Daniel Goldhagen and Kenya: recycling fantasy

Daniel Goldhagen’s book “Worse Than War” includes British colonial rule in Kenya in the 1950s among its case-studies of “elimination”. A close reading of the demographic evidence reveals the falsity of the argument, says David Elstein.

(This article was first published on 4 March 2010)

A scholar who makes large claims must expect to be held to an exacting test of accuracy. If it is failed, the integrity of the work is called into question. In this respect, Daniel Goldhagen’s treatment of events in colonial Kenya in the 1950s - which takes up thirty pages of his new book Worse Than War: Genocide, Eliminationism, and the Ongoing Assault on Humanity (Little, Brown, 2009) - deserves careful scrutiny.

Daniel Goldhagen made a name for himself with Hitler's Willing Executioners: Ordinary Germans and the Holocaust (Random House, 1997) in which he argued that millions of Germans - not just a hard core of evil Nazis - were directly involved in the extermination of Jews during the second world war. Worse Than War presents an overview of various historical and contemporary acts of “elimination”, whilst attempting to construct a way to prevent further “eliminationist” outbreaks (including genocide) in the future.

Most criticism of the book has focused either on the implausibility of its preventative schema or the disjunctions of scale of the various “eliminations” it considers. My concern is more limited: the author’s inclusion of Kenya as an episode of “eliminationism”. This is inaccurate, draws on unreliable and tendentious sources, and contravenes known facts which were in the public domain (including on Wikipedia) for two years before Worse Than War was published.

During the state of emergency in Kenya in 1952-60, British forces battled to suppress a violent uprising by radicals whose support-base lay among the country’s largest single ethnic grouping, the Kikuyu. The radical movement, which came to be known as “Mau Mau”, cemented support amongst its followers with bloodthirsty oaths and rituals. Daniel Goldhagen’s description of what happened in this period is emphatic. He refers, inter alia, to:

“Mass murder”; “mass slaughter”; “butchered victims”; “1.5 million Kikuyu incarcerated”; “tens of thousands killed (estimates range from 50,000 to 300,000)”; “perhaps a half million Kikuyu” driven into an “extensive and murderous camp system”; one million “restricted to deadly barbed-wire villages” where “the death tolls were enormous”; “worked to death”; “beaten to death”; “most of the Kikuyu dead resulted from British starvation policies”.

Goldhagen concludes his litany of British atrocities by saying they were “all, according to the British, in response to the putative threat and unsurpassable savagery of the Kikuyu liberation movement known as the Mau Mau. How many whites did the bestial Mau Mau kill? Thirty two?”

The acid question seems to invite only assent to the depiction of colonial perfidy and hypocrisy. But follow the trail of evidence, and a more complex pattern emerges.

The trail of detail

In the first place, when Goldhagen contrasts what he believes to be 300,000 Kikuyu dead with thirty-two white settlers murdered during the conflict, there is a missed connection. For the emergency was declared not in response to the killing of whites but after a series of fatal attacks on Kikuyu leaders who supported British rule - a reminder in turn that the Mau Mau uprising was as much a civil war within the Kikuyu as it was a revolt against British rule and British land policy in Kenya. Nearly all of the 1,800 known civilian victims of Mau Mau were Kikuyu, killed for refusing to take oaths, or in settlement of old scores, or for staying loyal to Britain (see Daniel Branch, Defeating Mau Mau, Creating Kenya: Counterinsurgency, Civil War, and Decolonization [Cambridge University Press, 2009]).

The most infamous of the period’s massacres took place at Lari, when seventy-four Kikuyu were cut down by a Mau Mau gang with machetes or burned to death. The immediate response by the Kikuyu Home Guard was to pursue and kill 150 local Mau Mau whom they blamed for the attack (several other Mau Mau were captured, tried for the murders, convicted and hanged). British troops played only a minor part in this revenge mission.

Goldhagen’s belief that the great majority of Kikuyu supported the Mau Mau may or may not be true, though out of a Kikuyu population of 1.5m there were never more than 20,000 armed Mau Mau fighters: equal numbers joined the Kikuyu Home Guard and the prison service.

In any case the Kikuyu, although the largest ethnic group in Kenya, represented only 20% of the country’s African population (it is not clear whether Goldhagen is aware of this). Few non-Kikuyu joined Mau Mau, and there is scant evidence that in Goldhagen’s own terms the colonial power targeted non-Kikuyu ethnic groups for “elimination”.

The suspicion of the Mau Mau was shared by the nationalist Kikuyu leader Jomo Kenyatta, who easily won Kenya’s first free election in May 1963 and became its first president when independence was achieved in December 1963; true, his cabinet included three former Mau Mau members, but Kenyatta denounced the movement as a “disease” that had needed to be suppressed.

More generally, there are grave doubts over Goldhagen’s principal source for his account of Kenyan “genocide”: namely, Caroline Elkins’s book Imperial Reckoning: The Untold Story of Britain's Gulag in Kenya (Henry Holt, 2005).

Britain’s Gulag: The Brutal End of Empire in Kenya (as the book was called in the United Kingdom) may have been written by a Harvard University professor and won a Pulitzer prize; but it was widely criticised even by sympathetic reviewers for its shrill comparisons between British policy in Kenya and the Nazi holocaust (see, for example, Neal Ascherson, “The Breaking of the Mau Mau” [New York Review of Books, 7 April 2005]). Some academic reviewers were more dismissive. Susan Carruthers of Rutgers University, who noted that Elkins had managed to confuse the Hutu and the Tutsi in Rwanda, said: “she proves the least reliable guide to history: this was not genocide - history is not well served by its sloppy invocation”.

Elkins’s cavalier approach to evidence is highlighted by the complete unreliability of her most notorious assertion: that there were some 300,000 “unaccounted for” Kikuyu at the end of the British campaign against the Mau Mau rebellion, as compared with the official figure of 11,503 Mau Mau killed in action. I was one of those who drew attention to these flaws of approach and detail (see “Tell me where I’m wrong” [London Review of Books, 2 June 2005]) and “The End of the Mau Mau” [New York Review of Books, 23 June 2005].

In these letters I demonstrate how Elkins manipulates her comparisons of Kenyan ethnic populations in the censuses of 1948 and 1962, covering the Mau Mau years. She chooses six ethnic groups. Comparing the Kikuyu to the other five would have shown a 60% increase for the Kikuyu from one census to the other, and a 51% increase for the other five. Elkins, however, chooses to treat the two groups with lowest growth - the Embu and Meru - as “Kikuyu” (on the grounds that they spoke the Kikuyu language), thus creating a contrast between a 42% growth for the combined “Kikuyu” and 61% for the remaining three groups (Kamba, Luo and Luhya). By this sleight of hand she creates a theoretical shortfall of 19%, which she then translated into her “300,000 unaccounted for”; a figure that Goldhagen uncritically recycles.

Goldhagen also treats as gospel all the anecdotal evidence assembled by Elkins - including accounts of events recalled at fifty years’ distance. Yet the senior British historian of Kenya, John Lonsdale, whom Elkins thanks profusely in her book as “the most gifted scholar I know”, warned her to place no reliance on anecdotal sources, and regards her statistical analysis - for which she cites him as one of three advisors - as “frankly incredible”.

The scope of violence

There is no doubt that many of the white settlers who joined in the military campaign to suppress the Mau Mau rebellion were deeply unpleasant and malevolent racists. The accounts of torture and murder of captured suspects – often given boastfully by the perpetrators themselves - are gruesome and almost certainly true. Some of those responsible were tried, convicted and imprisoned for their misdeeds, but the atmosphere of menace and hysteria that Mau Mau generated meant that much brutality and violence was either inflicted or condoned by British soldiers as well as their auxiliaries.

Moreover, the state of emergency declared in October 1952 allowed a mass round-up of any Kikuyu suspected of being or becoming Mau Mau members. The official total of detainees during these years is 80,000; the peak figure at any one time was reached in December 1954, when 70,000 were being held. Elkins, without offering any explanation, claims the true total across the years was between 160,000 and 320,000. David Anderson, whose rightly praised Histories of the Hanged: The Dirty War in Kenya and the End of Empire (Orion, 2005) was published at the same time as Britain’s Gulag, estimates 150,000; though, again, the only statistics his book provides happen to support the 80,000 figure. Goldhagen’s “half a million” is a number he just made up.

Whatever the actual number, the total swamped the British prison authorities. Conditions in the camps were poor, often lamentable. Outbreaks of typhoid, dysentery and other diseases, as well as diet-related illnesses like pellagra, caused hundreds of deaths. There were violent deaths too, not least - on both sides - in clashes between warders and well-organised prisoners. The Mau Mau camp leadership also routinely ordered the murder of fellow detainees who refused the Mau Mau oaths or were suspected of collaborating.

However, the only documented episode of multiple violent deaths happened at the very end of the detention process, in 1959, when eleven of the last hard-core prisoners at Hola camp in eastern Kenya were beaten and killed by African warders after they refused an order to work. The subsequent inquiry and political scandal almost brought down Britain’s Conservative government: which suggests that this must have been a very unusual event.

Caroline Elkins herself unwittingly provides evidence supporting this. She cites the Mwea camps between March 1957 and March 1958 as the very worst period for detainees. Yet one of her own interviewees confirmed that, three-quarters of the way through these twelve months, not a single detainee had been seriously injured, let alone killed, at Mwea: and there is no evidence of serious injury or death in the remaining three months. It should also be noted that the camps were regularly visited by outside parties, including the International Red Cross, and - in response to a flow of lengthy letters and petitions from detainees - Kenyan ministers and British parliamentarians such as Barbara Castle. None of these written complaints mentioned detainee deaths.

Moreover, however often white settlers might equate all Kikuyu with the Mau Mau, British officials were all too aware that protecting loyal Kikuyu and their cattle from Mau Mau mutilation was a central objective of the war. The policy chosen was to require the great majority of Kikuyu to live inside 800 fortified villages. This served the purpose of protection, but also ensured that no supplies from sympathisers reached the Mau Mau fighters in the forests.

These are the “barbed-wire villages” to which Goldhagen repeatedly refers. He follows Caroline Elkins in seeing this approach - which entailed undoubted hardships resulting from restrictions on normal farming and from violent attacks on villagers - as amounting to a genocidal policy. In fact, there was a kind of precedent in Kenyan history, in that (as John Lonsdale points out, citing the classic study by Louis SB Leakey of The Southern Kikuyu Before 1903) Kikuyu villagers in the 19th century used to protect themselves with stockades to fend off attacks from Maasai; the thinking behind the new fortified villages was thus not altogether alien to Kenyan experience. In any event, British forces had successfully used this “villagisation” policy to crush a communist insurgency in Malaya in the early 1950s, and it is hard to see how the Mau Mau rebellion could otherwise have been defeated so quickly. The militants in the forests, cut off from provisions and heavily outgunned, were defeated militarily in barely a year, though it took five more years for all the detainees finally to be released and for the emergency to be ended.

The claim of genocide

So where was the genocide? If it was not in the detention-camps, it must have been in the villages, inflicted by deliberate starvation and ill-treatment. It is certainly true that wartime conditions led to food shortages, which in turn led to malnutrition amongst mothers, undernourished babies and a rise in infant mortality.

Defenders of British policy would no doubt argue that this was the unwelcome but unavoidable consequence of a wartime situation, and that more Kikuyu lives were saved by protecting the loyal population and ending the conflict quickly than were lost in these indirect ways. Yet the key question is: what was the scale of this loss of life?

The decisive blow to the Elkins thesis was dealt by the late expert in Kenyan demography, John Blacker, who had been the key adviser to the Kenyan administration for the censuses taken both before and after independence.

Blacker published an important paper in 2007 (“The demography of Mau Mau: fertility and mortality in Kenya in the 1950s: a demographer's viewpoint” [Journal of African Affairs, 106/423, April 2007]). He exhaustively analysed all the relevant census data - from 1948 and 1962, but also from 1969. He noted that Elkins’s figure of 300,000 deaths “implies that perhaps half of the adult male population would have been wiped out - yet the censuses of 1962 and 1969 show no evidence of this - the age-sex pyramids for the Kikuyu districts do not even show indentations.”

Where Blacker found evidence of “excess deaths” - that is, more deaths than might have been expected above previously established annual rates - they were primarily amongst children under the age of 10, though the Kikuyu child-mortality rate still remained well below that of the Luo, and also below that of the Kamba and Luhya. This increase (of 26,000) he attributed to wartime exigencies: food shortages, malnourishment amongst mothers and their children, and the easier spread of infectious diseases inside the protected villages. He judged similar causes to have been responsible for a smaller number of “excess” adult female deaths: 7,000 in the period, equivalent to one or two each year in each protected village during the emergency.

During the emergency itself, John Blacker calculated a figure of 17,000 “excess” adult male deaths. The official figure of Mau Mau killed is 11,503, though this does not include the 1,090 who were hanged after being convicted of capital crimes. David Anderson estimates that at least 3,000 African police and soldiers (mostly Kikuyu) died, as well as the 1,800 civilians murdered by Mau Mau according to the official count (he thinks that figure almost certainly too low). Not all the civilians killed were male, or Kikuyu, and many hundreds died of disease in the worst-run detention-camps; but combining all these figures brings the tally to more than 17,000, leaving no room for any campaign of “elimination”.

Blacker’s estimate of 50,000 “excess” Kikuyu deaths across the 1950s is given added context by the effects of violence between the Kikuyu, the Luo and the Kalenjin in the month after the disputed elections of 27 December 2007 - when approximately 1,500 people were killed. It is relevant here that Blacker attributed two-thirds of the 1950s deaths to malnutrition and disease rather than violence, and that Anderson calculates that 5,000 of these were actually victims of the Mau Mau.

In summary, Blacker’s view is that the deaths associated with villagisation represented 0.003% of the Kikuyu population for each of the years of the emergency. So much for the “enormous death tolls”: far from being “eliminated” by a brutal and racist colonial power, the Kikuyu population grew from 1.026m in 1948 to 1.642m in 1962, at which point the British handed over power to a predominately Kikuyu government.

Blacker concludes that “there is no evidence to support the claims made by Elkins [which] are based on a misunderstanding of the data”. Before it was published, Blacker’s paper was read by both David Anderson and John Lonsdale. There could not have been a more decisive intervention in this debate, but it finds no place in Goldhagen’s work.

If Daniel Goldhagen had removed all reference to Kenya in Worse Than War, it would not in any way have affected his broad thesis. In retaining the section on Kenya, he exposes himself to the charge that he is the kind of scholar who is either unaware of the facts or prefers to exclude those which do not fit his thesis. Either way, he has done himself - and history - no favours.

About the author

David Elstein is Chairman of openDemocracy's Board. He is also Chairman of the Broadcasting Policy Group. He is a director of Kingsbridge Capital Advisors, and a supervisory board member of two German cable companies.

More On

Also in openDemocracy on calculating the cost of conflicts:

Jean Seaton, "The numbers game: death, media, and the public" (6 October 2005)

John Sloboda, “Tony Blair and Iraq: blithe ignorance” (9 July 2006)

Michel Thieren, “Deaths in Iraq: how many, and why it matters” (18 October 2006)

Michel Thieren, “Deaths in Iraq: the numbers game, revisited” (11 January 2008)

Martin Shaw, "Afghanistan and Iraq: western wars, genocidal risks" (24 July 2009)

John Sloboda, “The human cost of war: name before shame” (29 July 2009)