were also fighting on the Republican side, having the same reputation as fierce warriors. They were a conspicuous presence in Sukarno's personal guard and were even grouped together in the Pattimura Battalion, named after an early-nineteenth-century Moluccan leader who fought Dutch colonialism (Chauvel 1990:203;Smeets and Steijlen 2006:37).
In the escalating warfare of the late 1940s, the knil's Indonesian and particularly Ambonese rank and file were ascribed particular fierceness, and this view of them has persisted. Eagerness to take revenge for violence committed against fellow Ambonese during the Berdaulat/Bersiap period was often advanced as an explanation for this alleged fierceness, but as we will see, for the Moluccans newly recruited after the Second World War, this was most likely not a motive to enlist. Another explanation often advanced for this assessment of particular fierceness is that Ambonese, like other Indonesian troops under Dutch command, were disproportionally employed in potentially violent activ- 9 Kahin 1970:143;Chauvel 1990:201-3;Limpach 2016;McMillan 2005;Matanasi 2007. 10 Some 60% of the Moluccan soldiers who came to the Netherlands in 1951 had been recruited after the Second World War (Steijlen 1994:65-75). 11 According to Matanasi (2007:115), the term andjing nica was originally used for both Moluccans and other pro-Dutch Indonesians. 12 Kahin 1970:143;Chauvel 1990:201-3;Limpach 2016;McMillan 2005. ities that required familiarity with the terrain, and interrogations that required their specific linguistic and local knowledge. This line of argumentation makes more sense. Finally, assumed excessive warfare has often been explained by the fact that the Ambonese (like the Indo-Europeans) had everything to lose. Whereas the Dutch military were fighting an overseas war from which they hoped to return safely to their previous lives, their Indonesian mates were fighting for their way of life, as supporters of the colonial regime. For them and their families, defeat could spell disaster. This interpretation has credibility. Indeed, even after the transfer of sovereignty, Ambonese former knil military were involved in coup attempts against the Republic and particularly its policy of converting the federal state into a unitary state administered from Jakarta, Java. In January 1950, Moluccan soldiers led by former knil captain Raymond Westerling participated in the amateurish 'apra'13 coup against the federal government of Sukarno; in March 1950, Moluccan soldiers were prominently involved in a revolt led by Andi Azis in Makassar to prevent the inclusion of the federal state of East Indonesia into the Republic Indonesia; and, finally, in April 1950, Moluccan knil soldiers attempted to establish a sovereign Republik Maluku Selatan (rms, Republic of the South Moluccas) (Steijlen 1996a:38-40).14 The rms held out longer than the apra's and Andi Azis's coups. After the Indonesian army invaded the Moluccan islands in June, they were confronted by fierce resistance from former knil soldiers who had already been demobilized in Ambon and now formed the backbone of the rms army. In November, Indonesia regained control over the Moluccan islands, but rms guerrilla warfare continued in the interior of the Moluccan island of Seram until rms president Chris Soumokil was arrested in 1963 and executed for betrayal in 1966.
While more or less half of the Indonesian knil soldiers eventually changed sides during or immediately after the war, or were demobilized, a substantial number of former Moluccan soldiers, located outside the Moluccas and waiting for demobilization, supported the rms proclamation. Because of this and because the Indonesian and Dutch governments did not want trained soldiers to be demobilized on the Moluccas at that time, their demobilization became severely complicated. In 1951 the situation reached a stalemate, and eventually the Dutch government reluctantly decided to ship these roughly 3,500 men with their families to the Netherlands, some 12,500 people in all. 13 Westerling had organized a secret personal army called Angkatan Perang Ratoe Adil (apra, the Army of the Righteous King). 14 The proclamation referred to the South Moluccas, because the rms was proclaimed by the governing council of the Daerah (region) South Moluccas.
Upon arrival, the Moluccan soldiers were demobilized and collectively housed in camps dispersed throughout the country. Initially these Moluccans as well as the Dutch government thought that their stay would be temporary. In an attempt to put the rms on the international agenda, second-generation Moluccans initiated violent Moluccan political actions starting with arson of the Indonesian embassy in 1966; this was followed, in 1970, by the occupation of the residence of the Indonesian ambassador; train hijackings (1975 and 1977); and occupations of the Indonesian consulate (1975), a primary school (1977), and the provincial government office of Drenthe (1978).15 While this eruption of violent protest was short-lived, it may have revived the idea of a Moluccan 'martial race' ,16 both within their own community and in wider Dutch society. This may in turn have influenced the ways in which Dutch veterans presented their Moluccan, or 'Ambonese' , brothers in arms in their memories of the war in Indonesia, which were mainly written down in the 1980s and later (Oostindie 2015:315).

Testimonies: Ego Documents
The historiographies of both the Indonesian Revolution and the decolonization process-consisting of warfare and negotiations-have long been surprisingly separate domains. Dutch historiography on the decolonization of Indonesia has long shied away from addressing colonial and postcolonial violence, focusing even for the 1945-1949 years on the political and diplomatic processes rather than on the violence of warfare. It stands to reason to explain this myopia as an unwillingness to accept that violence and racism were part and parcel of Dutch colonialism, and that colonialism as such was not a phenomenon somehow detached from metropolitan history but integral to the national history of the supposedly progressive, liberal, and non-belligerent Netherlands.
In recent years, however, the Netherlands has witnessed a broad and mainly critical 'rediscovery' of colonial history, and particularly of the 1945-1949 war in Indonesia. Juridical procedures started at the behest of families of Indonesians randomly killed by Dutch troops forced the Dutch government to recognize, albeit grudgingly, responsibility for these killings, to make apologies, and 15 An overview of Moluccan history in the Netherlands is provided in Smeets and Steijlen 2006. A summary in English can be found in Oostindie 2011:85-8, 114-6, 164-70, 174-6, 218-9. See also Steijlen 2010:143-62;Steijlen 1996a and1996b. 16 Comparable with that of the Gurkhas in the British Imperial Army. The Gurkhas were also considered a 'martial race'; see Streets 2004. to pay compensation (Scagliola 2014;Lorenz 2014;Oostindie 2019). A growing number of historical studies and press publications has severely undermined the official Dutch position, dating from 1969, that such war crimes were only sporadic 'excesses' in a war otherwise conducted in a 'correct' manner (Brocades Zaalberg 2015:67-83;Limpach 2014Limpach , 2016Oostindie 2015). In 2017, finally, the Dutch government financed an extensive and independent research project on the war, with a focus on the description, analysis, and explanation of Dutch extreme violence.17 As this research progresses, one of the issues that has surfaced pertains to war crimes perpetrated by Indonesians in the Dutch army, and particularly Moluccans. This is a delicate issue. The belated recognition of Dutch excessive violence in Indonesia inevitably entails debates about Dutch racism and 'Othering' in colonial history-in this new context, paying attention to indigenous knil violence may well be used as an attempt at white-washing, or at least it may be interpreted as such. Within the Moluccan community in the Netherlands, which was in many ways itself a victim of decolonization, this issue is no less delicate. Debates within the community hover between pride and uneasiness about colonial military traditions and bitter resentment of Dutch post-war policies relating to the Moluccan knil soldiers as former supporters of the Dutch (Steijlen 2016a). Thus, there is all the more reason to address this issue with due caution.
So, what is the evidence? Dutch propaganda during the 1945-1949 war obviously did not dwell on military violence but rather on the peace-keeping task allegedly performed by the Dutch army, the knil included. According to this propaganda, their duty was to combat the enemy-routinely described as 'terrorists' or 'extremists'-in order to restore 'Order and Peace' for the benefit of the Indonesian population. The ethnic diversity of the army was hailed as exemplary. Thus, in the middle of the war, chief of staff Lieutenant-General H.J. Kruls wrote of 'camaraderie, often even friendship, even when one doesn't understand the other's language and can only interact through sign language and a few words.
[…] Brothers in arms, that's what they all are, in the Indies, no matter whether they're kl [Dutch army] or knil, whether their skin colour is white or brown' (Kruls 1947:90).
In Dutch military archives and the historiography subsequently based on this corpus, we do find ample evidence of the participation of indigenous and particularly Ambonese troops in the Dutch army, but at first sight there 17 The results of this project will be published early 2022; both authors are part of this research programme. See https://www.ind45-50.org/en (accessed 25-8-2021).
is little direct evidence or discussion of a link between these 'ethnic soldiers' and extreme violence. However, both military archives and contemporary press reports indicate that knil troops and particularly commandos (the so-called Red and Green Berets) accounted for a disproportionate share of Dutch excessive violence. This has been well-documented in subsequent historical studies (for instance, De Moor 1999;Limpach 2016). While the leadership of these units was Dutch, the rank and file were mainly locally recruited. In archived reports, the indigenous soldiers are rarely identified by their names, in contrast to Dutch personnel, and one can only guess at the ethnic background of the troops. In contrast, closer inspection of the list of names of casualties in the Dutch army-just below 5,000-suggests that at least one-third of these were born in the colony, not in the Netherlands (Litjens 2020). It is precisely the best-known and, for his violent actions, ultimately mostcriticized commander of the Dutch commando units, Captain Raymond Westerling, who gave a new twist to the pre-war image of the ever-loyal Ambonese. In his memoirs, published shortly after the war, Westerling spoke with great appreciation of his Indonesian, largely Ambonese troops, whose 'loyalty and affection' were beyond doubt and whose conduct in military operations had 'never disappointed me' (Westerling 1952:94; see also Venner and De Vries-Spoor 1982). Though he made a point of taking full and unapologetic responsibility, it is clear that Indonesian, and specifically Ambonese, soldiers were instrumental in his actions, which afterwards would be labelled as extremely violent or considered war crimes.
Westerling, who had staged his unsuccessful 1950 counterrevolutionary apra coup with a small band of mainly Ambonese troops, would not be the last to extol Ambonese belligerence, and hence helped to perpetuate a view of this group that dated back to the late nineteenth century. As early as 1954, the book Oom Ambon van het knil (Uncle Ambon from the knil) had paid tribute to the loyal Moluccan soldier, echoing the fierceness of the Ambonese during the Aceh wars as described by Zentgraaff in 1938 (Dames 1954;Zentgraaff 1938).
The body of ego documents subsequently produced by Dutch veterans of the war included recurring, but also ambivalent, indictments of Ambonese soldiers, as we will illustrate. But before discussing this evidence, mainly taken from the book Soldaat in Indonesië (Oostindie 2015, see also Oostindie 2018), we should make a caveat about this type of sources, and particularly the unique corpus of some 650 published ego documents featuring over 1,350 soldiers of all ranks and totalling over 100,000 pages in print analysed for present purposes (Oostindie 2015:310-9). Just like the Dutch governmental and military archives, the overwhelming majority of all ego documents in our database was written or recorded by Dutch men. This results in a dramatic imbalance, with many hundreds of Dutch soldiers and veterans commenting about the war, about their own and fellow soldiers' behaviour, and specifically about their indigenous comrades in arms-but only a very few out of the latter category providing an alternative perspective. The evidence emerging from this corpus of ego documents is therefore highly biased and should be interpreted with utmost caution-these are Dutch soldiers speaking about 'Ambonese' soldiers, hence producing decidedly one-way and possibly distorted (re-)constructions whose reliability cannot be taken for granted.
The analysis of this huge corpus through conventional methods by a group of researchers ranging from ma students to senior historians resulted in the book Soldaat in Indonesië, which discussed the entire trajectory of Dutch soldiers, from their recruitment in the Netherlands to their post-repatriation lives, which often included bitter memory-making, with most chapters dedicated to the issue of Dutch war crimes. Alongside the opening quotes of this article, the following citations offer a fairly representative view of Ambonese soldiers among Dutch veterans:18 I am convinced that I have been saved by the Ambonese fighting on our side. What brave fellows! jan wilting, soldier [The Ambonese were known as] hard-nosed men whose methods were rather crude. anonymous soldier And proudly the Ambonese showed their trophies suspended on their belts. These were the ears of their slain adversaries. While the last two citations point to the differing positions of the Moluccan and Eurasian soldiers, most citations conjure up the uneasy framing of extremely loyal and indispensable, but also violence-prone Ambonese. While this framing is discussed with great caution in Soldaat in Indonesië, the available data did little to enable an alternative reading of the corpus (Oostindie 2015:139-41). We are aware though that not only are ego documents inherently biased, but that there may also have been a bias in the way that the entire corpus was analysed. While the research group had been instructed to systematically enter all evidence of excessive Dutch army violence in a joint database, references to a host of other topics, including depictions of indigenous soldiers, were not noted down as systematically. Hence it could be assumed that the researchers had mainly copied the out of the ordinary, thus inadvertently helping to construct the 'Ambonese as violent' view themselves.
Only subsequently did we find the means to digitize the entire corpus in order to apply methods from the digital humanities and hence refine and expand the research.19 While this digitized corpus is now being used by several researchers with diverse queries, we used it specifically to sharpen our analysis of references to Ambonese soldiers. In order to retrieve relevant fragments, we 19 The digitization was done by the Centre for Digital Scholarship (cds) of Leiden University. Further methodological advice and search queries were taken care of by Peter Verhaar of the cds and by Stef Scagliola (University of Luxembourg and kitlv).
Bijdragen tot de taal-, land-en volkenkunde 177 (2021) 491-523 used a co-occurrence analysis. We created two reference lists. A first list (A) with the relevant ethnic descriptions (Ambonese, Moluccan, South Moluccan, but also Menadonese). A second list consisted of verbs, nouns, and adjectives that were in a way connected to possible ways of judging or valuing the Moluccan soldier as fierce or loyal, or relating them to violence or non-violence. In fact, we used two of these last lists, one compiled manually and a second compiled automatically. The A list was used to select all fragments in which Moluccans were mentioned. The computer-generated search did produce new fragments which, while not substantially altering the previous analysis, do allow us to make some additional observations. First, on the loyalty and reliability of the Ambonese: The Ambonese were very reliable and capable soldiers. piet van der wijst, soldier20 The Moluccans have rendered us excellent services in Indonesia. They knew the tactics of the adversaries and were acquainted with the nature and culture.
hille van leeuwen, soldier21 They were fighters in the vanguard, experienced and ruthless warriors […] hence fantastic allies. gerrit and jan kuipers, soldiers22 The Ambonese were reputed to be the most extreme and fanatic men in the knil. karel c. snijtsheuvel, Eurasian knil soldier23 The enemy hates [the Ambonese], because the secret propaganda to abandon and to betray the Dutch never works out but on the contrary only feeds the disdain of the Ambonese for the undisciplined, wild bands of terrorists. Next, the computer search resulted in a considerable number of by now familiar 'Othering' frames, commenting on both attitudes and actual behaviour: The Ambonese had taken revenge by taking over the police office by force and killing all Javanese policemen present. c. giebel, knil officer25 They wielded their klewangs [bladed weapon] rapidly and with resolve against the rebels. We have witnessed shocking scenes. piet van der wijst, soldier26 The Ambonese were wild and blood-thirsty after this massacre. They kicked and beat the prisoners until there was nothing but blood.
harry brummans, soldier27 [The Ambonese] were the fighting machines of the knil, the colonial army. Horror stories had it that they attacked with the klewang [bladed weapon] and then licked the enemy's blood from the blade. ger vaders, soldier28 The knil boys sometimes use ruthless methods.
[…] Moluccans, Madurese, Timorese, or Sundanese. Great soldiers, but tough. The commander makes short work of the prisoners. It is an extremely unpleasant sight. This evening they interrogated the prisoners. If they didn't want to talk, they were kicked, beaten and pounded in the stomach until they finally did talk. What can you, an ordinary soldier, do when your commander kills, tortures, or spares a life, on a whim? unknown soldier29 The soldiers from Timor and Ambon were so horribly wild by then, that they smashed everything they saw, women, children, men, boys, everything they saw.
[…] They killed everything they could. Our boys did not dare intervening for fear of being slain themselves. The Ambonese are particularly ruthless, they don't respect the concept of taking prisoners. They kill everything they see and it's the same for the other side. Things happened that can't be answered for, but in times of war one has another perspective than in times of peace.
han de haas, soldier31 A Japanese officer interviewed after the war about the battle at Palembang described the fights against the Ambonese as ones against demons. They really frightened the Japanese unit.
j. muller, sergeant knil32 The Ambonese, roughly 30 men, had destroyed the lot in a moment. Eight hundred out of 1,000 were killed by the klewang [bladed weapon].
jus wagter, soldier33 In sum, these citations do not really differ much from the ones previously presented in Soldaat in Indonesië. The Ambonese were respected for their loyalty and courage but slightly feared and at times-mainly after the factcondemned for their ruthless violence. Such accusations, however, were often voiced in a slightly apologetic tone, referring to the hardships of the Bersiap period and particularly the supposedly equally merciless behaviour of their opponents. At this point an observation must be made about a difference between the citations presented in Soldaat in Indonesië and the computer-generated citations. While the citations in Soldaat in Indonesië came with the context in which they were written, and were attributed a value by the researchers, the computer-generated citations came without context. We had to look at the original source to find the context. Thus, on closer inspection, the quote given above from Ger Vaders was not a first-hand observation but came from a book in which he recounts being taken hostage in a train-hijacking by Moluccans in 1975. This particular citation only pertains to his musings about the 'fathers' of the hijackers. Computer-generated citations do not come with important metadata, like the function of the quote in the main narrative, whether or not it is a first-hand memory or part of a description to understand other parts of the book. Thus, even more caution is requisite. 31 Penders 2016:40. 32 Muller 1995:37. 33 Klumper-Eleveld 2010:52.
We next considered linking the entire dataset to the years of publication of the relevant ego documents, in order to find out whether perhaps the passing of time might have influenced the ways in which veterans reported on their experiences with Ambonese fellow soldiers. We had two, slightly contrary hypotheses to test. On the one hand, we assumed that the violent political actions in the 1970s of a younger generation of Moluccans in the Netherlands (Steijlen 1996a:133-69;Steijlen 1996b) might have reinforced ideas about their fathers' violent nature among Dutch veterans. On the other hand, we hypothesized that as the years passed, there might also have been a tendency among Dutch veterans to refrain from the ethnic stereotyping of Indonesians, and particularly Ambonese, that was clearly more acceptable in the 1940s than, say, in the 1990s.
The database did not enable a computer-driven analysis here, if only because the year of publication of ego documents is an unreliable indicator. While most of the memoirs published decades after the war were indeed written much later, on closer inspection several publications were actually based on texts-diaries and correspondence-written during the war. The new fragments produced by the digital search did however enable us to broaden our heuristic search into the entire set of ego documents. The database yielded a modest number of references to the militant Moluccan actions of the 1970sbut rather than eliciting comments about an inherent tendency towards violence among the Moluccans, these inspired veterans to ponder the sad story of this community and to criticize the way in which the Dutch government had handled this episode. A strong example is the testimony of Gerrit van der Stelts (cited in Molegraaf 2009:67-8): I do not want to justify what happened in 1977 with the train hijacking at De Punt, but I sort of supported these boys. I could feel for them and in my opinion these boys were short-changed. The Ambonese could have joined the Indonesian army, where they would have made it because of their military skills. But they stayed loyal to the Dutch and once in the Netherlands, we left them standing in the cold.
We did find scattered evidence corroborating the second hypothesis: that as time progresses, there seems to be less explicit ethnic stereotyping, at least of (Christian) Ambonese; however, we did encounter more and occasionally more overtly denigrating or hostile allusions to the Muslim adversaries of the late 1940s.34 34 On Muslim adversaries, see Oostindie 2015:101-2. We found no negative stereotyping of the very small number of Moluccan Muslim soldiers fighting on the Dutch side.

Testimonies: Oral History
Over the past decades, Dutch researchers have built extensive oral-history collections based on interviews with (re)migrants from the Dutch East Indies as well as veterans from the last colonial war.35 Potentially, the resulting collections of veterans' testimonies could provide some nuance to the framing of the Ambonese in the corpus of published ego documents. Unfortunately, except for the interviews done earlier by the Moluccan Historical Museum and by Fridus Steijlen for the kitlv after the publication of Soldaat in Indonesië, the overwhelming majority of veterans interviewed were Dutch natives, so this corpus suffers from the same bias. The theme of the multicultural character of the colonial army and, particularly, of the loyalty and military prowess of Ambonese soldiers was discussed with some frequency in these collections, though interviewers would generally not push interviewees to comment on the latter issue. What comes out of these interviews seems to be quite similar to the arguments provided by veterans in published ego documents. However, because in these collections mainly civilians are interviewed, other stories emerge too, for instance, about the brave attitude of Moluccan knil personnel before the Japanese invasion: In 1941 we knew war was on its way.
[…] In the unit, some were afraid, but not so the Ambonese. That is a martial race, they were always loyal to the queen. A recurring story tells about Moluccan soldiers protecting European and Indo-European civilians against violence by Indonesian youth, or taking revenge: In Batavia we went to another camp, the 10th Battalion, where we stayed for a while. There they also started to shoot, and the camp was attacked. But it was defended by Ambonese. So, in fact, the Indonesians did not stand a chance. anita van der els-vertregt37 We escaped in time, but there were Indo-Europeans who were still out there […] I respect the Ambonese. They said: We are on duty until 2 pm or 4pm. After that we do not say anything, but we go on our own, we take a truck and we go to pick up that family. Indo-European boys were also involved, not only Ambonese. But I question whether we would have made it without the Ambonese, whether we would still be there.

c.s. brender à brandis38
In the 10th Battalion was an Ambonese sergeant, Mister Pieterse. And Mister Pieterse went to pick up women and children [who were in peril]. When I saw this, I called Mister Pieterse and told him that my mother, brother, and sister were also out there. He said okay and went to pick them up. ferry schram39 The British did not allow us anything. We did not like that at all, so we decided to kill all Indonesians in the neighbourhood of the camp. Everybody went out and we took policemen prisoner, because in Bandung the allies stayed in the northern part, the southern part was where the Indonesians were. In the northern part were some Indonesian police officers. We killed them one by one on the spot, or we took them with us to the camp. At the camp were Ambonese, […] they also returned to the barrack as military. They were so angry that they killed the Indonesians in front of the gate of the Tjihapit camp. Other stories tell about the fierceness of Moluccan soldiers during military confrontations with Indonesian adversaries: I was involved in several actions; I went with a navy ship and was attached to a knil unit. I was only a telegraph operator. I had a Tommy gun and a revolver but did not know what to do with them. But there were Ambonese, they called me Oom [uncle]. They said: 'jangan takut, kita jaga' . That means: do not be afraid, we will take care of you. j. bennewitz, telegraph operator41 A soldier went to the toilet in the middle of the night and he met a guy with a huge knife in his hands. So he ran away. But the guy ran too, he was a pelopper [independence fighter], and he ran into the barbed wire. The camp was guarded by Ambonese. They caught the guy, and I must say … we were curious and went to the officer on duty, also an Ambonese. Well, the way that guy was treated, that is … we thought it was nasty and yelled 'is that necessary?' , 'that looks like the ss' . It was our first confrontation with events that are sometimes denied among veterans. leo schipper, war volunteer42 There were enough excesses. They never said anything about it because back then it was normal.
[…] I witnessed a bunch of Ambonese interrogating four or five Indonesians. They used iron wire, which they held in fire, and then that is it. That is how it went. And whether these guys were guilty or not, it made me sick. But what to do? You would not stand up against a South Moluccan. He would take his klewang and kill you. They were nuts. They were good soldiers, but sometimes they went too far.

Moluccan Voices
Over the past decades, the Moluccan Historical Museum has built an extensive oral-history collection on life in the Moluccan islands, the Japanese occupation, Moluccan participation in the Netherlands-Indonesian conflict, the failed attempt to found a separate Moluccan state, the Moluccan's unaskedfor demobilization, and their migration to, and ambivalent integration in, the Netherlands. Not surprisingly, Moluccan loyalty to the Netherlands and the feeling that they were betrayed by the Dutch are frequently discussed in this collection. In contrast, the role of the Moluccan soldiers in the war seldom surfaces, although young Moluccans recall that at home their fathers talked all the time about their toughness during that conflict (see also Steijlen 2016b). It seems a public silence was created around this issue. We therefore decided to do a round of interviews among the last cohort of former Moluccan knil soldiers, with Steijlen this time asking them not only about their overall experiences during and after the war, but also specifically eliciting their comments on their war experiences and their interaction with Dutch soldiers. Following the preferences of the interviewees, these conversations were conducted mainly in Moluccan Malay, interspersed with Dutch. The language barrier is clearly part of the explanation as to why these men had not been systematically interviewed by Dutch researchers before. But more importantly, previously the focus was on the main frustration of Moluccan soldiers-that is, their unexpected discharge from the army upon arrival in the Netherlands in 1951 and the way they were treated after that-rather than on their experiences during the period 1945-1950.
Nineteen Moluccan former knil soldiers were interviewed out of a total of almost forty men we could trace-the others were already too old and sick to be interviewed, or unwilling.48 The majority of the men interviewed were 47 smgi interview 1032.1 track 18, dated 13-5-1997. 48 Interviews were held in 2016 and 2017. All interviewees were contacted through family members and local key persons in the Moluccan community. The interviews were conducted by Fridus Steijlen. At some occasions, interns or a colleague joined the interview.
Because of the small number of elderly people in the Moluccan community in the Neth-so-called soldadu muda, young soldiers who entered service after the Second World War; those serving before the war were called 'old soldiers' , or soldadu tua.
In view of the trope of traditional, even 'eternal' loyalty to Dutch colonialism, the first question was whether the Moluccans who joined the knil after the war were indeed motivated by such loyalty or opposition to the Indonesian Revolution. Most of the interviewed former knil soldiers denied this, arguing they enlisted simply because they wanted to make a living, cari makan, looking for food, or wanted to see something of Indonesia. Enlisting for the knil in the Moluccas was easy, because soon after the Japanese capitulation Dutch recruiters visited Moluccan villages to recruit young men, even far away from Ambon in the south-eastern islands of Kei and on the thinly populated island of Buru.49 I reported for duty with the knil because of the Dutch call for all young men in Ambon to serve because of the situation in Java. There was an emergency. I enlisted with the knil on 13 November 1946 and trained for 6 months. After that I was sent to Celebes-to Pare Pare, Watampone, and Mandar to patrol. Then Bandung ordered the military to Java to support the Dutch over there.50 Some knew a little bit about the situation on Java: I wanted to help our people who were taken prisoner by the bambu runcing [referring to young revolutionaries armed with sharpened bamboo sticks], to liberate them, because there were also many Ambonese captured and killed.51 Others underlined their fearlessness: But wherever the Moluccans are, they are not afraid. Moluccans are not afraid. Even if there is a war, we are not afraid of death.52 erlands and the possible sensitivity of the topic, we only use initials in this article and only mention the date of the interview. 49 Interviews J.H. 1-12-2016 and M.S. 11-7-2016. 50 Interview N.T. 7-7-2016. 51 Interview A.F. 14-7-2016. 52 Interview P.N. 28-7-2016.
The Moluccan veterans we interviewed in the Netherlands all served in the knil. They knew, however, that other Moluccans had joined the Republican army. Occasional confrontations were inevitable: We were with a knil battalion formed of only Ambonese. We were all Ambonese new soldiers (teken soldadu baru). We met other Ambonese [supposedly from the opposite side] who said to us: you are licking the arses of the Dutch-Kemong paleng jilat pantat Belanda. Sergeant Manuhutu responded. 'What are you saying?' He put the klewang on [the other man's] neck. I will never forget Sergeant Manuhutu. Ambonese fought Ambonese.53 The interviewed men confirmed rather casually that Moluccans were fighting on both sides, but apparently there was little room to change sides once the war had really started. So, it was mainly at the beginning of the war and after the transfer of sovereignty that Moluccan knil soldiers joined the Indonesian army.
As for the image that these Moluccan knil military were disproportionally violent, some veterans rhetorically answered that Moluccans could not possibly be more jago, fierce, than others, because they were strict and followed the army rules, pointing to the basic rules or their adherence to Christian values: In the army there was no difference, we had to act according to military discipline. That made us one.
[…] There were also Muslims in the army, but we had the military discipline, a book, a kind of code of conduct. With all the rules written in Malay.54 But they are afraid of God. Moluccans are first of all afraid of God. You cannot do anything wrong. Not with your mouth and not with your hands.55 In contrast to this position, children of the veterans share stories of their fathers boasting about violence. At one of the interviews, a nephew of the interviewee was present. His already deceased father had been a Green Beret, a commando. While his father had never spoken about it, he was sure his father was involved in a lot of killings. 53 Interview N.T. 7-7-2016. 54 Interview A.F. 14-7-2016. 55 Interview P.N. 28-7-2016. According to the same veteran, the Buginese from Makassar were just as prone to use violence: I've seen the Bugis from Makassar. That they would just kick people in the street. They wanted to hear stories, but if the information did not come out, then they would start beating.63 This veteran then proceeded to construct a sharp contrast between this ethnic group and the presumably more relaxed approach of Central Moluccans: The prisoners were men only. There were women in the houses, but we did not know whether they belonged to those men. We took the men. Another veteran pointed to violent Javanese knil soldiers, and was surprised at their violence because they were fighting 'their own': The Javanese knil soldiers then beat the tni soldiers. I have never hit anyone myself. They are also human beings. If you answered honestly, I would release you.
[…] If you looked closely, the Ambonese rarely hit. It was more the Javanese who beat people.65 These testimonies each seem to pass the buck on to another ethnic group. How does this compare to Dutch veterans claiming that the 'Ambonese' in particular were disproportionally violent? How was their communication at the time, how did they get along? In retrospect, most of the Moluccan veterans are positive about the Dutch military that they had to work with. But there was an obvious distance, culturally and in terms of language. The language barrier could be problematic.
One day I was on guard. I picked up the binoculars and saw many people approaching with weapons. It was the tni. I called the command post. I got my lieutenant, one who spoke to me in the Groningen dialect. I didn't understand a word. I thought: is this English? There were also English troops. The phone number was right. But I didn't get it. Were they English? Groningers? I was angry, I saw the tni coming. But I did not understand them on the phone. […] I put the phone down saying, 'damn it' .66 Moluccans saw the Dutch soldiers as newcomers without any experience, in need of protection. This corresponds well with the Dutch memories cited above. Young Dutch arrivals were called tentara susu, 'milk army' , or green. One of the Moluccan veterans referred to this label and said he and his Moluccan comrades were also called tentara susu, but he then said, 'we are Moluccans' , made a gesture of silencing, and 'then they did not say anything anymore' (Steijlen 2016b). Moluccan soldiers, so the veterans recalled, were considerate towards the young Dutch soldiers.
There was a situation with a group of Dutch people from Groningen. Sometimes they shouted 'Mommy, Mommy' . I said, you shouldn't call for your mother, you should fight. Whether you dare or not, you have to shoot! 65 Interview B.F. 30-5-2016. 66 Interview N.L. 24-6-2016.
If you don't shoot, you will die. These were students who had to go to Indonesia in the army. With no experience with weapons.67 As an example of the weakness or vulnerability of the Dutch soldiers, two veterans referred to the way the Dutch soldiers could easily be tempted by Indonesian girls, who might have served as agents of the enemy: The Dutch conscripts did not yet know the Javanese. The Dutch loved the girls. That's why they died easily.68 Do you know why many knil personnel were killed? They were seduced and lured by beautiful women. It would start with a 'come on' and then you wouldn't see them anymore. They never came back. But that [trick] did not work with the Ambonese.69

Conclusion
The present analysis is not nearly as multi-vocal as we would like it to be, if only because we have not been able to include Indonesian perspectives. During the war, Republican Indonesians were obviously hostile to compatriots still fighting on the colonial side. Moluccan or, more broadly, Indonesian military participation in the Dutch army is not a major theme in Indonesian historiography, which has tended to downplay internal rivalries during the Indonesian Revolution-whether of a regional, ethnic, religious, or political nature (McGregor 2007). If anything, suspicions about lingering Dutch-Moluccan aspirations for a Republik Maluku Selatan, rather than memories of Moluccan betrayal during the Revolution, feature in the historiography and perhaps also in broader contemporary thinking in Indonesia. So, what conclusions may we draw? First, the post-war construction of some sort of self-evident, traditional loyalty of Moluccans to Dutch colonialism, and particularly to the Dutch Crown, is largely a myth, already an invented tradition before the Second World War. On the other hand, many Moluccans had started to see themselves as loyal to the colonial regime and acted accordingly. Precisely for this reason the Japanese made Moluccan knil soldiers pows and 67 Interview P.T. 21-6-2016. 68 Interview B.F. 30-5-2016. 69 Interview N.T. 7-7-2016.
did not release them nor employ them in their own auxiliary troops, unlike Javanese knil soldiers (Chauvel 1990:189). After the Japanese capitulation, when the Dutch started to re-establish the knil, former Moluccan pows were mobilized. At the same time, new soldiers were recruited on the Moluccan islands. So, the Moluccan contingents in the post-war knil were both young soldiers (soldadu muda) and old hands (soldadu tua).
At the same time as the Dutch started to recruit for the knil, other Moluccans, especially on Java, joined the Republican cause and joined the Moluccan laskar (militias) in the Indonesian army. Most Moluccans who joined the knil between 1945 and 1949 did so primarily for pragmatic reasons, that is, to secure a job and flee from the Moluccan islands. Only a few were driven by knowledge of what was happening on Java, including the attacks on pro-Dutch Moluccans. At the end of the day, it is very difficult to ascertain whether these Moluccan knil recruits enlisted because of a promising career, were inspired by the ethnic soldiers' loyalty to the Dutch or were just seeking new opportunities after 3.5 years of oppression by the Japanese.
How about the violent image of Moluccan knil soldiers? There are plenty of indications that underline the idea of the more experienced and fierce Moluccan soldiers aggressively protecting civilians on the Dutch side, and introducing young Dutch conscripts to the perils of Indonesian life and helping them to survive the guerrilla war. Dutch veterans and civilians comment on this with a mixture of gratitude and awe. Within the Moluccan veteran community in the Netherlands, this warrior image was upheld with pride, and some of them explained that they had had to protect those new, unexperienced, and naïve young Dutch soldiers. This protection would definitely have added to the image of the brave and loyal soldier. At the same time, it can easily be imagined that this resulted in an assertive attitude among Moluccans, in order to show the Dutch servicemen that they were the guides-the true insiders regarding culture, language, and danger. They were the ones who really 'knew' the dangers and when it was necessary to resort to violence-in short, all the things that made them popular as local experts on the front line.
In previous decades, most interviews with Moluccan veterans focused on the frustrations around the rms, the unexpected discharge from the army upon arrival in the Netherlands, and the challenges of integration in the Netherlands. While Moluccan veterans talked about the cruelty of warfare in their own ranks and informally during community gatherings, they remained silent concerning this sensitive issue in public and in public communications. In our final round of interviews, the subject remained delicate and hence difficult to address. Perhaps part of the explanation is that those we interviewed were the youngest recruits, the soldadu muda, who were less intensively involved in ugly combat situations than the soldadu tua. The latter generation had already passed away when we did our last round of interviews.70 This does not necessarily eliminate the possibility that Moluccans were disproportionally involved in the excessive violence. As argued, the available evidence does provide some grounds to support this idea, but then again, the bias in this source material is all too evident. Most of the stories of excessive violence conducted by Moluccans were told by Dutch veterans. There may be a self-serving motive at play in blaming others for violence that can no longer be ignored, a common mechanism used to pass responsibility on to others, which in this case might be further tainted by racist ascription. So, while we cannot take these memories for granted, it is also too easy to simply dismiss them as fictional. The sobering conclusion is that the question remains open, and it is not evident whether it can be answered at all on the basis of empirical research.
Clearly, though, the interviews with the soldadu muda did not directly confirm a particular violent role of Moluccan soldiers in the violence. Indirectly the interviewees did underline the image of fearless soldiers and, in extension, their potential for violent action. As one Moluccan veteran cited above said with some bravado, 'don't mess with Moluccans!' .
As stated in the opening of this article, there is a prohibitive bias in the available sources, which our last round of interviews has not been able to correct. The interviews did provide new information though on the ways Moluccan soldiers looked at their relationship with the Dutch and how they perceived their own contribution. In the end, our research did not produce hard evidence to either reject or fully support the hypothesis that Moluccan soldiers were disproportionally involved in extreme violence, but it definitely counters the myth that their role in such violence should be explained by their inherently belligerent traditions. Although Moluccans themselves contributed to the longevity of the 'martial race' myth, explaining Moluccan attitudes and behaviour through Orientalist frames of ethnicity remains a dead end.
All of this does not diminish the fact that Moluccan recruitment in the knil and commandos and their deployment in on-the-ground, counter-guerrilla warfare and interrogations of Indonesian prisoners by the intelligence services put them, like other knil soldiers, on the spot where most of the Dutch violence we know of was carried out. So indeed, they were probably often used 70 Also, the majority of soldadu tua did not migrate to the Netherlands in 1951, as they had already returned to the Moluccan islands or were enlisted in the Indonesian army in 1950.
to do the dirty work for their Dutch superiors and same-rank mates. In addition, it does seem likely that as the war proceeded, these Moluccans' loyalty to the Dutch side became increasingly tenuous, which put them in an existential dilemma that was unfamiliar to the Dutch soldiers, who could simply return home. But again, we cannot assume a direct, much less a causal connection between this predicament and outbreaks of extreme violence. The elephant in the room in this discussion is not primarily any sort of assumed Moluccan ethnicity, but rather the chain of command and hence Dutch responsibility. Dutch politicians allowed the military commanders to conduct the war as they saw fit. In a continuous stream of ordinances, the latter urged the soldiers on the ground to stick to the rules of 'decent' warfare and to protect civilians as much as possible (Limpach 2016:603-5). In practice, however, excessive violence was condoned and was hence allowed to become a structural ingredient of the way in which the war was fought. Scapegoating one particular ethnic, and conveniently non-metropolitan, non-white group for a disproportionate share in this violence misses the more important point of ultimate responsibility, and impunity.71 And this, of course, is a bias that virtually all studies on the alleged 'ferociousness of ethnic soldiers' seek to question.