Tweet
In his book “Muslims and Minorities” Justin McCarthy attempts to prove (among other things) that the population of Armenians was less than that which could have supported the European and Armenian figures of approximately 1.5 million killed by the (Young) Turks in 1915, and that Armenians were clear minorities in the Eastern provinces. In my opinion he certainly fails on the first count as he himself shows in his book (though he weasel words justifications for concluding the opposite). On the subject of Armenian relative population in the east perhaps he is correct – but to a much lesser extent than he proposes. Certainly, the deliberate policy (denied by McCarthy – but well known and accepted elsewhere) of settlement of Muslims into Armenian areas for the preceding (at least) two centuries and the outflow of Armenians during the 19th and early part of the 20th century (to escape harassment, excessive taxation and massacre etc.) did result in a shift from majority Armenian population to majority Muslim (note Muslim - however not Turkish...) – further massaged by gerrymandering of political boundaries. In fact however, there were still significant populations of Armenians in many of these areas – particularly in certain districts where there still were Armenian majorities (and in fact Armenians outnumbered ethnic Turks in most all areas of Eastern Anatolia).
McCarthy seeks to discredit Armenian Patriarchal population counts (as I in fact did – before I read McCarthy). In fact, IMO he really shows them to be much more accurate than the proven inaccurate and obviously biased Ottoman “counts”. It is particularly interesting that where the Ottoman and Armenian figures match is in areas of western Anatolia where Ottoman control was the greatest (and there was no issue of a potentially independent Armenian State). Here the Ottoman figures were deemed most accurate (even by McCarthy) and lo and behold – they nearly match the Armenian figures. Yet in the east – where the Ottomans were concerned about Armenian nationalism – their figures for numbers of Armenians are much much lower than the Armenian counts. First -–there should be no reason why the Armenian figures would be accurate in the west – but inaccurate in the east. Secondly, there are plenty of reasons (deliberately and inadvertently) why the Ottomans would severely undercount Armenians in the East. The first is of course political motivation – to attempt to discredit the idea of an Armenian nation-state. Secondly, the Ottomans had very tentative administrative control over a large percentage of Armenians in the East – who for all practical purposes ruled themselves under the millet system. These Armenians did not want to be counted (taxation and other reasons) and certainly did not cooperate in such. Additionally, the Turks obviously did not wish to count many of these Armenians as well (most of these they had some difficulty collecting taxes from anyway). It is always difficult to count people in developing areas anyway – where communications and infrastructure are not good – and I am sure that the Kurds and other nomads in the east were also significantly undercounted – but combined with the political motivation (which McCarthy amusingly discounts) it is apparent what occurred. Particularly since the Armenian derived figures were shown to be reasonably accurate in western Anatolia. And – the European intelligence agencies developed figures which much more closely matched the Armenian figures – not the Ottoman ones – obviously they wanted the most accurate counts and they either independently derived figures which agreed with the Armenian counts – or they used the Armenian figures – trusting them more than the Ottoman ones – either way the conclusion is the same – and McCarthy’s attempts to prove otherwise ring hollow and biased. The following consists of excerpts from “Muslims and Minorities” which illustrate some of the points I have made above:
“Population registers were the basis of all Ottoman government population statistics.”
“Like all population records, the ottoman population registers contained errors.”
“The Ottoman population records on the Anatolian Muslims became reliable (my note: even this is very much disputed – in fact it is self-contradicted by McCarthy IMO – though perhaps they were improved over prior – though certainly not reliable) long before those on minority populations.”
“The overwhelming inaccuracy of the Van data leaves no choice but to draw the correction factors used on the Van population from other sources.”
“Since no major Armenian population survived in Anatolia at the time of the modern Turkish Republican censuses it is impossible to compare the (Armenian) Patriarch’s data to more modern material.”
“There were indeed Armenians killed in the troubles of 1895-6, but not even the most exaggerated count of mortality records such a loss for the entire Ottoman Empire.” (McCarthy discounts Armenians losses in earlier massacres and claims Armenian exaggeration of both their own numbers & losses – without providing any proof or giving us any real evidence as to what sources he bases such a statement on)
“It is doubtful that the Ottomans, suspicious as they were of Armenian community action and publicity, would have allowed the type of massive collection and checking of data necessary for such records.” (Is McCarthy denying the existence of the millet system and minority communities being allowed to take care of their own? Additionally, they often were in no position to deny such – and as the collection was done as part of the church functions this was left alone by the Ottomans – again McCarthy’s “conclusions” are self serving, biased, and IMO wrong)
“Detailed census and registration figures in a statistically underdeveloped area like the Ottoman Empire are usually somewhat incorrect due to undercounting, but the undercounts can be corrected.” (This later point is the basis for McCarthy’s figures – but by his own admission you can see that counts of Armenians were more likely to be purposefully or otherwise undercounted as compared to Muslims – and were much less accurate because the degree of cooperation with and control by the Ottoman authorities was much less in Armenian areas. Additionally, as McCarthy points out there were many factors for less accurately counting Armenians versus Muslims – such as the need to more accurately count Muslims for conscription purposes etc. – in any event it is obviously very difficult to accurately count people in areas not under full administrative control – particularly when those people may not have been interested in being counted. On top of this is also clear that the Ottomans were very interested in portraying the Armenians as a more minor population than they actually were to discourage Western advocacy of an independent Armenia etc. – thus they had much incentive to report less Armenians then there actually were – this motivation cannot be discounted – though McCarthy does just this thing – taking two pages to justify why, in his mind, the Ottomans were not deliberately undercounting Armenians – these justifications ring hollow IMO).
“…in the two eastern vilayets the level of undercounting of Armenians was considerably larger than that of Muslims.”
Certain European accounts….”give…accurate indications of the large number of Armenian churches and schools in an area such as Siirt Sancagi of Bitlis Vilayeti that officially (in the Ottoman counts) listed few Armenians, thus providing an indication that the Armenian population of Siirt was undercounted.”
“A central assumption made below is that the misreporting of age and sex for Armenians was approximately the same as for Muslims. Until extensive archival research has been done, there is no way to prove this assumption.”
Sivas Vilyet/eastern Anatolia: “This was the area of poorest Ottoman administrative control, poorest statistics, and the greatest divergence between Ottoman and Patriarchate statistics. Sivas, while still fairly far removed from the seat of central Ottoman government, was clearly much advanced statistically over Bitlis, Van, or Mamuretulaziz.” “Ottoman figures for the Armenian population of Sivas were very close to those of the Armenian Patriarchate.” (Is he contradicting himself here?)
“..the Ottoman records, as their collection methods improved, counted Armenians as a greater proportion of the Manuretullaziz population than they had previously done.” (an indication of much room for error as we can see) “The Ottoman figures for Marmuretulaziz were considerably better than those of Cuinet, who made large-scale errors in enumerating the Armenian population of Harput Kazasi.”
“The decision to use Ottoman statistics is based solely on their proven relative reliability, and on the proven unreliability of the Armenian statistics.” (note: his basis for this is his own conjecture regarding how each set was compiled – no direct evidence) “Intellectually and statistically, this is not a completely satisfactory approach…” (At least he admits this – though it is buried in the book which is presented with a falsely authoritative air)
“The undercounting of the Armenian population in Diyarbakir Vilayeti was approximately the same as that of the Diyarbakir Muslim population.” (how can he claim this with any certainty at all? In fact there are many many reasons for a greater undercount of Armenians, which he discounts on principle without knowledge).
Ezurum Vilayeti: McCarthy discusses that the Ottomans paid closer attention to this area then others and that Armenians had been a much larger percentage but had “migrated” to Russian territory. He indicates that his “correction factor” that he uses may not be valid here – though he uses it just the same.
Van and Bitlis: “..neither one of which has completely reliable Ottoman published statistics.” (though he claims Armenian figures are overcounts – how could this be known if the Ottoman figures are unreliable?) He does add however that “there will always be a great deal of uncertainty about these populations.” (which he adds to by publishing unverifiable, concocted figures and discounting the Armenian counts). “The Ottomans knew that they had undercounted (minority) populations in Van and Bitlis, and commented on this fact in published and secret documents. …the registers of Muslims were improved. The registers of non-Muslim population were not significantly improved.” “..Armenian men were also significantly undercounted.” (in addition to women and children – which McCarthy admits was a normal practice).
“There were more Armenian schoolchildren listed in Siirt than in Bitlis or Mus even though the (Ottoman census) listed much larger (Armenian) population numbers in the latter two sanaks.” (Indications of unpredictable unreliability of the figures) (McCarthy then justifies the use of [his own low] estimates in these cases as he admits the reported figures are unreliable). (note McCarthy admits that official counts indicated a substantially less Armenian presence in these areas than what was thought to exist – so why would the counts elsewhere be any different? I am sure that McCarthy only admits this in this case because the Ottoman figures can be proven wrong – where they cannot be proven wrong he uses them – just because they cannot be proven wrong does not mean that they are correct – and every indication is of greater undercount of Armenians then of Muslims – in every instance where this can be proven it is the case – thus is it not likely elsewhere as well?).
“..other indicators also point to an undercount of Armenians and other Christians in Van.” (his rational for this – emigration of Armenian males ??? [what about the rest of the family?] is almost comical – the hoops he jumps through to justify why these undercounts and discrepancies are an aberration). “The military exemption tax paid by Van non-Muslims in 1313 indicates a greater non-Muslim population than does the population records. Ottoman statistics indicate that the Ottoman government was gradually improving its counts of Muslims in Van Vilayeti, but that the enumeration of non-Muslims was improving little. The Muslim population, as recorded, would have had to have been increasing three times faster than the non-Muslim population. This was impossible (note: perhaps less were being massacred). “..the assumption made when beginning to analyze the Ottoman data on Van Armenians was that the undercounting of Van Armenians would be of the same magnitude as the undercounting of Muslims. By every experiment and test, this did not prove to be the case.” (ah ha!)
Cilicia: “While by no means completely accurate, the Ottoman figures can be corrected.” (can they really? Who really is to know? By whose methodology? Once you “correct” them are they still “census” figures? Or your own estimates – just as the figures you critique – in fact the Armenians Patriarchal figures may not be “estimates” but actual counts following birth, baptismal and death records kept by the church)
“It was in the provinces closest to central authority that Ottoman and Armenian estimates of Armenian population most closely coincided.” (In fact he shows that the Ottoman figures were greater than the Armenian ones in some cases. Does this not seem to indicate that the accuracy of the Armenian figures in the Eastern provinces may have been more accurate than the government figures? If they were shown to be accurate in the areas where government control and administration was strong – and counts were good – could it not be possible that the Armenian figures were good in the areas of less government control – and that there were far more Armenians than the Ottomans were admitting to? This seems to make much sense does it not?)
“There can be no doubt that Ottoman figures on Armenian population in eastern Anatolia were somewhat mistaken, as were Ottoman statistics on eastern Anatolian Muslim population. It must be asserted, however, that…Ottoman statistics did not selectively discriminate against Armenians.” (It seems to me that if we look at these discrepancies closely they did just that – thus McCarthy’s methodology, by his own - down in the fine print – admission, is flawed and biased against the Armenians in his totals. It seems that European intelligence sources agree more with the Armenian figures – why would this be so if the Ottoman figures are the more accurate as McCarthy contends?)
McCarthy seeks to discredit Armenian Patriarchal population counts (as I in fact did – before I read McCarthy). In fact, IMO he really shows them to be much more accurate than the proven inaccurate and obviously biased Ottoman “counts”. It is particularly interesting that where the Ottoman and Armenian figures match is in areas of western Anatolia where Ottoman control was the greatest (and there was no issue of a potentially independent Armenian State). Here the Ottoman figures were deemed most accurate (even by McCarthy) and lo and behold – they nearly match the Armenian figures. Yet in the east – where the Ottomans were concerned about Armenian nationalism – their figures for numbers of Armenians are much much lower than the Armenian counts. First -–there should be no reason why the Armenian figures would be accurate in the west – but inaccurate in the east. Secondly, there are plenty of reasons (deliberately and inadvertently) why the Ottomans would severely undercount Armenians in the East. The first is of course political motivation – to attempt to discredit the idea of an Armenian nation-state. Secondly, the Ottomans had very tentative administrative control over a large percentage of Armenians in the East – who for all practical purposes ruled themselves under the millet system. These Armenians did not want to be counted (taxation and other reasons) and certainly did not cooperate in such. Additionally, the Turks obviously did not wish to count many of these Armenians as well (most of these they had some difficulty collecting taxes from anyway). It is always difficult to count people in developing areas anyway – where communications and infrastructure are not good – and I am sure that the Kurds and other nomads in the east were also significantly undercounted – but combined with the political motivation (which McCarthy amusingly discounts) it is apparent what occurred. Particularly since the Armenian derived figures were shown to be reasonably accurate in western Anatolia. And – the European intelligence agencies developed figures which much more closely matched the Armenian figures – not the Ottoman ones – obviously they wanted the most accurate counts and they either independently derived figures which agreed with the Armenian counts – or they used the Armenian figures – trusting them more than the Ottoman ones – either way the conclusion is the same – and McCarthy’s attempts to prove otherwise ring hollow and biased. The following consists of excerpts from “Muslims and Minorities” which illustrate some of the points I have made above:
“Population registers were the basis of all Ottoman government population statistics.”
“Like all population records, the ottoman population registers contained errors.”
“The Ottoman population records on the Anatolian Muslims became reliable (my note: even this is very much disputed – in fact it is self-contradicted by McCarthy IMO – though perhaps they were improved over prior – though certainly not reliable) long before those on minority populations.”
“The overwhelming inaccuracy of the Van data leaves no choice but to draw the correction factors used on the Van population from other sources.”
“Since no major Armenian population survived in Anatolia at the time of the modern Turkish Republican censuses it is impossible to compare the (Armenian) Patriarch’s data to more modern material.”
“There were indeed Armenians killed in the troubles of 1895-6, but not even the most exaggerated count of mortality records such a loss for the entire Ottoman Empire.” (McCarthy discounts Armenians losses in earlier massacres and claims Armenian exaggeration of both their own numbers & losses – without providing any proof or giving us any real evidence as to what sources he bases such a statement on)
“It is doubtful that the Ottomans, suspicious as they were of Armenian community action and publicity, would have allowed the type of massive collection and checking of data necessary for such records.” (Is McCarthy denying the existence of the millet system and minority communities being allowed to take care of their own? Additionally, they often were in no position to deny such – and as the collection was done as part of the church functions this was left alone by the Ottomans – again McCarthy’s “conclusions” are self serving, biased, and IMO wrong)
“Detailed census and registration figures in a statistically underdeveloped area like the Ottoman Empire are usually somewhat incorrect due to undercounting, but the undercounts can be corrected.” (This later point is the basis for McCarthy’s figures – but by his own admission you can see that counts of Armenians were more likely to be purposefully or otherwise undercounted as compared to Muslims – and were much less accurate because the degree of cooperation with and control by the Ottoman authorities was much less in Armenian areas. Additionally, as McCarthy points out there were many factors for less accurately counting Armenians versus Muslims – such as the need to more accurately count Muslims for conscription purposes etc. – in any event it is obviously very difficult to accurately count people in areas not under full administrative control – particularly when those people may not have been interested in being counted. On top of this is also clear that the Ottomans were very interested in portraying the Armenians as a more minor population than they actually were to discourage Western advocacy of an independent Armenia etc. – thus they had much incentive to report less Armenians then there actually were – this motivation cannot be discounted – though McCarthy does just this thing – taking two pages to justify why, in his mind, the Ottomans were not deliberately undercounting Armenians – these justifications ring hollow IMO).
“…in the two eastern vilayets the level of undercounting of Armenians was considerably larger than that of Muslims.”
Certain European accounts….”give…accurate indications of the large number of Armenian churches and schools in an area such as Siirt Sancagi of Bitlis Vilayeti that officially (in the Ottoman counts) listed few Armenians, thus providing an indication that the Armenian population of Siirt was undercounted.”
“A central assumption made below is that the misreporting of age and sex for Armenians was approximately the same as for Muslims. Until extensive archival research has been done, there is no way to prove this assumption.”
Sivas Vilyet/eastern Anatolia: “This was the area of poorest Ottoman administrative control, poorest statistics, and the greatest divergence between Ottoman and Patriarchate statistics. Sivas, while still fairly far removed from the seat of central Ottoman government, was clearly much advanced statistically over Bitlis, Van, or Mamuretulaziz.” “Ottoman figures for the Armenian population of Sivas were very close to those of the Armenian Patriarchate.” (Is he contradicting himself here?)
“..the Ottoman records, as their collection methods improved, counted Armenians as a greater proportion of the Manuretullaziz population than they had previously done.” (an indication of much room for error as we can see) “The Ottoman figures for Marmuretulaziz were considerably better than those of Cuinet, who made large-scale errors in enumerating the Armenian population of Harput Kazasi.”
“The decision to use Ottoman statistics is based solely on their proven relative reliability, and on the proven unreliability of the Armenian statistics.” (note: his basis for this is his own conjecture regarding how each set was compiled – no direct evidence) “Intellectually and statistically, this is not a completely satisfactory approach…” (At least he admits this – though it is buried in the book which is presented with a falsely authoritative air)
“The undercounting of the Armenian population in Diyarbakir Vilayeti was approximately the same as that of the Diyarbakir Muslim population.” (how can he claim this with any certainty at all? In fact there are many many reasons for a greater undercount of Armenians, which he discounts on principle without knowledge).
Ezurum Vilayeti: McCarthy discusses that the Ottomans paid closer attention to this area then others and that Armenians had been a much larger percentage but had “migrated” to Russian territory. He indicates that his “correction factor” that he uses may not be valid here – though he uses it just the same.
Van and Bitlis: “..neither one of which has completely reliable Ottoman published statistics.” (though he claims Armenian figures are overcounts – how could this be known if the Ottoman figures are unreliable?) He does add however that “there will always be a great deal of uncertainty about these populations.” (which he adds to by publishing unverifiable, concocted figures and discounting the Armenian counts). “The Ottomans knew that they had undercounted (minority) populations in Van and Bitlis, and commented on this fact in published and secret documents. …the registers of Muslims were improved. The registers of non-Muslim population were not significantly improved.” “..Armenian men were also significantly undercounted.” (in addition to women and children – which McCarthy admits was a normal practice).
“There were more Armenian schoolchildren listed in Siirt than in Bitlis or Mus even though the (Ottoman census) listed much larger (Armenian) population numbers in the latter two sanaks.” (Indications of unpredictable unreliability of the figures) (McCarthy then justifies the use of [his own low] estimates in these cases as he admits the reported figures are unreliable). (note McCarthy admits that official counts indicated a substantially less Armenian presence in these areas than what was thought to exist – so why would the counts elsewhere be any different? I am sure that McCarthy only admits this in this case because the Ottoman figures can be proven wrong – where they cannot be proven wrong he uses them – just because they cannot be proven wrong does not mean that they are correct – and every indication is of greater undercount of Armenians then of Muslims – in every instance where this can be proven it is the case – thus is it not likely elsewhere as well?).
“..other indicators also point to an undercount of Armenians and other Christians in Van.” (his rational for this – emigration of Armenian males ??? [what about the rest of the family?] is almost comical – the hoops he jumps through to justify why these undercounts and discrepancies are an aberration). “The military exemption tax paid by Van non-Muslims in 1313 indicates a greater non-Muslim population than does the population records. Ottoman statistics indicate that the Ottoman government was gradually improving its counts of Muslims in Van Vilayeti, but that the enumeration of non-Muslims was improving little. The Muslim population, as recorded, would have had to have been increasing three times faster than the non-Muslim population. This was impossible (note: perhaps less were being massacred). “..the assumption made when beginning to analyze the Ottoman data on Van Armenians was that the undercounting of Van Armenians would be of the same magnitude as the undercounting of Muslims. By every experiment and test, this did not prove to be the case.” (ah ha!)
Cilicia: “While by no means completely accurate, the Ottoman figures can be corrected.” (can they really? Who really is to know? By whose methodology? Once you “correct” them are they still “census” figures? Or your own estimates – just as the figures you critique – in fact the Armenians Patriarchal figures may not be “estimates” but actual counts following birth, baptismal and death records kept by the church)
“It was in the provinces closest to central authority that Ottoman and Armenian estimates of Armenian population most closely coincided.” (In fact he shows that the Ottoman figures were greater than the Armenian ones in some cases. Does this not seem to indicate that the accuracy of the Armenian figures in the Eastern provinces may have been more accurate than the government figures? If they were shown to be accurate in the areas where government control and administration was strong – and counts were good – could it not be possible that the Armenian figures were good in the areas of less government control – and that there were far more Armenians than the Ottomans were admitting to? This seems to make much sense does it not?)
“There can be no doubt that Ottoman figures on Armenian population in eastern Anatolia were somewhat mistaken, as were Ottoman statistics on eastern Anatolian Muslim population. It must be asserted, however, that…Ottoman statistics did not selectively discriminate against Armenians.” (It seems to me that if we look at these discrepancies closely they did just that – thus McCarthy’s methodology, by his own - down in the fine print – admission, is flawed and biased against the Armenians in his totals. It seems that European intelligence sources agree more with the Armenian figures – why would this be so if the Ottoman figures are the more accurate as McCarthy contends?)
Comment