Arrow Left Arrow Right Camera icon set icon set Ellipsis icon set Facebook Favorite Globe Hamburger List Mail Map Marker Map Microphone Minus PDF Play Print RSS Search Share Trash Twitter Video Camera Youtube
Turkey and Armenia: Opening Minds Opening Borders
Turkey and Armenia: Opening Minds Opening Borders
Table of Contents
  1. Executive Summary
Report 199 / Europe & Central Asia

Turkey and Armenia: Opening Minds Opening Borders

Turkey and Armenia are close to settling a dispute that has long roiled Caucasus politics, isolated Armenia and cast a shadow over Turkey’s European Union (EU) ambition.

  • Share
  • Save
  • Print
  • Download PDF Full Report

Executive Summary

Turkey and Armenia are close to settling a dispute that has long roiled Caucasus politics, isolated Armenia and cast a shadow over Turkey’s European Union (EU) ambition. For a decade and a half, relations have been poisoned by disagreement about issues including how to address a common past and compensate for crimes, territorial disputes, distrust bred in Soviet times and Armenian occupation of Azerbaijani land. But recently, progressively intense official engagement, civil society interaction and public opinion change have transformed the relationship, bringing both sides to the brink of an historic agreement to open borders, establish diplomatic ties and begin joint work on reconciliation. They should seize this opportunity to normalise. The politicised debate whether to recognise as genocide the destruction of much of the Ottoman Armenian population and the stalemated Armenia-Azerbaijan conflict over Nagorno-Karabakh should not halt momentum. The U.S., EU, Russia and others should maintain support for reconciliation and avoid harming it with statements about history at a critical and promising time.

Turks’ and Armenians’ once uncompromising, bipolar views of history are significantly converging, showing that the deep traumas can be healed. Most importantly, the advance in bilateral relations demonstrates that a desire for reconciliation can overcome old enmities and closed borders. Given the heritage and culture shared by Armenians and Turks, there is every reason to hope that normalisation of relations between the two countries can be achieved and sustained. 

Internal divisions persist on both sides. Armenia does not make normalisation conditional on Turkey’s formal recognition as genocide of the 1915 forced relocation and massacres of Armenians under the Ottoman Empire. But it must take into account the views of Armenians scattered throughout the global diaspora, which is twice as large as the population of Armenia itself and has long had hardline representatives. New trends in that diaspora, however, have softened and to some degree removed demands that Turkey surrender territory in its north east, where Armenians were a substantial minority before 1915. 

Over the past decade, Turkey has moved far from its former blanket denial of any Ottoman wrongdoing. Important parts of the ruling AK Party, bureaucracy, business communities on the Armenian border and liberal elite in western cities support normalisation with Armenia and some expression of contritition. Traditional hardliners, including Turkic nationalists and part of the security services, oppose compromise, especially as international genocide recognition continues and in the absence of Armenian troop withdrawals from substantial areas they occupy of Turkey’s ally, Azerbaijan. These divisions surfaced in events surrounding the assassination of Turkish-Armenian journalist Hrant Dink in January 2007. That the new tendencies are gaining ground, however, was shown by the extraordinary outpouring of solidarity with Armenians during the Dink funeral in Istanbul and a campaign by Turkish intellectuals to apologise to Armenians for the “Great Catastrophe” of 1915. 

The unresolved Armenia-Azerbaijan conflict over Nagorno-Karabakh still risks undermining full adoption and implementation of the potential package deal between Turkey and Armenia on recognition, borders and establishment of bilateral commissions to deal with multiple issues, including the historical dimension of their relations. Azerbaijan has strong links to Turkey based on energy cooperation and the Turkic countries’ shared linguistic and cultural origins. Ethnic Armenian forces’ rapid advance into Azerbaijan in 1993 scuttled plans to open diplomatic ties and caused Turkey to close the railway line that was then the only transport link between the two countries. For years, Turkey conditioned any improvement in bilateral relations on Armenian troop withdrawals. Baku threatens that if this condition is lifted, it will restrict Turkey’s participation in the expansion of Azerbaijani energy exports. While Azerbaijani attitudes remain a constraint, significant elements in Turkey agree it is time for a new approach. Bilateral détente with Armenia ultimately could help Baku recover territory better than the current stalemate.

Outside powers have important interests and roles. The U.S. has long fostered Armenia-Turkey reconciliation, seeking thereby to consolidate the independence of all three former Soviet republics in the south Caucasus and to support east-west transit corridors and energy pipelines from the Caspian Sea. Washington was notable in its backing of efforts that kick-started civil society dialogue between Turkey and Armenia. The Obama administration is working hard at repairing the damage done to U.S. relations with Turkey by the war in Iraq. Although Obama repeatedly promised on the campaign trail to formally recognise the 1915 forced relocation and massacres of Armenians under the Ottoman Empire as genocide, he should continue to steer the prudent middle course he has adopted as president. The U.S. Congress, which has a draft resolution before it, should do the same. At this sensitive moment of Turkish-Armenian convergence, statements that focus on the genocide term, either to deny or recognise it, would either enrage Armenians or unleash a nationalist Turkish reaction that would damage U.S.-Turkish ties and set back Turkey-Armenia reconciliation for years.

U.S. support for Turkey-Armenia reconciliation appears to be mirrored in Moscow. Russian companies have acquired many of Armenia’s railways, pipelines and energy utilities and seek to develop them; Russian-Turkish relations are good; and Moscow is looking for ways to mitigate the regional strains produced by its war with Georgia in August 2008. If sustained, the coincidence of U.S.-Russian interests would offer a hopeful sign for greater security and prosperity in the South Caucasus after years of division and conflict. All sides – chiefly Armenia and Turkey but potentially Azerbaijan as well – will gain in economic strength and national security if borders are opened and trade normalised.

Istanbul/Yerevan/Baku/Brussels, 14 April 2009

Map showing casualty rates for provinces, cities, towns and specific curfews. CRISIS GROUP

Turkey’s PKK Conflict: The Death Toll

On 20 July 2016, the first-year anniversary of resumption of violence in Turkey's PKK conflict, Crisis Group made public an open-source casualty infographic in order to draw attention to the rising human cost, trace conflict trends and demonstrate how the tactics on the ground are evolving. The infographic is meant to function as a clear, impartial, and factual account of the conflict’s changing dynamics.

The failed coup attempt on 15 July 2016 has focused renewed attention on Turkey’s military and security apparatus at a time when violence is already high.

Exactly one year ago, on the afternoon of 20 July 2015, an Islamic State (IS) suicide bombing tore through the majority Kurdish town of Suruç in south-eastern Turkey, killing 33 and injuring over 100 people, mostly young activists en route to support reconstruction efforts in the Syrian Kurdish town of Kobani.

Just three hours later in the nearby province of Adıyaman, militants of the Kurdistan Workers’ Party (PKK) accusing Turkey of abetting the IS attack, killed 23-year old Müsellim Ünal, a corporal in the Turkish military. PKK militants killed five more security officials in the following week.

The two-and-a-half-year ceasefire between Turkey and the PKK – internationally listed as a terrorist organisation – had broken down. Since then, the PKK conflict has entered one of the deadliest chapters in its three-decade history. Over the past year, more than 1,700 people have been killed, according to Crisis Group’s interactive, open-source database of the casualties, published online in English and Turkish today.

A fragile ceasefire, which had already been faltering, collapsed as the region was engulfed in the unpredictable mayhem of clashes and security operations, resulting in the displacement of more than 350,000 civilians and massive urban destruction in some south-eastern districts. A year later, whole swathes of Turkey’s majority Kurdish south east have been devastated, bombings have struck at the heart of the country’s largest metropolitan centres, and the PKK conflict is inextricably linked with conflicts in the Middle East, especially the war in Syria.

Though dynamics have changed considerably, the past year contains echoes of what had been the worst period of the PKK conflict in the early 1990s. Then, information was hard to obtain and there were many obstacles to travel in the region. Now the problem is that information is tightly controlled, travel is limited for some, and it can take months to work out what exactly happened even in high-profile incidents.

Tracking the Data

International Crisis Group has long worked to track the rising cost of violence using open-source data, including reports from Turkish-language media, local Kurdish rights groups and the Turkish military. In the last bout of violence, from June 2011 to March 2013, Crisis Group was able to confirm the deaths of 920 people.

Crisis Group is now making public in real time the death toll since 20 July 2015, based on its open-source methodology. The goal of this project is to draw attention to the rising human cost of the conflict, trace the trends of the conflict and demonstrate how the tactics of the parties are evolving. We aim to present a clear, impartial, factual account of what is happening in the conflict, and where violence is taking place to guide policymakers, researchers and public opinion in Turkey and beyond. This is an important task, as casualty figures are generally politicised by all parties as they present competing narratives. This resource should be viewed within the framework of Crisis Group’s calls for de-escalation of the PKK conflict.

Crisis Group first started formally tracking the casualty toll of the 32-year conflict during the July 2011-March 2013 cycle of violence. We used the casualty information released by the PKK and the Turkish military for their own losses, relying on the fact that a relatively clear distinction between militants and civilians was possible in a conflict that mainly took place in rural settings. We were generally cautious about the use of figures presented by either side for the opposing party’s losses, since these were often inflated to mobilise support and legitimise further action.

As the fighting moved to urban areas during the current conflict cycle, the line between PKK militants and civilians became increasingly blurred. The difficulty of distinguishing between civilians and militants led Crisis Group to create an additional category for “youth of unknown affiliation”, covering male and female casualties, aged 16-35, who were killed in curfew zones or areas of clashes, but who were not claimed by the PKK’s military wing (People’s Defence Forces, or HPG) or its urban youth wing (Civil Protection Units, or YPS).

The visualised dataset for casualties since July 2015 utilises a more complex methodology, systematically tracking casualty data together with other indicators such as urban versus rural casualties; types of attacks targeting security forces; age of civilian casualties; breakdown of security force casualties into police, soldiers, village guards; breakdown of PKK militant deaths into HPG, YPS, and TAK (Kurdistan Freedom Falcons, the PKK’s affiliate responsible for attacks in the west of Turkey). The data also accounts for curfews enforced by security officials in majority Kurdish-speaking south-eastern districts, specifying the duration of the curfew and the minimum number of individuals confirmed dead during these operations.

21 July 2016
The fighting in the southeast, which reignited in July 2015, has killed 1,761 people, according to new figures released by the International Crisis Group. Middle East Eye

— Direnç Balık


One of the Most Violent Flare-ups in Three Decades

The July 2015 flare-up of the PKK conflict is one of the most violent episodes in its 32-year history. The most significant indicator for this is the number of state security force members killed, usually accurately reported by military sources and subsequently covered in Turkish media.

Based on its open-source casualty database, between 20 July 2015 and 19 July 2016, Crisis Group confirmed the deaths of at least:

– 307 civilians

– 582 security force members

– 653 PKK militants

– 219 “youth of unknown affiliation”

Since August 2015, security forces have imposed a total of 85 curfews of different durations in 33 majority Kurdish south-eastern districts to ensure government control over areas where members of the Kurdish political movement have declared self rule, and where PKK militant presence grew significantly.

The Turkish military and PKK both maintain their own count of the overall casualties during the last escalation cycle. The Turkish military on 24 May 2016 claimed that since July 2015, 4,949 PKK militants have been killed. The HPG, the military wing of the PKK, claimed on 2 May to have killed a total of 1,557 police and soldiers since July 2015, while for the same period Crisis Group confirmed a total of 465 security force members dead.

What the new infographic shows

The new infographic illustrates changing conflict dynamics in the course of the year, including more large-scale attacks in the west of Turkey, a rise in single PKK attacks creating more casualties, a return to rural clashes in the last two months, and a rise in HPG militant deaths – as opposed to YPS militants in June 2016. Notable also is that one-third of all recorded casualties are concentrated in the districts of Sur, Nusaybin and Cizre, of which the latter two are directly neighbouring Kurdish-inhabited areas of Syria.

  • While the bulk of the PKK conflict has remained highly localised in the country’s majority Kurdish south east, since January 2016 violence has increasingly spread to the west of Turkey. Two bombings by TAK struck Ankara on 17 February and 13 March 2016, killing a total of 38 civilians and 28 security officials. A TAK suicide bomber also detonated herself outside of the Great Mosque in Bursa, a city in north west Turkey, injuring thirteen on 27 April. This was followed by a May 2016 attack in Istanbul’s Sancaktepe district that narrowly missed a bus full of police. TAK also claimed responsibility for a car bomb attack on a police bus on 7 June in central Vezneciler district of Istanbul killing seven police officers and four civilians.
  • Between 1 February and 19 July 2016, of the 312 security force members killed, 150 were victims of improvised explosive device (IED) attacks (48 per cent), while in the four months prior (October 2015-January 2016) of the total of 115 security force deaths, 33 were from IED attacks (29 per cent). This suggests the PKK is engaging in more high-profile attacks, killing more security forces with single IED attacks, which has resulted in an increase in monthly security force casualties since March 2016 (a daily average of 1.8 between March 2016 and June 2016; compared to a daily average of 1.3 between November 2015 and February 2016).
  • Evaluated together with the urban vs. rural casualty information and the location of casualties, the data also indicates that the PKK is withdrawing its militants from urban centres and focusing its strikes in rural areas. May 2016 was the first month since November 2015 in which the number of total rural deaths was higher than total urban deaths (57 per cent of security force casualties occurred in rural areas in May 2016, as opposed to 33 per cent in April and 13 per cent in February). In June 2016, the ratio of rural security force casualties increased to 62.5 per cent. This clearly shows that the PKK is withdrawing from urban areas and shifting back to its traditional rural tactics. Besides intensified military operations of state security force members against militants in urban centres, growing public resentment among Kurds against PKK actions in restive districts around the country’s south east could be one of the reasons for this change in tactics.
  • 30 per cent of confirmed casualties have been in Cizre, Sur and Nusaybin. Over the last year, fighting has been focused in mainly four provinces: Şırnak, Diyarbakır, Mardin and Hakkari. The highest number of casualties were recorded in Şırnak’s Cizre district, followed by Diyarbakır’s Sur district, and Mardin’s Nusaybin district. Casualties in these three districts comprised 30 per cent of the total casualties since July 2015.
  • In March, April and May 2016, a sharp drop in the number of deaths among “youth of unknown affiliation” has paralleled a rise in casualties claimed by militias of the PKK’s youth wing, the YPS. Deaths among “youth of unknown affiliation” peaked in early February, when military operations in Cizre saw over 100 deaths in a single week. YPS deaths rose sharply in the following weeks, peaking in March and April. This could be an indication of two dynamics: either the PKK’s young militants are increasingly joining the YPS, or the PKK has begun reporting these deaths to give the appearance that its “resistance” is embraced by a broader group of youth in conflict areas.
  • Since February 2016, casualties suffered by YPS and HPG militants have been relatively close in numbers, an indication that the two PKK wings have been engaged in fighting (HPG/YPS killed: February at least 13/23; March at least 28/37; April at least 50/59; May at least 21/21). Casualty figures for June 2016 indicate that HPG casualties are rising, with the balance tipping considerably toward HPG militant casualties (HPG/YPS killed: June at least 45/3).
15 July 2016
Nor would a coup decisively end a revived Kurdish insurgency, which has claimed over 1,700 lives since July 2015, according to the International Crisis Group. Foreign Policy

— Noah Blaser


Overall Casualties in the Conflict since 1984

The PKK conflict in Turkey is commonly estimated to have killed around 30,000-40,000 people since 1984. The overall toll of the conflict is difficult to determine given the limitations on accurately verifying death claims. To reflect this ambiguity and minimise the risk of inflation or underestimation, caution in citing the overall death toll is in order.

Over the years, Turkish media and academia have all too often simply cited overall death toll figures voiced by politicians and military representatives. As described in a forthcoming report by Noah Arjomand, an academic at Columbia University, even these figures varied: in December 1997, Süleyman Demirel, then Prime Minister, said 37,000 people had been killed in the insurgency; in January 2005, then Prime Minister Recep Tayyip Erdoğan said the PKK had killed 40,000; in September 2008, the chief of general staff said it was 44,042; in February 2013, a parliamentary commission report estimated the total death toll at 35,576.

The same study shows that the breakdown of annual death toll figures announced by the parliament and gendarmerie (rural police) between 1984 and 2009 are divergent, too. The difference was highest after Turkey launched two large-scale operations into northern Iraq in 1997, accompanied by huge claims of PKK casualties in pro-government media: the overall toll reported by parliament was 3,419 killed, the gendarmerie said 8,234. Confirming casualties during this incursion was particularly difficult since, in contrast to reporting from Turkey’s south east, local human rights organisations did not have the means to verify deaths occurring during these military operations due to limited access to northern Iraq.

According to Turkish military figures compiled by Arjomand, the ratio of PKK militants reported killed by the military has tended to increase considerably compared to security force casualties in years of cross-border military operations against the PKK. For most years, the rate of PKK militants killed is, according to military-issued figures, two to four times higher than state security forces killed during low-intensity years of the conflict. However, a significant hike in the ratio can be observed during high-intensity conflict periods involving incursions of the military into northern Iraq, with the peak being the year 1997, when Turkish military claimed that 14.6 times more PKK militants than members of state security forces were killed.

In the current round of escalation that began in July 2015, the Turkish state claims to have killed around ten times more PKK militants than it has lost security forces. Similarly, the PKK also claims to have killed around ten times more state security forces than it has lost militants. Crisis Group’s casualty count of the last year reflects the security force-to-militant ratio at 1.16, while the same ratio was 1.75 for Crisis Group’s casualty count for the 2011-2013 escalation cycle. However, Crisis Group’s open-source methodology does not always allow for an accurate account of PKK militant deaths, which PKK-linked sources often only announce weeks, months, or even years later. Moreover, means of verifying PKK militant deaths in cross border incursions into northern Iraq are particularly limited. Despite these methodological limitations, it is safe to say that PKK claims for the number of state security forces killed, as well as state figures for PKK members killed are inflated. As a human rights activist in Diyarbakır told Crisis Group in January 2016: “Our cities in the region would be flooded with corpses and we would constantly hold funerals if the number the state announces for PKK members killed was correct”.

Given these limitations and ambiguities, the precise figure for the overall casualty toll of the conflict is impossible to confirm, whether by referring to Turkish state sources, or by using reports of human rights NGOs close to the Kurdish movement – who have collected data with differing methodologies for different years. Furthermore, the number of unresolved murders, extrajudicial killings and forced disappearances occurring especially in the 1990s – the conflict’s most deadly period – is unknown. In order to avoid the risk of using inflated figures or excluding unrecorded casualties, the safest option is to refer to the conflict’s overall toll as “tens of thousands of deaths”. Such caution should guide the use of any data on conflict casualties, where even seemingly accurate estimates may be flawed.