The ancient Maya
What a decade of lidar, ancient DNA, isotopes and drought records has done to the story
This is the shorter of two essays. It was written from nine books in the collection and some sixty papers in the library, read and reduced to notes and then written up from them. The sources are named as they are used; the papers and books themselves open behind the login, where a longer version of this essay quotes them at length and reproduces their figures. What follows is the current picture as those sources give it, which is a slice of the field and not the whole of it, with the sources’ own hedges kept.
The old story and the new one
The old story ran like this: maize farmers slowly built villages, then ceremonial centres, then, in the Classic period from about AD 250 to 900, a civilisation of priest-astronomers whose cities emptied in a mysterious collapse, leaving a remnant for the Spanish. Almost every clause has been overturned, and most of the overturning is in papers published since 2013.
The new story. Monumental building came first, before villages, before kings, before maize was the staple, at a scale the Classic cities never matched. The calendar and the script are older than was thought and were shared across southern Mesoamerica before the Maya made them their own. The Classic period was neither an empire nor a scatter of ritual chiefdoms but a durable political culture of many small kingdoms held in shifting hierarchies by a few overkings, into which a foreign power, Teotihuacan, reached in AD 378 and left its mark in the ground at Tikal. The lowlands held not a few hundred thousand people but somewhere between seven and sixteen million, farming a landscape that was terraced, drained and gardened rather than burned. The collapse was real, but it was not one thing: a political crisis around 810, a drought that cut rainfall by half, war, hurricanes, and a system fed by tribute that could not survive its own success, unevenly across the map, over a century. And the Maya did not vanish. Five million are alive, and their genomes run unbroken back to the people in the ground.
A plateau in Tabasco
The paper that most changed the picture found Aguada Fénix. In 2020 Takeshi Inomata’s team reported an artificial plateau in Tabasco 1,400 metres long and up to fifteen metres high, with nine causeways, built between 1000 and 800 BC: the oldest monumental construction in the Maya area and the largest in the region’s whole pre-Hispanic history. Five years later they showed it was the centre of a plan laid out on axes nine and seven kilometres long, marked by canals and fed by a dam, which they read as a cosmogram, a map of the ordered universe drawn on the earth. Beside Teotihuacan and Tikal at one scale, the site that came first is not the smaller.
What it lacked matters as much. Its builders show little sign of the marked inequality of the Olmec centre it resembles. Inomata’s conclusion is that communal work, not kingship, lies at the start of Maya civilisation. Daniela Triadan’s 2025 review puts the Maya beside Göbekli Tepe, Caral and Poverty Point as places where monuments came before settled farming. At Ceibal the same team had earlier shown a ceremonial platform of 1000 BC growing into a pyramid by 700 BC, not as a passive copy of the Olmec but as a participant in the invention, and a cemetery of 1100 to 800 BC beside it holds twenty-six people with no sign of rank among them. A figurine tableau in El Salvador and a sediment core in the Petén add two notes: monumentality “may arise independently of rulership”, and the Maya were altering their lakes a thousand years before the collapse usually blamed on environmental damage.
Coe and Houston’s survey, the ninth edition of Michael Coe’s The Maya, does not yet know Aguada Fénix, but it already had El Mirador, which it calls the earliest Maya capital, with a pyramid seventy metres high, and it already compared the failure of the Mirador Basin cities around AD 150 to the later collapse.
Older calendars, borrowed scripts
The 260-day count that Guatemalan daykeepers still keep has twice grown older since 2022. A painted “7 Deer” day sign from the earliest phase of San Bartolo’s pyramid, dated to 300 to 200 BC, is the earliest securely dated Maya calendar. The orientations of the Gulf Coast ceremonial complexes built between 1100 and 750 BC, Aguada Fénix among them, appear to encode the same count eight centuries earlier still. The Long Count and the Calendar Round were probably devised under Olmec or Isthmian influence, and the Isthmian script that shares signs with the Maya glyphs remains unread; a 2026 review of Martha Macri’s book on it argues that the one published decipherment mistook a punctuation mark for a suffix.
The Maya script was cracked in stages: the calendar by the 1900s; the phonetic principle by Yuri Knorosov in 1952, resisted for decades; and the discovery in 1960 by Tatiana Proskouriakoff, a museum curator, that the stelae recorded the lives of real rulers rather than priestly abstractions. Everything in the next section rests on that discovery. On the calendar and the 2012 fuss, David Stuart’s book is the one to read: the thirteenth bak’tun turned in December 2012 as the arithmetic required, the one ancient text that names the date says nothing of destruction, and its final glyphs are too damaged to read with confidence.
What lidar did to the map
Airborne laser scanning removes the forest from the data, and it has changed the scale of everything. The 2018 survey of northern Guatemala found 61,480 structures in 2,144 square kilometres and extrapolated seven to eleven million people across the central lowlands in the Late Classic, with 106 kilometres of causeways and, alongside them, defensive works that point to large-scale conflict. The sceptics said lidar sees false positives and that the surveys were flown over known sites. The answer came from lidar flown for a forest-carbon project in Campeche with no archaeological bias at all: it found the same densities and a whole unrecorded city, Valeriana, and its authors wrote that anyone waiting for an empty hinterland to offset the numbers “is running out of places to look”. The aggregated estimate now stands at 9.5 to 16 million, and the same data show elite masonry houses spaced at regular intervals among ordinary ones, in town and country alike, as if placed to oversee the flow of goods. Under the sea floor of a Belize lagoon, pole-and-thatch houses that leave no mound remind us the counts are a floor.
A society of kings
For the politics of the Classic, Simon Martin’s Ancient Maya Politics is the argument of record. His question is why a world of many small kingdoms persisted for six centuries without any of them becoming an empire. His answer is a hegemonic system: kingdoms each notionally sovereign, bound into hierarchies under a few overkings, above all Tikal and the Snake dynasty at Calakmul, in which war aimed at tribute and clientage rather than territory, marriage bound rivals, and a shared moral order kept the competition within bounds. He counts at least fifty-seven royal emblems, some 275 uses of the overking’s title over 473 years, sixteen of them by women, and 165 recorded wars. He is doubtful of “divine kingship”, finds no Venus-timed warfare, and shows that the supposed spike of war before the collapse rests on two unusual monuments. What the texts cannot give is the economy, which he compares to dark matter: known only by its effects. Case studies agree that governance varied from place to place, and that a palace at Yaxnohcah was no grander than its neighbours.
The Entrada
On 16 January AD 378 a lord named Sihyaj K’ahk’ arrived at Tikal on the day its king died, and a year later a boy king was installed whose father the texts name as Spearthrower Owl. For sixty years scholars argued whether this was a Teotihuacan takeover or a local fashion for Mexican style. Two excavations have settled the physical side: lidar at Tikal shows a precinct that reproduces Teotihuacan’s Ciudadela at about a third of its size and on its orientation, with Mexican-style incense burners and two sacrifices beside green obsidian; and an altar in an elite house is painted with Teotihuacan’s Storm God in a technique the excavators call fully metropolitan, the work of a foreign enclave. Whether it was conquest or an outpost is still debated, since commoners’ lives barely changed; what is not debated is that Tikal rose afterward. The genomes agree: Copán’s founding dynasty of 426 brought an influx of highland Mexican ancestry, and Teotihuacan itself housed a foreign quarter with prior ties to the Maya.
What sacrifice was for
The Classic Maya idea of the person, reconstructed by Houston, Stuart and Taube in The Memory of Bones, extended beyond the skin: a carved portrait shared its subject’s essence, breath and soul were one, time itself was made by a primordial sacrifice, and a captive could become food for the gods. Sacrifice was a mechanism, and the newest evidence shows it run to a calendar. Sixty-four children from a mass burial beside the Sacred Cenote at Chichén Itzá, sequenced in 2024, were all boys; a quarter were close kin; two pairs were identical twins, the heroes of the Popol Vuh. They were deposited in pairs at regular intervals over four centuries. The old assumption of virgin girls is dead. The living Maya of a village nearby are their relatives, with one addition: genes for resistance to the diseases the Spanish brought.
How they were fed
The idea that the Maya lived by slash-and-burn maize and that the forest could not bear them was the frame in which the collapse was explained. Two long reviews have taken it apart: the plant remains show root crops, tree crops, orchards, gardens, terraces and drained fields, capital invested in land over generations, and more resilience than the maize-failure models predict. Maize was not even the first staple; it became one when limestone, which underlies the whole peninsula, supplied the alkali for nixtamalisation, and its pollen rises in dry periods as a crop of last resort. Salt kitchens under the Belize lagoons scaled up with the inland cities and produced a surplus; markets, once denied, are painted on the walls at Calakmul.
The collapse
Coe and Houston give the consensus: warfare, overpopulation with environmental damage, and drought, not all at once. Martin gives its shape: a rapid crisis around 810 when most kingdoms raised their last monuments, then a century’s coda under a new elite with foreign-sounding names, until the last Petén date in 910. At Ucanal, someone in the early ninth century burned two dead kings and their jade in public, which the excavators read as the dismantling of an old regime and a pivot to something new. The drought has now been measured: the gypsum of Lake Chichancanab records rainfall cut by half on average and by seventy per cent in the worst years.
But drought alone does not explain it, and the record is clear on why. At Mayapán, drought raised civil strife between 1400 and 1450 and the city fell, yet the region’s institutions endured to the Spanish. Hurricanes ran above average from 700 to 1450. The most integrative model has climate, conflict and economies of scale interacting to build the cities and then to unmake them. The genomes record the population falling around 1,200 years ago. Where the people went, Coe and Houston call one of the great mysteries.
After, and before
The north went on: Chichén Itzá, whose Toltec phase is now dated after its Maya one rather than alongside it, then walled Mayapán until the 1440s. The Spanish found no single authority to topple, so the conquest took from 1542 to 1697, and disease did what arms could not. Yet Mam villagers registered as converts rebuilt their plaza to the shape of their ancestral city, and Ixil houses were painted with pre-Columbian figures into the eighteenth century.
The rediscovery began with John Lloyd Stephens, who cleared and measured forty-four sites in 1841 and 1842 and argued, against the learned opinion of his day, that the cities were built by the ancestors of the Maya around him. The genomes from Chichén Itzá are the final proof that he was right. His question about the glyphs, “Who shall read them?”, was answered a century later.
What is still open
The sources are candid, and a synopsis should be too. The Classic economy is nearly invisible. The Mirador Basin’s early collapse is unexplained. The ninth-century newcomers are untested. The population is somewhere between three and sixteen million depending on the estimate. Women and children are almost absent from the record. The Isthmian script is unread. And a new railway has cut 1,500 kilometres through the forest past 1,300 recorded sites.
My note —
Sources
Books: Coe and Houston, The Maya (ninth edition); Martin, Ancient Maya Politics; Stuart, The Order of Days; Proskouriakoff, Maya History; Houston, Stuart and Taube, The Memory of Bones; Montgomery, How to Read Maya Hieroglyphs; Stephens, Incidents of Travel in Yucatan; Marcum, The Fallen Stones.
Papers, by first author and year: Inomata 2013, 2020, 2025; Triadan 2025; Burham 2025; Szymański 2025; Birkett 2023; Stuart 2022; Šprajc 2023; Robinson 2026; Canuto 2018; Ford 2018; Auld-Thomas 2024; Estrada-Belli 2023, 2025; Britton 2025; McKillop and Sills 2022, 2025; Kupprat 2025; Houston 2021; Ramírez 2025; Wade 2019, 2020; Murray 2025; Aguirre-Samudio 2025; Barquera 2024; Curry 2024; Vance 2019; Locker 2025; Negrin 2024; Morell-Hart 2023; Fedick 2024; Voorhies 2024; Islebe 2022; Fisher 2023; Horowitz 2024; Halperin 2024; Evans 2018; Kennett 2022; Sullivan 2022; McCool 2025; Price 2021; Weidele 2016; McAvoy 2026; Castillo 2022; Źrałka 2020; Ortega 2022. All are in the library, which opens with my login.