THE POPULATION OF THE OTTOMAN ARMENIANS

Justin McCarthy

Nineteenth and early twentieth century European and American commentators had no idea how many Armenians lived in the Ottoman Empire. Yet this did not stop them from valiantly presenting statistics on Armenian population. Their estimates ranged from fewer than one million to three million, and their supposed methods of collection were as varied as the estimates themselves. American missionary reporters simply increased the numbers of Armenians, often doubling them. Other writers printed Ottoman statistics with completely uninformed "corrections". Ottoman population statistics were sometimes deliberately falsified, either to make a political point or simply because the writer would not admit that he did not know.[1] Europeans who took the simple expedient of asking Armenians their group's numbers drew up many estimates. Travellers constructed total population numbers by multiplying the numbers in the villages and towns they had observed. Enterprising scholars found what they felt were a sure way to statistical truth when they took all available estimates, then averaged them. Hundreds of estimates were made, each as unreliable as the next.[2]

The many estimators necessarily failed, because the only way to know a population's number is to count it. Reliable data on the population of Ottoman Armenians only could have come from one of two sources-the Ottoman population registration system or the Armenian Church. The former was a well-ordered system of registers of population, which were both updated routinely at the local level and in nation-wide updates (erroneously called censuses). The latter, many assumed, must have kept records of Armenians, analogous to Western European baptismal records.

Knowing that Armenian Church figures would have carried weight in the deliberations of the post-World War I peace conferences that were creating a new Armenian state, Armenian apologists duly presented "Armenian Patriarchate Statistics". As might be expected, the Patriarchate Statistics gave a very high figure for the Armenian population, but that was not enough to make the case that Armenians were a majority of the population in eastern Anatolia. The Patriarchate Statistics therefore made every effort to divide the Muslim population. The category "Muslim" was broken down into Turks, Circassians, Persians, Laz, Gypsies, Sedentary Kurds, and Nomadic Kurds. Under "Other Religions" were listed "Kizilbashis, Zazas, Chareklis, and Yezidis". The intent was obviously to show that there were more Armenians, counted at 1,018,000, than Turks, set at 666,000. This was an understandable tactic, one that avoided the question of whether these ethnic groups would have preferred Muslim or Armenian rule. As will be seen below, however, even these figures could only be derived by adding considerably to the Armenian numbers and subtracting from the number of Turks. An even more telling point, though, is the question of how the Armenians could ever have known how many Muslims were ethnic Kurds, Turks, Circassians, etc. The Ottomans had never counted the population by ethnic group or any category other than religion, and no one else had counted the Muslim population at all. In 1912, no one knew the ethnic groups of Ottoman Muslims, surely not the Armenian Patriarch.

The "Patriarchate Statistics" claimed six vilayets (provinces) as "Turkish Armenia", excising the portions of those provinces that, it was admitted, had very few Armenians.[3] Although exact geographic designations were not given in the Patriarchate Statistics, it is possible to ascertain approximately the Ottoman provinces that corresponded to the "Patriarchal" borders and compare Ottoman and Armenian statistics. The Ottoman figures in Table One have been corrected for undercounts of children and adult females, but this does not affect the proportion of Armenians and Muslims, because the same correction factor is applied to all religious groups. Thus, the Armenian population is assumed to have been as much undercounted as the Muslim-a necessary procedure which, however, adds selectively more to the Armenian population,which was more urban and thus better enumerated. In order to be comparable to Ottoman registration data, the various Muslim ethnic groups in the Armenian figures have been added together as Muslims.

In fact, the Armenian Patriarchate Statistics were pure invention, and the dubious source of the so-called Patriarchate Statistics reflects on their accuracy. No one at the Armenian Patriarchate in Istanbul has ever claimed them.[4] Indeed, there is no record of registration statistics on Armenian population ever having been kept by the Patriarchate. Unlike the situation in Western Europe, no baptismal or marriage records were generally taken or retained, much less centrally collated. Even had there been such records, how would that have given population figures on Muslims and other Christians? These "Patriarchate figures" did not appear at the Patriarchate, nor in Istanbul, nor even in the Ottoman Empire. They appeared in Paris on 1913, in a book written by Marcel Leart (a pseudonym of Krikor Zohrab).[5] It does appear that the Armenian Patriarchate may have collected estimates of Armenian population from Armenian bishops, but these numbers are different than those of the Zohrab inventions and do not include any data on Muslims. As seen in Appendix Three below, they are only slightly different than the figures in Table Two, which are Ottoman counts corrected for under-enumeration of women and children.[6]

To find the actual Armenian population of the Ottoman Empire one must rely on the type of governmental population statistics that are the staple of demographic studies throughout the world. The principle is the same everywhere: the only way to know a population is to actually count it. The Ottoman Government counted its population through a registration system that recorded births, deaths, and migration. Unlike the Patriarchate figures, the Ottoman population figures were used for internal governmental purposes, not propaganda. Although the modern Ottoman population registration system began in the 1830s, not until just before World War I did the Ottomans publish any of the data in a western language, only in Ottoman Turkish, indicating the published statistics were not meant to affect foreign opinion. Internal population documents, never intended to be seen outside the administration and only recently found in archives, and those published statistics were consistent with each other. In short, the intent of the Ottoman Government was to produce the type of usable, accurate population statistics that were seen in other countries.

As in most developing countries, particularly Middle Eastern countries, the Ottomans consistently undercounted women and children. The undercounts have been corrected through the use of standard demographic techniques.[7] These, it might be noted, do not favor any one sub-group of the population. Armenian population is corrected (statistically increased) at the same rate as that of Muslims and other groups. Table Two gives the Ottoman population figures corrected for the undercount of women and children.[8] Figure One is a graphic representation of the Armenian and Muslim populations in Anatolia, Istanbul, and Edirne Vilвyeti, the region that contained all but a small amount of the Ottoman Armenian population.

Table 2

"Ottoman Armenia"

Despite the continued European use of the term, demographically there was no "Ottoman Armenia". The area claimed as Armenia, the "Six Vilayets" (Figure Two) was only 17% Armenian. Indeed, if all the Armenians of the Ottoman Empire and the Russian Southern Caucasus took up residence in the Six Vilayets, Muslims would still have outnumbered Armenians by 40%. In the entire region claimed by Armenian nationalists-Erivan and Kars provinces of Russia, the Six Vilayets, and Cilicia (Adana Vilвyeti)-21% of the population was Armenian, 73% Muslim.

The political implications of the population numbers in what was called Armenia cannot be avoided. If an Armenia had been created in the entire area, it would perforce have been a state in which a distinct minority ruled over a majority, or a state in which that majority had been expelled. Alternatively, a smaller state that drew in Armenians from other regions could have been created. This is what occurred, but only after a sanguine disaster for both the Armenians and the Muslims.

Distribution of the Ottoman Armenians

As can be seen in Table Two, Armenians were distributed throughout all of Ottoman Anatolia, with small Armenian communities in Ottoman Syria. Armenians were also well distributed throughout the Russian domains in the Southern Caucasus. Given the geographical poverty of their homeland, it was natural for large numbers of Armenians to leave for better opportunities. Much of this migration took place before the last quarter of the 19th century. Ottoman statistics for the period from 1878 to 1914 indicate only slight Armenian migration to most of western Anatolia and Ottoman Europe.[9] The exception was Istanbul. A considerable number of Armenian males came to Istanbul for work. In the first half of the 19th century, Ottoman records showed that more than one-third of the Armenian males in Istanbul were bachelors who had come as workers.[10] While later publications of population do not offer this data, it is obvious that the phenomenon must have continued. It is known, for example, that 24% of the Armenian males (and virtually no females) in the city of Istanbul in 1882 lived in inns or shops, not in houses or apartments, a sign that they had come to the city to work.[11] Armenian population in Istanbul also showed a markedly higher number of males than females, much higher than could be accounted by an under count of women, another indication of migration for employment. Judging by standard practice in the Middle East, the settled Armenian population of Ottoman Europe and north-west Anatolia was probably augmented when male migrants found employment and sent for their wives and families.

The Ottoman Armenians did not have the compact population mass upon which national states normally were founded. The density (population per square kilometer) of Armenian population reflected the migration of the Armenians from their homeland. The densest Armenian populations were in İzmit and Istanbul, far from eastern Anatolia (Figure Three). Yet the Armenian migrants did not simply blend into the overall population of their new provinces. In the core Armenian provinces of eastern Anatolia Armenians were well distributed, with sizeable numbers of Armenians in almost all districts (kazas). In other regions they tended to settle together in certain localities, often in larger cities. As mentioned above, Armenians were largely an urbanized population. Population figures for most cities are not available, but in the Six Vilayets of eastern Anatolia 38% of the Armenian population lived in the districts that included the largest towns. In the rest of Anatolia, 45% of the Armenians lived in the districts of the largest towns. By comparison, throughout Anatolia only 20% of the Muslim population lived in these same districts. The Armenian population was twice as urbanized.[12]       

Armenian Population Prior to 1912

Although they had not been a majority in any province of the Ottoman Empire since well before 1800, Armenians were a larger proportion of the population of eastern Anatolia at the beginning of the 19th century than they were in 1912. The main reason the proportion of Armenians decreased was out-migration.

In each of the Russo-Turkish wars of the 19th century, large numbers of Armenians took the side of the Russians when the Russians invaded Ottoman territory in north-eastern Anatolia. The Russians were each time forced to abandon part of their conquests by the terms of the peace treaties that ended the wars. As the Russians retreated, large numbers of Armenians joined them, fearful of retribution from Muslims. Others left to take advantage of tax incentives and free land offered them by the Russians. The Russians had expelled hundreds of thousands of Muslims from their conquests in the Southern Caucasus; they encouraged Armenian in-migration to repopulate the land.

Numbers of the Armenians who left Ottoman for Russian territory are very rough estimates, because the migrants were very seldom counted. Perhaps 50,000 Anatolian Armenians left for the Russian Southern Caucasus during and immediately after the Russo-Turkish War of 1828-29. They joined an estimated 40,000 Armenians who had migrated as a result of the 1827-29 Russo-Iranian War.[13] The Crimean War generated further Armenian migration, but numerical data is very limited. Anecdotal evidence in consular reports indicates that an estimate of a further 50,000 Armenian migrants after the Crimean War is reasonable. After the 1877-78 War, an estimated 25,000 Armenians left the Ottoman dominions for regions that had been conquered by the Russians.[14]

Armenian out-migration continued at a far lower rate after the 1877-78 War. This was primarily economically-based movement. Armenians, particularly those in the Black Sea Region and Erzurum Vilвyeti, left for better opportunities in the Russian Empire. Others fled unsettled conditions in the south-east. Numbers of these migrants are unknown, but an estimate of approximately 1,000 per year is consistent with known demog­raphic data.[15]

Armenian migrants to the United States were relatively well counted. Robert Mirak has examined U.S. immigration statistics and found that 66,000 Armenians came to America from the Ottoman Empire from 1834 to 1914 (Table Three). Connections with the American missionary establishment and open immigration laws facilitated Armenian in-migration. Other countries, many of which were to welcome Armenians during and after the First World War, were not targets of mass migration before the war. According to Armenian figures, before World War I, only 1,000 Armenians went to Canada and Latin America and fewer than 5,000 Armenians immigrated to Europe.[16]

(Source: Mirak, "Armenian Emigration".[17])

Deaths in war and civil upheaval were not the most significant cause of decrease in the Armenian population of the Ottoman Empire before World War I.  Civilian mortality accompanied the Russian invasions of Anatolia in 1828, 1855, and 1877, and the Armenians were naturally a part of that mortality. Unless reliable statistical evidence of differential loss is found, it must be assumed that the proportion of Armenian deaths was approximately the same as that of Muslims and other groups, leaving the figure for the Armenian proportion of the population unaffected. Armenian mortality (like Muslim mortality) in the rebellions of Kurdish tribes in 1834, 1836, 1847, and 1879 is unknown. However, the effect of civil disorder was most likely that the Armenian population did not grow, just as the Muslim population did not grow, in the period before 1878.

Outside of wartime, the greatest amount of Armenian mortality took place in the provinces of Istanbul, Van, Erzurum, Adana, and Haleb (northern) during Armenian demonstrations and rebellions in the 1890s and in the Adana rebellion of 1909. The number of deaths, both Muslim and Armenian, has been the subject of much contention. The careful study of Kamuran Gьrьn indicates that 20,000 Armenians died in the events of the 1890s.[18] Between 17,000 and 20,000 Armenians died in the 1909 rebellion in Adana.[19]  

Table Four projects the Armenian population of the Ottoman Empire (1912 borders) back to earlier dates, taking into account mortality from war and civil disturbance and out-migration as delineated above. Outbreaks of plague and cholera have not been calculated, because their effect on eastern Anatolia is largely unknown. These figures are crude estimates, based on available information. They should not be taken as anything but approximate. Too much is unknown of the environment of the early 19th century to make more accurate projections.[20]

The increase of the Armenian population from 1880 to 1912 was primarily due to the general lessening of mortality that was seen by all groups of Ottoman Anatolia and Europe after the 1877-78 Russo-Turkish War. Improvements in civil order and transportation extended life spans and, in particular, decreased infant mortality. Prior to 1880 the cause of Armenian population decrease was primarily migration to the Russian Empire and the United States. Due to unsettled conditions, rebellions, and warfare, Armenian natural increase before 1878 must have been at or near zero, as it was for other population groups.

World War I and the Turkish War of Independence

During World War I, the Armenians of Ottoman Anatolia were caught up in the worst disaster to strike the modern Middle East. Estimates of the number of Ottoman Armenians who died in World War I and the Turkish War of Independence that followed have grown larger over the years. At the end of the wartime period, Armenian representatives to the Paris Peace Conference and Armenian publicists estimated that 600,000 to 800,000 had died. In recent times, the figure 1.5 million is frequently asserted, sometimes rising to 2.5 million in less temperate pronouncements. Even the smaller of these numbers would mean that nearly all the Armenians died, a contention belied by the existence of those making the estimates.

The method of finding the actual number of Armenian dead is deceptively simple-subtract the number of Armenians who survived the wars from those who were present at war's beginning. This does not strictly provide statistics of mortality, but rather "population loss", because an unknowable number of children were born in the wartime period and a number of adults and children   would   have   died   of   natural   causes   under   normal circumstances.   The   result,   however,   is   a   good   surrogate   for mortality.

It is commonly believed that the Armenians of Ottoman Anatolia were nearly all deported to the Arab Provinces, and that high Armenian mortality was a result of the deportation. This was not the case. Because some deportees who were sent to Syria and Iraq moved to Egypt and Europe during and after the wars and some returned to Anatolia, it is impossible to estimate the number of the deportees with absolute accuracy. It can be seen, however, that the largest group of Armenian refugees were those who fled to the Southern Caucasus. These were not deported to Syria or Iraq. They fled north in three waves: The Russian Army invaded eastern Anatolia in May of 1915, relieving the Armenians of Van, who had seized the city from the Ottomans. When the Russian Army was temporarily forced to retreat from Anatolia, the Armenians of the region the Russians had conquered accompanied them. The Russians returned in 1916, conquering most of eastern Anatolia, and many Armenians returned to their homes. In 1918, the Ottomans advanced, and Armenians departed for the Southern Caucasus once again. Many of these returned after the Ottomans surrendered to the Allies in October of 1918, but they left once again when Turkish Republican forces retook the region in 1920. The 400,000 refugees in the USSR in Table Five were the survivors of a much larger group. Contemporary accounts indicate that the refugees starved to death in great numbers, even being forced to resort to cannibalism. Well in excess of 500,000 must have gone north. In addition, many, perhaps most, of the Armenians who went to Europe and the Americas were never deported. Those who fled to Iran were likewise not deported. It can thus be seen that most Anatolian Armenians were not deported, although their fate as refugees was misery and death.

More Armenians were forced migrants from the eastern Anatolia war than were deported, and they unquestionably suffered higher mortality. Muslims joined in their suffering. When the Russians and Armenians advanced it was the turn of the Muslims to flee. More than a million Muslims were forced migrants.

There were 1,465 million Armenians in Ottoman Anatolia in 1912, before the wars began. (This does not include the 28,000 Armenian residents of Southern Haleb Province, which became part of Syria after the world war nor the Armenians of Istanbul Province and Ottoman Europe, who were neither killed nor deported during World War I, although some were conscripted.) At wars' end, 881,000 remained alive, a loss of 584,000, or 41%.[21] Most of these were victims of the war fought between the Muslims and Armenians between 1915 and 1920, directly or indirectly through starvation and disease. To put the Armenian loss into perspective, it should be noted that the Muslims of the war zone suffered equally horrific loss: The Muslim population of the Van Province decreased by 62%, that of Bitlis by 42%, that of Erzurum by 31%. Not coincidentally, these were the provinces of greatest conflict between Ottoman and Russian armies and between Muslim and Armenian civilians.

(Source: McCarthy, Muslims and Minorities)

The massive mortality in Anatolia was the product of total war in which no quarter was given, as well as years in which no crops were harvested and disease ravaged populations already ravaged by hunger. All shared starvation and disease; each side killed the other mercilessly. It is no wonder that death tolls were so high. Those who elevate the mortality of one group or ignore the mortality of another mistake the lesson of the times, which is not of persecutors and the oppressed, but of general inhumanity.

Appendix One. Russian and Armenian Statistics

The Russian Empire took only one actual census in the Southern Caucasus. Like all population counts in the region, it was imperfect. Its figures, however, were the best available, because they were the only ones based on actual enumeration of the population. Though the discrepancy was less than that seen for the Ottoman Empire, the actual figures and "Armenian Patriarchate" figures for the Southern Caucasus were quite different.

Some of the discrepancy between the two sets of data can be explained by the seventeen years between the two statements, but not very much. Under the best of conditions, the Armenian population might have increased by 300,000 in those seventeen years. This would still leave a discrepancy of more than 350,000. In-migration of Armenians in the period could only account for a small portion  of the  difference. As  there were  no independent Armenian sources of information on the population (and no one has asserted that the Russians ever allowed an army of Armenian census-takers to spread out across the region), it must be assumed that the Armenian Patriarchate Statistics on the Southern Caucasus, like those for the Ottoman Empire, were simply invented. Interestingly, the author of the Patriarchate Statistics did not much undercount Muslims in the Southern Caucasus. Armenian numbers were artificially inflated, but Muslim numbers left relatively unchanged.

Appendix Two: Armenian Migration to the Southern Caucasus

Except for major migrations resulting from wars, there is very little evidence on migration of Armenians from the Ottoman Empire to Russia. It is known that this migration existed, but no registration figures have been found. Most likely the Armenian migrants were never counted. This has not kept some from making fanciful estimations: Supposedly a yearly average of 10,000 Armenians went to Russia from 1845 to 1870, 15,000 a year from 1870 to 1900, and a total of 150,000 from 1900 to 1914.[22] Ignoring the fact that such a massive migration would have been mentioned in Ottoman documents and European consular reports, which it is not, statistical analysis proves it to have been impossible.

In order to analyze various statements on migration to Russia, a simple computer program projected the Armenian population of the Russian Southern Caucasus from the figures in the 1897 census back to 1826 (immediately before the Russians conquered the Erivan Province-"Russian Armenia"). The program assumed various rates of natural increase (i.e., without major migration). When the figures for migration estimated in this article (1828-29 War: 50,000; Crimean War: 50,000; 1877-78 War: 25,000; and yearly migration of 1,000 in other years), the resulting population for 1826 is reasonable-between 395,000 Armenians (.01 yearly growth rate) to 613,000 Armenians (.005 growth rate) in the Southern Caucasus. The actual figures lay somewhere in between these two extremes. If the fanciful estimations of migration as given above are used in the program, there would have been only between 26,000 (.01 rate) and 144,000 (.005 rate) Armenians in the entire Southern Caucasus in 1826. These figures are absurdly small,   far   below   any   estimates,   including   those   of Armenian scholars.[23]

The 1897 Russian census supports these conclusions: The census registered 10,187 of the inhabitants of the Erivan Province, 38,664 of the Kars Province, and 101,066 in the entire Southern Caucasus as born in the Ottoman Empire. Figures for Armenians alone are not available, and these figures include many non-Armenians.

The only inference that can be drawn is that the estimates of large-scale migration from the Ottoman Empire to Russia outside of years of war and immediately following war are totally unrealistic. As stated in this article, the estimate of migration in "normal" years of 1,000 is speculation, but it is speculation that fits the demographic possibilities. It thus cannot be far from the truth.

Appendix Three. Armenian Patriarchate Statistics

A set of Armenian Patriarchate Statistics, drawn from the Armenian Archives, was published by Raymond H. Kevorkian and Paul B. Paboudjian in 1992.[24] It appears to be far superior to the other "Patriarchate Statistics". According to this source, the Armenian Patriarch collected this data from Armenian bishops throughout the Ottoman Empire in 1913. The most likely explanation of these statistics is that an order went out to the Armenian bishops, who took Ottoman statistics and local information from Armenian parishes, correcting them for perceived undercounts. As might be expected, this produced some exaggeration. However, comparison to Ottoman data corrected by demographic techniques does not show great differences. For some cases, the corrected Ottoman data shows more Armenians than does the Armenian data. It is to be hoped that a more complete set of the Armenian data will be published, as the authors promise. It should be possible to give a better evaluation of this data. It would be especially valuable to investigate material on how the data was collected.

None of the conclusions drawn in this article on Armenian population would be upset if the slightly higher figures in these patriarchate statistics were true, with the exception of the number of Armenians who were lost during the World War I period, adding approximately 250,000 to the death toll. In fact, these figures provide support from an Armenian source for the above analysis of the demographic place of Armenians in the Ottoman Empire.

Figure 1

Figure 2

Figure 3

Стр.| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 |