Borobudur — Lost Knowledge and Anomalies

About Borobudur — Lost Knowledge and Anomalies

The hidden panels themselves are sealed; the photographs are all that remain accessible. Between 1890 and 1891, the Yogyakarta court photographer Kassian Cephas (1845-1912) produced 164 large-format glass-plate negatives of the carved andesite reliefs at the base of Borobudur — 160 panels of the Karmavibhanga sequence plus four corner orientation shots — and those plates, now archived at the Karmawibhangga Museum on the Borobudur grounds, are the only public window onto a register of the monument that no living visitor has been able to walk past since 1885 in any sustained way and that has been entirely sealed since 1973. To stand at the southeast corner of the candi today is to see four of the 160 panels exposed in a small archaeological cutaway. The remaining 156 are buried under the encasement foot, exactly where the Sailendra builders, or their immediate successors, left them more than a thousand years ago. Every reading of those 156 panels — every reproduction in a Karmavibhanga monograph, every photograph in a museum catalogue, every art-historical analysis of the cause-and-effect ethics they depict — passes through Cephas. The plates are, in the strictest sense, the record. The reliefs themselves are gone from view.

This page tracks what is genuinely missing or hidden at Borobudur, what was only colonially "lost," and what fringe claims survive close reading of the evidence. The site is unusually well-documented for a monument of its age — the Karangtengah inscription of 824 CE names the Sailendra patrons Samaratungga and Pramodhawardhani and ties the building to a Mahayana sanctuary called Jinalaya — so the kinds of "lost knowledge" that flourish around less anchored sites cannot really take root here. What survives is more interesting: a buried register of carvings, a colonial reclearing wrongly remembered as discovery, a documented modern bombing, a restoration that traded one kind of preservation for another, and a 2003-2004 expedition that proved the monument's ship reliefs are working blueprints rather than stylized art.

The anchor anomaly is the buried foot. In 1885, the Dutch engineer J. W. IJzerman, then chair of the Yogyakarta Archaeological Society, peeled back a section of the lowest visible terrace and found a complete decorated register concealed beneath the encasement that forms the present base of the monument. The hidden register, since identified as the Karmavibhanga ("differentiation of karma") cycle, runs the full perimeter of the sub-foundation in 160 panels and depicts the law of cause and effect in scenes of everyday Javanese life — markets, hunts, childbirths, lawsuits, executions, monastic offerings — paired with their hereafter consequences. It is the only narrative register at Borobudur drawn from this particular Mahayana sutra cycle, and it is the only one almost entirely sealed from view.

Five years after IJzerman's exposure, the colonial administration commissioned Cephas — a Javanese Christian convert, court photographer to the Sultan of Yogyakarta, and the first native member of the Royal Netherlands Institute of Southeast Asian and Caribbean Studies (KITLV) — to document the buried register before it was reclosed. Working between 1890 and 1891 with wet-plate equipment hauled across the temple's south flank, Cephas produced 164 plates: 160 frame-by-frame photographs of the panels and four corner reference shots. The full set survived the twentieth century; high-resolution scans circulate today through the Karmawibhangga Museum and through the Photo Dharma archive maintained by Anandajoti Bhikkhu. For most working purposes — Buddhist studies, Javanese art history, even art-historical close reading — the Cephas plates are now treated as the canonical edition of the panels.

Why is the register sealed? The question is unresolved and the debate is genuinely live. One camp argues theological intent: the Karmavibhanga depicts the realm of kamadhatu (desire), which is the lowest of the three Buddhist realms in the Borobudur cosmology, and a deliberate concealment of this register would mirror the practitioner's renunciation of desire as they ascend the upper terraces. On this reading, the encasement is a teaching, not a flaw. The other camp argues structural intent: the encasement was added — possibly within decades of the original construction — to buttress a base that was sliding outward under the weight of nine ascending terraces and a central stupa, and the Karmavibhanga panels were sealed not for spiritual pedagogy but because the engineering required mass at the foot. The BRIN (Indonesian National Research and Innovation Agency) AMERTA Journal article "Identification of Karmawibhangga Reliefs at Candi Borobudur" reviews both positions without choosing between them, and the UNESCO 1973-1983 restoration documentation maintained the southeast cutaway specifically so that future analysts could continue testing the question. Both camps have evidence; neither has won. Treat the debate as open.

The 1973-1983 UNESCO campaign reburied the four panels that had been exposed and confirmed the rest as inaccessible until and unless a future restoration justifies opening the foot. For the working life of every visitor reading this page, the Cephas plates are the access route. There is a methodological lesson in the Cephas archive worth pausing on. The 164 plates are wet-collodion glass negatives, exposed under tropical conditions on a slope still partly covered in vegetation, with light angled from the south for most of the working day. They are technically rough by the standards of later twentieth-century archaeological photography, yet in the cases of certain water-damaged panels at the southwest corner they remain more legible than the panels themselves now appear under the cutaway lighting. For the Karmavibhanga register, the photographic record is the record.

The colonial story credits Sir Thomas Stamford Raffles, lieutenant-governor of Java during the brief British interregnum (1811-1816), with the rediscovery of Borobudur in 1814. The credit is misplaced. Raffles personally never visited Borobudur; he learned of the monument from local informants in Semarang, dispatched the Dutch military engineer H. C. Cornelius to investigate, and read Cornelius's report later. Cornelius assembled a working crew of about 200 men in late 1814 and spent roughly two months cutting back undergrowth, burning vegetation off the upper terraces, and digging out the lower galleries from a deep mantle of volcanic ash and topsoil. The resulting site survey was sent to Raffles in early 1815 and entered the European literature through Raffles's The History of Java (1817).

The villagers of Bumisegoro, the settlement adjacent to the monument, were the proximate local source — the candi sat in their landscape and they guided the colonial party to the buried hill. Local knowledge of the candi was continuous: it appears in eighteenth-century Javanese chronicles as a place of ill omen and pilgrimage, and the surface stones — the heads of the upper Buddhas, projecting through the brush — were visible to anyone who climbed the hill. The historian John N. Miksic, who has written more on Borobudur in English than perhaps any other living scholar, notes flatly: "the locals were already well aware of its presence." The pattern holds across the colonial-rediscovery genre. Petra was never lost to the Bdul. Khajuraho was never lost to the villagers near whose fields T. S. Burt rode in 1838. Borobudur was never lost to Bumisegoro. What was lost — and what European narrative kept reasserting in the language of discovery — was the colonial archive's own awareness of these sites, not the sites themselves.

The longer concealment is real, however, and predates the British. The political center of Mataram had already shifted eastward under Mpu Sindok around 929 CE, well before the major Merapi event around 1006 CE that further depopulated the Kedu Plain. The reasons for the shift were probably plural — political pressure from Sriwijaya, court succession dynamics, and earlier volcanic activity — and historians still debate which factor was decisive. What is not in doubt is that by the eleventh century the political center of gravity had left the Borobudur landscape, monastic patronage had ceased, and the monument began the slow eight-century process of being reabsorbed into jungle. By 1814, when Cornelius arrived with his clearing crew, Borobudur was a hill with stones in it. The 800-year burial is an authentic loss; it is also a mainstream-explained one. There is no need to invoke deliberate concealment, religious suppression, or any of the ahistorical narratives that sometimes attach to other "lost" cities. A monument the size of Borobudur, in a tropical climate, on a fertile volcanic plain, will be reabsorbed by the landscape inside a few centuries if no one tends it. The 1814 "rediscovery" is mainly a colonial misnaming overlaid on this real but ordinary process.

On the morning of January 21, 1985, nine bombs detonated across the upper terraces of Borobudur, damaging nine of the 72 stupas on the three circular platforms. There were no fatalities. The blasts were precisely timed and placed. Of 13 devices planted overnight on January 20-21 across the upper three terraces, nine detonated; four were recovered intact and supplied investigators with usable forensic evidence. Investigators recovered enough physical and documentary evidence to charge a small group of perpetrators within two years.

This is the point at which the public record on Borobudur most often goes wrong. Many secondary sources — including travel writing, journalism, and even some academic surveys — attribute the bombing to "Jemaah Islamiyah" or "JI." That attribution is anachronistic. Jemaah Islamiyah did not exist as a named organization until roughly 1993, when Abdullah Sungkar and Abu Bakar Ba'asyir formally founded it in exile in Malaysia. The 1985 Borobudur bombing was the work of an earlier and distinct cell led by Husein Ali Al-Habsyi, an Indonesian preacher of Hadhrami Arab descent whose group was rooted in the older Komando Jihad and Darul Islam (DI) lineage that traced back to S. M. Kartosuwirjo's mid-twentieth-century West Java insurgency. Other cell members charged included Abdulkadir Ali Al-Habsyi, Achmad Muladawila, Abdulkadir Braja, and Ibrahim Djawad. Husein Ali Al-Habsyi was sentenced to life imprisonment in 1986 and granted clemency by President Habibie in 1999. The cell's stated motivations were revenge for the September 1984 Tanjung Priok killings of Muslim protesters and opposition to the Suharto government's enforced Pancasila ideology — grievances that locate the attack squarely in the older Komando Jihad / DI tradition rather than the post-1993 JI organization. Future writers covering this site should not repeat the JI attribution.

Restoration began almost immediately. The damaged stupas were reassembled between 1985 and 1991, drawing on the same anastylosis methodology and the same stone catalogue that the 1973-1983 UNESCO campaign had produced. The bombing remains the only documented modern attack on the monument and one of the few cases anywhere in the world in which an extensively photographed and catalogued World Heritage site was rebuilt from internal records within a few years of damage. The repaired stupas are not visibly distinct from their neighbors at viewing distance. Internal Borobudur conservation reports from the late 1980s note minor color differences in the replacement stone where original blocks were unsalvageable, and a small fraction of the perforated lattice work on the affected stupas was rebuilt with newly milled andesite to the original specifications from the 1973-83 catalogue. The bombing damage is still in the monument; it is just no longer apparent to the visitor.

The campaign that defines modern Borobudur ran from 1973 to 1983 under the leadership of the Indonesian archaeologist Soekmono, with funding and technical coordination from UNESCO and roughly 27-28 contributing nations. The total cost — frequently overstated in popular sources at $25 million — was approximately $6.9 million, drawn from international contributions and Indonesian government funds. Soekmono's project employed about 600 personnel over the decade, including stonemasons, conservators, surveyors, chemists, and laborers, and used the anastylosis method: dismantle the monument stone by stone, treat the underlying foundations, install new drainage, and reassemble each stone in its catalogued position with conservation-grade fill where original material was missing.

The block count is itself a contested figure. The UNESCO primary documentation — including The Restoration of Borobudur (unesdoc.unesco.org/ark:/48223/pf0000143333) and How Borobudur was saved (unesdoc.unesco.org/ark:/48223/pf0000053924), the latter authored by Soekmono himself — uses the phrasing "over a million stones dismantled and recatalogued." Some secondary sources cite 1.6 million; others cite 1.2 million. The honest answer is that the precise count varies depending on whether one includes infill blocks, replacement stones, and the foot encasement masonry, and the UNESCO primary phrasing is the safest one to quote.

What was preserved is well-known: the monument's structural stability against the moisture and seismic stresses that had been progressively destabilizing the upper galleries; the legibility of the relief panels, which had been weathering rapidly under tropical rain; and the iconographic program, including the five-Buddha (Pancha Tathagata) directional system, with four orientations housing the cardinal Dhyani Buddhas across the 504 statues and Vairocana at the center. What was lost is less commonly named. The original drainage system — a network of channels cut into the foundation that had carried monsoon water down through the structure for a millennium — was replaced with a modern lead-and-concrete system. The organic patina that had accumulated on the relief surfaces from a thousand years of biological growth, oxidation, and devotional touch was chemically treated and in many places altered. The sub-foundation moisture history, which carried information about the building's original hydrology, was overwritten by the new drainage. None of these losses invalidate the campaign — without it the monument would now be in advanced collapse — but they are real, and a fair assessment of the restoration names them as part of the trade.

The most important recent demonstration that Borobudur's reliefs encode functional knowledge, not stylized art, came from the sea. Among the 1,460 narrative panels in the visible galleries (the 160 hidden Karmavibhanga panels are counted separately), ten depict large ocean-going outrigger vessels of a distinctive double-outrigger configuration: two slender hulls flanking a central deck, twin tripod masts, square sails, lashed-plank construction. Maritime historians had argued for decades that these reliefs documented the vessels of the Sailendra-era Indian Ocean trade, but the argument was conjectural; no surviving vessel of the type existed, and the reliefs were typically read as artistic representations rather than measured drawings.

The British maritime archaeologist Philip Beale, working with the Australian shipwright Nick Burningham and the Indonesian master builder As'ad Abdullah, set out to test the proposition by building one. Between 2002 and 2003, the team constructed a double-outrigger vessel approximately 19 meters overall (17.29 m keel), with a beam of 4.25 m and bungor hull planking — christened Samudra Raksa, "Defender of the Seas" — at Pantai Kuwaru on Java's south coast, working entirely from the dimensional and structural information visible in the Borobudur reliefs. The build relied on traditional Indonesian boatbuilding techniques, lashed-plank hull construction, and rigging configurations interpolated from the carved evidence. On August 30, 2003, Samudra Raksa sailed from Jakarta. Over the following six months, the vessel crossed the Indian Ocean from Jakarta to the Seychelles (September 12, 2003), Comoros, Madagascar, around the Cape of Good Hope to Cape Town (January 5, 2004), and arrived at the Ghanaian port of Tema near Accra on February 23, 2004 — a route consistent with the early-medieval Cinnamon Route and broader Indian Ocean exchange that Sailendra-era seafarers are now believed to have run.

The expedition's significance is methodological. The reliefs functioned as a working ship blueprint. The proportions held; the rigging worked; the vessel sailed across an ocean. Whatever else Borobudur encodes, its sculptural program includes at least one body of decoded technical knowledge — Indonesian naval architecture as practiced in the eighth and ninth centuries — that is now back in the world because someone built from the carvings. Samudra Raksa is housed today in a dedicated pavilion on the Borobudur grounds, a few hundred meters from the panels she was built from.

One last item lives in the popular literature and is worth naming briefly. Inside one of the perforated stupas on the upper terraces — sometimes identified as stupa 32 in the popular literature, though the exact placement has shifted across restoration episodes — sits a Buddha statue that appears, to the casual eye, partly carved or unfinished. A nineteenth-century reading proposed that the figure was deliberately incomplete: a statement, perhaps, about the unattainability of nirvana or the inadequacy of representation. That reading still circulates in guidebooks. The current scholarly consensus is much more prosaic: the figure is almost certainly a damaged or displaced statue from elsewhere on the monument, placed in stupa 32 during one of the historical reassemblies. Borobudur's other 503 Buddha statues are uniformly finished. The "unfinished Buddha" reading does not survive close inspection.

Several fringe claims circulate around Borobudur and deserve direct response. The "pyramid math" or cosmic-encoding genre — which proposes that Borobudur's dimensions encode the speed of light, planetary orbital ratios, or pre-Sailendra geodetic constants — does not survive contact with the textual record. The Karangtengah inscription of 824 CE is a contemporary epigraphic source, written in Old Javanese and Sanskrit on a stele found in the Kedu region of Central Java, that names the Sailendra royal sponsors and identifies the sanctuary as Mahayana Buddhist. The three-realm cosmology of the monument (kamadhatu, rupadhatu, arupadhatu) and the five-Buddha mandala (the Pancha Tathagata directional system) are mainstream Mahayana iconography, not lost-civilization geometry. Borobudur's encoded meaning is the encoded meaning of its tradition; that is enough. A related claim — that Borobudur is an Atlantean or pre-Sailendra megalith later refaced by Buddhists — is ruled out by the same epigraphic and stratigraphic evidence. The "magnetic anomaly" sometimes attributed to the site reduces to the local volcanic geology of the Kedu basin, where iron-bearing andesitic stone routinely produces minor magnetic readings; this is not a site-selection mystery.

What is anomalous at Borobudur is what this page has named: a genuinely buried narrative register, a colonial misnaming of local continuity, a documented and well-attributed modern bombing, a preservation campaign with real losses inside its real wins, and a sculptural program detailed enough to launch a working ship across the Indian Ocean. That is more than enough to read on a single hill in Central Java.

Significance

Borobudur's significance in the lost-knowledge and anomaly conversation is unusual for a heritage site of its scale: most of what people call mysterious about it is actually well-documented, and most of what is genuinely hidden is hidden in ways that the documentary record can name. The monument resists the usual fringe overlays because its dating, patronage, and theological program are anchored in a contemporary inscription — Karangtengah 824 CE — and because its iconography is recognizable Mahayana Buddhism rather than something stitched from speculation. This makes it a useful test case for how the lost-knowledge category should be applied with discipline.

The buried Karmavibhanga register anchors the case because it is the rarest kind of archaeological finding: a complete narrative cycle of a major monument that is publicly known to exist, photographically documented in full, and yet inaccessible to direct viewing under nearly all conditions. The 156 sealed panels are not destroyed; they are, in a precise sense, withdrawn. The Cephas plates are the standing edition. Whether the withdrawal was theological (renunciation of kamadhatu as the practitioner ascends) or structural (mass at the foot to stabilize the upper terraces) is a debate that has not closed and probably will not close without a future restoration. Both possibilities should be held open.

The colonial-rediscovery question demands correction. Raffles did not rediscover Borobudur; H. C. Cornelius's clearing crew did not rediscover it either; the Bumisegoro villagers led Cornelius to a hill they had always known. The pattern recurs across the imperial archaeology of the long nineteenth century, and naming it correctly aligns Borobudur with the corrected accounts of Petra (the Bdul) and Khajuraho (the village near T. S. Burt). The 800-year jungle burial after 1006 CE is real; the 1814 "discovery" is mainly a misnaming.

The 1985 bombing matters because it happened and because the public record routinely attributes it to the wrong organization. Jemaah Islamiyah did not exist as a named entity until roughly 1993; the perpetrators were Husein Ali Al-Habsyi's cell, in the older Komando Jihad and Darul Islam lineage. Getting this right matters for the site's history.

The 1973-1983 UNESCO restoration sets the tradeoff frame: preserved and altered in the same ten years. The Samudra Raksa expedition supplies the living-evidence frame: the carvings as a working blueprint that put a Sailendra-era ship back across the Indian Ocean. Together these threads describe a site whose mysteries are the well-named kind, and whose real anomalies reward close looking rather than fringe overlay.

Connections

Borobudur — the parent entity. This sub-page focuses specifically on what is genuinely lost, hidden, or wrongly remembered about the monument. The parent covers the standalone history, construction sequence, Sailendra patronage under Samaratungga and Pramodhawardhani, the three-realm cosmology of kamadhatu, rupadhatu, and arupadhatu, the Pancha Tathagata mandala that organizes the 504 Buddha statues by direction, and the candi's place in Central Javanese Mahayana practice. Read the parent first if you have not yet seen the standard account; this page assumes that frame and works at the documentary edges.

Borobudur — Astronomical alignments — the sibling B1 page. The cardinal-orientation, equinox-sunrise, and Pawon-Mendut three-temple-line questions, including Giulio Magli's 2017 zenith-passage hypothesis, are handled there in full. The two pages are deliberately separate so that the alignment debate (which has its own evidentiary structure, mostly observational and metrological) does not get tangled with the lost-knowledge debate (which is mostly archival, epigraphic, and political).

Borobudur — Comparisons to other sites — the sibling B9 page. Cross-site material — Borobudur next to Angkor Wat, Bagan, the wider Indianized Southeast Asian temple complex, and the megalithic monuments of mainland and insular Southeast Asia — lives there, including the Mannikka metrological work on Khmer cubits and the chronology that places Borobudur three centuries before Angkor Wat. This page does not attempt cross-site comparison; that is the B9 lane.

Angkor Wat — the Khmer cosmic-mountain temple of mainland Southeast Asia (1113-1150 CE), built under Suryavarman II as a Hindu Vaishnavite monument and reconsecrated as Theravada Buddhist from the late thirteenth century onward. Useful as the closest large-scale Indianized temple to Borobudur in the regional record, though the Khmer-Sailendra political relationship is itself contested and the iconographic vocabularies (Vaishnavite cosmology versus Mahayana mandala) diverge sharply once one reads past the surface "stone temple complex in Southeast Asia" parallel.

Angkor Wat — Comparisons to other sites — useful as the parallel sub-page on the Khmer side, particularly for readers tracking how the lost-knowledge and rediscovery genres play out at the other major Southeast Asian Indianized monument. The colonial-rediscovery problem at Angkor (Henri Mouhot 1860, the Cambodian villagers who guided him) follows the same pattern named here at Borobudur: local continuity, European misnaming, and an eventual archive correction. Reading the two together makes the pattern legible across Southeast Asia and clarifies that the misnaming is structural to nineteenth-century imperial archaeology, not idiosyncratic to one site or one expedition.

Angkor Wat — Astronomical alignments — the sibling alignment page on the Khmer side, included here for readers who want to compare how alignment claims and lost-knowledge claims interact differently at the two monuments. At Angkor the alignment evidence is metrologically richer (Mannikka's Khmer-cubit work) and the lost-knowledge record is thinner; at Borobudur the relationship runs in the opposite direction.

Further Reading