Şanlıurfa Province · Southeastern Türkiye
Rings of carved T-shaped pillars — older than the pyramids — raised by hunter-gatherers, then buried.
The evidence
Göbekli Tepe in depth
Göbekli Tepe was recorded and dismissed before it was understood. A joint Istanbul University–University of Chicago survey noted the mound in the 1960s, but read its scatter of broken limestone as a medieval cemetery and moved on. The hill kept its secret for another thirty years.
In 1994 the German archaeologist Klaus Schmidt re-examined the site and recognised the flint and worked limestone for what they were. Excavations began in 1995 as a joint project of the German Archaeological Institute (DAI) and the Şanlıurfa Museum. After Schmidt's sudden death in 2014, the project continued under the DAI — today led by Lee Clare — with Istanbul University and the Turkish authorities, and the site was inscribed on the UNESCO World Heritage List in 2018.
Göbekli Tepe is not one moment but a stack of them. The older phase — what excavators call Layer III, belonging to the Pre-Pottery Neolithic A, around 9500 BCE — is the famous one: large round and oval enclosures ringed by benches, their walls set with T-shaped pillars, and two much larger pillars standing at the centre of each.
Above and later sits Layer II, of the early Pre-Pottery Neolithic B: smaller, more rectangular rooms with smaller pillars. The shift from round to rectangular that plays out here is one of the defining architectural changes of the whole Neolithic. Radiocarbon dates place the site's long life across roughly the tenth and ninth millennia BCE — well over a thousand years of building, filling, and rebuilding.
The main excavated enclosures are labelled A through D. Enclosure D is the best preserved and most complete: its two central pillars (Pillars 18 and 31) are unmistakably human figures, with arms, hands, belts and a fox-pelt loincloth, and its walls carry the site's most famous carving — the Vulture Stone, Pillar 43. Enclosure C is ringed by boar imagery and, unusually, a predator carved in high relief on Pillar 27. Later buildings such as Building F show the site's continuing life.
Crucially, the enclosures are not identical: each has its own combinations of animals and its own character, which is part of why a single, tidy explanation of the site keeps failing.
The prevailing reading is that Göbekli Tepe was a monumental gathering place — somewhere scattered hunter-gatherer communities came together to build, feast, and mark shared beliefs. Dietrich and colleagues (2012) made the case that large-scale feasting was central, pointing to huge quantities of animal bone and food-processing equipment. Human remains complicate and deepen the picture: Gresky, Haelm and Clare (2017) documented deliberately modified human skulls, evidence of a Neolithic 'skull cult,' and the site's art is saturated with death and the dangerous animal.
Whether to call it a 'temple' is debated — the word imports later ideas of priesthood and organised religion that we cannot confirm. What is not in doubt is that this was a place built for something beyond mere survival, by people who had not yet taken up farming.
The enclosures were found filled with rubble, soil, and debris, which sealed and preserved them. Schmidt argued this was deliberate backfilling; more recent work (Clare and others) has questioned how intentional it really was, suggesting much of the fill may have accumulated through slope wash, collapse, and repeated use. It remains genuinely unresolved.
And Göbekli Tepe is far from fully dug. Geophysical survey shows many more enclosures still lying underground — only a fraction of the hill has been excavated. This is not a closed story; it is one of the most active frontiers in world archaeology.
Explore this site
68 pages — structures, carvings, discoveries, debates and guides. Jump straight in.
The wider world
Visit
Small-group and private tours to Göbekli Tepe and across the Taş Tepeler landscape, led by people who know the ground.
Full bibliography: the Taş Tepeler reference library →
Last reviewed 21 July 2026. Something wrong?
Last reviewed 21 July 2026. Something wrong?