A Glimpse into

the Armenian Patriarchate Censuses

of 1906/7 and 1913/4

The size and composition of the population of the Ottoman Empire has been disputed for over 100 years.(1) The primary sources used to document the various assertions have included Ottoman government statistics, Armenian Patriarchate statistics and estimates by numerous contemporary observers—each with strengths and weaknesses. The subject of this article will be the Armenian Patriarchate statistics. Detailed records from the Patriarchate have largely been ignored to date. I aim to show that they can be used for meaningful analysis and are an indispensable resource.

Background

The Armenian community in the Ottoman Empire maintained over 2,000 churches—the great majority Armenian Apostolic, but Catholic and Protestant as well. Baptisms, marriages and deaths were recorded, but almost all such records were destroyed during the genocide.(2)

In addition to the recording of vital events, the church periodically undertook the task of enumerating the Armenian population via a census. The Armenian National Constitution (1862) created a census department within the Bureau of the Patriarchate. The census was used for taxation, as well as for determining representation in the national political and religious assemblies.(3)

While scholars have made use of previously published summaries compiled from the census registers, actual registers have never been analyzed or even been known to exist. The absence of detailed records has led some to question how the summary tables were generated(4); however, some actual registers have survived.

Register from the 1906/7 Armenian Patriarchate Census (click on picture for larger view)	Register from the 1906/7 Armenian Patriarchate Census (click on picture for larger view)

Images 1 and 2 are pages from registers compiled in 1906/7 and 1913/4. Both samples are records for the same houses on Khan Street (Han) in the Mouhsine Khatoun (Muhsine Hatun) district of Istanbul. The registers are contained in the archives of the Patriarchate and were microfilmed by the Mormon Church.

Analysis

The existence of the registers allows for a better understanding of how they were compiled, as well as of some of the strengths and weaknesses inherent in the summaries. What follows is only casual treatment, but hopefully it is enough to pique the curiosity of other researchers.

The first item to note is that the census was not a continuous register from which periodic summaries were tabulated. Each register does not appear to have been compiled simply by updating the previous register with the vital events occurring in the intervening years. A summary for the 1906/7 census for Pera indicates that the results were in fact summarized by street and district and, thus, were the foundation for aggregation.

The column headings are district, street, number, first name, last name, occupation, father’s name, mother’s name, native city, and year of birth (note that the years of birth are according to the hijra calendar). The data and years collected correspond closely to the Ottoman registers. This raises some interesting questions given that Ottoman registers containing Armenians have not yet come to light.

Detailed data of this kind can be analyzed for quality and consistency. For instance, age misreporting is a common error found in censuses. Slightly better results are achieved by asking for the year of birth instead of age; yet, it is still common to observe heaping at years ending in certain digits. Table 1 is compiled from a sample of 2,300 individuals in the 1913/4 census:

It is interesting to note that women displayed heaping in years ending in 0 and 6, while 5 and 6 were most common for men. These results are not sufficient to form conclusions as the heaping may result from actual events. For instance, a large number of people reported births in the years immediately following the 1877/8 Russo-Turkish War.

Table 2 summarizes the same sample into 5-year age groupings indicating an undercounting of children under the age of 10. Demographers have made use of stable population theory to estimate the degree of undercounting by comparing the enumerated population to standard model life tables.

Extreme care is called for, though, as such methods can easily lead to the masking or removing of the impact of actual events on the age structure of the population. It cannot be emphasized enough that one must understand the history of the region under analysis before conclusions can be drawn. That one must also understand the situations where stable population theory is applicable is also self-evident.

In the sample, males accounted for 51 percent of the total population. While superficially one might expect a 50/50 male to female ratio, it is difficult to interpret such results. Yet, it is known that Istanbul contained a large Armenian male migrant population. Possibly the ratio was further impacted by a greater natural female life expectancy or the massacres of the late 1800’s resulting in more male deaths than female. The tilt in the age structures implies greater female life expectancy; however, such conclusions are premature without further analysis.

Other areas open to exploration are the prevalence of certain occupations, the disparity in ages between spouses, the composition of households, the origin of the population by gender, etc.(5) For instance, a conclusion drawn from producing the above tables was that husbands were generally significantly older than their wives.

It would be fascinating to explore the population by age, gender, and native city. Istanbul served as an economic center for Armenians, but in the last years of the empire this may not have played as great a role as other cities and countries served as economic magnets for Armenians.

Conclusion

Population figures originating from the Armenian Patriarchate have come under harsh criticism, particularly from those who deny the Armenian Genocide and, thus, attempt to validate a much lower pre-genocide Armenian population.

That the Patriarchate underrepresented Muslims is problematic, yet the Patriarchate had no means of counting the Muslim population. Justin McCarthy accepts that Armenians were undercounted to a greater degree than Muslims, yet this has not diminished the value of Ottoman statistics in his analysis. While it is not known how the Patriarchate arrived at the Muslim population, it is clear that the Patriarchate had the incentive and means to enumerate the Armenian population and did so.(6)

We now have a glimpse into the process used by the Patriarchate to compile data on the Armenian population. In addition, the information available is in greater detail than previously known and, thus, allows for an assessment of quality and an easier comparison with other sources. Much tedious and technical work needs to be done, but from what is available thus far, it is apparent that the data presented by the Armenian Patriarchate is a valuable and required resource for analyzing the Armenian population of the Ottoman Empire.

Endnotes

1) Throughout the 19th century, observers would speculate on the population of the Ottoman Empire, its changing characteristics, and prospects for the future. There were, however, great disparity in the figures and a general belief that an accurate accounting was impossible given the data available. More recent sources of interest include Justin McCarthy, Muslims and Minorities: The Population of Ottoman Anatolia and the End of the Empire (New York: New York University, 1983); Kemal H. Karpat, Ottoman Population 1830–1914: Demographic and Social Characteristics (Madison: University of Wisconsin Press, 1985); Levon Marashlian, Politics and Demography: Armenians, Turks, and Kurds in the Ottoman Empire (Cambridge: Zoryan Institute, 1991); Raymond H. Kevorkian and Paul B. Paboudjian, Les Armeniens dans l’Empire Ottoman a la veille du genocide (Paris: Les Editions d’Art et d’Histoire ARHIS, 1992); and Robert H. Hewsen, Armenia: A Historical Atlas (Chicago: University of Chicago Press, 2001).

2) Outside of churches in Istanbul, the only known records I have personally worked with are the baptisms for the period 1902 through 1915 from the church of St. Gregory the Illuminator in Gesaria (Kayseri). In addition, some tax registers from the Van region have survived from the 1800’s. The archives at the Armenian Patriarchate in Istanbul may contain addition records.

3) H.F.B. Lynch, Armenia: Travels and Studies (New York: Armenian Prelacy, 1990), vol. 2, pp. 449–467.

4) McCarthy offers a number of criticisms, including the presumed absence of any records (Muslims and Minorities, pp. 47–57). McCarthy is also concerned that some of the figures have been rounded and were used for political objectives. He considers some problems “insoluble,” such as that population by age and sex were not published. In summary, McCarthy invalidates the Armenian Patriarchate statistics primarily on the inaccuracy in the estimate of the Muslim population. McCarthy exhibits a favorable bias toward Ottoman government statistics that often compromises his collective works.

5) An interesting study of Istanbul Muslim households was done by Alan Duben and Cem Behar in Istanbul Households: Marriage, Family and Fertility, 1880–1940 (Cambridge: Cambridge University Press, 2002). Duben and Behar make use of Ottoman censuses. A valuable project would compare the results of the 1906/7 Ottoman census with the 1906/7 Armenian Patriarchate census, including a comparison of the registers for the same districts. Such a comparison may go a long way in understanding the controversy in population estimates.

5) Justin McCarthy, Muslims and Minorities: The Population of Ottoman Anatolia and the End of the Empire (New York: New York University, 1983), pp. 54–57.

6) Souren Papazian, Odyssey of a Survivor (Maryland: Jensen Press, 2002), p. 37. Beyond the known available census and vital records, Papazian describes his role in carrying out the 1913 census in the village of Havav in the district of Palu.