At Göbekli Tepe, a limestone ridge in southeastern Turkey’s Şanlıurfa Province, hunter-gatherers raised T-shaped pillars weighing up to 20 tonnes and standing nearly six metres tall — roughly the height of a two-storey house — around 9600 BCE. That is about 6,000 years before the first sarsens went up at Stonehenge, and about 7,000 years before the Great Pyramid at Giza. The builders had no metal, no wheeled vehicles, no draft animals, and no pottery. They had flint, obsidian, antler, rope, timber, and each other.
The mechanism, reconstructed from decades of excavation across the Taş Tepeler region, was brutally simple: quarry the pillars directly from the bedrock of the same plateau, dress them with harder stone hammers, drag them a few hundred metres on wooden sledges and rollers, and tip them upright into pre-cut sockets in the limestone floor. The scale is the shock. The technology is not.
The site that broke the timeline
Göbekli Tepe sits about 15 kilometres northeast of the city of Şanlıurfa, atop a bald hill that looks, from a distance, like a very ordinary mound. It was catalogued in 1963 during a joint Istanbul–Chicago survey and dismissed as a medieval cemetery. In October 1994 the German archaeologist Klaus Schmidt returned, recognised the tips of buried megaliths poking through the soil, and the following year opened the first systematic trenches in cooperation with the Şanlıurfa Museum. What his team uncovered forced a rewrite of the standard Neolithic sequence, in which farming came first and monumental architecture came thousands of years later, once societies could feed a specialist class.
Göbekli’s oldest ringed enclosures — labelled A through D — date to the Pre-Pottery Neolithic A, roughly 11,500 years ago. The people who built them were still hunting gazelle, aurochs, and wild boar and gathering wild cereals. There is no evidence of domesticated grain at the site’s founding. As Archaeology Magazine reported in its 2024 survey of the Taş Tepeler, more than 20 related sites now scattered along a 125-mile arc of the Taurus foothills share the same repertoire: T-shaped pillars, rounded enclosures, stone benches, and carvings of leopards, snakes, vultures, and human figures with skeletal ribs.
Where the stone came from
The pillars did not travel far. The plateau on which Göbekli Tepe stands is itself a limestone shelf, and quarries pocked with half-finished pillars sit within a few hundred metres of the enclosures. One unfinished monolith still lying in its bed measures about 7 metres long and would have weighed an estimated 50 tonnes had it been raised — larger than any pillar successfully erected on the site. It cracked, apparently, during extraction, and the workers walked away.
Extraction relied on a technique used across the Neolithic Near East. Workers scored a rectangular outline around the intended pillar using flint picks and stone mauls, cutting narrow channels a few centimetres deep. They then drove wooden wedges into the channels and soaked them with water. As the wood swelled, it split the stone along the scored line — the same principle Egyptian quarrymen would use thousands of years later on granite obelisks. Fire-setting, in which a bonfire is lit against a rock face and then doused, may also have been used to fracture the limestone along natural bedding planes.
Limestone is soft enough that this worked. It has a Mohs hardness of roughly 3 to 4. Flint, at around 7, easily cuts it. The pillars at Göbekli show tool marks consistent with pecking — the repeated striking of the surface with a hard hammerstone to remove millimetre-thin flakes until a smooth face emerges. On a 20-tonne pillar, that is months of work by a team.
Carving without metal
The reliefs are the part that seems, at first, impossible. Pillar 43 in Enclosure D — the so-called Vulture Stone — carries a crowded panel of birds, a scorpion, a headless human figure, and abstract symbols carved in low relief, some of them barely a centimetre proud of the pillar face. The fox on Pillar 10 is rendered mid-leap, its tail curled along the stone’s edge. These are not scratches. They are considered, three-dimensional images cut into a boulder taller than the sculptor.
The toolkit that produced them has been recovered in quantity from the site’s fill: chert and flint blades, burins, scrapers, polished basalt axes, and grooved stone abraders. Experimental archaeology has shown that a knapped flint burin can incise limestone at roughly the rate a modern steel chisel does on the same rock. The carving is slow but entirely feasible. Fine detail — the eye of a bird, the scale of a snake — was likely worked with obsidian flakes, which take an edge sharper than surgical steel.
The broader picture of Neolithic tool sophistication has been reinforced by unrelated finds elsewhere. Analyses of Bronze Age subsistence practices in the Levant have found that flint tools continued to be used alongside early bronze implements for generations. Stone did not lose to metal quickly. It lost slowly, over centuries, and only in specific applications.
Even the era’s medical work speaks to how much people could do with worked stone. Neolithic skulls with trepanation holes — surgical openings cut into living bone with flint scrapers — have been recovered from sites across Europe and Anatolia, many of them showing signs of healing, meaning the patient survived the operation. If a Neolithic hand could open a human skull cleanly enough for the bone to regrow around the cut, the same hand could carve a leopard into limestone.
Moving 20 tonnes without a wheel
The transport problem is where most people’s imagination stalls. A 20-tonne block is roughly the weight of a loaded city bus. There is no wheel in the Göbekli toolkit — the earliest evidence for wheeled vehicles anywhere comes from about 3500 BCE, some 6,000 years too late. There are no oxen; cattle were not yet domesticated in the region when the earliest enclosures went up.
Ethnographic parallels and experimental archaeology point to the same answer everywhere it has been tried: wooden sledges, log rollers, and a lot of people. Experiments with moving multi-tonne stones on timber sledges over greased wooden rails have demonstrated that a 20-tonne pillar over a few hundred metres of gently sloping bedrock could be moved by perhaps 200 people over several days — well within the labour available to a Neolithic community that could assemble hundreds of workers for feasting events, as the animal bone deposits at Göbekli suggest they did.
The route from quarry to enclosure at Göbekli is short and slightly downhill. Rollers of oak or juniper — both plentiful in the region’s original forest cover — would have reduced friction to something a large crew could overcome with ropes made of twisted plant fibre. Water or animal fat may have been used as a lubricant on the timber tracks. None of this requires invention beyond what any settled Neolithic community already possessed.
Raising the pillars
Setting a 5.5-metre pillar upright in a socket cut into bedrock is the final trick, and the one where the archaeology is clearest. The floors of the enclosures at Göbekli are the limestone bedrock itself, dressed flat, with rectangular sockets carved directly into the rock. The two largest central pillars in each ring stand in these sockets. The remaining pillars — typically around eleven arranged around the perimeter in the best-preserved enclosure, Enclosure D — were held upright by dry-stone walls built up around their bases as the enclosure was constructed.
The lifting method was almost certainly the ramp-and-tip technique used at later megalithic sites worldwide. The pillar is dragged base-first up to the socket. Workers dig a sloped pit behind the socket and lever the pillar’s base into it using timber pry-bars. Ropes tied to the pillar’s head and hauled from the opposite side then tip the stone upright, its base pivoting into the socket like a key into a lock. Once vertical, dry-stone packing wedges it in place. Gil Haklay and Avi Gopher of Tel Aviv University, in a 2020 architectural analysis published in the Cambridge Archaeological Journal, showed that the centre-points of the three great enclosures B, C and D form a near-perfect equilateral triangle with sides of about 19.25 metres — implying deliberate geometric planning of the complex before a single pillar was raised.
The people, not the technology
The bottleneck at Göbekli was never engineering. It was organisation. Feeding, coordinating, and housing a workforce of hundreds — assembled seasonally, likely from multiple bands — is a social problem, and it is the social problem that the Taş Tepeler sites appear to have been solving.
Necmi Karul of Istanbul University, who has coordinated the Taş Tepeler project since its formal launch in 2021 and directs the excavations at Karahantepe, roughly 35 kilometres east of Göbekli Tepe, argues that Karahantepe’s mix of small dwellings, quarries, and communal buildings shows the sites were living settlements as well as ritual centres. At Karahantepe, one subterranean chamber — Structure AB — contains eleven phallic pillars carved directly from the bedrock, ringed around a carved anthropomorphic head that watches over the space; another structure has stepped benches around a fallen central pillar. The buildings differ, but the labour investment is uniform: enormous.
Animal bone assemblages from Göbekli — dominated by gazelle, aurochs, and wild ass, with cut marks consistent with mass butchery — suggest large communal feasts. The current best guess is that the enclosures were built during periodic gatherings, perhaps once a year, when dispersed hunter-gatherer bands converged on the ridge, worked stone for weeks, ate together, and left. Over generations, the enclosures grew.
What lies underneath
Göbekli Tepe has only been partly excavated. Geophysical surveys using ground-penetrating radar and magnetometry have identified at least 15 more ringed enclosures still buried beneath the mound, and recent subsurface scanning at related Taş Tepeler sites has detected comparable structures across the region. Only about 5 per cent of Göbekli itself has been dug. Whatever the builders were doing, they were doing it on a scale still largely underground.
They were also, deliberately, burying it. Each enclosure at Göbekli was eventually filled — packed with rubble, animal bone, and flint debris — and sealed. Mehmet Özdoğan of Istanbul University has argued that this intentional burial marked the end of a structure’s ritual life, a decommissioning rather than an abandonment. Whole buildings were pensioned off. The pillars, some of them the tallest objects the builders had ever seen, went back under the earth their descendants would farm.
Why 9600 BCE, and not later
The most durable puzzle at Göbekli is not how but when. The dates — established by radiocarbon on organic material in the fill and on plaster fragments — put the earliest enclosures firmly in the Pre-Pottery Neolithic A, before the domestication of wheat, barley, sheep, or goats in this part of Anatolia. Klaus Schmidt argued, before his death in 2014, that the site itself may have driven the transition to agriculture: the need to feed the construction crews forced the intensive management of wild cereals, which eventually tipped into cultivation. Others, including researchers profiled in Archaeology Magazine’s 2021 overview of the site, are more cautious, pointing out that the earliest domesticated grains appear in the region several centuries after the first enclosures went up.
Either way, the sequence is now reversed from the textbook version. Monumental architecture came first. Agriculture, villages, and pottery followed. The T-shaped pillars were not a product of civilisation. They were part of what produced it.
Recent editorial work at Science Blog has looked at how much of prehistoric human behaviour turns out to be more sophisticated than the archaeological consensus assumed a generation ago — from evidence of deep-time astronomical observation to the surprising social complexity of pre-agricultural bands. Göbekli sits at the extreme end of that revision. The people who raised it knew how to plan geometrically, quarry, carve, transport, and coordinate hundreds of workers thousands of years before any of the standard prerequisites for such feats — cities, kings, writing, metal — existed.
What a hand and a piece of flint can do
Stand today in Enclosure D, at the centre of the ring, and the two tallest pillars flank you at head height and above. Their T-shaped tops are the stylised heads of anthropomorphic figures — you can see the arms carved along the shafts, hands meeting at the belt, a fox pelt hanging from the loincloth on Pillar 18. They have been standing there, or lying buried where they fell, for about 11,500 years. The people who set them into their sockets did it with rope, timber, muscle, and a very good understanding of levers.
Every stone at Göbekli was moved by someone who did not know what a wheel was. The Great Pyramid, when it went up, was closer in time to the first iPhone than to the day the last worker at Göbekli Tepe piled a final basket of rubble against the buried head of a limestone giant and walked home.