Tiwanaku & Puma Punku

What is actually there

Puma Punku is a terraced platform, part of the larger Tiwanaku complex on the Bolivian Altiplano near Lake Titicaca, at nearly 3,850 metres. Radiocarbon dating of material from the earliest construction fill places the start of building at roughly 536 to 600 CE. Tiwanaku itself grew into a city of perhaps 20,000 people at its height around 800 CE, and was abandoned some centuries before the Inca arrived and found it in ruins.

What survives is fragmentary and heavily disturbed. The Spanish chronicler Pedro Cieza de León visited in the 1540s and described precision stonework he could not explain. Since then the site has lost material to earthquakes, colonial-era quarrying for building stone and railway ballast, and target practice by the Bolivian military. Most of what you see today is not in its original position.

Two kinds of stone were used, from two different places. The largest blocks, including a red sandstone slab of about 131 tonnes, came from a quarry roughly ten kilometres away. The more finely worked stone, including the H-blocks, is andesite from the Copacabana Peninsula on the far side of Lake Titicaca, some 90 kilometres away, almost certainly moved at least part of the way by reed boat.

Nobody disputes that the H-blocks are precise. The argument is entirely about how.

The H-blocks have deep, uniform channels, flat to within about a millimetre across their length, with sharp internal corners and almost no visible tool marks. In 2018 Alexei Vranich led a project that 3D-printed 140 andesite pieces and 17 sandstone slabs at scale, from more than a century of accumulated field measurements, and physically reassembled the structure on a tabletop. It confirmed what the popular accounts get right: the blocks share a small set of standardised dimensions, repeated across pieces that were shaped weeks or months apart.

Unfinished blocks at local quarries show the method: pounding with stone hammers to rough out the shape, then grinding and polishing with flat stones and sand. Cramps made from a copper-arsenic-nickel alloy locked finished blocks together, and the channels for them survive. None of this requires metal tools that can cut andesite. It requires an enormous amount of skilled, patient labour, which is a different claim.

The sky here

The site's most famous claim comes from the sky, not the stones. The Kalasasaya, a rectangular platform lined with standing pillars, is widely read as a sun-oriented temple. An Archaeology magazine field note on Tiwanaku says its remains pinpoint the solstices, equinoxes, and the zenith and nadir passages of the sun with astonishing accuracy, and suggests that the pillars of the western balcony wall form a solar calendar based on where the sun sets. I have not found the measurements behind that statement, so I treat it as a reading, not a result.

In the 1920s the Austrian engineer and self-taught archaeologist Arthur Posnansky took the astronomy much further. As usually reported, he measured the sightlines of the Kalasasaya, derived the angle of the earth's axial tilt that they implied, which drifts slowly over a 41,000-year cycle, and worked backward to a construction date of around 15,000 BCE. In his book Tihuanacu: The Cradle of American Man (1945, 1957) he argued that Tiwanaku was the origin of civilisation in the Americas.

The method is real astronomy. The problems are around it. The Kalasasaya has been damaged and reconstructed more than once since the sixteenth century, which makes trusting its present alignment to arc-second precision a stretch. No stratigraphic or artefactual evidence supports occupation at 15,000 BCE, and the Altiplano at that date, in the last glacial period, would have been a harsher place to build than it is now. As far as the chronology I have read goes, every datable object from Tiwanaku is consistent with the first millennium CE. Later work on the site, including by Kolata, disputes the deep-antiquity claims, and I have not read each objection in the original.

What is easy to miss in the debunking is that Posnansky was not a crank working alone. He spent decades on careful documentation of the site's architecture and sculpture, and his photographs and maps are still a valuable record. Real fieldwork combined with a wrong conclusion is much of why the 15,000-year date still circulates.

What is still open