Petra — Lost knowledge and anomalies

About Petra — Lost knowledge and anomalies

Call it "the Treasury" and you have already smuggled in a story. Al-Khazneh — the Arabic name visitors learn first — translates roughly as "the treasure" or "treasury," anchored in a Bedouin tradition that the urn at the facade's apex held a pharaoh's gold and that bullets fired into the stone might crack it open and release the hoard. The pockmarks on the urn are real; the gold was never there. The label has nonetheless stuck for two centuries of Western tourism. It is also, almost certainly, wrong about the building's purpose, and the gap between what the popular name assumes and what the archaeology actually shows is the most useful entry point into Petra's contested chambers.

The same problem repeats at scale. The "Royal Tombs" carved into the eastern cliff face are called royal because nineteenth-century travelers thought they looked the part — no inscription inside any of the four major facades names a king. "The Monastery" preserves an Arabic name (Ad Deir) for a building that may never have been monastic in any sense and that may not have been a tomb either. "The Lost City" is the third familiar phrase, and it is the most misleading of all: Petra was lost only to Europeans, not to the Bdul Bedouin families who lived in its caves continuously into the late twentieth century. Each conventional name encodes a guess that hardened into a fact, and each one hides a question the evidence has not closed.

What follows is a tour through the contested chambers — places where Petra's familiar labels and the underlying scholarly record diverge, and where honesty about what is actually known matters more than narrative tidiness.

The Treasury's facade is iconic, and its purpose is not settled. The leading hypothesis treats Al-Khazneh as a royal tomb, most likely for Aretas IV (reigned ca. 9 BCE – 40 CE), based on dating of the architectural style, parallels with other Nabataean elite tombs at Petra and at Hegra in northwest Arabia, and recent burial finds at and beneath the structure. But the facade carries iconography — Isis-like figures, dancing Amazons, eagles, lion-griffins, Dioscuri-like riders flanking the lower order — that several scholars read as temple imagery rather than as funerary ornament, and the interior rooms preserve niches whose proportions and placement are unusual for a typical Nabataean rock-cut tomb. The interior is also, by the standards of Petra's larger tomb complexes, surprisingly bare: three plain chambers, no carved sarcophagi or loculi visible, no inscriptions on the walls. Readings that propose a temple, a shrine to a deified ruler, an archive, or a ceremonial banquet space remain in active circulation, and have been defended in print as recently as the early 2020s.

Three excavations have shifted the weight of the evidence toward a funerary reading without closing the alternatives. In 1996, a small burial crypt was identified in front of the facade during routine survey work. In 2003, the Jordanian archaeologist Suleiman Farajat led an excavation that recovered eleven skeletons in a chamber on the Treasury's left flank, the first substantial burial assemblage tied to the structure. In October 2024, a joint Jordanian–American project led by ACOR Executive Director Pearce Paul Creasman, with University of St Andrews geophysicist Richard Bates conducting the ground-penetrating radar survey, in partnership with Jordan's Department of Antiquities, identified a sealed chamber directly beneath the facade and excavated it as part of the Discovery Channel program Expedition Unknown, hosted by Josh Gates. They recovered twelve articulated skeletons together with grave goods of bronze, iron, and ceramic, including a ceramic vessel found in the hand of one individual; sediment and material from the tomb enclosure dated the construction window to roughly the mid-first century BCE through the early second century CE — a window that includes the reign of Aretas IV but does not isolate it.

That 2024 find is genuinely significant. Complete early-Nabataean burials are rare — most Petra tombs were emptied, robbed, or reused over the long period of the city's decline and post-antique occupation, and articulated skeletons in primary deposit are unusual. The location beneath the most famous facade in the site reframes what a "Treasury" is, and the date range fits a royal-tomb reading. It does not, however, eliminate the temple-iconography problem on the upper register, nor does it explain the unusual interior niches, nor does it identify the occupants. Twelve burials in one sub-chamber is not the same as a single named royal interment, and the distribution of the bodies — including individuals who appear to be of varying ages — has not yet been published in detail. The current honest framing is the one most cautious scholars use in print: consensus tilts toward a royal tomb, most plausibly of Aretas IV, while preserving temple, archive, and ceremonial readings as live alternatives. A definitive identification — by inscription, by linkage of grave goods to a named ruler, by isotopic or genetic analysis of the recovered remains — has not happened yet.

In December 1993, ACOR archaeologists excavating a sixth-century Byzantine church on Petra's northern ridge — what is now called the Petra Church — found a side room filled with carbonized papyrus. The fire that destroyed the church in the early seventh century had charred roughly 140 scrolls into something close to charcoal, fragile enough that the conservators initially feared the find was unsalvageable. A Finnish team led by Jaakko Frösén began work on the texts in 1994–95, with conservation, unrolling, and digital imaging carried out over the following years at ACOR Amman.

The story of the Petra Papyri is sometimes told as if it were an ongoing enigma — a mass of unread documents waiting for a breakthrough. It is one of the most successful carbonized-papyrus conservation projects of the modern era, on the same methodological line as the Herculaneum work without the same scale or fame. The texts — Greek-language administrative and legal documents from the household archive of a sixth-century Petra resident named Theodoros, son of Obodianos — were published in five volumes between 2002 and 2018 by ACOR. Volume V, edited by Antti Arjava, Jaakko Frösén, and Jorma Kaimio, closed the series with thirty-seven new texts (numbered 50–87 in the corpus) and two re-editions of texts originally published in Volume I, twenty-five years after the discovery. ACOR held a public celebration of the final publication in November 2018.

What the papyri reveal is a working sixth-century Petra: property transfers, marriage contracts, tax records, dispute settlements, a small landed family negotiating inheritance and church dealings inside the framework of late Roman and early Byzantine provincial law. They place the city firmly inside the Roman provincial system as Palaestina Tertia, and they document a community still doing routine legal and economic business in Greek roughly two centuries after the conventional "decline" narrative would have Petra emptying out. They do not reveal a lost esoteric tradition or a hidden Nabataean library — they post-date the Nabataean kingdom by half a millennium, and the language of the archive is Greek, not Nabataean Aramaic. The triumph here is conservation and scholarship, not unsolved mystery, and the corpus is now fully available to specialists in print and through ACOR's publication channels. Anyone framing the Petra Papyri as a still-unread cache of enigmas is misrepresenting a closed scholarly project.

Petra suffered a series of severe seismic events in late antiquity. The 19 May 363 CE earthquake — the same event documented across the southern Levant by the so-called "Cyril of Jerusalem letter" and tracked in the DEADSEAQUAKE.info Cyril Quakes catalog — destroyed Qasr al-Bint, the so-called Great Temple, the Small Temple, and the Temple of the Winged Lions, along with regional sanctuaries at Khirbet et-Tannur and Khirbet edh-Dharih. The water management infrastructure that had made Petra possible as a desert city — the cisterns, channels, and dams along the Siq and across the wadi system — was significantly damaged, and Kenneth Russell's foundational 1980 paper in the Bulletin of the American Schools of Oriental Research argued that the 363 event marked the practical end of public paganism in southern Transjordan. A second major quake in 551 CE caused additional damage to what remained, and further shocks in 419 and 747 CE are recorded in the regional seismic record.

The trouble is that the two events — and to a lesser degree the later ones — are not always cleanly separable in the dirt. At several Petra structures, paleoseismic samples and associated material cluster within roughly two and a half centuries on either side of the 363 and 551 events, and post-363 reoccupation, salvage of dressed stone for reuse, and 551 collapse can fold into one another in the stratigraphic record. ACOR's work at the Temple of the Winged Lions has identified more than two thousand individual stratigraphic units across the structure, and even with that resolution, certain destruction phases cannot be confidently assigned to one quake or the other. The pottery and coin sequences typically allow a "before 363" terminus — material in primary deposit beneath the destruction layer dates no later than the mid-fourth century — but what they often cannot do is distinguish 363 collapse from 551 collapse on top of disturbed 363 layers, especially in areas where post-363 inhabitants used the ruined structure as a quarry or dump.

The Temple of the Winged Lions illustrates the issue at fine grain. The southwest quadrant of the temple precinct appears to have been used as a debris dump while the structure was being cleared after 363, with pottery in that fill consistent with mid-fourth-century use. But subsequent occupation phases on top of that dump complicate any attempt to read the deposit as a single 363 destruction event, and ACOR's published stratigraphic analysis is careful to note where the boundaries blur. The honest framing is that the 363 event was catastrophic and well-attested, the 551 event was real and added meaningfully to the destruction, and the boundary between them in the field record is sometimes blurry. Treating that blur as a settled stratigraphic narrative — or assigning every collapsed wall in Petra to 363 because 363 is the better-known earthquake — misrepresents the data.

Four monumental facades on the eastern cliff above the city center are conventionally called the Royal Tombs: from north to south, the Urn Tomb, the Silk Tomb, the Corinthian Tomb, and the Palace Tomb. The label is nineteenth-century, descriptive of scale and ornament, not derived from inscriptions. None of the four contains an interior text naming its occupant from the Nabataean period. The ascription to specific Nabataean kings rests on stylistic dating and parallels with dated facades elsewhere in Petra and at Hegra, all of which place the structures broadly in the Aretas IV (9 BCE – 40 CE) through Malichus II (40 – 70 CE) window, with no firm claim for Rabbel II (70 – 106 CE) at any of the four despite occasional speculation in the older literature.

The Urn Tomb illustrates the problem in detail. Jean Starcky, working from architectural and historical considerations, read it as the tomb of Malichus II, who died in 70 CE. Andreas Schmidt-Colinet, working from a closer architectural analysis of the facade orders and decorative program, argued for Aretas IV instead. Neither identification is firm, and both rest on stylistic chains rather than direct evidence. The tomb was repurposed as a Byzantine church in 447 CE, an event recorded in a Greek inscription on the rear wall — the only firm name attached to the structure belongs to its second life as a Christian space, not its first. The Palace Tomb, the largest and northernmost facade in the group, is even less securely attributed; its scale and the fact that its upper registers were partly built up in masonry rather than rock-cut have generated their own debates about whether it was finished, whether it was originally intended as a tomb at all, and how it relates to the Corinthian Tomb immediately beside it.

The pattern across the four facades is the same. Phrases like "traditionally called the Royal Tombs" or "attribution remains uncertain" are not hedging; they are the most accurate description of the evidence. The structures are large, they are elaborate, and they almost certainly served the Nabataean elite — what they are not is identified.

Ad Deir — "the Monastery" — sits on a high plateau a steep climb of roughly eight hundred steps above the city, its facade roughly the same scale as Al-Khazneh but cruder in detail, almost certainly later in date and likely commissioned during or after the reign of Rabbel II. The conventional reading treated it as a tomb-monument, on analogy with the larger rock-cut facades elsewhere on site. Three lines of evidence have pulled the favored interpretation toward a cult banqueting hall instead, dedicated to the deified Nabataean king Obodas.

The first is an inscription found in 1991 during cleaning of a structure near Ad Deir, recording a dedication "in memory of 'Ubaydu son of Waqihel and his associates, the symposium of Obodas the god." Obodas — most plausibly Obodas I, who reigned ca. 96–85 BCE (early first century BCE), and was deified after his death — is the Nabataean king most clearly attested as a recipient of cult, and a separate inscription at Avdat in the Negev preserves a similar dedication. The Greek loanword marzeha (rendered into Greek as symposion) names a private religious banqueting brotherhood with a defined membership, attested across the Levant in Aramaic and Greek inscriptions from the Persian period onward. The 1991 text places such a brotherhood in the immediate vicinity of Ad Deir.

The second is the interior of the Monastery itself: a single large cella, roughly square, with low benches running along the side walls. The benches suit ritual reclining banquets — the typical posture for a marzeha — and fit poorly with funerary use, where a tomb chamber would normally include carved niches, loculi, or sarcophagus emplacements. None of those features is present. The third is the leveled forecourt directly outside the facade, an unusually large open space cut and partly built up to accommodate gatherings, suitable for the assembly that a banquet implies and consistent with theatrical or processional use as well.

Some scholars still hold for a tomb-monument or a mixed function — for example, a tomb that incorporated banqueting rituals for the deceased — and the building may have been adapted as a Christian space in the Byzantine period, when the Arabic name "the Monastery" likely originated. The point is that the favored reading has shifted toward Obodas-cult banqueting hall, but the question is not closed and the alternatives are still defended in print. Once again, the popular name encodes a reading the underlying evidence does not require.

Petra is older than the Nabataeans. The summit of Umm el-Biyara, the high mesa west of the city center, preserves a 7th–6th-century BCE Edomite domestic settlement of stone-built longhouses, excavated by the British archaeologist Crystal-M. Bennett in three seasons (1960, 1963, 1965) and finally published in book form by Piotr Bienkowski in 2011 as part of the Levant Supplementary Series. The site was unwalled, its houses arranged in irregular rows along the natural contours of the mesa top, and its material culture — pottery, small finds, an Edomite seal impression — fits the late Edomite kingdom rather than the later Nabataean trading culture.

Tawilan, on the ridge above Wadi Musa east of Petra (Bennett, four seasons 1968–1970 and 1982), and Buseirah — biblical Bozrah, the Edomite regional capital, located north of Petra (Bennett, five seasons 1971–74 and 1980) — fill out the regional picture. Together they document a coherent Edomite settlement system in the 7th and 6th centuries BCE, with a hilltop refuge or domestic settlement at Petra itself, an agricultural village above the spring at Wadi Musa, and a fortified administrative center at Buseirah. Bienkowski's 1992 edited volume Early Edom and Moab is the standard regional reference.

What the Edomite layer does not show is direct settlement continuity into the Nabataean period. The Iron II material at Umm el-Biyara is sealed and ends in the 6th century BCE; the Nabataean material at Petra represents a different population settling in the same landscape after a substantial gap, with Nabataean tombs and inscriptions appearing in the 4th and 3rd centuries BCE. The honest framing is coexistence in a region with shared trade routes and gradual cultural intermixing, not a transmission of city, language, or rulership from Edom to Nabataea. The "lost knowledge" version of the story — secret Edomite techniques carried forward into Nabataean hydraulics or carving — has no archaeological support, and the temptation to invent such continuities should be resisted. What Petra preserves at Umm el-Biyara is an earlier, separate culture in the same place, not a hidden ancestor of the rose-red city.

By the early eighth century, Petra's Byzantine population had thinned, the Petra Church had burned, and the city ceased to function as an urban center. For roughly eleven centuries, no European traveler is known to have entered the site and left a written record of having done so. Crusaders built castles at Petra itself — al-Wu'ayra and al-Habis — in the twelfth century, after which the rock-cut city drops out of the European record almost entirely until the early nineteenth century.

On 22 August 1812, the Swiss explorer Johann Ludwig Burckhardt, traveling for the Association for Promoting the Discovery of the Interior Parts of Africa and disguised as an Arab Muslim pilgrim under the assumed name Sheikh Ibrahim ibn Abdallah, was traveling through Wadi Musa on his way south. He told his local guides that he had a vow to make a sacrifice at the supposed tomb of Aaron — Harun in Arabic — on Jebel Haroun, the high peak that rises west of the city. The route to Jebel Haroun runs through Petra. Burckhardt talked his way past suspicious guides, walked through the Siq, glimpsed the Treasury and the major facades on the eastern cliff, and emerged on the far side, having become the first modern European to document the ruins. He published his account in Travels in Syria and the Holy Land, posthumously, in 1822.

The phrase "lost city" travels with this story, and it is the single most overstated framing in the popular literature on Petra. The site was not lost. The Bdul Bedouin tribe — roughly 140 families by the late twentieth century — lived in Petra's caves and tombs continuously through the period of European absence, until 1985, when the Jordanian government relocated them to the new village of Umm Sayhoun on the ridge above the site, coinciding with Petra's UNESCO World Heritage inscription. Burckhardt himself recorded a substantial Bedouin presence on his arrival — he wrote of meeting people in the wadi who knew the ruins well and who guided his movement through them, however reluctantly. The Bdul oral tradition preserves a continuous association with the site that long predates 1812.

What 22 August 1812 marks is not the rediscovery of a forgotten place but the first documented European visit since late antiquity. The accurate phrasing is "lost to the West" or "rediscovered by Europeans" — never "forgotten" or "lost" in any unqualified sense. The Bdul knew exactly where Petra was; they were living in it. Writing the history of the city otherwise inverts the relationship between a continuously inhabited landscape and the much narrower history of when European travel writers happened to notice it.

Significance

Petra concentrates a particular kind of archaeological problem: a site that is enormously famous, enormously photographed, and still contested at the level of basic identifications. The question is rarely whether something exists. It is what each thing was for, who it belonged to, and which destruction layer ended its use. The popular labels — Treasury, Royal Tombs, Lost City, Monastery — are confident where the underlying record is not, and the gap between confidence and evidence is itself part of what makes Petra worth studying carefully.

The Treasury debate matters because it is a test case for how Nabataean elite religion and political authority fit together — a question the 2024 sub-Treasury burials tilt toward funerary use without closing the temple-iconography or interior-niche objections. A site that lets you watch consensus shift without forcing closure is more honest than one that pretends the answer was always known.

The Petra Papyri matter as a benchmark for what successful conservation of a fragile carbonized corpus looks like, and as a corrective to the standard "decline" narrative — sixth-century Petra was still doing routine legal and economic business in Greek under late Roman and early Byzantine administration.

The earthquake stratigraphy matters because honest field reports preserve the ambiguity between the 363 and 551 events; popular accounts collapse it. Readers who care about how archaeological knowledge is actually made benefit from seeing where the data resists tidy assignment.

The Royal Tombs and Monastery reframings matter because they show how labels become facts. "Royal Tombs" was a nineteenth-century descriptor that began to function as an identification. "Monastery" preserves a Byzantine adaptation while obscuring what may have been the building's original purpose as a banqueting hall for the cult of a deified king.

The Edomite layer matters because the temptation to invent continuities — secret Edomite knowledge, hidden carving traditions — is strong, and the evidence does not support it. Coexistence in a shared landscape is the honest framing; transmission is not.

And the lost-city framing matters most of all, because it is the place where colonial-era narrative and the lived reality of an indigenous population diverge most sharply. The Bdul never lost Petra. Burckhardt was the first European to document it, not the first human to find it. Saying so is the difference between writing the history of a place and writing the history of when Europeans noticed the place.

Connections

Petra's contested chambers do not stand alone. They sit inside a wider conversation about how monumental sites get named, who gets credit for "discovering" them, and how destruction layers and reuse phases get sorted in the field record. A few directions to follow.

Within Petra: The parent overview at Petra covers Nabataean trade networks, water management, the rock-cut facades, and the Byzantine and post-antique phases as a single integrated picture. The astronomical alignment question — whether Al-Khazneh, the Urn Tomb, and other facades encode solar or solstitial orientations, and what the methodologically careful claims actually are — sits in a separate companion piece at Petra astronomical alignments. That page disentangles the verified Bedal/Belmonte work from the enthusiast overclaims, and is the right next stop if the Treasury's iconography or the Urn Tomb's orientation interests you.

Other rock-cut and monumental traditions: Egypt's relationship to monumental rock-cut architecture is the closest comparison. Abu Simbel, with its Ramesside facade and interior chambers carved into a sandstone cliff, offers a parallel for what scale-of-ambition looks like in a different cultural setting and four centuries earlier. The Valley of the Kings — pharaonic tombs cut into a desert wadi, identified by inscription and king lists — is the inverse of Petra's Royal Tombs problem: there, attribution is firm because the texts survive; here, the texts do not. Karnak Temple and Luxor Temple show what a built temple-and-archive complex looks like at scale, and provide a useful contrast for the Treasury function debate — Egyptian temple architecture has clear iconographic and spatial signatures that the niche-and-cella debate at Al-Khazneh does not resolve.

Other contested-function sites: Baalbek raises a parallel set of questions about Roman-period reuse of older sacred ground, the boundary between temple and ceremonial complex, and the difficulty of dating extreme-scale stonework. Eridu in southern Mesopotamia, traditionally identified as the first Sumerian city, illustrates how stratigraphic reuse of a single sacred precinct over millennia produces problems of phase assignment that resemble Petra's earthquake-layer issues at a different scale.

Cities lost to the West but not lost in fact: Mohenjo-daro, the Indus Valley city excavated beginning in the 1920s, was unknown to European archaeology until the twentieth century but was a marked feature in the local landscape long before. The pattern — local knowledge preserved, Western "rediscovery" celebrated as the discovery — is the same pattern that misnames Petra as a lost city. Reading them together makes the colonial-rediscovery framing visible as a frame, not a fact.

Further Reading