• 1 These areas were assigned to counties known as sancaks in Turkish. Counties contained districts, ca (...)
• 2 J. Bryce, 1896, p. 471.

1The Sasun massacre of 1894 was the first in a series of incidents of mass violence against Armenians in the Ottoman Empire. The massacre marked the beginning of the widespread and more systematic Hamidian massacres that swept across most Armenian-populated areas1 over the course of the following three years, claiming the lives of tens of thousands of civilians in the process, with reoccurrences of violence in 1904 during the second incident in Sasun and in 1909 during urban riots and pogroms in Adana. The massacre was seen by many contemporaries as marking the launch of Sultan Abdülhamid II’s new, and more systematic, policy directed against his Armenian subjects, which British historian and liberal politician James Bryce chose to call a “policy of deliberate oppression.’2

• 3 See, for example, B. Adjemian and M. Nichanian, 2018; R.G. Suny, 2018; Idem, 2001.
• 4 BOA, Y.EE. 97/53 (24 August 1894), Directive to the Sultan’s aides-de-camp.
• 5 Report by Mr. Block, Inclosure in no. 536, 25 December 1894.

2It is not the purpose of this study to look critically into the question of the Sultan’s responsibility for the instigation and upsurge of violence, a historiographical debate that remains open among the specialists of this period.3 However, because the sending of government troops resulted in a large number of deaths among the civilian population, it should be mentioned that the Sultan gave his aides-de-camp full freedom (in Turkish, mezuniyet-i kamile) of action so as to ensure that the “rebels” were destroyed.4 British embassy correspondence containing a confidential report prepared by Adam Block, chief dragoman and oriental second secretary, confirmed that “there seems no reasonable doubt that the general opinion in well-informed circles in Constantinople that the orders emanated from Yıldız is correct”. The report specifically noted that “no provincial governor would dare to commit such acts without instructions”, and had vali Hassan Tahsin, the governor of Bitlis, done so, “he would have been instantly dismissed; on the contrary, the Sultan maintained him in his post”. Block added that in a heated exchange over the responsibility for the massacre, apparently with an official from the Yıldız Palace, Zeki Kolaç, better known as Zeki Paşa, the Commander of the Ottoman Fourth Army Corps who was widely blamed for the massacre, had said: “I did not approve of these proceedings, but don’t let them press me or make me speak out, for I still have the Sultan’s orders by me written by [the Sultan’s chief clerk Ahmed] Süreyya.”5

• 6 J. Verheij, 1999, pp. 81-91.
• 7 M. Polatel, 2016, p. 180.
• 8 E. Gölbaşı, 2018, p. 5.

3If the violence in Sasun, as far as can be judged from these contemporary reports, was administered from the top down, the violence during the Hamidian massacres that ensued was a bottom-up Muslim backlash against Armenians, triggered by widespread anxieties that hierarchies privileging Muslims would be curtailed if reforms in the eastern provinces were implemented.6 If in Sasun the use of “extraordinary terror” (in Turkish, dehşet-i fevkalade) was ordered directly by the Porte7 and most of the killings were accomplished with the involvement of regular army units, the violence during the Hamidian massacres was perpetrated by soldiers, militias, tribesmen, and ordinary Muslims who, according to the contention advanced by the Ottoman authorities, “simply responded with violence” to the acts of provocation and aggression on the part of Armenians closely connected with revolutionary groups in order to eliminate the threat or punish the provoking party.8

• 9 See, for example, H. Vardanyan, 1964.
• 10 A.J. Toynbee, 1915, pp. 73-74.

4Evidence for the emergence of a “policy of deliberate oppression” is found in the period after the 1878 Treaty of Berlin. The European Powers’ intervention to secure reforms in the six eastern Ottoman provinces, where Armenians formed a particularly large portion of the population, became a cause célèbre known as the “Armenian Question”. According to numerous commentators, the Ottoman state, alarmed by this dramatic development, intensified oppression by raising taxes and arousing resentment against Armenians among the Kurds.9 The Armenians of Sasun who, as related by young English historian Arnold Toynbee during World War I, had “led a semi-independent, almost self-sufficing existence for centuries to the chagrin of the Ottoman government and the envy of [their] less prosperous neighbours the Kurds”10, became particular targets.

• 11 For the Armenian interpretation, see authors mentioned in Table B. For the Turkish interpretation s (...)
• 12 Hallward to Currie, Inclosure 2 in no. 339, 06 November 1894.

5Well over a century after the Sasun massacre, uncertainty still surrounds the number of its victims. In part, this can be attributed to differing interpretations of the nature of the incident. Most Armenian authors describe it as an “uprising” or “heroic battle” and, as heralds of the victimised side, over-report the numbers of killed. Most official Turkish historians and apologists for the “Turkish” view of history characterise it as a “revolt” or “rebellion” and, as mouthpieces of the perpetrators of violence, underestimate the death toll.11 In reality, the incident was a social protest which deteriorated into state violence. Hardly can it be qualified as an “uprising” or “battle”, because, for resistance fighters or combatants, the peasants were very poorly armed, bitterly divided and too few in number. Nor was it a “revolt” or “rebellion”, because no evidence was found showing that the protesters’ primary motivation was to overthrow the existing political order. As attested by Cecil Marsham Hallward, the British Vice-Consul in Van who was dispatched to Sasun to investigate allegations of massacre: “There was no insurrection, as was reported in Constantinople; the villagers simply took up arms to defend themselves against the Kurds».12

• 13 Shipley to Currie, Inclosure in no. 160, 27 July 1895.
• 14 Report of the Consular Delegates attached to the Commission appointed to inquire into the Events at (...)

6The root cause of the uncertainty about the number of casualties lies in the fact that the Commission Appointed to Inquire into the Events at Sasun, which was entrusted with the task of investigating the Sasun “affair”, offered very few figures. Consular delegates from Britain, Russia and France, who served as observers on the Commission comprised of Ottoman officials, stated in the Joint Report that the number of deaths could not be determined. British Commissioner Hammond Smith Shipley, who was additionally tasked with verifying the reports from British consular officers in Muş, Erzurum and Van, thus mandated with the responsibility of monitoring the circumstances surrounding the massacre, admitted that one of the points on which his colleagues were unable to arrive at any conclusion was the number of victims.13 By and large, consular delegates were only able to confirm several localities that were attacked and to conclude that “Armenians – men, women, and children – were during the events the object of repeated pursuit on the part of soldiers, Kurds and gendarmes, who wounded or killed, without distinction of age and sex, all who fell into their hands”.14

• 15 Hallward to Currie, Inclosure 2 in no. 339, 06 November 1894.
• 16 Ibid., Memorandum from Currie, Inclosure 1 in no. 294, 01 November 1894.

7Another reason for the uncertainty about fatalities lies in the fact that the Ottoman members of the Commission had thrown up all sorts of obstacles to their European colleagues’ efforts to discover the truth. Hindrances included delaying the start of the Commission’s work until the end of January 1895 (almost six months after the massacre had taken place), preventing Armenian witnesses from appearing before the Commission, and creating difficulties for commissioners as they sought reliable information, such as when, in November 1894, Vice-Consul C.M. Hallward was so closely watched by police agents that the natives were afraid to be seen speaking to him.15 To make matters worse, Ottoman officials at the local level often showed reluctance to co-operate with the European delegates in undertaking field visits to the areas of massacre. In one instance, Hallward was prevented from visiting Sasun on the grounds that he might carry a cholera infection to the troops and that his presence there could cause an insurrection, even though he had no intention of visiting the troops and the danger of an “insurrection” was, after the massacre, utterly illogical and unrealistic.16

• 17 Shipley to Graves, Inclosure 1 in no. 31, 18 June 1895.

8The real number of deaths will probably never be known. Commissioner H.S. Shipley admitted that the “fact of the massacres is indisputable, though we shall never be able to estimate the number killed”.17 Whatever the veracity of these words, until now there has been no comprehensive inquiry into the number of victims intent on marking out the scope of statistics and making estimates based on available information. From a methodological perspective, this work is an exercise in the integration of quantitative data into qualitative analysis. It is therefore not intended as a case study of violence but as an attempt to recalculate fatalities resulting from violence in order to establish an argument of significance. By significance, I mean that it is not enough just to count the number of deaths during a mass atrocity event. An atrocity is significant when it is so “large” as to warrant not only the label “massacre” but also a historic turning point.

• 18 M. Sandelowski, 2001, p. 230.

9Critics may argue that fatality figures are a limited, if important, data point that little affects the ability of historians of violence or genocide scholars to study patterns of violence or understand atrocity endings. Yet, as Margarete Sandelowski aptly points out, “numbers are integral to qualitative research, as meaning depends, in part, on number”.18 Numbers come into play when there is a need to document what is known about a mass atrocity event; to verify and, when appropriate, to validate a researcher’s interpretations and conclusions about atrocity endings; and to address the theoretical implications of violence for re-thinking national and regional histories. That being said, an interim question that requires some elaboration at this point is why fatality figures matter.

• 19 J. Maxwell, 2010, p. 478.
• 20 S. Peláez, 2014, p. 42.
• 21 E. Towner, 2011, p. 635.

10The use of numerical data in qualitative research studies has been controversial. On the one hand, numbers, as Joseph Maxwell contends, are “a legitimate and important sort of data for qualitative researchers as they give precision to statements about the frequency, amount, or typicality of particular phenomena”. On the other hand, however, they “do this at the cost of stripping away everything but the quantitative information and are thus necessarily complementary to qualitative information rather than substituting for it”.19 One of the major concerns about numerical narratives of atrocities is that fatality figures do not properly quantify suffering. Sol Peláez claims that in the “accounts of horror, […] violence appears as abstract violence through the distance of geography and of the numbers».20 Another concern is that “fatal flaws exist within the numbers themselves” because, according to Emil Towner, “data are limited in scope (that is, time and location) and are ultimately based on educated guesses, rather than precise counts”.21

• 22 M. Newcomb, 2010, p. 181.
• 23 B. Conley, 2017, p. 15.

11Despite such concerns, numerical narratives provide important perspective on the size of an atrocity. Moreover, they represent a faith in the ability of numbers to impose a sense of objectivity and believability in the instances when words fail to adequately convey the horror of an atrocity. It has been argued that a reliance on statistical measurement appears in the definition of genocide which places emphasis on the “intent to destroy, in whole or in part, a national, ethnical, racial or religious group”. While the “intent” aspect caused considerable trouble for criminal courts, the “in whole or in part” verbiage seems to necessitate the use of statistics.22 As Bridget Conley rightly points out, “the evidence base for civilian fatality figures that defines an ending can fundamentally alter the narrative about a mass atrocity event”.23 In addition, the use of statistics as the main narrative of an atrocity can help undermine counterarguments put forth by deniers of atrocity.

• 24 M. Newcomb, 2010, p. 178.
• 25 S. Peláez, 2014, p. 39.

12Statistics, in the view of Matthew J. Newcomb, have the potential to “function as informative images, extended experiences or symbols of missing context that promote at least more information gathering, and occasionally service or testimony”.24 Peláez goes further, noting that “the counting – in numbers and in writing – is the only way […] violence becomes visible”.25 Statistics, to put it another way, are used as arguments of size or scale which are often necessary to establish the significance of an atrocity. In this study, statistics were deemed necessary to establish the significance of the massacre that took place in 1894 in the mountains of Sasun and launched an appalling pattern of violent acts, eventually leading to the Armenian Genocide in 1915.

• 26 Catalogue codes appearing in the footnotes, passim, refer to the following. BOA: Cumhurbaşkanlığı D (...)

13The available data were collected, inter alia, from the Annex to the Joint Report, abbreviated to Anx to JR in Table D and the Appendix “Lists of affected villages”; Archives of the Armenian Patriarchate of Constantinople held at the AGBU Nubar Library (Paris) and used in Raymond Kévorkian’s 2001 study; the 1902 Patriarchate census data collected by Gegham Ter-Karapetian; British Foreign Office documents containing correspondence from British diplomatic and consular offices in the Ottoman Empire, which in the original records appear under alphanumeric codes starting with the letters “FO” but in this paper are fully cited in the footnotes and abbreviated to BD (British Dispatches) in Table D and the Appendix; contemporary Armenian periodicals Ardzagank, Droshak, Moorch and Mshak; the Yıldız Collection on Sasun held at the Ottoman Archives; and Ottoman correspondence held at the Presidential Directorate of State Archives of Turkey (Türkiye Cumhuriyeti Cumhurbaşkanlığı Devlet Arşivleri Başkanlığı).26

• 27 E.A. Brayley Hodgetts, 1896, p. 60.
• 28 See “The Villages of Sasun Destroyed by the Turks in 1894”, in E.A. Brayley Hodgetts, 1896, pp. 111 (...)

14Among contemporary sources, the one that came closest to addressing the subject was Moorch, whose August 1895 no. 8 edition contained a list of villages destroyed during the massacre. Unfortunately, the list cannot be viewed as exhaustive because it only supplied the number of households per affected village. Even less exhaustive is a narrative record of travels through Turkey and the Caucasus published by Edward Arthur Brayley Hodgetts, The Daily Graphic’s correspondent for Armenia, which appeared in the British press in the first half of 1895 along with a series of news stories covering the incident. Evidently, the number of fatalities could not have come under Brayley Hodgetts’ observation because he set off on his journey nearly five months after the incident and the evidence he received from a group of survivors from Sasun, whom he interviewed in Tiflis, was, in his view, “not so trustworthy”.27 In addition, the village list that the author drew up could not be used for computations because it was based on data drawn from peasants or from hearsay and only contained the names of the localities that were destroyed.28

• 29 For an example of analysis focused on missionary work in Sasun, see O. Miller, 2015.
• 30 F.D. Greene, 1895, pp. 11-12.
• 31 Ibid., p. 14.
• 32 New York Times, 16 December 1894.
• 33 This Report was included in a correspondence sent to the British Foreign Office. For Knapp’s author (...)
• 34 Report of Sassoon Affairs, by a resident at Bitlis, Inclosure in no. 78, 18 January 1895, p. 80.

15Offering an important perspective on the size of atrocity in Sasun are the missionary reports published as narratives, the most notable of them belonging to Frederick Davis Greene and George Perkins Knapp of the American Board of Commissioners for Foreign Missions (ABCFM).29 The focus they placed on communal suffering and loss provided a depth of understanding and compassion. However, the fatality data they reported in the first months after the violence were replete with thousands, even tens of thousands, of Armenian dead. Writing in October, F.D. Greene reported that “six thousand is a low figure – it is probably nearer ten”.30 The following month, he wrote that “[sober] estimates vary from 3,000 to 8,000 for the number of persons massacred” and that “[w]ild estimates place the number as high as 20,000 to 25,000”.31 In December, ABCFM missionaries stated in the American press that “thousands were slain, some state 6,000 others 10,000”.32 Then in January 1895, in the “Report of Sassoon Affairs, by a resident at Bitlis”, which is believed to have been prepared by G.P. Knapp,33 the author wrote: “I think the truth is not far from 10,000; 6,000 I consider a safe minimum estimate and should not feel like contradicting one who asserted 16,000 had perished”.34

• 35 L. Qosyan, 2014.
• 36 N. Hakobian, 2016.

17In the recent specialised literature on Sasun, the data extracted from Moorch no. 8 (1895) were examined and compiled into a consolidated text by Lilit Qosyan. Regrettably, however, her paper does not extend beyond providing an overview of data and a reprint of the list.35 Basing his work on a richer dataset, Narek Hakobian endeavoured to set forth the archive materials held at the Holy Synod at Etchmiadzin within a single text in an effort to provide a comparative overview of the reported fatalities. However, despite its documented research, the work falls short of our expectations for fatality data analysis. It is needless to look for an estimation of fatalities in the paper because it was not the task to which the author set himself in his study.36

• 37 J. Verheij, 1998, p. 242. Idem, 1999, p. 83.
• 38 R. Kévorkian, 2001, p. 50.
• 39 R. Morris, 2001.
• 40 M. Polatel, 2016, pp. 179-180.

18In two impressive works on the Hamidian massacres, Jelle Verheij provided several estimates from Ottoman military authorities and foreign observers regarding the number of casualties in Sasun.37 However, the author attempted no in-depth analysis of the number of victims because this subject lay outside his area of inquiry. Raymond Kévorkian, on the other hand, did suggest in a village-by-village presentation of the demographic situation in Sasun before 1894 and in 191438 that the number of fatalities was probably closer to one or two thousand. However, this figure was produced using a rough, yet not unreasonable, assumption on the basis of a single source, the Archives of the Patriarchate, and not the statistics derived from multiple datasets. Similarly, through the use of a single dataset, Rebecca Morris provided limited demographic information on the sources used in the Joint Report, as well as on the affected settlements involved. The author’s meticulous examination allowed her to reaffirm that no conclusions could be drawn from the Joint Report regarding the number of victims.39 The number of deaths was touched upon, albeit only briefly, in a paper by Mehmet Polatel. Yet, because his work is a case study of violence, the author chose to place greater emphasis on the analysis of the massacre in the context of the Porte’s handling of the unrest, the practices employed by local civilian and military officials, and the workings of the Commission.40

• 41 J. McCarthy et al., 2014, p. 188.
• 42 N. Gülmez, 2006, p. 731.

20In a study that stands out as the most noteworthy among works minimising the casualties, Justin McCarthy, Ömer Turan and Cemalettin Taşkiran put the number of killed Armenian defenders, whom they identified as “dead fighters”, at unbelievably low figure of 97. McCarthy and his Turkish co-writers noted that judging by what they could verify from the Commission testimony, the number of those who, in their carefully calibrated words, “died as an indirect result of fighting – the old, the young, and women», was at most 150.41 The examination of the Joint Report reveals that McCarthy and his co-writers arrived at this number by way of deliberate misuse of the incomplete figures drawn up by the European delegates. Apparently contented, they chose not to corroborate the figures with further evidence. Another example of the blatant distortion of findings of the Joint Report to justify killings by placing the blame for the “rebellion” squarely on the Armenians is the article by Nurettin Gülmez. To illustrate the number of deaths, the author chose not to avail himself of primary sources but quote incorrect figures from Russian General Staff Colonel Vladimir Mayewski.42 Along with arbitrary interpretation of facts, the reliance on unverified data has fundamentally undermined the author’s effort.

• 43 M. van Bruinessen, 1992, pp. 133-202.

21The period from the early to mid-1890s was marked by an upsurge of violence against the Armenian peasantry in the six eastern provinces of the Ottoman Empire. The removal of local Kurdish beys, or princes, the result of successful military expeditions carried out by the Ottoman state, and the failure of the Porte to replace Kurdish chiefdoms with effective local governance created a power vacuum in which Kurdish ashirats, or tribal groups, competed for influence and wealth. Armenian peasants, who were among the principal victims of this competition, had to make incremental extortion payments, often enforced by violence and intimidation.43

• 44 A. Sipahi, 2016, p. 818.
• 45 T. Altıntaş, 2017, p. 314.
• 46 V. Ketsemanian, 2018, p. 738.

22In the 1890s, to resist the entrenchment of the practice of violence in their home communities, Armenian peasants took up arms to defend themselves, while during the same period the Armenian revolutionary groups intensified their activities.44 With the emergence of such groups, the Hunchakian party being one of the earliest, “legal institutionalisation within the framework of the Ottoman administration ceased to be the only mode of Armenian political organising”.45 In the early 1890s, Hunchakist cells infiltrated the hinterlands of the empire. When party members distributed flyers in the central provinces calling for resistance against Abdülhamid’s regime, the action resonated in the eastern provinces, provoking clashes in the mountains of Sasun.46

• 47 L. Nalbandian, 1963, pp. 124-125.

23Until the early 1890s, the European Powers’ pressure on the Porte to implement reforms in the six eastern provinces remained sporadic and largely ineffective. However, when Armenian villages in Sasun were attacked and civilians were killed, the Powers grew more responsive to the plight of the Armenians. On 11 May 1895, they confronted Abdülhamid with the “Project of Reforms for the Eastern Provinces of Asia Minor”, a document intended to hold the Sultan accountable for his reluctance to implement reforms. Events took a dramatic turn on 30 September 1895 when the Hunchakists organised a rally in Constantinople to denounce the regime’s failure to launch reforms and to demand justice for the victims of the Sasun massacre. In response, the regime encouraged pogroms in the residential quarters across the city targeting their Armenian inhabitants. The consequences were horrendous. Prisons were overcrowded and scores of dead bodies were collected from the streets of Constantinople.47

• 48 Report of the Consular Delegates attached to the Commission appointed to inquire into the Events at (...)
• 49 H.F.B. Lynch, 1901, p. 158.
• 50 BOA Y.EE 151/2 (01 August 1894), Coded Telegram from the provincial authority of Bitlis to the Mini (...)
• 51 Hallward to Currie, Inclosure 2 in no. 339, 06 November 1894; Report of Sassoon Affairs, by a resid (...)
• 52 Hallward to Currie, Inclosure 2 in no. 339, 06 November 1894.

24In August, the inhabitants of “seven or eight mahalles in the nahiye of Talori consisting of seventy or eighty houses, according to official sources”,48 refused to pay government taxes unless the state protected them against unruly Kurdish extortionists to whom they paid an illegal tax khafir. Also known as kafir, hafir or khafirat, this tax was exacted unrestrictedly by the Kurdish chieftains, nominally for “protection” against the marauding Kurdish tribes, despite appeals that Armenians had repeatedly lodged with the government. The protest was exaggerated by vali Hassan Tahsin into a “rebellion”.49 On 1 August, the vali reported that Armenians had mustered a fighting force that was engaged in killing and torturing the Muslims and harboured plans to murder the kaymakam (district governor) of Sasun and attack the troops stationed at Shenik.50 A few days later, the Kurds from the Bekiran and Badikan tribes launched an attack in the direction of Shenik and Semal. After they were repulsed, the Kurds refused to resume attacks unless helped by the troops. The soldiers from the Bitlis and Muş garrisons, who were present in the area, joined the attacking mobs reportedly disguised as Kurds.51 This time the attack was more successful. The defenders retreated toward Gelieguzan, where fierce fighting broke out, and then fell back to the slopes of Mount Andok. Mingled with the soldiers, the Kurds went on to the offensive and besieged Andok on 13 August. The Armenians were able to hold them off until, on 26 August, the bulk of government troops that arrived from Muş the day before captured their positions and committed some of the greatest excesses of the entire army operation that lasted from 18 August until 10 September.52

• 53 BOA, Y.EE. 153/142 (30 August 1894), Correspondence from the Fourth Army Command to the Yıldız Pala (...)
• 54 BOA, Y.EE. 97/53 (30 August 1894), Confirmation from a Sultan’s aide-de-camp for the dispatch of th (...)

25On 30 August, Zeki Paşa informed Yıldız Palace that the Armenians flocked in the mountains, most of whom were women, children or unarmed male peasants, offered no resistance to the troops and that the number of insurgents among them was limited to Hampartsum Boyadjian, one of the defence leaders known by the pen name “Murad”, and a small group of his comrades-in-arms, called “outsiders” in Zeki Paşa’s report, who had arrived from Tiflis.53 Despite this, the authorities chose not to intervene or halt the violence. On the contrary, after receiving the report, the centre reiterated its original orders, in even stronger terms, to continue with the pursuit of the Armenian peasants, whom the authorities still indiscriminately referred to as “rebels”.54

• 55 Report of Sassoon Affairs, by a resident at Bitlis, Inclosure in no. 78, 18 January 1895, p. 76.
• 56 A political party founded in 1890, also known as Dashnaktsutyun.
• 57 K. Sassouni, 2013, p. 519.
• 58 Memorandum on the Joint Report of the Consular Delegates to the Sassoun Commission of July 20, 1895 (...)

26After the butchery at Mount Andok, the troops marched southwards and, on 4 September, went on a killing spree in the villages of Talori. There as well, ghastly outrages brought about a multitude of civilian deaths.55 Having nearly exhausted their meagre ammunition, most Armenians were doomed whether they surrendered or took to flight. Karo Sassouni, a renowned author and activist of the Armenian Revolutionary Federation56 in the first half of the twentieth century (see Table A and the paragraph that follows it), claimed that after the fighting ended, villagers were found dead in the canyons, the victims of Kurdish gangs and Turkish soldiers.57 As if to pre-empt Sassouni’s statement, Commissioner H.S. Shipley reported a year after the incident that “the Armenians were absolutely hunted like wild beasts, being killed wherever they were met, and if the slaughter was not greater, it was solely owing to the vastness of the mountain ranges of that district which enabled the people to scatter, and so facilitated their escape”.58

• 59 Their Armenian equivalents are, respectively, Khoulp, Khiank, Psank, Gavar, Shatakh, and Talvorik. (...)
• 60 Hallward to Currie, Inclosure 2 in no. 339, 06 November 1894.
• 61 E.A. Brayley Hodgetts, 1896, p. 95.
• 62 M. Polatel, 2016, p. 190.

27Concentrated in the area of the massacre were Armenian populations of the kazas of Kulp, Hiyan, Psank, Sason, and Şatak and the nahiye of Talori.59 Vice-Consul C.M. Hallward revealed that “the Kurds did not take any large share in the massacre”, adding that “at the instigation of the soldiers, they certainly did take part in it, though the worst excesses seem to have been committed by the former”.60 British reporter E.A. Brayley Hodgetts concurred that “[t]he Kurds actually refused to have any share in the horrors which followed; for these the Turks are alone responsible”.61 In an effort to absolve themselves of responsibility for the crime, the Ottoman authorities dismissed atrocities against Armenians as nonsense, claiming that “insurgents” evacuated the villages that were burnt and transferred the civilians to the nearby mountains before the fighting erupted.62

• 63 Memorandum on the Joint Report of the Consular Delegates to the Sassoun Commission of July 20, 1895 (...)

28Estimates of the number of killed range from about 1,000 to as many as 11,870, with most of the deaths occurring in the period from 26 August to 10 September when main government troops were on the scene (see Table A). Shipley suggested that for “some 23 villages so destroyed”, the “total number of victims, including those who perished from want or exposure, would amount to some 900 in all”. Yet he admitted that the estimate was based in some part on his visits to the villages of Shenik, Semal and Gelieguzan only. For other villages, primarily those in Talori, Shipley relied on the visit to that nahiye in early October 1895 of Charles Seymour Hampson, the British Vice-Consul in Muş, whose report he regarded as forming a supplement to the Memorandum on the Joint Report drawn up by himself.63

• 64 Idem, p. 76.
• 65 Hampson to Graves, Inclosure 2 in no. 320, 18 August 1895.

29Hampson put the number of victims at between 20 and 60 per village. Accordingly, taking 40 as an average for each of the 23 villages destroyed, Shipley arrived at the figure of about 900 killed. He noted, however, that “the total of Armenian losses could scarcely even be approximate but are little better than mere guesses” and that if the same proportion were maintained for the villages the delegates had not visited, “the losses of the Armenians would be even greater than 900”.64 Hampson maintained that it had been impossible in most cases to obtain any definite information concerning death tolls and seemed certain only that large numbers had perished. In his own words, “from twenty to fifty from each village appeared to be the idea of the number who lost their lives”.65

• 66 Hallward to Currie, Inclosure in no. 283, 09 October 1894.
• 67 Hallward to Currie, Inclosure 2 in no. 339, 06 November 1894.

30Vice-Consul C.M. Hallward, who was dispatched to the area to investigate allegations of the massacre, claimed that “[t]here must have been 1,000 or more killed”,66 noting in his next dispatch that the details were principally collected from soldiers who took part in the killings.67 In both dispatches, Hallward referred to the figure of about 4,000 killed in 12 villages that had been supplied to him by a local Armenian. Because he admitted that “there [could] be no doubt that a large majority of the population of some twenty-five villages perished”, the total, in his reckoning, might be as much as 8,000. Whether this figure was at all near the mark, Hallward could not affirm, but noted in passing that in Bitlis he heard “a far higher estimate given”.

• 68 Memorandum by Consul Graves, Inclosure 2 in no. 36, 27 December 1894.
• 69 E.A. Brayley Hodgetts, 1896, p. 100.

31Meanwhile, British Consul in Erzurum Robert Wyndham Graves noted that “about 32 villages are said to have been destroyed” and that during the 23 days prior to 10 September betwen 3,000 and 4,000 persons had been killed.68 Brayley Hodgetts maintained that after the soldiers conducted the Armenian escapees back to Sasun, the local priests made a list of the missing. According to the author, it “was found that, roughly, about 4,000 of the original population had survived, and that 12,800 were missing”. He noted, however, that he could “not think that all these were killed, but so the peasants believed”.69

• 70 This information was obtained from an Armenian informant who believed that of that figure about 3,0 (...)
• 71 This figure appeared in a verse bearing the author’s pseudonym Lerents. See Gaghapar no. 2 (1894), (...)
• 72 This figure appeared in a memorandum presented to Lord Kimberley, British Foreign Secretary. See Ms (...)
• 73 Archives of the Armenian Patriarchate of Constantinople, cited in R. Kévorkian, 2001, p. 49.
• 74 K. Sassouni, 2013, p. 519.

• 75 Ibid.
• 76 In fairness, however, it must be said that the figure of 10,000 killed was not a uniquely Armenian (...)

32Of these sources, Karo Sassouni stands out as particularly important. A relative of Armenak Ghazarian, one of the defence leaders known by the pen name “Hrayr Dzhoghk”, Sassouni no doubt heard the retelling of the incident while growing up in the village of Aharonk in Sasun, where he was born in 1889. Sassouni revealed that it was decided in the aftermath of the massacre, apparently in the inner circle of Armenian party and community leaders, that the figure of 10,000 dead should be circulated “for advocacy purposes”. In the period since, he maintained, “the figure had remained unchanged in the European and Armenian chronicles”.75 In fact, most Armenian authors used this figure as the basis of their narratives about the outcome of the massacre (see Table B).76

• 77 A. Babakhanian (Leo), 1925, p. 104.
• 78 V. Petoyan, 2005, p. 127.
• 79 K. Taroyan, 1966, p. 16.
• 80 T. Aghayan, 1976, p. 125.
• 81 R. Hovhannisyan, 1981, p. 245.
• 82 S. Poghosyan, 2008, p. 482.

• 83 BOA Y.EE 158/17 (08 September 1894), Correspondence from the Fourth Army Command to the Yıldız Pala (...)
• 84 Coded telegram of 16 September 1894 from the Commander of the Fourth Army to the Department of Gene (...)
• 85 Listed were Shenik, Semal, Gelieguzan, Aghpi, Hartk, Shoushnamerk, Shenist, Yeghgart, Khotsotsvank, (...)
• 86 Coded telegram of 18 September 1894 from the Commander of the Fourth Army to the Department of Gene (...)

33Putting to one side Commissioner H.S. Shipley, who remarked that his tally of Armenian deaths would have been greater than 900 had the delegates been able to visit more localities, the lowest figure was supplied by Zeki Paşa. On 8 September, the general reported that the number of insurgents killed was around 1,000 and that another 1,000 were expected to die of wounds received in battles.83 On 16 September, he sent a telegram to the Department of General Staff, which stated that after the rebels in Talori joined other groups of insurgents, “their number multiplied to over three thousand”.84 Zeki Paşa listed Armenian villages whose inhabitants “were coerced into collaboration with the revolutionaries”, implicitly suggesting that they were attacked by the troops.85 In another telegram, sent two days later, Zeki Paşa wrote that “several gangs consisting of some 40 or 50 brigands each have been completely eliminated” and the “losses suffered by the insurgents add up to a thousand”.86

• 87 Chermside to Currie, no. 22, 25 March 1895.
• 88 Moorch no. 1 (1895).

34In March 1895, Colonel Herbert Charles Chermside, British Military Attaché in Constantinople, heard that another telegram sent by Zeki Paşa sometime in or before January 1895, which admitted that 1,722 Armenians – men, women, and children – were killed, had been removed from the Department of General Staff.87 This information was based on hearsay and thus cannot be verified reliably. Interestingly, Moorch concurred, based on information the editors received from Constantinople on 4 January 1895, that the Ottoman War Ministry removed all the telegrams sent by Zeki Paşa, among them the one in which “he admitted that 1,720 Armenians were killed in Sasun”.88 If there indeed was a second telegram sent by Zeki Paşa, the general’s figure would surprisingly match up to the one reported by Sassouni.

• 89 Memorandum communicated to Mr. Block, Inclosure 2 in no. 324, 17 November 1894.

35In November 1894, two months after the cessation of violence, an unsigned memorandum from the British embassy, apparently submitted to the Sultan, mentioned “3,000 individuals killed”.89 More recent authors failed to settle the festering imbroglio involving fatality figures. Some offered numbers based on British consular documentation. Others chose to use numbers that are proximate to Sassouni’s figure (see Table C).

• 90 J. Haslip, 1958, p. 217.
• 91 R. Melson, 1982, pp. 487-488.
• 92 C. Walker, 1990, p. 142.
• 93 H. Kaiser, 1998, p. 9.
• 94 R. Hovannisian, 2004, pp. 219-220.
• 95 P. Balakian, 2004, pp. 54-56.
• 96 G. Bennett et al., 2008, p. 322.
• 97 D. Rodogno, 2012, pp. 191-194.
• 98 C. Laderman, 2019, p. 12.

• 99 Massis, February 5 (17), 1881.

36To facilitate estimations, preference was given to sources that contained data on the number of inhabitants rather than households. In cases where sources offered household counts only, it was deemed fit to use the average number of members per household to draw an estimate of the total number of inhabitants per village. In cases where sources provided no data on the number of victims but reported household numbers, eight was used as an average number of members per household, as suggested by Bishop Vahan Ter-Minassian (Partizaktsi), a commissioner at the Patriarchate in the 1860s and 1870s.99

• 100 Report of Sassoon Affairs, by a resident at Bitlis, Inclosure in no. 78, 18 January 1895, p. 75.
• 101 Droshak no. 1 (1895).

37Looking at the data in the Appendix, the number of fatalities for Gelieguzan, put at 40 by the Annex to the Joint Report, may seem rather low when compared to Shipley’s figure of 900 for the 23 localities destroyed, because massacres in this large village group, whose population across data sources ranged from 1,500 to 5,600 depending on the number of mahalles, potentially could have produced a greater number of victims. The figure is also incompatible with the information found in the “Report of Sassoon Affairs, by a resident at Bitlis”, which stated that there were only two survivors from one particular Gelieguzan household of 52 members.100 This being said, Droshak’s figure of 3,000 killed at Gelieguzan is clearly overstated.101

• 102 This number is not supported by other sources, most of which indicate that there were from 50 to 75 (...)
• 103 Moorch no. 10 (1894).
• 104 Report of Sassoon Affairs, by a resident at Bitlis, Inclosure in no. 78, 18 January 1895, p. 80.

38Moorch suggested that, except for a number of escapees, most of the inhabitants in 360 houses in Gelieguzan, 106 in Shenik, 85 in Semal, 165 in Alianc,102 78 in Ishkhandzor, and 66 in Aghbi were killed.103 Once the total number of households in these villages, rounded up to 760, is multiplied by eight to give a total population, three quarters (provided, certainly, that the quantifier “most” denotes “almost all”) of the estimated 6,080 inhabitants equates to about 4,560. This figure is similar to the tally of 4,200 killed in Talori and on Mount Andok, where the inhabitants of the said villages retreated, which appears in the “Report of Sassoon Affairs, by a resident at Bitlis”: “There is good reason for believing that 2,200 were killed in Dalvorik [Talori] and vicinity. Then about 2,000 on Mount Andok».104

• 105 Moorch no. 10 (1894).
• 106 Because we have as fairly reasonable cut-off points Sassouni’s 1,700, Zeki Paşa’s possible 1,722, a (...)

39It can be reasonably estimated, however, that both figures – 4,560 and 4,200 – are impossibly high, particularly when seen in the context of clashes in August, which reportedly produced more than 770 victims. Notably, Moorch reported 42 deaths on 2 August, 100 on the 3^rd, eight between the 4^th and the 6^th, 200 or more on the 8^th, 236 on the 13^th, 184 on the 15^th (when the Turks sounded a parley but then killed a group of negotiators who came down the mountain slopes), and an unspecified number before the 24^th.105 Assuming that close to three times as many villagers died in late August and early September, when the most ferocious wave of violence was unleashed with the direct involvement of troops,106 the total comes to 2,310, which constitutes roughly half of 4,560 or 4,200.

• 107 List of the villages of Sassoon and of Kharzan, Inclosure 2 in no. 133, 31 January 1895; and Graves (...)
• 108 Droshak no. 1 (1895).

40One other source placed in Table D is Patriarch Izmirlian’s list containing 52 destroyed localities and household numbers per village. Included in the list are many of the villages of Garzan, a kaza located in the sancak of Siirt which, according to Consul R.W. Graves, “have not suffered in the recent disturbances”. The consul pointed out other inaccuracies, such as the counting of Talori as one village of 700 houses and not as a nahiye consisting of several mahalles, the total number of houses in which, by his estimate, was far less than 700.107 Oddly, British dispatches supplied no data for Talori, where a series of fierce clashes and killings occured. The number of deaths placed by the Annex to the Joint Report at 14 in Talvorik (Egutoun) and 10 in Spghank is improbably low. On the other hand, the death toll of 5,000 in Talori given by Droshak is a gross overstatement.108

• 109 T. Martirosyan, 2021, Shatakh and Tsovasar, Talvorik, and Ishkhandzor sections.
• 110 A-Do, 1912, p. 122.

41To produce a lower bound on the number of deaths, one fifth was deemed fit for estimations. Most village populations in Şatak, Ishkhandzor and Talori, the areas which were most severely affected by violence, ranged from 80 to 350 people,109 with the average amounting to about 200. It seemed fair, therefore, to apply Shipley’s average of 40, constituting one fifth, or 0.2 in decimal notation, of 200, uniformly to estimations. This said, Vice-Consul C.S. Hampson’s estimate of 20 to 50 persons per village who lost their lives gives an average of 35, which is very close to one fifth, 0.175 to be exact. Significantly, contemporary Armenian statistician A-Do reported that in Semal, which had about 70 households or 560 inhabitants at the time of the massacre, 70 men and 50 women and children had been killed.110 The total of 120 dead amounts to about one fifth, 0.214 to be exact.

• 111 It cannot be ascertained which wards the Joint Report regarded as mahalles of Talori.
• 112 Mshak no. 27 (1895).
• 113 Annex to the Report of the Consular Delegates attached to the Commission appointed to inquire into (...)
• 114 Hampson to Graves, Inclosure 2 in no. 320, 18 August 1895.

42Because Moorch failed to supply data for Talori and the Annex to the Joint Report chose to put the number of houses at about eight to ten per ward,111 figures were collected from the Patriarchate and C.S. Hampson. For Egutoun, the Patriarchate listed 25 houses, which is substantially similar to the 30 to 40 houses put by Alexander Kalantar, the editor-in-chief of Mshak.112 The low per ward numbers in the Annex to the Joint Report almost certainly came from Ottoman sources, because cited in the text was the mutasarrıf (chief administrator) of Genç who supplied an unbelievably low figure of 67 households or 278 inhabitants for the nahiye.113 Ottoman figures get flimsier when compared with Hampson’s statistics, which counted 255 houses in seven of Talori’s mahalles,114 or those of Kalantar, which counted 200 to 220 houses across the nahiye. In Table D, in the sub-column “houses” under the sub-heading “Moorch no. 10 (1894)”, Hampson’s figure is quoted for Talori since it is hard to imagine that in one of the most densely Armenian populated nahiyes there could be only four persons living in a house, as proposed by the Turkish mutasarrıf.

• 115 Droshak no. 12 (1894).

45For the affected localities on the Plain of Muş, one source which supplied fatality data is the October 1894 edition of Droshak. The editorial indicated that “in Muş, 270 Armenian villagers had been killed and between 400 and 500 wounded”.115 However, these figures cannot be reliably verified. Through the examination of data, it has become known that most of the villages on the plain were besieged and plundered and some were destroyed, with only a few cases of killings reported (see the column “Mshak no. 133 (1894)” in Table D).

• 116 Given the approximate methods of calculation, this count should be treated as a very rough estimate (...)
• 117 List of the villages of Sassoon and of Kharzan, Inclosure 2 in no. 133, 31 January 1895; and List o (...)
• 118 Ardzagank no. 105 (1894).

47The results of numerical computations suggest that the estimated total for the number of victims during the massacre lies in the range between 1,663 and 2,231, with the average standing at 1,946.116 The incident engulfed more areas than previously believed, one being outside Sasun proper. The violence spread to Şatak (with Tsovasar), Talori, Sason (with Ishkhandzor), Kulp, Hiyan, Psank and Muş. This brings the total number of affected areas to nine. Consequently, the number of affected localities, which most sources extracted from the Joint Report and the statistics collected by the Armenian prelacies,117 appears to be higher than previously estimated. In addition to the villages that were attacked, destroyed or burnt, there were localities that were either besieged or plundered or indeed subjected to both forms of violence. The violence in its various forms targeted or had a direct impact on 85 Armenian localities in Sasun, affecting 2,435 households or approximately 19,480 people. Expanding well beyond the hub of protest, the violence affected 21 Armenian localities in Muş which were attacked on 27 August by the Kurdish mobs, apparently on the orders of the government.118 This brings the total number of affected villages to 106.

• 119 These were: Girgo (Grigor Moseyan) of Shenik, Gelou or Gyalouci Peto (Petros Srgeyan) of Gelieguzan (...)

48The Annex to the Joint Report under-documented the killings and thus underreported the numbers of killed. British dispatches supplied data for selected localities only or included reports containing inflated estimates of missionaries. Conversely, by claiming that most of the inhabitants of the localities in Şatak and Ishkhandzor had been killed, Moorch over-reported the death toll. The veracity of the number of dead estimated at 1,000 by Zeki Paşa becomes questionable when viewed in the context of most other reported figures or the average obtained from current estimations. It is known that there were about 20 Armenian group leaders involved in fighting.119 If each of them is assigned to a “gang of some 40-50 brigands”, the resulting total would surprisingly come to about 1,000. However, it is hard to imagine that every one of Zeki Paşa’s victims was a “brigand”. The Joint Report indicated that many unarmed villagers were killed while in flight and a number of peasants were killed while seeking protection from troops.

• 120 This figure is arrived at on the basis of information provided by Divisional General Abdullah Paşa, (...)
• 121 V. Mayewski, 1904, p. 26 (Strategic Study).
• 122 Calculated on the basis of number of households in the seven or eight mahalles of Talori (see Repor (...)
• 123 In the fevered imagination of Ottoman military officials at the local level, all Armenian males wer (...)
• 124 BOA Y.EE 156/13 (03 September 1894), Correspondence from provincial authorities of Bitlis to the Yı (...)
• 125 This slight difference is due to the fact that Moorch no. 10 (1894) reported 304 riflemen from Egut (...)
• 126 Droshak no. 1 (1895).

49The deployment of troops totalling up to 2,850 men,120 reinforced with ancillary weapons and aided by Kurdish mobs, was hardly conditioned by a military necessity to suppress a minor social protest. If the average number of houses per village, placed at 20 by Mayewski for the province of Bitlis,121 is multiplied by eight, the average number of members per household suggested by Partizaktsi, the total number of Armenians living in the protest area would come to 1,400 or 1,600.122 Of these, one quarter at most, or between 350 and 400, would have been fighting-age males.123 In his report to Yıldız, vali Hassan Tahsin admitted that the number of “brigands” at the start of the protest stood at 120 and then rose to 1,000, involving peasants from remote villages.124 By the end of August and throughout early September, when main government troops came on the scene, their numbers had reportedly grown to 1,955 or 1,960.125 Of these, only 700 had rifles, mostly obsolete; all others fought with hatchets, daggers and other cold weapons.126 While the ratio of troops to defenders illustrates that the Armenians of Sasun displayed exceptional bravery fighting off a considerably superior force, it also provokes the question as to why such a large force was used against protesters.

• 127 E. Gölbaşı, 2018, p. 50.
• 128 R. Douglas, 1976, p. 125.
• 129 Memorandum on the Joint Report of the Consular Delegates to the Sassoun Commission of July 20, 1895 (...)
• 130 E.A. Brayley Hodgetts, 1896, p. 96.

50The answer lies in the Porte’s “propensity to deploy violent and coercive measures to ‘solve’ the Armenian Question for good”127 as the desired end result of the Sultan’s “policy of deliberate oppression”. Mehmed Said, sometime Sultan’s first secretary, quoted a statement made by Abdülhamid to British Ambassador Philip Henry Wodehouse Currie that “the Armenian Question must be settled not by reform but by blood”.128 One indicator attesting to such a propensity is Sassouni’s figure, which suggests that the number of murdered civilians was 2.4 times larger than the number of killed fighters. Shipley insisted that “the Turkish authorities, whether they instigated the attack or not, were responsible for it; [the attack] took place with their knowledge and consent, as is shown by the fact that the soldiers, sent nominally to keep order, sided with the Kurds, and so contributed to the ruin of the Talori Armenians”.129 In turn, E.A. Brayley Hodgetts observed that Zeki Paşa “drew a cordon of troops round Mount Andok and then proceeded to bombard it without any notice or warning”.130 Here again a question arises as to why Zeki Paşa devised a plan of attack in which the manoeuvring of Kurds was intended to prevent Armenian civilians from escaping. If his orders were to quell a disturbance, why would heavy gunfire under his command destroy villages, rendering large parts of the region uninhabitable? And what sort of troops would ransack the cattle, farm tools and worldly possessions of peasants?

• 131 Polatel, 2016, pp. 192-193.
• 132 BOA Y.EE 153/94 (28 August 1894), Correspondence from the Yıldız Palace to the Commander of the Fou (...)
• 133 BOA, Y.EE. 97/53 (30 August 1894), Confirmation from a Sultan’s aide-de-camp for the dispatch of th (...)
• 134 Moorch no. 10 (1894).
• 135 Memorandum on the Joint Report of the Consular Delegates to the Sassoun Commission of July 20, 1895 (...)
• 136 The Outlook, vol. 52, no. 8 (24 August 1895), p. 301.

51There is evidence to suggest that the Porte saw the unrest in Sasun as “an opportunity to teach the Armenians a lesson and to establish a central authority in the region”.131 When, by the end of August, the Ottoman authorities realised that the protesters were not all “rebels” but ordinary villagers, the centre sent in more troops, heavily armed and reinforced by mountain artillery, thereby allowing more civilian deaths to occur. A directive sent from Yıldız to the High Command of the Fourth Army in the run-up to the deployment of troops instructed Zeki Paşa, in no uncertain terms, to take effective measures to “violently quell the insurgency” and to ensure that “such incidents are not repeated in the future”.132 Another directive explicitly ordered the pursuit of fleeing Armenians and their eradication, including those seeking asylum, so as to impart a powerful lesson to Armenian “evildoers”.133 At around the same time, Moorch gave voice to a widely held view among the locals that Ottoman troops’ intervention in Sasun was intended to “do away with semi-autonomous existence of the Armenians”.134 This view has been affirmed by H.S. Shipley, who stated that the real cause of the sending of troops lay in the fact that the Ottoman authorities were unable to break the feeling of independence which had survived among the Armenians of Sasun.135 As if to confirm this view, at some time after the wholesale butchery in Sasun, Zeki Paşa was reported as saying: “Not finding any rebellion, we cleared the country so that none should occur in the future”.136

• 137 Report of the Consular Delegates attached to the Commission appointed to inquire into the Events at (...)
• 138 Memorandum on the Joint Report of the Consular Delegates to the Sassoun Commission of July 20, 1895 (...)
• 139 H. Poghossian, 1985, p. 219.

52There is, it seems, an element of cross-connection between “teaching a lesson” and the Ottoman drive to centralise power in Sasun. The Ottoman state’s mode de faire in “addressing” the grievances of their Armenian subjects led to “the absolute ruin of the region” which, in the words of the contemporaries who had a mandate to investigate the incident, could “never be regarded as a measure proportionate to the punishment even of a revolt”.137 Speaking with a full sense of responsibility, Shipley admitted that “it was not so much […] the suppression of a pseudo-revolt, which was desired by the Turkish authorities, as the extermination, pure and simple, of the Gelieguzan and Talori districts”.138 Not long after the “lesson was taught”, Ottoman military outposts were installed within a remarkably short time in Shenik, Gelieguzan and Egutoun,139 tightening the central government’s grip on Sasun.

• 140 According to the statistics of the Patriarchate, on the eve of the massacre there were nearly 40,00 (...)

53The excessive and indiscriminate use of force against villagers who refused to pay an onerous illegal tax resulted in a disproportionally high number of victims, most of whom were defenceless civilians. In percentage terms, the average obtained from current computations equates to about five percent of the total Armenian population of the kaza of Sason (Gavar)140 where most deaths occurred as a result of the fiercest fighting and subsequent killings. While the focus of this study was almost entirely on computations based on available statistics, it can safely be concluded that the punitive operation carried out by the Turkish troops and their Kurdish auxiliaries was incommensurate with both the nature and geography of the Armenian protest.