The fabrication of the Aboriginal death toll
There have been three serious attempts to estimate the total number of Aborigines killed by Europeans in frontier conflict since 1788. By serious attempts, I mean calculations made by people who are recognised as historians and who use historical evidence and argumentation. They have been made by Henry Reynolds in his book The Other Side of the Frontier (1981); by Richard Broome in his article “The struggle for Australia: Aboriginal-European warfare 1770-1930” in the anthology edited by McKernan and Browne, Australia: Two Centuries of War and Peace (1988); and by Robert Murray in “What really happened to the Kooris?” in Quadrant, November 1996.
Each of these analyses arrives at the same figure, a total of 20,000 Aborigines killed, mainly by gunshot. Depending on different estimates for the total population of Aborigines over time and place, this figure suggests that between five and twenty per cent of all Aborigines in the colonial period died violently by white hands. This is nowhere near as great as their death from disease in the early nineteenth century, but it is still a disturbing total. If true, then notwithstanding the fact that some particular massacre stories have been invented, it would justify the unforgettable phrase about European-Aboriginal relations used by Justices Deane and Gaudron in their 1992 Mabo judgement: “a legacy of unutterable shame”.
In The Other Side of the Frontier, Henry Reynolds argues that those Aborigines who died defending their territory outnumber, in both relative and even absolute terms, the number of white Australian war dead.
Twenty thousand blacks were killed before federation. Their burial mound stands out as a landmark of awesome size on the peaceful plains of colonial history. If the bodies had been white our histories would be heavy with their story, a forest of monuments would celebrate their sacrifice… In parts of the continent the Aboriginal death toll overshadows even that of the overseas wars of the twentieth century.
This figure of twenty thousand Aboriginal dead was used by Reynolds then, and has been repeated by him many times since, to depict the Australian pastoral frontier as a scene of “open warfare”. It now entered the established record and become part of the received wisdom of Australian historiography. It is the current consensus among Australian historians. In the Oxford Companion to Australian History, this is the figure that is cited under the category of massacres, written by Richard Broome. The magnitude of the number justifies him treating the phenomenon as a significant feature of Australian history. “Massacres now refer to mass killings of defenceless Australians, but historically referred to the large-scale slaughter of Aborigines,” Broome records. “Massacre was a marked feature of the frontier. Australia was settled by Europeans as Western ethnocentrism meshed with racial ideas … Aboriginal deaths by massacre formed a significant proportion of the total killings.” Broome also notes the political weight that the concept of these mass killings has come to bear today.
Many Australians now see recognition of massacre as a prerequisite for reconciliation. What was once unmentionable is now openly discussed. This surely is a measure of our maturity as a nation. Now, it needs to be discussed critically, as a measure of our impartiality.
I certainly agree with Broome that this issue needs to be discussed openly, critically and impartially. It was in this spirit that, after reading Rod Moran’s book on the events at Forrest River, I went back to the existing estimates of the number of Aborigines killed in Australian history to see how they had been compiled. I had long presumed that the confidence of the rhetoric used by Reynolds and Broome — “a landmark of awesome size”, “a marked feature of the frontier” — meant that historians had compiled something like a list or tabulation of the number of Aborigines who had been observed, or at least reported, killed during the nineteenth and early twentieth centuries. Instead, to my surprise, I quickly discovered that, apart from a small number of local area studies, nothing of the kind exists. When you trace the references from which the overall estimates are calculated, you find that the original sources on which they are based themselves openly admit that they have little idea what the actual figures were and that they are doing no more than guessing what they might have been. Moreover, when you look at the arithmetic behind the national calculation you find its assumptions are designed to produce not an impartial figure but as large a total as possible, no matter how much fudging is required. I will illustrate this by discussing the sources on which each of the three main Australia-wide estimates have been based.
The unsubstantiated guesswork of Henry Reynolds
The earliest of these estimates by Henry Reynolds is the one that has set the agenda ever since. It has never been challenged, despite the fact that it is based on a remarkably cursory discussion and itself acknowledges there is a large hole in its own calculations. Here is the count he offers in The Other Side of the Frontier:
Calculating the Aboriginal death toll is much more difficult. Conflict is better documented in Tasmania than anywhere else in the country and Ryan’s estimate of 800 is possibly more accurate than any other we can make. Green has accounted for 102 Aboriginal deaths in his segment of Western Australia and Prentis for 100 in the northern rivers district of New South Wales. Christie has recently argued that the whites killed 2000 blacks during their occupation of Victoria while Reynolds suggested that as many as 8000-10,000 Aborigines died violently in Queensland. For the continent as a whole it is reasonable to suppose that at least 20,000 Aborigines were killed as a direct result of conflict with the settlers.
Now, the numbers cited in this passage come to a total of 11,000-13,000, not 20,000. Given that Reynolds cites no further studies of Aboriginal deaths in any of the regions outside those he mentions, and does not provide any methodology with which to estimate them or to extrapolate from those he does cite, his claim that it is “reasonable” to simply add another 7,000-9,000 to reach the national total is unwarranted. We cannot simply assume that, because Aborigines were killed in some parts of Australia, the same happened in all the other parts at about the same rate. Without either evidence or argumentation, Reynolds cannot simply fill in the figures any way he likes. To do so, is to act in a way that is most unreasonable.
When we look at the figures on which his 11,000-13,000 are based, things get even worse. For Tasmania, Reynolds cites Lyndall Ryan’s 1976 Macquarie University PhD thesis, later published as the book The Aboriginal Tasmanians (1981). In the book, Ryan’s total is actually 700, not 800, but let us look at how she arrives at this figure. She writes:
In 1823 the estimated population of the Big River, Oyster Bay, North Midlands, North East and North tribes was about a thousand. By 1832, 156 had been captured, 50 lived with sealers, and 27 lived with settlers. Of the remainder, 280 were recorded shot, which leaves some 480 unaccounted for. It seems that even on the Tasmanian frontier only about one-third of the Aborigines killed were recorded and that a more realistic total would be about 700, or nearly four times as many as the 176 Europeans killed by the Aborigines.
In other words, Ryan is arguing that, because she cannot account for 480 Aborigines –that is, almost half the population — she is justified in presuming that some 420 of them must have been killed by Europeans. This sort of reasoning is completely unacceptable. If she does not know what happened to them, the most Ryan can say is that she does not know what happened to them. Moreover, the actual size of the population from which they are presumed missing was only a rough estimate and could itself be wildly astray. The truth is that all she can legitimately claim is that 280 of them were shot. (I am assuming, of course, that she does have adequate evidence for each of these deaths). Rather than an Aboriginal-to-European death ratio of four to one, the most that the historic evidence of Tasmania allows her to argue is a ratio of 280 to 176, or 1.6 to one. This is a point worth remembering because, as we will see below, it is ratios of this kind that Reynolds and others have themselves used to estimate death rates on the mainland.
For Victoria, Reynolds cites another thesis, M. F. Christie’s 1978 Monash University PhD. I have not been able to access this but Richard Broome has seen it and had this to say about it:
Michael Christie claimed 2000 Aboriginal deaths for Victoria, a ratio of 33.3 to 1 [i.e. black deaths to white]. However, his figures are not based on careful counts and estimates, but on a calculation from Edward Curr’s very high 1886 guesstimate that fifteen to twenty-five per cent of Aboriginal deaths on the frontier were by rifle. With no justification (given the mildness of the Victorian frontier), Christie opts for twenty-five per cent or more, and alleges there were 2000 deaths.
The estimate of 100 deaths in the Northern Rivers of New South Wales is from a 1972 Macquarie University MA thesis by M. D. Prentis, which I have not seen. However, the figure of 102 deaths in Western Australia is from a 1979 journal article by Neville Green, which is readily available. Of the five studies cited by Reynolds, this (apart, possibly, from Prentis) is the only one that produces a credible total. Green’s study is of violent conflict between Aborigines and settlers between 1826 and 1852. He provides a table that lists the date each conflict occurred, its location, a description of the incident, the number of dead and wounded whites and Aborigines from each incident, and his documentary source. As well as 102 Aborigines killed, there were 25 settlers and soldiers killed by Aborigines at the same time, a ratio of four to one.
The final study on which Reynolds’ figure of 20,000 dead relies is his own work on Queensland, where he claims a total of 8,000-10,000 deaths, that is, about half the total for the whole continent. The reference he provides for this is the anthology he edited in 1978 entitled Race Relations in North Queensland. Reynolds has written three of the sixteen essays in this compilation, one of which, “The unrecorded battlefields of Queensland”, contains his estimate of Aborigines killed during frontier conflict. This estimate, however, is not easy to find. You can read this essay, and indeed the whole volume, and not find any mention at all of 8,000-10,000 Aboriginal deaths in either the text or the footnotes. However, towards the end of this essay you find the following statement: “Aboriginal suffering was vastly disproportionate to that of the settlers.” This statement has a footnote that reads, in full:
Aboriginal fear and insecurity was, we must assume, infinitely greater than that of the settlers. Their death rate may have been ten times more than that of the Europeans.
This is the only evidence Reynolds provides in this essay about the overall scale of Aboriginal deaths in Queensland. So where does the figure of 8,000-10,000 killings come from? This requires some detective work from the reader.
Having discovered this footnote, you can go back through the essay and try to work out its significance. The only totals of deaths provided in the text are those for Europeans. In fact, the principal purpose of the essay is to record not Aboriginal deaths but Europeans killed by Aborigines. Reynolds says that he and Dr Noel Loos went through newspapers, official records and correspondence, and personal reminiscences to compile a register of frontier deaths of Europeans, as well as of the small number of Chinese, Melanesians and “so called ‘civilised Aborigines’ who had been incorporated in the advancing wave of settlement.” He acknowledges “the uncertainty and unevenness of available records made definitive accounting impossible”. However, he still records a tally of 400 deaths of Europeans and their allies in Southern and Central Queensland from the beginnings of pastoral expansion in 1841 until the 1870s, and between 420 and 440 in North Queensland from the first settlement at Bowen in 1861 to the passage of the Aboriginal Protection Act in 1897. He puts the total number of whites killed by Aborigines at between 800 and 850 in the fifty years it took to occupy Queensland.
When he came to write The Other Side of the Frontier, Reynolds decided to combine the figure of 800-850 white deaths with his footnoted guess that the ratio of black deaths to those of whites “may have been” ten to one. So he multiplied 800-850 by ten. This is the origin of the book’s claim of 8,000-10,000 Aboriginal deaths, and is the sole reference from which this total has been compiled. Even on his own arithmetic, of course, the figure should have been 8,000-8,500 deaths, but Reynolds obviously has no qualms about plucking another fifteen hundred bodies out of the air to make the total look even bigger.
In the Australian Journal of Politics and History in 1976, Reynolds and Loos had published an earlier version of the same article, then entitled “Aboriginal Resistance in Queensland”. Again, this is not a study of Aboriginal deaths but of how many white settlers were killed by Aborigines. And again, they provide a figure of 800-850 Europeans dead. Similarly, there is nothing about the number of Aborigines who were killed until the very last sentence of the essay, which says: “Their [Aboriginal] attacks on life and property increased the material and psychological costs of pioneering but triggered massive retaliation which left many thousands dead and gave rise to a tradition of white violence which lived on into the twentieth century long after all overt resistance had been crushed.” This statement, too, has a footnote that reads, in full:
It is impossible to do anything but guess at the number of Aborigines killed by Europeans. We think Aboriginal deaths in frontier violence outnumbered those of the settlers by at least ten to one.
There is no evidence or further argumentation provided to support this guess.
The two brief footnotes in these two articles constitute the sole basis for Reynolds’ claims about the total number of Aboriginal deaths in Queensland. In other words, they provide no real authority for his figures at all. They are not claims that should be treated seriously by historians. They are unsubstantiated guesswork. What is more, in their initial publication in 1976 and 1978 Reynolds himself acknowledged them as such by relegating them to footnotes and by offering neither evidence nor argumentation to support them. However, by the time he had a contract with Penguin Books for The Other Side of the Frontier, he had obviously changed his mind and felt that the case he wanted to make needed bolstering. So he dropped all reference to “guesses”, “assumptions” and “may have beens” and adopted a bold and aggressive new stance. What he initially acknowledged were unfounded speculations from “uncertain and uneven” records were transformed overnight into authoritative statistics.
If you read through Reynolds’ essay “The unrecorded battlefields of Queensland” you discover that, while he provides plausible evidence that many nineteenth century Queenslanders feared attack by blacks and went about armed as a result, there is very little evidence of any Aborigines being killed. In fact, he provides only one record of this actually happening. This is a report in a Brisbane newspaper of an incident at Rockhampton in 1861 where, after a white woman was attacked, the native police were called in to disperse the local Aboriginal camp and killed five in the process. Even though the essay’s purpose is primarily to calculate European deaths, we might have expected more anecdotal evidence about Aboriginal deaths in a piece ostensibly about the “unrecorded battlefields” of Queensland. The fact that it is almost devoid of evidence of this kind adds further weight to the conclusion that it is worthless as an authoritative source for the claims made in The Other Side of the Frontier.
Overall, then, the references Reynolds provides for his total of twenty thousand Aboriginal deaths in Australia do not establish his claim. At most, they provide evidence that directly verifies 280 shot in Tasmania, 102 dead in Western Australia, 100 killed in New South Wales and five victims in Queensland.
In September this year, I presented this data to a Quadrant seminar in Sydney where it attracted considerable publicity in the news media. Reynolds replied in the press by admitting that his figure of twenty thousand dead was only a guess. However, he said he had spent ten years researching the subject and his guess was both educated and conservative. He had done “a mountain of research” and had published his reading in a 25-page bibliography. Because his work rested on such a great body of reading, people should trust his judgement. This response, however, displays a strange misunderstanding of the historian’s role. Historians should not have to ask people to take them on trust. No matter how much reading they claim to have done, their job is to actually put their evidence on the public record where their readers can assess its plausibility and other scholars check its authenticity. This is especially so in the case of unlawful killings, which some want to play down and others exaggerate. Not only do they have to actually present their evidence, but historians should also tell their readers whether their sources come from direct observations or from rumours several times removed, whether the reports they use were contemporary with events or made months or years later, and whether their informants were indifferent observers or had axes to grind. Neither Reynolds’ text nor his bibliography makes any of these distinctions. Historians who have no comment to make about the reliability of their sources on contentious issues, let alone those who deceive their readers about the content of their own references, are not entitled to be taken on trust. If Reynolds wants to continue to talk about a state of “open warfare” on the pastoral frontier, he has to actually put up his evidence. Until he does, no one should believe him.
Richard Broome’s ratios and formulas
After estimating the total number of European deaths at the hands of Aborigines at around 1,800 between 1770 and 1930, Richard Broome advances two types of argument about Aboriginal deaths. The first derives from regional black-to-white death ratios. He cites regional studies from north-west Victoria, the Western District of Victoria, Gippsland, Moreton Bay, plus Neville Green’s Western Australian tally, Lyndall Ryan’s Tasmanian figure, and the estimates by Henry Reynolds and Noel Loos for Queensland. The regional studies provide black-to-white death ratios ranging from 3.9 to 1 in Tasmania, to 58.3 to 1 in Gippsland. After describing the Gippsland study as “atypical”, Broome also raises some doubt about the Queensland figures.
Loos has recently suggested that for North Queensland, where the fighting was particularly fierce, a ratio of 10 to 1 might be “so conservative as to be misleading”. However, all the other regional ratios, except sparsely settled Gippsland, are well below 10 to 1. Therefore to adopt a 10 to 1 ratio continent-wide would seem generous. Such a ratio gives an Aboriginal death toll of 18,000 to 20,000.
Broome then goes on to advance a second kind of argument of his own which endorses the Reynolds and Loos total. He says that in pre-contact Australia there were 586 Aboriginal groups with an average of 550 people each — a total pre-contact population of about 320,000. Broome argues that for a death toll of 20,000, each group would have had to lose an average of 34 people to white violence. He goes on:
Most large-scale killings or massacres carried off ten to twenty people. Thus each group’s average death toll would represent perhaps one or two large losses, and half a dozen small losses. This does not appear an excessive average distribution of violence, given the large losses in some districts. Perhaps 20,000 may be a little too low a total.
I will address Broome’s two analyses in reverse order. The second one is extraordinary, coming from an historian. All he is offering is a mathematical formula of how to reach the figure of 20,000 deaths from “an average distribution of violence”. This is an argument that assumes the conclusion first, and then calculates what might have happened for that conclusion to be true. Broome provides no evidence that the majority of the 586 Aboriginal groups suffered violence from whites. As I noted above, the evidence he does provide comes not from the whole of the continent but from only eight regions. Moreover, some of these regional studies, as I showed earlier in the case of Reynolds and Loos, do not themselves provide any evidence of widespread killing of Aborigines at all, let alone massacres.
Indeed, the record presented in some of the regional studies questions the whole notion of mass killings being common. In his tally of 102 Western Australian deaths, Neville Green shows there were only two events where Aborigines were killed in large numbers: the Battle of Pinjarra in 1834 where he records fifteen Aborigines dead, and an incident at Vasse in 1841 where seven were killed. For the rest, the great majority were single killings involving either personal conflict between individuals or settlers fighting off intruders on their homesteads. In her more recent study of Victoria’s Western District, A Distant Field of Murder (1990), Jan Critchett confirms this pattern. She counts a total of 200 Aborigines killed by whites in about seventy separate incidents. Most of these homicides were in ones and twos, while there were only three events that involved mass killings.
This pattern, it deserves noting, seriously questions the decision by the editors of the Oxford Companion to Australian History, Graeme Davison, John Hirst and Stuart Macintyre, to discuss frontier conflict under the label “massacres”. It is under this category that the question of violence between Europeans and Aborigines is primarily discussed in their volume. According to the regional studies, however, the majority of deaths on both sides resulted from individual conflicts. The phenomenon of mass killings was rare and isolated. To discuss all frontier conflict under the name of “massacres” is to falsify and wilfully exaggerate the downside of Australian history.
Moreover, the Oxford Companion’s entry massacres, itself written by Broome, omits any discussion of the mass killing of Europeans by Aborigines. There were a number of them in the early colonial period, including:
- six shepherds killed on Terry and Tindell’s station at Wellington in 1824;
seven employees of William Lawson killed and scalped at Bathurst in 1824;
- ten members of William Faithfull’s overland trek to new pastures killed near Benalla in 1828;
- the twenty six survivors of the ship Maria, wrecked near the mouth of the Murray River in 1840 killed and some badly mutilated by the Milmenrura tribe in what is known to other historians as the “Coorong Massacre”;
- seven stockmen on the Bogan River run of William Lee killed in 1842;
- eight members of the Fraser family murdered, and the women raped, by the Jiman tribe at Hornet Bank Station, Queensland, in 1857;
- nineteen Europeans, including six children, killed on Cullinlaringoe Station, Queensland, in 1861.
None of these events are mentioned under the general heading of “massacres” and only the Faithfull Massacre is recorded anywhere else in the Oxford Companion. As I demonstrated in Part I, this volume provides an inaccurate and exaggerated treatment of the most-publicised of the mass killings of Aborigines by the colonists. Its almost complete omission of mass killings of colonists by the Aborigines is further evidence that this book’s treatment of frontier conflict will deceive readers who want a truthful account. Unfortunately, this reference book is likely to sit on the shelves of school and university libraries for the next twenty years, misleading the next few generations of young people about the true nature of their history.
Let us return to Broome’s first method of calculation, the use of ratios of black deaths to those of whites to reach a national total of Aboriginal killings. Even though he shows some scepticism about the use of these ratios in all regions, by treating them seriously Broome is lending credence to the same, flawed methodology used by Reynolds and Loos. The latter were trying to derive an arithmetical formula they could apply to those areas where there were no historical studies of Aboriginal deaths. However, the most reasonable position to take is that, in the absence of proper, empirical research, we cannot presume any number of deaths of either Aborigines or Europeans. For all we know, the pattern of events in the unrecorded regions may have been radically different to other parts of the country. All we can legitimately say in such cases is that we do not know.
A second reason to reject the use of these ratios is that they vary so enormously, as Broome himself recognised. The disparities he cited — from 4:1 in Tasmania to 58:1 in Gippsland — were bad enough but, when we consider my original point that the Tasmanian figures should actually be 1.6 to 1, the whole issue degenerates beyond repair. Who is to say which particular ratio along such a wide spectrum should be applied to areas that have not been researched? On what basis should one ratio be selected in preference to another? No one could ever say with any plausibility. To apply this methodology to extended periods of time, across different regions, in circumstances where most deaths had their own individual causes, is an abuse of historical methodology. It is an absurd procedure and it reflects badly on Australian historiography that its practitioners have been able to get away with it, uncriticised, for the past twenty years.
Robert Murray’s “guesstimates” and “projections”
Despite the inadequacy of its predecessors, the third major calculation of Aboriginal deaths, conducted as recently as 1996 and published in Quadrant in November that year, uses an even cruder methodology. It was made by Robert Murray who, even more than Broome, defers to the conclusions of Henry Reynolds. Here is the passage where Murray makes his tally and where he openly acknowledges that his figures are composed almost entirely of “guesstimates” and “projections”.
Counting numbers of deaths can seem hard-hearted but it is a standard way of assessing any tragedy. On some very rough guesstimating, I suggest a figure of around 1,000 for Victoria and 500 for Tasmania of Aborigines killed in frontier conflict. Projecting this, a figure of 2,000 to 3,000 seems likely for New South Wales and 1,000 for South and Western Australia. This gives a likely total for all southern Australia of about 5,000. Henry Reynolds puts the Queensland total at 8,000 to 10,000, mostly killed between 1850 and 1900 … Projecting again, a figure of 5,000 for the Northern Territory and the northern half of WA seems feasible. This adds up to the total of 20,000 Aboriginals killed (over more than a century), estimated by historians such as Henry Reynolds and Richard Broome, using somewhat different routes.
Needless to say, in a debate already sullied by the crudity of its calculations, Murray does not feel called upon to offer any explanation, mathematical or otherwise, behind the methodology of his “projections”. We are expected to accept them just as they are, devoid of either evidence or reasoning.
The most disappointing feature of Murray’s article, however, is not so much its emulation of the prevailing standards but the fact that, elsewhere in the same piece, he combines this with some of the most sensible comments made in the whole debate. He rejects wild accusations about “genocide” and quite rightly dismisses rumours and urban myths about the murder of Aborigines through handouts of poisoned flour. And rather than rushing to smug moral judgements, as both Reynolds and Broome so readily do, he puts the Australian events into a broader historical context, which sees them as characteristic of every period and every society in the world where hunter-gatherers have come up against expansion by pastoralists.
Nonetheless, Murray’s arithmetical calculations remain the most cavalier of the three. I can underline this with another example. This is how he calculates his Victorian “guesstimate” of 1,000 deaths.
Henry Gyles Turner, in his History of the Colony of Victoria (1904) counted a total of 350 Aboriginal and fifty white deaths from frontier conflict, but acknowledges the inadequacy of evidence about Aborigines. A rough calculation is to double his number and add another couple of hundred for Gippsland. This produces a rubbery total of perhaps 1,000 slayings for Victoria …
Before this passage, Murray provides a brief literature review of all the regional Victorian studies. This shows that the figure of 200 deaths he suggests adding for Gippsland is plausible, but that is all. The only other regional total he cites is Critchett’s 200 from the Western District, which Murray says, on the basis of unnamed “other estimates”, should really be expanded to 500. His state total for Victoria, then, is not “rubbery” but reckless.
As I observed above, one of the most notable features of this whole issue is the almost complete lack of any criticism of the principal advocates. This is not because there is no one in Australia with expertise in historical demography or mortality statistics. On the contrary, on almost any other historical issue you could name, numerical calculations as suspect as those used in this debate would bring down a hail of disapproval, especially from economic historians. Rather than criticism, what characterises this issue is the degree of collusion involved and the protection offered to each other by the historians concerned. For instance, Richard Broome is very well aware of the flimsy nature of the figures for Aboriginal deaths in Queensland offered by Henry Reynolds. Broome actually cites the lone footnote in the 1976 AJPH article, where the unsupported 10:1 ratio of black-to-white deaths is first put forward. Yet Broome writes this up as: “Henry Reynolds and Noel Loos, after exhaustive research on Queensland, have suggested 10 to 1 as a likely ratio.” In saying this, Broome could hardly have been unaware that the “exhaustive research” in the article he cites is not an investigation into Aboriginal deaths at all, but a calculation of the number of Europeans killed by Aborigines.
How many Aborigines, then, died in frontier violence? Until we have much better evidence, it is impossible to say and historians should eschew the temptation. To guess at or make up the figures is irresponsible. To do so it to fail the historian’s duty to his profession, to his readers and to the people whose lives he is portraying.
In the media debate in September over this question, there were at least seven academic historians who wrote articles or letters denouncing the press reports of my findings. None of them bothered to obtain the original paper beforehand — an instructive indicator of their commitment to scholarship. The most revealing comment was made by Bain Attwood of Monash University who wrote: “Most of the historical sources that might have enabled us to enumerate the number of Aboriginal people killed on the frontier have, for various reasons, either never existed or have since been lost or destroyed.” Attwood went on to claim that “very little historical interpretation is verifiable in any strict sense” and that historians arrive at the truth on the basis of a “scholarly consensus”.
Now, the notion of a scholarly consensus might be acceptable if there was sufficient evidence to support it. However, if the evidence “never existed” or can no longer be found, then the consensus can owe nothing to scholarship. It is no more than a shared ideological position. To imagine that one can arrive at conclusions without evidence, but simply on the basis of an agreement between those currently in the field, is to abandon historical methodology in favour of politics — he who has the numbers determines the truth. Unfortunately, this postmodernist assumption now dominates the teaching of history at our universities.
I have argued here that previous efforts to formulate ratios or projections have all been failures. Any future attempt is likely to be fraught with the same methodological problems. The only reliable way to proceed is by careful region-by-region studies. Ironically, Henry Reynolds’ sometime co-author and colleague at James Cook University, Noel Loos, has provided a good model in his study of whites killed by Aborigines in North Queensland between 1861 and 1897. In an appendix to his book Invasion and Resistance (1982), Loos provides a table of the settlers and their employees reported killed by Aborigines in this region. He records the names, dates, locations, references and a number of other details about each incident. He also ranks the reliability of the reports, rating each as “acceptable”, “probable” or “possible”. He lists 381 deaths of whites in this period as subjects of acceptable reports, 30 as probable and 59 as possible. By providing full references for his sources, he allows other researchers the opportunity to check his claims for authenticity. This is the way that an empirical historian should proceed. As I indicated above, the handful of regional studies of this kind undertaken to date have shown that mass killings of Aborigines were rare and isolated phenomena. Most violent deaths were in ones and twos and each incident had its own specific cause that does not permit easy generalisation, let alone the accusation of wilful genocide that is now bandied about with such disdain for evidence.
There is one good, general reason why we should expect the eventual compilation of regional studies to produce a very much smaller tally of violent Aboriginal deaths than the twenty thousand now claimed. Ever since they were founded in 1788, the British colonies in Australia were civilised societies governed by both morality and laws that forbade the killing of the innocent. The notion that the frontier was a place where white men could kill blacks with impunity ignores the powerful cultural and legal prohibitions on such action. For a start, most colonists were Christians to whom such actions were abhorrent. But even those whose consciences would not have been troubled knew it was against the law to murder human beings, Aborigines included, and the penalty was death. Those on the pastoral frontier knew there would always be someone likely to report them, as happened at Myall Creek where the alarm was raised by the station overseer. The seven men hanged at the Sydney gallows in 1838 were a grim proclamation of this reality.
Moreover, modern historians who repeat rumours of large-scale massacres rarely explain, indeed they rarely think to ask, how those responsible disposed of the bodies without being detected. At Myall Creek, the killers burned the 28 corpses for all the following day but there was still flesh on them and some individuals could be personally identified. The site itself was advertised for miles around and for weeks afterwards by the birds of prey and carrion hovering overhead. The notion popularised by Henry Reynolds that ten thousand Aborigines met violent deaths in Queensland between 1841 and 1897, with very few of their bodies ever discovered or reported, is inherently implausible.
The questions therefore arise: why haven’t more historians done proper, empirical studies and why have the leading figures hurried in to paint the story in terms so bleak? Moreover, why have the latter felt compelled to resort to such a degree of numerical exaggeration? Part III of this essay will examine their motives. It will argue that, rather than being a product of academics in the last twenty years, the distortion of this history began in the early nineteenth century. It was begun by missionaries who wanted to keep Aborigines separate from the white population. It was later taken up by government officers who wanted to establish “protectorates” and reserves for Aborigines and who accepted the missionaries’ theories and evidence. Massacre stories provided the rationale for the notion of separatism that came to dominate policy towards Aborigines for most of the twentieth century. The academics who entered the debate more recently had some newer, political motives but their main role has been to perpetuate this long tradition.