Not "no good explanation." No explanation. The tools that made them, by every conventional account, could not have made them. The civilisation credited with building this place did not possess what the stones themselves require.

Puma Punku — "Door of the Puma" in Aymara — sits on the Bolivian Altiplano near the southern shore of Lake Titicaca. Altitude: approximately 3,800 metres. At that height, the air holds barely sixty percent of the oxygen available at sea level. Every act of labour costs more. Every kilogram carried farther.

The site is part of the Tiwanaku archaeological complex, a ceremonial and urban centre that dominated the Altiplano roughly between 400 and 1000 CE. Tiwanaku includes the Akapana pyramid, the Semi-Subterranean Temple, and the Gateway of the Sun. Puma Punku sits about a kilometre southwest — close enough to share a cultural lineage, distinct enough that researchers keep treating it as its own problem.

What you find there today is apparent chaos. Multi-ton blocks scattered, toppled, half-buried in sediment. But look at any single fragment. The chaos dissolves into something more unsettling than destruction.

The surfaces are flat. Not roughly flat. Flat to tolerances a modern machinist would recognise. The corners are sharp. The interior angles precise. The grooves — cut into stone harder than granite — repeat at identical widths and depths, block after block, as though stamped from a mould.

Archaeologists see a ruin. Engineers see something different. They see components.

The site's most famous artefacts are the H-blocks — named for their shape when viewed from above. Carved from red sandstone, they feature a complex geometry of recesses, channels, and protruding elements that interlock like a three-dimensional puzzle. Each H-block is roughly the same size. Their interior features repeat with remarkable consistency across multiple examples.

This is the detail that arrests engineers. The H-blocks are not unique artistic expressions. They are standardised. The recess in one block corresponds precisely to the protrusion of its neighbour. Channels are uniform in width and depth. Flat surfaces are level. When researchers measure interior angles and groove dimensions across multiple blocks, the tolerances are tight — tight enough to suggest not individual craftsmanship, but a controlled, repeatable production process.

The word that surfaces repeatedly in engineering assessments is modular. These are not monoliths carved to be admired. They are components designed to fit together, to create something larger, something structural. The interlocking design provides lateral stability without mortar — allowing the assembled structure to resist seismic forces. On the earthquake-prone Altiplano, this is not decorative. It is an engineering solution.

Several blocks also feature precision-cut rectangular slots and cleanly drilled holes, apparently designed to receive metal clamps — likely copper or bronze. Similar clamp-joinery appears at sites in Greece and Egypt. At Puma Punku, the execution is notably precise: slots uniform, drill holes clean, some featuring I-shaped channels where molten metal was poured to join adjacent blocks.

The flat surfaces on many blocks approach what machinists call ground flatness. Running a straight edge across the larger faces reveals minimal deviation over distances of several feet.

But here is the deeper implication. To produce modular, interlocking components with standardised dimensions, you need more than skilled hands. You need a design system — a way to specify dimensions, communicate them across multiple workers or work teams, and verify that the finished product conforms to specification. You need something functionally equivalent to engineering drawings and quality control.

Whether the Tiwanaku accomplished this through templates, measuring cords, or reference stones, the result speaks to an organisational and technical sophistication that goes well beyond what is commonly attributed to pre-Columbian South American cultures.

The H-blocks are the signature puzzle. The material they are carved from is the lock.

Puma Punku's stones come in two primary types. Red sandstone is massive, impressive in scale, and relatively workable. Andesite is another matter.

Andesite is a fine-grained volcanic rock. Its hardness typically ranges from 6 to 7 on the Mohs scale — comparable to, and in some specimens exceeding, granite. The darker, denser andesite at Puma Punku approaches the hardness range of diorite, which sits at 7 to 8 on Mohs. Steel tools register around 6 to 6.5. A modern steel chisel will barely scratch the surface of the harder andesite blocks. It will not carve them with precision.

The Tiwanaku civilisation, as conventionally understood, did not possess iron or steel tools. Their metallurgy included copper, arsenic bronze, and tin bronze. Bronze sits at roughly 3.5 to 4.5 on Mohs. You cannot cut andesite with bronze. You can abrade it, slowly, using harder intermediary materials — quartz sand, corundum — but this is a laborious process that produces characteristically rounded edges and uneven surfaces.

What we see at Puma Punku on many andesite fragments is the opposite. Sharp internal corners. Clean flat surfaces. Crisp geometric incisions.

Mainstream archaeology offers several responses. Stone-on-stone percussion and grinding using quartzite or andesite cobbles is well-documented in Andean tradition. Abrasive sand and water, combined with patient skilled labour, can achieve impressive results. Experimental archaeology has confirmed this — in principle. The limitation is that experimental results, while impressive, rarely match the crispness of the best Puma Punku specimens. The finest work involves grooves only a few millimetres wide cut into andesite, interior corners with radii suggesting a rotating cutting tool, surfaces flat over distances no grinding stone naturally produces.

Alternative researchers have proposed stone-softening techniques — plant extracts or chemical processes that temporarily soften stone for easier working. The idea appears in Andean oral traditions. Peruvian priest and scholar Jorge A. Lira claimed to have identified a specific plant capable of this. No such process has been scientifically verified. It remains speculative. But the claim originates from indigenous Andean knowledge traditions — a source that deserves honest consideration even where verification is absent.

More exotic proposals — ultrasonic cutting, lost mechanical technologies, prior advanced civilisation — lack direct physical evidence. They explain the anomaly by multiplying the mystery.

The stones remain. And they remain difficult to fully account for.

Orthodox dating places Puma Punku's construction around 536–600 CE, based on radiocarbon dates from organic material found in soil layers associated with the construction phases. This is supported by ceramic analysis, stratigraphy, and the broader Tiwanaku archaeological context. The conventional framework is coherent and evidence-based.

But radiocarbon dates organic material. Not stone. The dates tell us when organic material was deposited near the stones — not when the stones were carved or first assembled. If the site was built on an older foundation, or if the stones were reused from an earlier structure, the dates would reflect the later activity, not the original construction.

Arthur Posnansky, a Bolivian-Austrian engineer who spent decades studying Tiwanaku in the early twentieth century, proposed a dramatically different chronology. Using archaeoastronomical methods — specifically, the alignment of structures with astronomical phenomena and the known rate of change of Earth's axial tilt — he argued that the complex dated to approximately 15,000 BCE, or earlier. Professional archaeologists have widely criticised his methods, noting calculation errors and unreliable assumptions about original structural orientations given the site's disturbed condition.

Yet Posnansky's work persists. Partly because the site itself keeps producing friction with the conventional timeline. If Puma Punku was built around 600 CE by the Tiwanaku, it should sit seamlessly within a rich context of Tiwanaku-period artefacts and habitation evidence. And it does — but some researchers note that the precision stonework at Puma Punku is qualitatively different from other Tiwanaku constructions. The Gateway of the Sun, the Semi-Subterranean Temple, and the Akapana pyramid are impressive. None of them exhibit the same modular, machine-like precision as the H-blocks. That difference needs explaining.

Some alternative researchers also point to geological evidence suggesting catastrophic flooding across the Altiplano at the end of the last Ice Age, roughly 11,000–12,000 years ago. If Puma Punku predates that event, the site's demolition pattern might not reflect gradual decay or deliberate dismantling — but sudden, violent destruction by water. This remains a minority view, unsupported by mainstream geological consensus for this specific site.

The honest position: the conventional dating is the best-supported by available evidence and methodology. The alternative chronologies are intriguing and unproven. The gap between them is not emptiness. It is a space filled with legitimate questions.

Ancient sites typically decay gradually. Roofs collapse. Walls lean. Vegetation encroaches. Puma Punku looks like it was hit.

Multi-ton blocks lie scattered far from any plausible original position. Some are flipped. Some are buried under metres of sediment. The scatter pattern suggests not abandonment but catastrophe.

The most conventional explanation is earthquake. The Altiplano is seismically active. A sufficiently powerful event could topple and scatter a mortar-free stone structure. Over centuries, scattered blocks would be further displaced by erosion, sediment accumulation, and human removal.

That human removal is well-documented. After the Tiwanaku civilisation declined around 1000 CE, the site was quarried by later peoples, including the Inca. After the Spanish conquest, colonial settlers carried blocks away for churches, bridges, and farm walls. The railway running near the site was reportedly built in part using stone taken from Tiwanaku and Puma Punku. Centuries of scavenging progressively dismantled whatever survived the initial destruction.

The catastrophic flood hypothesis, favoured by some alternative researchers, proposes that a massive water event — possibly connected to the draining of an ancient, larger Lake Titicaca or to Ice Age meltwater pulses — swept across the Altiplano. Proponents point to the chaotic scatter, sedimentary deposits consistent with water-borne material, and what they describe as water-erosion patterns on certain stones. Critics note these features can be explained by other processes. The flood hypothesis lacks the specific geological evidence that would lift it from speculation.

A third possibility, less discussed: deliberate destruction. Some ancient sites were intentionally dismantled — ritual "killing" of a sacred space by conquering peoples, or by the builders themselves. No specific evidence supports this at Puma Punku. But absence of evidence is not evidence of absence when the surface has barely been excavated.

What we are looking at is a fraction of what once existed. The blocks that remain are survivors. Whatever story they told in their assembled form, we now have only scattered syllables.

New tools have been brought to Puma Punku in recent decades. Some of the results have deepened the mystery rather than resolved it.

Ground-penetrating radar (GPR) surveys have revealed subsurface anomalies suggesting the visible ruins are only a portion of a much larger complex. There appear to be buried walls, foundations, and possibly entire structures beneath the current ground surface. The full scope of Puma Punku may still be underground.

3D laser scanning applied to individual blocks has produced precise digital models, measuring dimensions, angles, and surface qualities to millimetre accuracy. These scans have confirmed what hand measurements suggested: the H-blocks exhibit standardisation that is genuinely remarkable. Internal dimensions repeat with deviations typically measured in fractions of a millimetre across different blocks. This level of consistency raises a direct question about process that the standard account of patient, hand-skilled labour has not fully answered.

Chemical and petrographic analysis has confirmed the stone identifications and helped narrow down probable quarry sources. The primary red sandstone quarry lies roughly ten kilometres from the site. The andesite source is believed to be on the shores of Lake Titicaca or the Copacabana peninsula — forty to ninety kilometres distant depending on the specific quarry.

Some researchers have investigated magnetic anomalies and the acoustic properties of the stones and spaces. These are early-stage investigations. Claims about unusual magnetic readings or resonant frequencies at Puma Punku should be held lightly. But they connect to a legitimate and growing field of inquiry into the acoustic and energetic properties of ancient stone structures — research that has produced intriguing results at Newgrange in Ireland and the Hal Saflieni Hypogeum in Malta. Whether similar phenomena appear at Puma Punku is an open question.

Alexei Vranich of the University of Pennsylvania conducted major excavations at the site in the early 2000s. His reconstruction work suggests Puma Punku was a large stepped platform surmounted by gated enclosures, with the H-blocks forming part of the interior architecture. His work places the site firmly in the Tiwanaku cultural context and supports the conventional dating. He does not deny the stonework is extraordinary. He argues it is explicable by known Andean techniques, executed with skill and organised labour.

The mainstream framework is the most evidence-based account currently available. It has gaps. Those gaps are not trivial.

Puma Punku is not alone. It belongs to a global catalogue of ancient stone sites that resist comfortable explanation.

At Sacsayhuamán, near Cusco in Peru, enormous blocks — some exceeding 100 tonnes — are fitted together with a precision that does not permit a razor blade between joints. The stones are polygonal with complex interlocking shapes. They have survived five centuries of earthquakes. The Inca are credited with this work, but some Spanish chroniclers recorded that even the Inca attributed the oldest stonework to an earlier people.

In Egypt, the precision of the Great Pyramid's casing stones and the granite surfaces of the King's Chamber raise parallel questions. Engineer Christopher Dunn studied ancient Egyptian drill holes extensively. Core samples show spiral grooves suggesting cutting rates that, he argues, exceed what bronze tools and abrasive sand can achieve. The argument is contested. The drill holes are not.

At Göbekli Tepe in Turkey, megalithic construction dating to approximately 9500 BCE pushed the accepted timeline for monumental architecture back by several thousand years when it was recognised. It demonstrated, directly, that assumptions about what hunter-gatherers could achieve were wrong.

The common thread is not shared culture or shared style. It is a capability gap — a discrepancy between the tools we believe were available and the results that were actually achieved. Each site has its own context, its own culture, its own explanations. But the recurrence of the same basic friction — how did they do this, with what they had? — invites a broader question about whether our model of technological development is missing something.

This is not an argument for a single lost civilisation spanning the globe. It may simply be an argument for the depth of human ingenuity — and for the possibility that sophisticated technical knowledge has been independently developed, and lost, more than once. The archaeological record preserves stone and sometimes metal. It does not preserve wood, rope, leather, or knowledge. A civilisation could achieve extraordinary things and leave behind only the stones, every perishable element of their toolkit gone.

That is not a fringe claim. That is the plain nature of what archaeology works with.

It would be dishonest to frame this as a contest between a defensive mainstream and a courageous fringe.

Professional archaeologists who have worked at Puma Punku acknowledge the stonework's exceptional quality. Vranich's excavations produced a rigorous reconstruction of the site's probable form. Mainstream archaeology does not dismiss Puma Punku. It explains it — with known Andean techniques, stone hammers, abrasive sand and water, bronze chisels for softer materials, patient skilled labour over extended periods, and standardisation achieved through templates.

This explanation is reasonable. It may be correct.

But it is incomplete in specific areas. The sharpness of interior corners on andesite. The fineness of millimetre-wide grooves. The scale and consistency of surface flatness across separate blocks. These features strain the explanatory reach of stone-on-stone grinding and abrasive sanding. Experimental archaeology has demonstrated the techniques in principle. It has not fully replicated the finest Puma Punku work in the specific materials.

Graham Hancock and Brien Foerster are among the most prominent alternative voices on the site. Both argue the gap between experimental results and actual artefacts is significant and is being glossed over. They propose various explanations — lost tool technologies, older dates, a vanished advanced civilisation. The evidence behind these specific conclusions varies from suggestive to absent.

There is an important distinction here. The observation — the stones are harder to explain than we readily admit — stands on firmer ground than most of the conclusions drawn from it. Acknowledging the gap does not require accepting any particular theory about what fills it.

The healthiest position is productive tension. Hold the mainstream framework as the most evidence-based account available. Hold the gaps as real and unresolved. Let both be true at once, and resist the pressure — from either direction — to collapse the tension prematurely.