Somewhere in a climate-controlled archive, a bundle of knotted strings sits in a preservation case. I've always found looking at one to be a bit surreal. To most eyes it looks like a frayed, colourful tassel – the kind of thing you might find on the end of a heavy curtain. To a trained Quipucamayoc of the fifteenth century CE, it was a document. A record. Possibly an entire story. We're still not entirely sure which, and that uncertainty is one of the most haunting puzzles in the history of human knowledge.
The object is a quipu. And the civilisation that created it ran 4,000 kilometres (2,485 miles) of coastline, mountain, and jungle without a single written alphabet.
Image: Claus Ableiter
A superpower without a page
At its peak in the early sixteenth century, the Inca Empire – Tawantinsuyu, "the four regions together" – stretched from what is now southern Colombia down to central Chile, encompassing somewhere in the range of ten to twelve million people across extraordinarily diverse terrain. It had roads, armies, storehouses, a sophisticated taxation system, and a postal relay network of trained runners called chasquis who could carry information at remarkable speed.
What it did not have was a written language in any conventional, Western sense. No alphabet, no syllabary, no ink on page. For a long time this puzzled European scholars who treated a phonetic alphabet as a strict prerequisite for what they categorised as "civilisation." But the Inca hadn't failed to invent writing. They had found a completely different, three-dimensional solution to the same problem.
How the knots actually worked
The basic structure of a quipu is a primary horizontal cord from which subsidiary strings hang vertically, with further strings sometimes branching off those. What made it a recording system rather than simply a piece of weaving was the extraordinary density of encoded information embedded in its physical properties. Specially trained officials known as Quipucamayocs – the word translates roughly as "those who know the quipus" – controlled the creation and reading of these objects.
In the accounting quipus, the type of knot mattered: a long knot, a single knot, and a figure-eight knot each carried distinct numerical values. The position of the knot along the string corresponded to a decimal place – units, tens, hundreds, thousands. The colour of the string carried meaning, with different hues associated with different categories of information: people, livestock, goods, tribute. Even the ply direction of the string – whether it twisted left or right – added a further binary layer of data.
It was a genuinely effective information system. Spanish administrators in the decades immediately following the conquest actually kept using quipus for practical record-keeping in some regions, because the system worked and the people who understood it were available. Colonial bureaucrats, whatever they thought of Inca culture in the abstract, knew good accountancy when they saw it.
More than a ledger
For most of the modern era, the scholarly consensus stopped there: quipus were brilliant calculators, but ultimately numerical. That consensus has been eroding.
Researchers like Harvard anthropologist Gary Urton have been unpicking a different theory. They've noticed that some quipus possess far more combinatorial complexity than pure accounting would ever require. Urton identified up to seven binary features in quipu construction, generating a potential sign inventory of over a thousand distinct units. For context, Sumerian cuneiform used roughly the same number of signs. The implication is that at least some quipus may have encoded not just numbers, but narrative – histories, administrative correspondence, perhaps something closer to literature.
We have evidence for this from early Spanish colonial chronicles. Writers like Pedro de Cieza de León and El Inca Garcilaso de la Vega recorded Indigenous informants claiming that trained specialists could "read" histories, genealogies, and battle accounts directly from the strings. These sources require some care – Garcilaso published his accounts in 1609, more than seventy years after the conquest, interpreting a fractured system through a complex colonial lens. But if they were even partially right, the quipu wasn't a calculator. It was a library.
So where are the books?
What the Spanish burned
When the conquistadors arrived in the 1530s, they dismantled the Inca state with ruthless efficiency. Quipus presented a particular problem: the Spanish couldn't read them, couldn't audit them, and deeply distrusted them. By the 1580s, Catholic authorities at the Third Council of Lima formalised that suspicion into policy, decreeing that quipus containing religious or historical records were idolatrous and required destruction. The burning accelerated.
The scale of what was lost is difficult to sit with. Of the hundreds of thousands of quipus that must have existed to administer an empire of that size, approximately 900 to 1,000 survive today, scattered across collections in Peru, Europe, and North America. Many are damaged, decontextualised, and separated from the oral traditions that made them fully legible.
You often hear comparisons to the burning of the Library of Alexandria. But in some ways this is worse. When Alexandria burned, the Greek alphabet survived elsewhere. When the Spanish burned the quipus, they didn't just burn the books – they destroyed the decoding system. The Quipucamayocs who carried that knowledge in their minds were gone within a generation.
The strings we still can't read
We can still extract numbers from many surviving quipus – the accounting system is substantially understood. What we cannot read are the histories. Researchers like Sabine Hyland have found what appear to be phonetic elements in certain rare quipus, hinting at a possible breakthrough. And unlike with Egyptian hieroglyphics, it isn't that no translations exist – Spanish scribes did record what Quipucamayocs read aloud, in documents called papel de quipu. The real tragedy is that no one has yet been able to definitively match a single surviving Spanish transcript to a single surviving physical quipu. The parallel text exists. The physical link between them has been lost.
I think about that a lot. The Inca built and administered a superpower using knotted strings, and five centuries later we're still staring at those strings, locked out. It's a reminder that human knowledge doesn't always look the way we expect it to. But more than that, it's a reminder of what conquest actually costs – not just in the moment, but in the permanent, irreversible silencing of the things a civilisation considered worth remembering. The quipu's silence isn't a failure of the technology. It's what happens when you erase the people who knew how to make the strings speak.