For adherents of lost technology, and for the ancient-astronaut programmes that made the ruin famous, Puma Punku is the proof case: andesite faces so flat, inner corners so sharp, and H-shaped blocks so alike that they read as machined, beyond what a people without iron could make. Archaeologists at Tiwanaku do not deny the precision. What they deny is the step from precision to an outside maker, and the argument runs through tool marks, quarry sources, metal cramps, dates, and the building's place in a local tradition. The question itself is old. Writing in the Penn Museum's Expedition magazine (https://www.penn.museum/documents/publications/expedition/47-2/Reed%20Boats.pdf), Alexei Vranich, Paul Harmon, and Chris Knutson recall that Cieza de León passed through in the mid-sixteenth century and was told by local Aymara that the place was older than the Inca and built by a race of giants. They put the modern puzzle plainly: how could a civilization without iron tools, the wheel, draft animals, pulleys, or cranes move thousands of tons of stone so far?
The UNESCO listing for Tiwanaku: Spiritual and Political Centre of the Tiwanaku Culture (https://whc.unesco.org/en/list/567/) sets the scene. The city lies near the southern shore of Lake Titicaca at 3,850 metres. It began as a small settlement and became a planned city between 400 and 900 A.D. The listing names Pumapunku among the structures of the ceremonial core, beside the Semi-underground Temple, the Kalasasaya, and Akapana, and describes ashlars carved accurately and a complex underground drainage system. It says the people of the city perfected the technology for carving and polishing different stones. Around them it places almost 50,000 raised fields, colonies reaching valleys and coast, and a decline from the eleventh century. On that record the precision belongs to a society with farms, pottery, textiles, and metals, not to a stone yard dropped from elsewhere.
Radiocarbon narrows the window. Erik Marsh, Vranich, and colleagues, modelling 102 radiocarbon dates in PLOS ONE (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0288798), place the start of Pumapunku construction around A.D. 580, the last use of quarried, carved stones around A.D. 630, and the end of construction around A.D. 710. After that, they write, masonry used minimally modified reused stones, so perhaps only one or two generations of masons made most of Tiwanaku's iconic stonework of flat planes, geometric corners, and precision fitting.
Jean-Pierre Protzen and Stella Nair do not wave the tool problem away (https://repositoriodigital.bnp.gob.pe/bnp/recursos/2/html/las-piedras-de-tiahuanaco/297/). They write that no construction tools have been excavated or identified at Tiwanaku, that iron was unknown in the Andes before the Spanish arrived, that the few copper or bronze chisels in museums look too small for building work, and that their own tests with modern bronze were not effective on hard stone. They cite Denys Stocks, whose flint tools carved a fair Egyptian glyph in granite where steel chisels suffered, and ask whether Tiwanaku masons used similar stone tools. What they did find is on the stones. Large andesite blocks still raw from the quarry show splitting and rough dressing: one meant to become four smaller blocks, another being parted into two slabs. These carry the marks and dents diagnostic of hammerstone work. They report a hematite hammerstone with use wear, found out of context, and quartzite cobbles on the valley floor that could have served. The answer is partial, and they say so.
Moving the stone was tested, not assumed. The Expedition team reports that the red sandstone came from a quarry 10 km away and the green andesite from the Copacabana peninsula, 90 km across the lake. Aymara boat-builders made a 15-metre boat from about 1.8 million totora reeds, and a crew of 25 to 40 people levered a 9-ton stone to the water before it crossed the lake.
The I-shaped cramps are where metal enters. A 2017 Scientific Reports paper on Andean copper metallurgy by Anja Eichler and colleagues (https://pmc.ncbi.nlm.nih.gov/articles/PMC5282569/) reports that most of 208 metal artefacts excavated at Tiwanaku sites were tin bronze or a copper-arsenic-nickel alloy, and that this unique ternary bronze was also used for I-shaped architectural cramps setting stone blocks. It illustrates one such cramp from Puma Punku. The alloy and the tradition of cramps, the authors write, were restricted to the Tiwanaku culture.
Vranich's 2018 paper in Heritage Science (https://www.nature.com/articles/s40494-018-0231-0) dates Tiwanaku to A.D. 500–950 and describes about 150 pieces of finely cut andesite around 17 sandstone slabs, from small blocks to the much heralded H stones, 97 by 99 by 55 centimetres and about 600 kilograms, and one block of about 20 tonnes from a quarry across the lake. Studies confirm that the andesite and sandstone were unfinished; looting left no block in place. His team turned 150 years of field measurements into 3D prints at 4 percent scale, sorted pieces by height and by clamp grooves for metal fastenings, and rebuilt part of the andesite building. Its form, he writes, matches earlier nearby sites such as Chiripa, which he offers against claims that the building has no local precursors.
Puma Punku's precision is not in dispute; its maker is. Adherents read flat andesite faces, sharp inner corners, and repeated H-blocks as machining beyond a people without iron, and some reach for lost technology or visitors. UNESCO places Pumapunku inside a planned Tiwanaku city of 400–900 A.D., and radiocarbon puts its construction at about A.D. 580–710. Protzen and Nair find hammerstone marks on raw quarry blocks, Aymara builders moved a 9-ton stone across the lake by reed boat, and the I-shaped cramps are a copper-arsenic-nickel bronze known only from Tiwanaku. Vranich's 3D prints rebuild part of the andesite building on a plan with local precursors. Which tools did the finest work is still unknown.
What Stays Contested
On these pages: a UNESCO-listed Tiwanaku city of 400–900 A.D. with Pumapunku in its core; construction dated to about A.D. 580–710; raw quarry blocks with hammerstone marks and no excavated building tools; a 9-ton stone moved by reed boat; I-shaped cramps in a bronze restricted to Tiwanaku; a partial 3D reconstruction with local precursors. Left open: the exact tools behind the finest work. No outside builder is baptized, and no adherent is mocked for asking how it was done.