Çatalhöyük Lost Knowledge and Anomalies

About Çatalhöyük Lost Knowledge and Anomalies

If the building Mellaart called Shrine VII.21 had been excavated by anyone but James Mellaart, the leopard-and-volcano interpretation would never have entered the archaeological record. The painting on the north wall — a red-orange irregular shape above a row of nested boxes — was published in 1964 as the world's oldest landscape, the twin peaks of Hasan Dağ erupting over a town. That reading has hardened into Çatalhöyük's most-circulated image. A second reading existed from the start: the irregular shape is a leopard pelt, the boxes are a stylized geometric panel, and the volcano was never there. Both readings stay live. The mural is only one of four anomalies on this site that no excavation season has resolved, and that the closure of active digging in 2017 has now likely sealed.

## The painting that won't resolve

The Hasan Dağ mural sits behind glass at the Museum of Anatolian Civilizations in Ankara. Its assigned date is roughly 6600 BCE, mid-Level VII of the East Mound. Mellaart published it twice — in *Anatolian Studies* in 1964 and again in his 1967 monograph *Çatal Hüyük: A Neolithic Town in Anatolia* — as a landscape depicting Hasan Dağ in eruption above the village. The argument was simple. The shape has two peaks. Hasan Dağ has two peaks. Hasan Dağ is visible from the mound on a clear day. The dotted spray rising from the upper edge reads as ash or pyroclastic flow. The grid below reads as the rooftops of Çatalhöyük itself, viewed from above — which would, if accepted, make the painting both the oldest landscape and the oldest map.

The alternative reading has been on the table almost as long. Stephanie Meece, in a 2006 *Anatolian Studies* paper, argued the grid is a panel of geometric ornament common in Çatalhöyük's repertoire and the irregular blob above it is a leopard pelt, the spotted carnivore skin that recurs across the site's iconography on figurines, plaster reliefs, and other paintings. In Meece's reading, no landscape is depicted. The image belongs to the same symbolic system as the leopard reliefs in Building 80 and the seated figure flanked by felines from Shrine A.II.1 — the painting is a cosmological emblem, not a window onto a horizon.

The interpretive deadlock is genuine because the visual evidence works in both directions. Pelts in Çatalhöyük iconography are spotted; the mural shape is dotted. Hasan Dağ has twin peaks; pelts spread on a wall produce two lobes. The dating is consistent with a known eruption window — a separate question handled on the alignments page — but the iconographic question is older than the geological one and not settled by it. Meece's leopard reading does not require the painting to depict anything beyond the wall it sits on. Mellaart's landscape reading requires the painting to do something nothing else in the site's visual culture is known to do: represent the world outside the village. No third interpretation has displaced either. The painting is now the most-reproduced single image from Neolithic Anatolia, and what it shows remains undecidable on iconographic grounds alone. That is the anomaly. A site this well-published should not have an unresolvable image at its center.

What is striking is how cleanly the deadlock holds even after the geological work. The Schmitt et al. 2014 *PLOS ONE* paper, which dated zircons from a Hasan Dağ pumice flow to the right millennium for the painting, removed one objection to the volcano reading — namely that no eruption was known to be near in time. It did not establish the painting as a depiction of that eruption. A leopard pelt painted in the same window of years would be equally consistent with the geology. The geology is necessary for the volcano reading but not sufficient. The iconographic question — whether the people who painted that wall meant a mountain or a pelt — is the same question it was in 1964. No subsequent decade of work has narrowed it.

## Bucrania: three readings, no winner

The second deadlock is structural to the site itself. Çatalhöyük houses regularly held bucrania — the horn cores and frontal bones of wild aurochs, embedded in benches, plastered into walls, and hung above thresholds. Some buildings carried more than one set. Some carried whole rows. The horns came from animals slaughtered in feasting events sometimes generations before the house was built; isotope work has shown the cattle came from the surrounding plains, not the mound itself. What the bucrania *meant* is the question every successive director of the site has had to take a position on, and three positions are now on record without one displacing the others.

The first reading is Mellaart's. In *Çatal Hüyük* (1967), bucrania are emblems of a male principle subordinate to a female principal deity — the bull as the son and consort of the goddess he reconstructed from the seated figurines and the splayed-figure reliefs. The figure on the leopard throne, in this reading, is the Mother; the bucrania on the walls are her offspring. The framework drew on James Frazer and on Marija Gimbutas's developing thesis on Neolithic goddess religion, and it gave the site its popular identity as a goddess-worshipping town for the next four decades.

The second reading came from Dorothy Cameron in the 1980s. Cameron, working from anatomical comparison rather than iconography, argued that the splayed reliefs Mellaart had read as goddesses giving birth above bucrania were instead anatomical depictions of the female reproductive system — uterus, fallopian tubes, ovaries — with the bucranium reading as the same form, horns as tubes, skull as uterus. In her framing, the wall was teaching reproductive anatomy. This argument has stayed in the literature without becoming the site's canonical reading. It is cited in obstetric history papers more than in archaeological monographs, but the visual fit is real enough that no one has decisively closed it.

The third reading belongs to Ian Hodder and the Çatalhöyük Research Project, which directed the site from 1993 through the close of fieldwork in 2017. Hodder's *Religion in the Emergence of Civilization* (Cambridge, 2010) and *Entangled* (Wiley-Blackwell, 2012) develop the bucrania as memory anchors — features attached to specific houses to mark continuity with ancestral feasts and with the living dead beneath the floor. Hodder's framing folds bucrania into his "history houses" thesis: certain buildings were rebuilt repeatedly on the same footprint, accumulated more burials, and carried more elaborate installations, and bucrania were one of the markers that flagged a building as one of these continuity nodes.

No one of the three readings has won. Cameron's anatomy argument and Mellaart's goddess argument and Hodder's memory argument all sit on top of the same horns. The symbolic grammar that linked horn to meaning at Çatalhöyük does not survive in any text, any descendant tradition, or any cross-cultural parallel close enough to settle the question. A reader looking for which is correct will not find it in the published record.

What the three readings share is that each is internally coherent. Mellaart's Mother-and-son framing accounts for the placement of bucrania near splayed-figure reliefs and for the regular pairing of cattle horns with the female figurines recovered from house contexts. Cameron's anatomical reading accounts for the visual symmetry between the bucranium and the splayed reliefs, with their concave-bowl-and-flanking-curves geometry. Hodder's memory-anchor reading accounts for why bucrania concentrate in particular buildings rather than appearing uniformly across all houses, and for why those particular buildings show longer rebuild sequences and higher burial counts. Each reading explains something the other two underexplain. None explains everything, and the evidence does not select among them. The Çatalhöyük Research Project's published position by the close of fieldwork was that Hodder's framing was the one consistent with the broader pattern of the site, but the project did not claim to have refuted Mellaart's or Cameron's readings — only to have offered a frame that did more work across more of the data. That is a position worth holding accurately. It is not a verdict.

## The 741 Skeletons and the Kinship Surprise

Çatalhöyük's third anomaly is buried under its floors. The site's burial pattern is one of its most-cited features: subfloor primary inhumations, the deceased flexed and wrapped and placed in pits cut through the plaster of occupied houses, the floor replastered over them, life continuing above. At least 741 individuals have been excavated from stratified Neolithic contexts at Çatalhöyük between 1993 and 2017, with 471 of those representing primary sub-floor inhumations within houses (the remainder include secondary deposits, scattered remains, and outdoor contexts). Some buildings carried ten or more burials and a small number carried up to thirty. Mary Pearson and the human remains team — Christopher Knüsel, Scott Haddow, Joshua Sadvari, and others under the project's bio-archaeology program — built the standing skeletal corpus that every subsequent genetic and isotopic study has drawn from.

The assumption riding underneath this pattern was that the people buried in a house were the household. Sub-floor burial reads as ancestor-anchoring; ancestors are usually kin; the people sharing a floor were therefore an extended family. That assumption ran the interpretive frame for almost forty years.

In 2021, Reyhan Yaka and colleagues at Middle East Technical University, the Max Planck Institute for the Science of Human History, and the University of Copenhagen published *Variable Kinship Patterns in Neolithic Anatolia Revealed by Ancient Genomes* in *Current Biology* (vol. 31, issue 11, DOI 10.1016/j.cub.2021.03.050). Working from sixty sampled bones, the team produced 22 new genomes from Çatalhöyük and Aşıklı Höyük, integrated with previously published genomes for a 59-genome dataset across four Neolithic sites (Boncuklu, Aşıklı, Çatalhöyük, Barcın). Preservation in the Anatolian climate is brutal, and the usable yield was a fraction of the bones screened. The Çatalhöyük subset focused on subadults interred within and around houses. Across the pairs they could test, close biological kin were rare. People sharing a floor at Çatalhöyük were not, by and large, parent and child or full siblings.

The direction of the inference matters. The study did not show that no one buried in a house was related to anyone else buried in that house. It showed that the rate of close biological kinship among co-buried individuals was lower than the model of "the household buries its own" predicts, and lower than the rate seen at Aşıklı Höyük two thousand years earlier on the same plain. Whatever bound a Çatalhöyük household was something other than parentage. Adoption, fosterage, ritual co-residence, or membership in a corporate group defined by something the genome cannot read all become live possibilities. The interpretive frame for sub-floor burial — what it meant, who was eligible to be placed there, who decided — has to be rebuilt around that finding rather than around the older assumption.

It is one of the rare cases in the site's literature where a single paper genuinely closed a door. The "family house" reading of Çatalhöyük's burial pattern, taken for granted from Mellaart through the early Hodder years, is no longer supportable in its strong form. What replaces it is open.

The finding is also worth holding alongside the Pilloud and Larsen 2011 *American Journal of Physical Anthropology* dental-morphology study, which had already pointed in the same direction a decade earlier. Pilloud worked from non-metric dental traits — the genetic signal expressed phenotypically in tooth crown morphology — and found that buried co-residents at Çatalhöyük were less biologically similar than the household model predicted. The 2021 genomic paper confirmed genomically what the teeth had already suggested phenotypically: the people sharing a floor were not, predominantly, biological kin. Two independent lines of evidence, separated by ten years and a different methodology, converge on the same answer. That convergence is what gives the closure its weight. The strong form of the family-house reading was wrong, and the correction does not depend on a single dataset.

## The settlement with no streets

The fourth anomaly is the shape of the place itself. Çatalhöyük East at its peak — the Middle phase, roughly 6700-6500 BCE — held somewhere between three thousand and eight thousand people, depending on which residential-density estimate is used. (Kuijt's 2024 reassessment in the *Journal of Anthropological Archaeology* argues for a much smaller average occupancy of six to eight hundred at any one moment, with the higher figure representing maximum capacity rather than typical population. The range stays open.) Whatever the headcount, the shape is the anomaly. The houses pack against each other with no gaps. There are no streets. There are no public squares. There are no formal doorways at ground level. People entered their homes through holes in the roof, climbed down a ladder set against the south wall above the hearth, and lived inside a mudbrick honeycomb whose only exterior surface was its rooftops.

Ian Hodder, in *Entangled* (2012) and across the Çatalhöyük Research Project's published volumes through 2017, framed this as a deliberate organizing principle rather than a stage in a developmental sequence. The site held that form for more than a thousand years. Houses were rebuilt on the same footprints. New houses extruded outward but retained the rooftop-entry, no-street pattern. The architecture was stable across generations because the social form it carried was stable across generations.

That social form had no successors. When occupation shifted to the West Mound after roughly 6000 BCE — Çatalhöyük West, the Early Chalcolithic phase — houses became larger, more independent, often two-storey; sub-floor burial disappeared; mural painting gave way to decorated pottery; the honeycomb broke up. Some alleyways and sector boundaries appear in upper levels, but rooftop entry continued in many West Mound contexts — there was no wholesale shift to streets and doors. The same site, the same plain, the same population sequence: a different urban grammar within a few centuries. By the time the cities of southern Mesopotamia rose two thousand years later, no one was building Çatalhöyük-style settlements anywhere. The pattern that organized the largest known Neolithic town in the Near East for a millennium produced no architectural lineage.

The rooftop-entry pattern carries social weight that needs naming. Approach to a Çatalhöyük house went up before it went down. A visitor crossed the rooftops of the visitor's neighbors to reach a particular ladder. The communal surface was the roof; the private space was the room below. Privacy as later urbanism understands it — a wall between the household and the street — was structurally absent. Surveillance was structurally constant. Whoever was on the rooftops saw who came and went from which ladder. Whatever the form rewarded socially must have rewarded a community willing to be visible to itself, with the dead under the floor and the neighbors overhead. The Larsen et al. 2019 *PNAS* synthesis of the project's bio-archaeology found rising rates of interpersonal violence in the later phases of the East Mound — cranial injuries from blunt-force impacts on the increase in the last few centuries before the West Mound shift. The closeness eventually frayed. But it had held for a millennium first.

This is the deepest of the four anomalies because it is the one that touches the question of what was being preserved by all the rest. The bucrania, the wall paintings, the sub-floor burials, the rooftop entries — all of them held together inside a single way of organizing settled human life. That way did not transmit. Whatever the people of Çatalhöyük understood about how to live packed against each other without streets and without privacy, they did not pass it on in a form recoverable from later sites. The town is not an early version of something that came after. It is an experiment that closed. What replaced it on the same plain — the West Mound's streets and ground-level doors — is recognizable as the kind of urbanism every later Near Eastern town used. The honeycomb is not.

## What Çatalhöyük refuses to tell us

Four anomalies, no resolutions. A painting that reads as either landscape or pelt and stays both. Bucrania with three serious interpretations and no decisive evidence among them. A burial pattern whose underlying social logic the ancient DNA has just shown was not what everyone thought it was. A settlement form that organized a thousand years of life and produced no descendants. The site has been excavated almost continuously since 1958, has hosted some of the most sophisticated bio-archaeology and genomics programs anywhere in the Neolithic world, and has now, with the close of active fieldwork in 2017 under the Çatalhöyük Research Project, likely been read as carefully as it will be in this generation. (Active fieldwork closed in 2017; the project formally wound down in 2018, with management transferred to Turkish institutions thereafter.) The anomalies are what remains. They are not gaps awaiting more data. They are the structural record of a way of life that produced no writing, no decipherable iconographic system, and no direct heirs — and that the spade and the genome together can describe but not translate.

Significance

Çatalhöyük's anomalies matter because they set the limit on what a fully excavated, well-published Neolithic town can tell its successors. The site has had nearly seven decades of continuous attention — Mellaart from 1958 to 1965, the Çatalhöyük Research Project under Ian Hodder from 1993 through the close of fieldwork in 2017, multiple independent specialist programs threaded through both. It has produced one of the densest Neolithic publication records in the world: a dozen monograph volumes, hundreds of journal papers, an open-access archive of season reports, full skeletal and genomic datasets. By any external measure it is the best-documented Neolithic town on the planet. And four of its central interpretive questions remain genuinely open after all of it.

The leopard-or-volcano mural shows that even a single image, well-preserved and repeatedly studied, will not yield a decisive reading when the symbolic system that produced it has died. Mellaart's eruption-and-village interpretation and Meece's leopard-pelt-and-pattern interpretation cannot both be right, and neither can be falsified from the painting alone. That is the shape of much of what survives from before writing.

The bucrania question demonstrates the same problem at structural scale. Goddess-cult, anatomy diagram, memory anchor — three readings hold the same evidence in three frames that do not converge. The frames track the interpretive priorities of the decades that produced them more than they track the Neolithic mind. There is no neutral standpoint from which to choose.

The Yaka et al. 2021 *Current Biology* paper is the rarer thing: an actual closure. The default assumption that sub-floor burials at Çatalhöyük were the household burying its own kin is no longer supportable. What replaces it stays open, and its openness changes the reading of every prior site report that took kinship for granted. Genomics has the power to break old frames cleanly while leaving the new frame underspecified.

The no-streets pattern matters because it is the one anomaly that touches the question of cultural transmission. A way of organizing settled life that produced no descendants is a reminder that what survives from antiquity is not the full set of solutions humans found to the problem of living together. It is the small subset that happened to propagate. Çatalhöyük's particular grammar — rooftop entry, dense pack, sub-floor dead, bucrania on the walls — held one community together for more than a thousand years and then ended. The library reads that ending as data, not as failure: evidence that the human archive is partial by design.

Connections

Direct site context. Start with the parent page Çatalhöyük for the basic chronology — Pre-Pottery Neolithic and Early Pottery Neolithic phases, roughly 7400 to 6000 BCE on the East Mound — and for the framing of bucrania, sub-floor burials, the no-streets layout, and the Mellaart and Hodder excavation sequences. Two companion sub-pages handle adjacent angles. Çatalhöyük Astronomical Alignments covers the Hasan Dağ mural's volcano-eruption interpretation in detail, the Schmitt et al. 2014 PLOS ONE eruption-dating paper, and the archaeoastronomy programs that have looked for solar or stellar alignments in the buildings and found no consistent pattern. Çatalhöyük Comparisons to Other Sites sets the town against its Pre-Pottery Neolithic peers across Anatolia and the Levant, runs the dense-urbanism comparison, contrasts the burial pattern with Newgrange and other passage-grave traditions, and unpacks the goddess debate around Mellaart, Gimbutas, and the later Hodder critique of both.

Regional Anatolian neighbors. The Tas Tepeler sites are the closest relevant context for what Çatalhöyük's symbolic vocabulary may have grown out of, even though they predate the town by roughly two thousand years and sit several hundred kilometers to the east. Göbekli Tepe carries the same pattern of figurative carvings whose meaning died with their last fluent speakers — the T-pillars, the headless man under the vulture on Pillar 43, the foxes and scorpions and snakes whose symbolic grammar is now unreadable. The kinship between Göbekli Tepe's figurative work and Çatalhöyük's wall paintings is not one of direct descent but of a shared regional problem: how to read iconography without the texts. Karahan Tepe extends that problem with the standing pillared chamber and the carved heads emerging from bedrock — figurative work even harder to read than Göbekli Tepe's, in a sister site that the same DAI team is still excavating.

Later Neolithic peer. The Megalithic Temples of Malta sit at the other end of the Neolithic — roughly 3600-2500 BCE — and show what a different community did with the same problems Çatalhöyük faced: how to anchor the dead, how to mark continuity across generations, how to organize ritual space. The Maltese answer (free-standing megalithic architecture, hypogeum burial, distinct figurine tradition) is so different from the Çatalhöyük answer that the contrast itself is informative. Both communities produced no direct architectural descendants, and both left iconographic systems no later text decoded.

Further Reading

• **Primary excavation reports and synthesis.**
• Ian Hodder, ed., *Çatalhöyük Excavations: The 2009-2017 Seasons*, Çatalhöyük Research Project Series Volume 12 (London: British Institute at Ankara, 2021). The closing volume of the project's published excavation record. Covers the final field seasons, the wrap-up of Building 132 and adjacent spaces, and the synthesis chapters that summarize what the 1993-2017 program established.
• Ian Hodder, *Entangled: An Archaeology of the Relationships between Humans and Things* (Wiley-Blackwell, 2012). Hodder's theoretical synthesis, with extended treatment of Çatalhöyük as the central worked example. The history-houses framing and the bucrania-as-memory-anchors reading sit here.
• Ian Hodder, ed., *Religion in the Emergence of Civilization: Çatalhöyük as a Case Study* (Cambridge University Press, 2010). The cross-disciplinary volume that consolidated the project's position on Çatalhöyük's symbolic system, with contributions from religious studies and cognitive archaeology alongside the excavation team.
• James Mellaart, *Çatal Hüyük: A Neolithic Town in Anatolia* (Thames & Hudson, 1967). The original popular synthesis that established the goddess-cult reading and published the Hasan Dağ mural as a landscape. Now read primarily as a historical document — many of its interpretations have been superseded — but indispensable for understanding how the site entered the literature.
• **The kinship paper.**
• Reyhan Yaka, Igor Mapelli, Damla Kaptan, et al., "Variable Kinship Patterns in Neolithic Anatolia Revealed by Ancient Genomes," *Current Biology* 31, no. 11 (June 2021): 2455-2468.e18, DOI 10.1016/j.cub.2021.03.050. Sixty bones scanned across Çatalhöyük and Aşıklı Höyük; usable genome-wide data from twenty-two individuals; the paper that broke the family-house assumption.
• **Skeletal and isotopic context.**
• Marin Pilloud and Clark Spencer Larsen, "'Official' and 'Practical' Kin: Inferring Social and Community Structure from Dental Phenotype at Neolithic Çatalhöyük," *American Journal of Physical Anthropology* 145, no. 4 (August 2011): 519-530. The dental-morphology study that anticipated the 2021 genomic finding — buried co-residents at Çatalhöyük showed less biological similarity than the household model predicted, on phenotypic grounds, a decade before the aDNA could confirm it.
• Clark Spencer Larsen, Christopher J. Knüsel, Scott D. Haddow, et al., "Bioarchaeology of Neolithic Çatalhöyük Reveals Fundamental Transitions in Health, Mobility, and Lifestyle in Early Farmers," *Proceedings of the National Academy of Sciences* 116, no. 26 (June 2019): 12615-12623. The synthesis paper from the project's bio-archaeology team — health markers, dental wear, residential mobility, evidence of interpersonal violence in the later phases.
• **The Hasan Dağ debate.**
• Stephanie Meece, "A Bird's-Eye View — of a Leopard's Spots: The Çatalhöyük 'Map' and the Development of Cartographic Representation in Prehistory," *Anatolian Studies* 56 (2006): 1-16. The leopard-pelt counter-reading of the mural Mellaart published as a landscape.
• Axel K. Schmitt, Martin Danišík, Erkan Aydar, et al., "Identifying the Volcanic Eruption Depicted in a Neolithic Painting at Çatalhöyük, Central Anatolia, Turkey," *PLOS ONE* 9, no. 1 (January 2014): e84711. The geological work that re-opened the volcano reading by dating a Hasan Dağ eruption to the right window. Detailed treatment of this paper sits on the alignments sub-page.
• **Online archive.**
• Çatalhöyük Research Project Archive Reports (catalhoyuk.com, 1993-2017). Open-access season-by-season reports from every specialist team — human remains, faunal, archaeobotany, ceramics, lithics, figurines, paintings. The granular record underneath the published monographs.