The Black Book

Probably the most authoritative source on the Armenian numbers is the so-called Black Book of Talat Pasha, found among the possessions of the former Ottoman Interior Minister and Grand Vizier after he was assassinated by an Armenian terrorist in Berlin in 1921.¹ The documents have been claimed as proof, from the architect of the genocide himself, that a million Armenians were subject to “race extermination” by the Ottoman government.² In fact, the 77-page black-cover notebook or journal, which does not seem to have been written in Talat’s hand, says nothing whatsoever about his intentions with the deportations.³

What the Black Book records is that about 1.1 million Armenians, out of a pre-war population given as 1.5 million, were “missing” from the Ottoman registry in 1917,⁴ with no indication of whether they had been deported, killed, migrated to Russia, or defected to join the Russian or French volunteer units, and “significant numbers of Armenians [listed as ‘missing’ in the Black Book] turned up later”.⁵

The Black Book’s provincial breakdown shows a wide variance in the application of deportation. While the Eastern Six Provinces were almost entirely cleared of Armenians,⁶ the proportion of Armenians remaining in the fifteen Central and Western Provinces ranged from 98% to 15%. The overall average of Armenians remaining in the Centre and West of Anatolia was about 45%, and seven of those provinces retained half or more of their Armenians.⁷ To the extent this bears on the genocide debate, it lends weight to the non-exterminatory interpretation of Ottoman conduct.

There being no evidence in the Ottoman archives of a decision to annihilate the Armenians, the genocide case is built by inference.⁸ An oft-made argument is that the deportations, instead of being a response to the Armenian rebellion in alliance with invading Russians in the east, were “neither a military necessity nor a contingency of war”,⁹ and this is demonstrated by deportations far from the warzone in Anatolia. Such an indiscriminate policy, the argument goes, indicates that the deportations were cover for a premeditated attack on the Armenians “as such”, in the language of the Genocide Convention. The thesis is troubled on its face by the simultaneous landing of the British and French—who had their own connections with the Armenian committees—at Gallipoli, extending the warzone from east to west, plus the wholesale exemption from deportation of Armenians in Constantinople, Smyrna (Izmir), and Aleppo.¹⁰ And the Black Book, assuming its authenticity,¹¹ adds further evidence against the thesis.

Armenian Deportees and Deaths

The Black Book suggests 250,000 to 280,000 Armenians were never deported or were relocated within Anatolia.¹² Walter Rössler, the German Consul in Aleppo, reported in December 1915 that 500,000 Armenians had been “spared from being deported”. Rössler’s evidence is problematic. An eyewitness he might have been, Rössler’s bias for the Armenian cause was well-known, a possible incentive to minimize the figure. More importantly, Rössler obviously did not gather this data personally, and he did not claim otherwise. Rössler said his un-deported number was “estimated”, and the only sources cited in his cable are “circles which must be taken seriously, as they are usually better informed than others”.¹³ Still, there are contemporary scholars who consider 500,000 a plausible total for Armenians not deported out of Anatolia.¹⁴

As to the numbers who were deported, Arnold Toynbee, a British intelligence operative responsible for wartime propaganda against the Ottomans, estimated in a 1917 product that there were 1.2 million deportees.¹⁵ Boghos Nubar, the Armenian representative in Europe, said the total was 600,000 to 700,000 in 1918.¹⁶ Turkey has also given the 700,000 figure.¹⁷ And Johannes Lepsius, the famous German Protestant missionary advocate for the Armenians, estimated 1.4 million in 1919.¹⁸ At the present time, Donald Bloxham is an outlier in implying there were 1.5 million deportees.¹⁹ Taner Akçam, relying on the Black Book, writes, “one can confidently say that the number of deported Armenians was around 1.2 million”,²⁰ and this is generally considered the high-end figure. On the low end, Fuat Dündar, using the same Black Book, has estimated 800,000,²¹ and Raymond Kévorkian estimates 850,000.²² All of these are specialist scholars on the genocide-recognition side of the debate. Most non-specialists either give the 800,000 to 1.2 million range,²³ or pick a number from within it; around 900,000 seems to be popular of late.²⁴

A reason to suspect that this range for the deportees is too high is that there was a mass-movement of Armenians out of the Six Provinces—especially Van, Bitlis, and Erzurum—before the deportations were ordered and several more waves of refugees up to 1918 as the Ottomans and Russians fought bitterly on a shifting front-line, a significant cause of the “missing” tally in the Black Book.²⁵ Toynbee and Lepsius do try to take some account of the refugees, though their estimates hover around 200,000, and the lower estimates in modern scholarship just for the Armenians who departed into the Russian Caucasus—without factoring in those who emigrated to Western Europe or the United States—range from 250,000 to 400,000.²⁶ Some demographers put the figure at half-a-million.²⁷

When it comes to the fatalities, Toynbee’s 1916 “Blue Book”, his most enduring messaging product, said an estimate of 600,000 Armenians killed in 1915 would “not be far wrong”.²⁸ Lepsius claimed that German diplomats estimated “between 800,000 and one million” in October 1916. A commission in 1919 created by Mustafa Arif Deymer, the Interior Minister of the post-war Turkish government, which was set on vilifying its predecessor and operated under Allied occupation, estimated that 800,000 Armenians were killed during the deportations.²⁹ The Armenian leadership in 1919 claimed: “One million Armenians have perished during this War”.³⁰ It is possible to see the approximate convergence of these numbers as supporting evidence for their validity. It is also possible to see the convergence as evidence of cross-contamination among actors in the same milieu.

For the sake of completeness: The Turkish government has said that on the available evidence “it can be concluded that the number of the Armenian losses was around 300,000”.³¹

Armenian Population Statistics

What would provide some solidity is reliable denominators—that is, numbers for the total Ottoman-Armenian population before the First World War and after its end. But alas, there are only competing claims on this issue, too.

The Ottoman census department recorded 1.3 million Armenians in 1914, and some sceptical demographers have gone as high as 1.7 million.³² The Armenian Patriarchate recorded 2.1 million Ottoman-Armenians in 1913, and some Armenian historians have gone as low as 1.9 million.

In 1919, the Armenian National Council of Constantinople estimated there were 1.25 million Ottoman-Armenian survivors, and George Montgomery, an American official at the Paris Peace Conference, presented an apparently detailed study showing 594,000 Armenians remaining in the rump Ottoman realm, with 450,000 in Russia and 60,000 in Iran, a total of 1.1 million.³³ The same year, Lepsius estimated 845,000.³⁴

The Turkish government has suggested a survivors’ figure of 980,000,³⁵ and Armenian activists tend to favor a figure around 600,000.³⁶ Demographers on both sides of the debate have produced surprisingly consistent estimates of 836,000³⁷ and 881,000³⁸ Armenian survivors.

What To Make of It All

From these numbers, it would be possible to calculate an Armenian death toll anywhere from 50,000 to 1.5 million.

It can be pointed out that the Armenians in 1919 were trying to maximize their survivor tally to argue to U.S. President Woodrow Wilson that they were a nation of sufficient size to qualify for his promised self-determination. But excluding numbers because they were politically generated is not going to leave much data. Decisions on inclusion and exclusion are intrinsic to data analysis, of course, but it makes nonsense of the idea that turning to the numbers eliminates subjectivity.

Take Dr. Montgomery’s overall figure, which is so close to the Armenian Council estimate and would allow a similar hyper-low-end fatality estimate. Is that going to be excluded, too? Or used in a weighted average? Or will Montgomery’s number for Armenians within Ottoman territory be used instead? Any one of these decisions could be justified, and then the process repeats for the next datapoint and the one after that. Dozens of decisions—large and small, conscious and unconscious—taking every analyst on a different route through the garden of forking paths. It is interpretation all the way down, and the demographers cannot rescue us from this because they are working with the same uncertain datapoints and making their own decisions on whether and how to use them.

Overall, then, in the Armenian case we see the usual pattern of incredibly spongy figures, with the underlying numbers marred by hearsay and agenda, and conclusions that depend on layers of assumption and interpretation, none of which can be taken for granted. These numbers might be all we have, but it does not make them any better. Attempting to work with these statistics provides so many researcher degrees of freedom in selecting a way through that it amounts to choosing your own adventure. The honest answer to the empirical numbers questions surrounding the Armenian catastrophe is that we do not know and probably never will.

FOOTNOTES