Göbekli Tepe

The mound sits on a limestone plateau with no spring on the hill; water came from rain, caught in cisterns cut into the bedrock⁠[11]⁠[6]. Schmidt distinguished three layers. Layer III, the oldest, holds the large circular enclosures, lettered A to D in the main excavation area and others found by geophysics across the mound. Layer II holds smaller rectangular buildings with one or two short pillars or none. Layer I is the plough soil⁠[1]⁠[2].

Enclosure D is the best preserved: an oval about 20 m across, eleven pillars in its wall and two central pillars (P18 and P31) about 5.5 m tall, each with arms, hands, a belt and a fox-pelt loincloth carved in relief⁠[2]. Pillar 43, the "Vulture Stone", in the same enclosure, carries vultures, a scorpion, a headless man and rows of V marks and boxes, and is the most argued-over object at the site⁠[12]⁠[13]. The pillars were quarried from the plateau around the mound, where unfinished pillars still lie⁠[2].

The buildings stand within a settlement. The DAI team has documented rectangular domestic structures, rock-cut cisterns and channels for collecting rainwater, hearths, and more than 7,000 grinding stones, mortars and stone vessels on the mound⁠[4]⁠[6]. The special buildings were repaired and rebuilt after damage, some of it apparently from earthquakes⁠[14].

Dating rests on radiocarbon. A sample of wall plaster from the ring wall of Enclosure D (KIA-44149) gave 9984 ± 42 BP, 9745–9314 cal BC at 95.4%; charcoal from the same enclosure clusters at 9664–9311 cal BC. A humus horizon over Layer II gave 8241–7795 cal BC as a limit for that layer. The team concluded that the Layer III enclosures were not all contemporary: D was built around the mid-10th millennium, C may be younger, and A younger still⁠[3]. Earlier dates on carbonate crusts that formed on the stones after burial only say when the stone was already covered⁠[21]. The centre points of enclosures B, C and D form a near-equilateral triangle, which Haklay and Gopher read as evidence they were laid out together⁠[22]; this sits uneasily with the staggered radiocarbon dates and is not resolved.

The sequence of use is now understood as longer and more complicated than Schmidt described. The buildings were modified, pillars were moved and re-used, walls were rebuilt, and the fills formed over time rather than in one burial event⁠[6]⁠[8]. Karul reports that Karahan Tepe was settled from about 9400 BC into the 8000s BC, with circular buildings first, then rounded rectangles, then rectangles with sharp corners⁠[10].

The economy was hunting and gathering throughout the main period. The absence of domesticates and the scale of wild cereal processing make the site central to the question of how farming began in the Fertile Crescent⁠[4]⁠[7].

Schmidt's reading is that the enclosures were open-air sanctuaries built and used by mobile hunter-gatherers who lived elsewhere, that the central pillars are supernatural beings, and that the need to feed the builders pushed people toward cultivation⁠[2]⁠[15]. It predicts no houses, hearths or burials on the mound and a wide catchment of visitors.

Banning's reading is that the enclosures were roofed communal houses: the lithics, bone and grinding tools are what a settlement leaves, and the Pre-Pottery Neolithic did not separate sacred from domestic space⁠[5]. The excavation team replied that the pillar buildings lack the hearths, floors and installations of houses and are far larger than any dwelling⁠[23].

The current DAI position lies between them: a permanent settlement of hunter-gatherers whose monumental buildings were communal and ritual structures, roofed, repaired and used for generations, within a wider network of similar settlements now being excavated under Taş Tepeler⁠[6]⁠[7]. The cisterns, domestic buildings, grinding stones and now the burials support it⁠[4]⁠[8]. Pillars inside an ordinary house at Karahan Tepe, one carrying a face, show the same symbolism in a domestic setting⁠[19].

Astronomical readings treat the reliefs as star maps or calendars. Sweatman and Tsikritsis read Pillar 43 as a record of a comet impact dated to about 10950 BC⁠[12]; Sweatman later counted V marks on the same pillar as a 365-day calendar⁠[24]. The excavators object that the pillar's date is 700 to 1,000 years younger than the proposed event, that pillars were moved and re-used, that the buildings were probably roofed, that there is no reason to assume modern constellations, and that the reading uses a few pillars out of more than sixty⁠[13]. No peer-reviewed archaeological study supports the calendar reading.

Claims that the site was built by a lost civilisation or records the biblical Eden have no evidential basis; the technology is Neolithic flint and stone, the dates are secure, and there is a 9,000-year gap to any written text⁠[13].

The temple-versus-settlement question is now mostly about degree and would be settled by publishing the full plan and stratigraphy of the mound: how many domestic buildings, how they relate in time to the pillar buildings, and whether people lived here year-round. Seasonality studies on animal teeth and plant remains, and strontium and oxygen isotopes on the new burials, would show whether the dead grew up locally⁠[8]⁠[6].

The chronology would be settled by many more short-lived samples from secure construction contexts across all enclosures, published as a Bayesian model, so that the order of A, B, C and D and the relation to Karahan Tepe are no longer inferred from a few plaster and charcoal dates⁠[3]⁠[10].

The astronomical readings would be tested by a blind study of the whole corpus of reliefs, not a chosen few, with the pillars' original positions and roofing taken into account⁠[13]. Taş Tepeler's monographs for Karahan Tepe, Sayburç and the other sites, still largely unpublished, are the single biggest gap⁠[25].