Karahan Tepe Comparisons to Other Sites

About Karahan Tepe Comparisons to Other Sites

When archaeologists place Karahan Tepe alongside its peers in the published ancient-sites corpus, the comparison frame breaks before any axis is drawn. Every other ceremonial complex on the list — Eridu, Caral, Çatalhöyük, the Megalithic Temples of Malta, Stonehenge, the Great Pyramid — is a post-agricultural monument, built by people whose food supply was already domesticated. Karahan Tepe is older than agriculture in this region. Its eleven phallic stelae carved from bedrock in the Pillars Shrine (Structure AB) — ten worked directly out of the floor, plus one free-standing serpent-form upright — its eighteen T-pillars in Structure AD's Great Ellipse, and its carved human head emerging from the western wall were shaped by foragers who had not yet planted grain. Necmi Karul, the Istanbul University prehistorian who has directed full-scale excavations at the site since 2019 (Karul, "Buried Buildings at Pre-Pottery Neolithic Karahantepe," Türk Arkeoloji ve Etnografya Dergisi 82, 2021, pp. 21-31), has called the buildings "intentionally filled in," and that fact alone — monumental architecture buried by its builders — puts the entire comparison terrain on its head. Most of the corpus tells the story of cities that became ruins. Karahan Tepe is a ritual landscape that was deliberately killed, a millennium before there were cities at all.

The five comparison axes below try to take that asymmetry seriously rather than smoothing it over. Karahan Tepe shares meaningful features with a few specific sites — its sister sites in the Tas Tepeler network, the later Anatolian Neolithic at Çatalhöyük, the rock-cut traditions of Cappadocia and Petra, and the worldwide phenomenon of skull and body cults. With every other site in the corpus, the more honest comparison is contrast.

The closest real peer is Göbekli Tepe, roughly 40-60 km west of Karahan Tepe across the same limestone plateau in Sanliurfa Province, southeastern Turkey. The two sites belong to a regional grouping the Turkish Ministry of Culture and Tourism formally designated Tas Tepeler ("the stone hills") in its 2021 Şanlıurfa Neolithic Research Project — a network of approximately twelve Pre-Pottery Neolithic sites including Göbekli Tepe, Karahan Tepe, Sefer Tepe, Sayburç, Ayanlar Höyük, Hamzan Tepe, Taşlı Tepe, Kurt Tepesi, Harbetsuvan Tepesi, Çakmaktepe, Yeni Mahalle, and the now-submerged Nevalı Çori. The dates fall in roughly the same window: Göbekli Tepe c. 9600-8000 BCE, Karahan Tepe c. 9400 BCE for its earliest known phases. Both sites use the T-shaped pillar — vertical limestone stelae with a perpendicular cap that, in Klaus Schmidt's reading, anthropomorphises the stone — and both show systematic ritual backfilling.

The differences inside that close kinship are what make the comparison productive. Göbekli Tepe was excavated under Klaus Schmidt from 1996 until his death in 2014. Schmidt's synthesis Sie bauten die ersten Tempel appeared in 2006 (English edition Göbekli Tepe: A Stone Age Sanctuary in South-Eastern Anatolia, ex oriente, Berlin, 2012). His enclosures at Göbekli Tepe are dry-stone walled circles into which the T-pillars were set as quarried, transported elements; the pillars are placed in built rooms. Karahan Tepe, under Bahattin Çelik's 1997 discovery survey and 2000 Neo-Lithics publication (vol. 2/00, no. 3, pp. 6-8), and under Karul's 2019-onward excavations, has revealed a different building grammar. The ten in-situ pillars in Structure AB plus the freestanding serpent-form upright that completes the count of eleven were carved in situ — the bedrock was lowered around them by approximately 2.5 m, leaving the pillars rising directly out of the stone floor that birthed them. Karul's term in the 2021 paper is "buried buildings," but the more precise architectural distinction is subtractive versus additive monumentality: at Karahan Tepe the architects removed stone to release the structure, at Göbekli Tepe they assembled it.

The iconographic registers also diverge. Göbekli Tepe's most reproduced imagery is animal: the foxes, vultures, snakes, and scorpions of Pillar 43 and the high-relief boars and aurochs on Enclosures C and D. Karahan Tepe's signature imagery is anthropomorphic and somatic — the carved face cut into the western wall of the Pillars Shrine, the eleven phallic uprights, and the 2.3 m emaciated male statue holding its phallus with both hands, discovered in September 2023 (Daily Sabah, "1st painted, lifelike human sculpture found in Türkiye's Göbeklitepe, Karahantepe," September 2023; Heritage Daily, "New monumental statues discovered at Göbeklitepe and Karahantepe," October 2023). At Göbekli Tepe the human is mostly implied by the T-pillars; at Karahan Tepe the human face is physically present, looking back at whoever entered the chamber. Hugh Newman and JJ Ainsworth's December 2021 observation that winter-solstice sunrise light enters the Pillars Shrine through a porthole stone and tracks across the carved head for roughly 45 minutes is consistent with this somatic emphasis: the architecture is built around an embodied gaze rather than a pictorial program.

One important caveat: approximately 5% of Karahan Tepe has been excavated as of recent seasons. Bahattin Çelik's surface surveys mapped roughly 266 in-situ pillars across the northern and eastern slopes — most of them still under the soil. The comparison between the two sites is necessarily provisional, and the relative weight of "T-pillar enclosure tradition" versus "rock-cut chamber tradition" inside Tas Tepeler may shift as more of Karahan Tepe and its smaller siblings (Sefer Tepe, Sayburç) come out of the ground.

Çatalhöyük, roughly 800 km west of Karahan Tepe in central Anatolia and dating to c. 7100-5700 BCE (with Bayesian-modelled chronology placing the Eastern mound c. 7400-6200 BCE and later occupation continuing on the Western mound to c. 5700 BCE), is the next major Anatolian Neolithic site in the corpus. The temptation is to read it as a continuation of the Tas Tepeler tradition. Two facts complicate that reading.

First, Çatalhöyük is post-agricultural and post-pottery. The plant remains, the painted ceramics, the obsidian-trade economy, and the densely packed mud-brick houses all date from a world where domestication had been operating for one and a half to two millennia. By the time Çatalhöyük's earliest levels were laid, the people of Sanliurfa had already buried Göbekli Tepe and Karahan Tepe under controlled fill. Second, Çatalhöyük has no T-pillars and no separate ceremonial structures. Ian Hodder, who directed the Çatalhöyük Research Project from 1993 to 2018, argued in the chapter "History houses: a new interpretation of architectural elaboration at Çatalhöyük" (in I. Hodder, ed., Religion in the Emergence of Civilization: Çatalhöyük as a Case Study, Cambridge University Press, 2010, pp. 163-186, co-authored with Peter Pels) that what James Mellaart had originally classed as "shrines" are in fact especially elaborate houses — domestic spaces that became loci of accumulated memory because they were rebuilt on the same foundations and held larger numbers of sub-floor burials. There is no ritual building at Çatalhöyük that is not also a house.

This is exactly the comparison E.B. Banning extended back to Karahan Tepe's older sister site. In "So Fair a House: Göbekli Tepe and the Identification of Temples in the Pre-Pottery Neolithic of the Near East" (Current Anthropology 52, no. 5, 2011, pp. 619-660), Banning argued that the Pre-Pottery Neolithic monumentality at Göbekli Tepe is better read as elaborated domestic architecture rather than dedicated temples — a Pre-Pottery Neolithic anticipation of Hodder and Pels's history-house thesis. The Karahan Tepe excavators have not accepted that reading. Karul's 2021 paper consistently calls Structure AB and the phallic-stelae chamber "special buildings" set apart from domestic life, and the absence of obvious hearths, cooking debris, or domestic refuse in the excavated chambers points the other way. The wall-fused head of the Pillars Shrine, however, complicates Banning's case against Schmidt as much as it complicates the "history-house" reading: a chamber whose rear wall is a face is not easily described as either a temple of the gods or a house of the ancestors. It is a third kind of space.

The honest comparison axis between Karahan Tepe and Çatalhöyük is therefore a transition. By the early 7th millennium BCE, the dedicated ritual-monumental architecture of the Tas Tepeler era had given way, in this region, to elaborated domestic architecture in which the ritual was woven into the house. Whether that transition reflects deliberate substitution, slow drift, or two unrelated traditions remains open. Karul, Hodder, and Banning all read the same material; the disagreement about what Karahan Tepe is is genuine, not rhetorical.

Where the comparison genuinely breaks is the corpus's three "earliest urban" sites: Eridu in southern Mesopotamia (founded c. 5400 BCE in the early Ubaid period, with seventeen superimposed temples below the later E-Abzu of Enki), Caral in Peru's Supe Valley (Caral's monumental construction radiocarbon-dated to c. 2627-2000 BCE within the wider Norte Chico horizon of c. 3000-1800 BCE), and Mohenjo-daro in the Indus Valley (c. 2500-1900 BCE). All three are routinely cited as "the first city" in their respective traditions. All three are roughly 4,000 to 7,000 years younger than Karahan Tepe. And all three are post-agricultural by definition — they sit on top of long histories of cereal domestication, animal husbandry, and stored surplus.

The conventional explanatory model that links monumental architecture to surplus agriculture treats Eridu, Caral, and Mohenjo-daro as confirming cases. Sumer's barley fields fed the labour that built Enki's temples; the maize-and-anchovy economy of the Peruvian coast funded the platform mounds and sunken plazas of Caral; the Indus floodplain underwrote the granaries, baths, and grid streets of Mohenjo-daro. The architecture follows the food. Karahan Tepe is the dissenting case. Its builders were Pre-Pottery Neolithic A/B foragers. The earliest evidence localising einkorn-wheat domestication to Karacadağ Mountain, roughly 30 km from Karahan Tepe, comes from the DNA-fingerprinting study by Heun et al. (Science 278, no. 5341, 1997, pp. 1312-1314); the earliest morphologically domesticated einkorn in the regional archaeobotanical record (Çayönü, Cafer Höyük) calibrates to roughly 8650-7950 BCE — placing it several centuries to a millennium after Karahan Tepe's earliest known monumental phases. The architecture does not follow the food here. Either it precedes the food, or it pulls the food into being.

This is why E.B. Banning's Current Anthropology challenge matters even when one rejects its conclusions. Banning is right that calling pre-agricultural monumental buildings "temples" imports assumptions from later, agriculturally-supported civilisations that may not apply. He is also right that the absence of domestic detritus is not by itself proof of dedicated cult use — Pre-Pottery Neolithic households could have cleaned their elaborate buildings. Where the case for ritual primacy gets stronger at Karahan Tepe specifically is the deliberate burial of the structures — the buildings were filled in by their builders before the site was vacated, the same pattern Klaus Schmidt documented at Göbekli Tepe. A house that is ritually buried while it still functions as a house is a strange house. The pattern fits poorly with any economic-domestic explanation.

The contrast with Caral, where Ruth Shady's excavations from the mid-1990s onward (after her 1994 identification of the site, with systematic excavation beginning in 1996) have documented platform mounds, sunken circular plazas, and quipu cordage in a setting that is unambiguously settlement, is instructive. Caral has no buried buildings. Eridu has a continuous temple sequence rebuilt on top of itself for four millennia. Mohenjo-daro was abandoned, not killed. The repeated cycle of construction, use, ritual destruction, and burial that Schmidt and Karul both identify at Tas Tepeler sites is — so far — a phenomenon of pre-agricultural monumentality, not of post-agricultural urbanism. They were born, served their purpose, and were deliberately killed.

Karahan Tepe's distinctive subtractive technique — carving structures down into living bedrock rather than building them up out of quarried blocks — invites comparison with two later rock-cut traditions in the corpus: Derinkuyu and the Cappadocian underground cities in central Anatolia, and Petra in southern Jordan. Both share the basic strategy of removing stone to make space, and both are sometimes cited as "rock-cut traditions" alongside Karahan Tepe. The materials, the dates, and the functions tell a different story.

Derinkuyu's underground city is carved into volcanic tuff — pyroclastic ash deposits laid down by the Mount Erciyes and Mount Hasan eruptions of the late Miocene onward (roughly the last 9 million years), and compressed into a soft, workable rock that hardens on exposure to air. Bronze and iron hand tools cut tuff easily; the resulting cavities are stable enough to support multi-story carving. The earliest plausible phases at Derinkuyu are tentatively associated with the Phrygians (8th-7th century BCE) and possibly Hittites (c. 1800 BCE), with the bulk of the city excavated and expanded by Greek-speaking Christians in the Byzantine period for refuge from Arab and later Seljuk raids. The function is defensive habitation: ventilation shafts, stone rolling-doors, stable rooms, kitchens, and chapels for populations of perhaps 20,000 hiding from above-ground threats. Karahan Tepe shares neither the material (limestone, not tuff), the date (8,000-10,000 years older), nor the purpose (ceremonial, not defensive). The "underground" framing also misleads: Karahan Tepe is not subterranean. Its chambers were carved down into a ridge that is itself above the surrounding plain, and were open to the sky during use; they are now buried only because their builders filled them in.

Petra's case is closer in technique but still distant. The Nabataean rock-cut monuments — Al-Khazneh (the Treasury, early 1st c. CE under Aretas IV Philopatris), Ad Deir (the Monastery, mid-1st c. CE), and the dozens of tomb facades along the Siq — are carved into vertical sandstone cliffs rather than excavated downward into a horizontal bedrock surface. The technique is exterior facade-cutting; the purpose is funerary and commemorative; the date is roughly 8,500 years younger than Karahan Tepe. Shaher Rababeh's analysis (How Petra Was Built: An Analysis of the Construction Techniques of the Nabataean Freestanding Buildings and Rock-cut Monuments in Petra, Jordan, British Archaeological Reports International Series 1460, Archaeopress, 2005) describes a workflow in which the Nabataeans first cut a working terrace into the cliff face, then dressed the facade from the top down using metal tools — a sequence that has no analogue in the Pre-Pottery Neolithic toolkit available to the Karahan Tepe builders.

Where the rock-cut comparison does sharpen something specific is the conceptual move common to all three: the architect imagines a finished form already present inside the unworked stone and removes the rest. The remove-stone-to-find-the-form approach is rarer in human building than additive masonry, and it is striking that the earliest known instance of it at monumental scale shows up at Karahan Tepe without any preceding tradition that we know of. Petra and Derinkuyu inherit a developed Near Eastern stone-working repertoire; Karahan Tepe predates that repertoire by eight to ten millennia.

The bedrock face that closes the Pillars Shrine's western side sits inside a wider Neolithic Anatolian preoccupation with the head as a ritual focus. Two specific peer cases are well documented and worth setting against it.

Julia Gresky, Juliane Haelm, and Lee Clare published the foundational Göbekli Tepe study: "Modified human crania from Göbekli Tepe provide evidence for a new form of Neolithic skull cult" (Science Advances 3, no. 6, 28 June 2017, e1700564, doi:10.1126/sciadv.1700564). Gresky's team analysed three partially preserved human skulls from Schmidt's excavations. All three carried artificial modifications — deep grooves cut into the cranial vault with flint tools, and on the best-preserved specimen a deliberately drilled perforation. The pattern matches no known contemporaneous skull-cult practice. The authors' interpretation, stated cautiously, is that the skulls were modified either to venerate ancestors shortly after death, or to display recently dispatched enemies; the grooves may have served as channels to suspend the crania on cords. Either way, the head was the relic, and it was being worked.

Çatalhöyük's plastered-skull tradition is the better-known comparator. A plastered skull was recovered from a Çatalhöyük burial cradled in the arms of an adult female, and Hodder's project documented an extensive practice of skull retrieval, secondary burial, and curation across the site's Neolithic levels. Plastering reconstructs the soft tissue of the face directly onto the bone, producing a portrait of the deceased that is then re-buried, displayed, or passed down. Across the wider Pre-Pottery Neolithic B Levant — Jericho, 'Ain Ghazal, Tell Aswad, Tepecik-Çiftlik — plastered skulls are interpreted (cautiously, since written sources do not exist) as evidence of an ancestor cult woven into the household.

The Karahan Tepe carved head is a different object in this same constellation. It is not a worked human skull. It is a sculpted limestone head, integrally part of the chamber wall, facing the entrance. The functional logic, however, is the same: a face is made the focal point of a ritual space, and the architecture is organised around it. What Karahan Tepe substitutes for the post-mortem cranium of Göbekli Tepe and the plastered skull of Çatalhöyük is a permanent, immobile, stone-carved equivalent — a face that cannot be moved, lost, or curated, only entered into. This may be the single most distinctive feature of Karahan Tepe inside the broader Anatolian Neolithic skull-cult sphere. The head is fused to the building. The building is fused to the bedrock. The bedrock is fused to the hill. The relic is no longer separable from the place.

Hugh Newman has linked the bearded 2.3 m male statue from Karahan Tepe (discovered in the September 2023 season) to the same somatic-ritual register, noting the figure's emaciation, the explicit phallic gesture, and the squared-off chin extension that resembles much later Egyptian pharaonic false beards. The transmission claim is speculative and depends on chains of inheritance that cannot be archaeologically demonstrated; the more conservative reading is that emaciation and beard styling are independently invented signals of ritual specialism in widely separated traditions. The comparison is worth keeping on the table because of how unusually anthropomorphic Karahan Tepe's iconography is by Pre-Pottery Neolithic standards, not because the Egyptian connection has been proven.

Karahan Tepe sits in an awkward but specific place in the corpus. With Göbekli Tepe it is genuinely a sister site — same region, same century-bracket, same T-pillar tradition, same ritual-burial pattern, with a divergent architectural grammar and a more anthropomorphic iconography. With Çatalhöyük it is a chronological predecessor whose elaborated ritual buildings appear to have been replaced, two thousand years later in the same broad Anatolian region, by elaborated houses. With the corpus's earliest urban centres — Eridu, Caral, Mohenjo-daro — it is an outlier that breaks the agricultural-surplus model of monumentality. With the rock-cut traditions of Petra and Derinkuyu it shares a subtractive technique whose lineage we cannot trace. With Göbekli Tepe's modified crania and Çatalhöyük's plastered skulls it shares a ritual focus on the face, expressed in stone rather than bone.

None of these comparisons collapse Karahan Tepe into a familiar category. What they outline instead is a site that was already strange in its own century — pre-agricultural, subtractive, somatic, deliberately buried — and that becomes stranger as the comparison set widens. The next several decades of excavation under Karul and the Tas Tepeler project will determine which of these axes hold up and which were artefacts of the small fraction so far exposed.

Significance

Karahan Tepe matters in the comparison network because it forces every neighbouring site in the corpus to defend its place on the timeline of monumentality. The agricultural-surplus model that comfortably explains Eridu, Caral, and Mohenjo-daro cannot account for Karahan Tepe; the elaborated-house model that Hodder and Pels (2010) developed for Çatalhöyük, and that E.B. Banning (Current Anthropology 52:5, 2011, pp. 619-660) extended back to Göbekli Tepe, fits the deliberately-buried Pillars Shrine awkwardly at best. Necmi Karul's 2019-onward excavations have produced an irreducibly specific object — a chamber whose rear wall is a carved human face, whose floor was lowered into bedrock to release eleven phallic stelae, and whose builders sealed it with controlled fill before walking away — that does not fit the temple, the house, or the underground-city categories the rest of the corpus offers.

Connections

Karahan Tepe — the parent entity. This sub-page focuses on cross-site comparisons; the parent covers Karahan Tepe in standalone depth, including the phallic-stelae chamber, Newman and Ainsworth's December 2021 solstice observation, and the 2.3 m (≈7.5 ft) seated male statue discovered in 2023.

Göbekli Tepe — the closest peer in the Tas Tepeler network, c. 9600-8000 BCE, sharing the T-pillar tradition and ritual backfilling but differing in additive (built-up) versus Karahan Tepe's subtractive (carved-down) construction.

Çatalhöyük — central Anatolian Neolithic, c. 7100-5700 BCE, where Hodder and Pels's "history-house" interpretation replaced Mellaart's "shrine" reading and where ceremonial life was woven into domestic architecture rather than housed in dedicated structures.

Eridu — southern Mesopotamian Ubaid-period temple sequence (founded c. 5400 BCE), the corpus's earliest post-agricultural ceremonial complex and a clean foil to Karahan Tepe's pre-agricultural monumentality.

Caral — Norte Chico ceremonial capital in Peru's Supe Valley (Caral's monumental construction c. 2600-2000 BCE within the wider Norte Chico horizon c. 3000-1800 BCE), an Old World/New World parallel for monumentality-from-surplus that highlights Karahan Tepe's anomalous date.

Mohenjo-daro — Indus Valley urban site (c. 2500-1900 BCE), the third "earliest city" foil; abandoned without ritual burial, contrasting Karahan Tepe's deliberate decommissioning.

Derinkuyu and the Cappadocian Underground Cities — central Anatolian rock-cut tradition in volcanic tuff, used for defensive habitation rather than ceremony, and roughly 8,000+ years younger than Karahan Tepe's bedrock chambers.

Petra — Nabataean rock-cut sandstone facades in southern Jordan (mainly 1st c. BCE-2nd c. CE), the corpus's other major subtractive tradition; useful as a contrast in material, technique, and chronology.

Stonehenge — Wiltshire, c. 3000-1500 BCE, the most-cited later megalithic peer; sometimes claimed as a Karahan-Tepe descendant in popular sources, but the chronological gap and lack of T-pillar form make the architectural lineage speculative.

Megalithic Temples of Malta — c. 3600-2500 BCE, an independent Neolithic temple-building tradition that — unlike Karahan Tepe — fits the agricultural-surplus model cleanly.

Further Reading