This analysis by Akın Ünver, Kadir Has University, published by FT, is interesting:
Turkey’s fragile Kurdish peace process, which has been in place with varying levels of success since 2009, effectively ended in June 2015, when both sides resumed hostilities. Since then, hundreds of people — civilians, Turkish soldiers and Kurdish fighters — have been killed, a level of violence that has destabilised the country and strengthened the resolve of the Kurdish separatists.
The exact reason for the breakdown has been buried by contradictory narratives.
While the Turkish government has emphasised its growing security concerns, the other side sees the breakdown of talks as a personal, political choice by President Recep Tayyip Erdogan.
Neither version fully explains how and why this historic opportunity for peace crumbled. If Turkey and the Kurdistan Workers Party, or PKK, are to resume talks in the future, however, it will be important to understand the circumstances of the collapse.
To find out, I carried out a geospatial data analysis to see how Kurdish groups were discussing the threats among themselves. Using a supercomputer to collate and match two years’ worth of Kurdish conflict-related actions, we cross-referenced these against mentions on social media in several languages to see when and where violent activities occurred.
The resulting data are surprising. First, if measured solely by the cessation of violence inside Turkish borders, the peace process appears to have been mostly successful in 2014 and early 2015. Rather than inside Turkey, most of the focus of the PKK and other armed Kurdish groups had been on Iraqi Kurdistan and northern Syria, where they were gaining territory and influence in the fight against Isis.
That suggests that the immediate reason given for calling off the peace process — the killing of three Turkish policemen by the PKK in July 2015 — was not the only catalyst. To be sure, the PKK had not fully disarmed and withdrawn from Turkish territory, as had been agreed. There had also been sporadic violence in eastern Turkey.
But, on the whole, the ceasefire seems to have held remarkably well, which suggests a similar detente could be replicated in future. In the first half of 2015, Kurdish armed groups were almost completely passive in Turkey and were most active in northern Syria, fighting in Kobani, Hasakah and Aleppo — as part of an effort to create what now looks like a Syrian Kurdistan.
Even so, the Turkish government interpreted the killing of the policemen as evidence that the PKK, encouraged by its Syrian affiliate’s gains in Syria, was not serious about keeping to the terms of the ceasefire.
Turkey considered the peace process officially over. Ankara had its jets carry out an aerial bombardment across the PKK’s stronghold in Qandil in Iraq and began large-scale security operations in urban centres where the PKK was entrenched.
For the PKK, the scale of this response confirmed its own long-held suspicion that calling off the talks was a political gambit by Mr Erdogan. The president was accused by Kurdish politicians of creating a “diversionary war” to gain support from nationalist conservative voters for a referendum to give his ceremonial presidency the executive powers he had given up after three terms as prime minister.
The impact of calling off the peace process was immediate. The data show that, after a tactical focus in early 2015 on the main population centres of northeastern Syria, including dams and bridges on the Euphrates river and the oil-rich Hasaka region, almost all the Kurdish armed groups converged and expanded in Turkish territory.
In late 2015, violence inside Turkey escalated. PKK-affiliated groups carried out terrorist attacks and deployed a youth militia to try to hold the centres of major Kurdish towns in southeastern Turkey, prompting a response from the military that claimed hundreds of lives.
So, the question remains, was it a good idea to end the peace process — and to do that in July 2015, when the violence inside Turkey seemed to have ebbed? There is only one thing both sides agree upon: in the months before the collapse there was not much negotiating going on. Our data show there was not much fighting either.
The security situation is much worse now. With urban warfare taking its toll on civilians, the cost of all this fighting will be questioned. While Turkish voters are sensitive to national security threats, such prolonged violence and mounting casualties will inevitably lead to reflection and questioning.
On the Kurdish side, renewed warfare serves as fodder for disenchantment and radicalisation. Some will join the PKK, some might leave Turkey and some — frustrated with both sides — may join Islamist radical groups.
More warfare will also weaken the pro-Kurdish Peace and Democracy Party (HDP), a welcome development for Mr Erdogan, who sees the party as both an existential and political threat.
In the end, the data and recent history show us both sides have a maximalist understanding of their own interests and an overconfidence about what they can accomplish through prolonged fighting in urban terrain. With miscalculated capabilities and without a reference point to measure success, the day when both sides return to the negotiating table may still be far away.