COVER STORY

How to read Inca We thought they left no written records. Now we’re untangling the truth, says Daniel Cossins

HE Incas left no doubt that theirs was a these devices was called a khipu (pronounced sophisticated, technologically savvy key-poo). We know these intricate cords to be Tcivilisation. At its height in the 15th an abacus-like system for recording numbers. century, it was the largest empire in the However, there have also been teasing hints Americas, extending almost 5000 kilometres that they might encode long-lost stories, from modern-day Ecuador to Chile. These were myths and songs too. the people who built Machu Picchu, a royal In a century of study, no one has managed estate perched in the clouds, and an extensive to make these knots talk. But recent network of paved roads complete with breakthroughs have begun to unpick this suspension bridges crafted from woven grass. tangled mystery of the , revealing the But the paradox of the Incas is that despite all first signs of phonetic symbolism within this sophistication they never learned to write. the strands. Now two anthropologists are Or did they? The Incas may not have closing in on the Inca equivalent of the bequeathed any written records, but they . That could finally crack the

JONNY WAN JONNY did have colourful knotted cords. Each of code and transform our understanding of >

29 September 2018 | NewScientist | 33 a civilisation whose history has so far been The Inca royal palace contains details of more than 900 of them. told only through the eyes of the Europeans of Machu Picchu There are all sorts of varying factors in who sought to eviscerate it. (right) remained khipus: the colour of the strings, the structure The Spanish conquistadors, led by hidden to the Spanish of the knots and the direction in which they Francisco Pizarro, first encountered the conquistadors – as did were hitched. Having spent countless hours Incas at the start of the 1530s. They were their system of writing poring over them, Urton began to think that awestruck by the magnificent stone cities, in khipus, or knotted binary differences in these features might be the gold and treasure. But as the Spanish cords (below) encoding information. For example, a basic began to take over the and knot tied in one direction could mean “paid”, impose their own customs, they became while in the other it would mean “unpaid”. equally enthralled by the way the society By 2012, he had developed a more specific was organised. hypothesis, proposing that the direction in The Incas governed the 10 million people which knots were tied, the colours of the in their realm with what amounted to a strings, or some combination of the two, federal system. Power was centred in Cusco, corresponded to the social status of the in the south of what is now , but spread people whose tributes they recorded, and through several levels of hierarchy across a even individuals’ names. Without a khipu series of partially self-governing provinces. translation, however, the idea looked destined to remain untested. “ Break the khipu code and Then in 2016, Urton was browsing his personal library when he picked out a book we might finally read an that contained a Spanish census document indigenous Inca history” from the 1670s. It was what the colonists referred to as a revisita, a reassessment of There was no money and no market six clans living around the village of Recuay economy. The production and distribution in the Santa valley region of western Peru. of food and other commodities was centrally The document was made in the same region controlled. People had their own land to and at the same time as a set of six khipus in farm, but every subject was also issued his database, so in theory it and the khipus with necessities from state storehouses in were recording the same things. exchange for labour, administered through Checking it out, Urton found that there an impressive tribute system. were 132 tribute payers listed in the text and Historians have argued variously that 132 cords on the khipus. The fine details fitted

the Inca empire was a socialist utopia or an CREATIVE GEOGRAPHIC HOPKINS/NATIONAL LEE RALPH too, with the numbers on the cords matching authoritarian monarchy. But no one disputes the charges the Spanish document said had its efficiency. “It was an extraordinary There are reasons to think khipus may He demonstrated that each row of knots at a been levelled. It seemed to be the match he system,” says , an anthropologist record other things, including stories and certain height denoted units, tens, hundreds had been looking for. at Harvard University. “Administratively myths – the sort of narrative information and so on. That made sense, fitting with the Even so, Urton was struggling to pick speaking, it was very sophisticated and it that many cultures write down. De la Vega was decimal system the Inca used to divide up apart the detail of the connections between seems to have worked well.” among many chroniclers who hinted as much, groups for tribute purposes. the Santa valley khipus and the Spanish Key to that success was the flow of reliable writing in one passage that the Incas “recorded documents. He ended up letting a Harvard data, in the form of censuses, tribute accounts on knots everything that could be counted, undergraduate student named Manny and storehouse inventories. For that, the Incas even mentioning battles and fights, all the Hard knot to crack Medrano take a look. He turned out to have relied on the khipumayuq, or the keepers of embassies that had come to visit the Inca, The discovery sparked a wave of interest in the perfect complement of skills for the the khipus, a specially trained caste who could and all the speeches and arguments they had khipus. By the 1990s, though, we still had job. He was a native Spanish speaker and, tie and read the cords. uttered”. True, he was prone to ambiguity and no idea what the numbers meant. “Say you majoring in economics, he was a whizz The majority of surviving khipus consist contradictions. But about a third of the khipus read off the number 76 – what does it refer with spreadsheets. Medrano painstakingly of a pencil-thick primary cord, from which in collections seem to have a more elaborate to?,” asks Urton. generated tables of the khipu data and hang multiple “pendant” cords and, in turn, construction than the others, as if they To answer that, you would ideally have a combed through them in search of matching “subsidiaries”. The Spanish described how contain a different sort of information. translation of a khipu into a familiar language. patterns. This year, he and Urton showed for they were used to record all manner of For decades the point was moot, however, It would be an equivalent of the Rosetta the first time that the way pendant cords are information. The poet Garcilaso de la Vega, because no one could read any of them. stone, which contained a translation of tied onto the primary cord indicates which son of an Inca princess and a Spanish The first hints of revelations from khipus Egyptian hieroglyphics into ancient Greek clan an individual belonged to. conquistador, noted in a 1609 account that came in the 1920s, when anthropologist and unlocked that picture language. In the “It is a really important achievement,” they had “an admirable method of counting Leland Locke analysed a bunch of them absence of that, Urton has spent the last 25 says Jeffrey Splitstoser at George everything in the Inca’s kingdom, including housed at the American Museum of Natural years tracking down and carefully digitising Washington University in Washington DC, all taxes and tributes, both paid and due, History in New York. He noticed that the the details of every khipu he could find in who specialises in khipus from the Wari which they did with knots in strings of knots are organised in rows almost like museums and private collections across the empire that preceded the Inca. “It gives us

different colours.” ©THE TRUSTEES OF THE BRITISH MUSEUM beads on an abacus (see diagram, page 37). world. Today, his Khipu Database Project a new way to interpret these sources. Gary >

34 | NewScientist | 29 September 2018 29 September 2018 | NewScientist | 35 has made things a lot more tractable.” Yet the could this be a syllabic writing system?’,” question of whether the khipus also contain says Hyland. She has since hypothesised that stories still hung there. the khipus contain a combination of phonetic Urton was not the only one trying to find symbols and ideographic ones, where a meaning beyond numbers and names symbol represents a whole word. in khipus. Sabine Hyland, an ethnographer at Earlier this year, Hyland even managed to St Andrews University in the UK, has spent the read a little of the khipus. When deciphering past decade searching in the central Andes for anything, one of the most important steps communities with enduring khipu traditions. is to work out what information might be She starts by looking for mentions of khipus repeated in different places, she says. Because in archives, before travelling to remote the Collata khipus were thought to be letters, villages in the hope they might have survived. they probably encoded senders and recipients. The strategy tends to be more miss than hit, That is where Hyland started. She knew from but in 2015, Hyland’s persistence paid off. the villagers that the primary cord of one of Having seen a documentary about her work, the khipus contained ribbons representing a woman in , Peru, got in touch about the insignia of one of two clan leaders. the khipus in the remote village of San Juan She took a gamble and assumed that the de Collata, where she grew up. After months ribbons referred to a person known as Alluka, of negotiations with the community, Hyland pronounced “Ay-ew-ka”. She also guessed that was invited to see two khipus. Villagers believe the writer of this letter might have signed them to be narrative epistles created by local their name at the end, meaning that the last chiefs during a rebellion against the Spanish three pendant cords could well represent the

JORDI BUSQUE/NATIONAL GEOGRAPHIC CREATIVE BUSQUE/NATIONAL JORDI in the late 18th century. By that time, the syllables “ay”, “ew” and “ka”. people spoke Spanish too, so there are corresponding written records. The khipus were kept locked away in an Tangled mystery underground chamber in the village church. Assuming that was true, she looked for cords Hyland and her husband were the first on the second khipu that had the same colour outsiders to lay eyes on them, and she was and were tied with the same knot as the ones she had tentatively identified on the first “ This writing system is three- khipu. It turned out that the first two of the last three cords matched, which gave “A-ka”. influenced by the alphabet, I still think it’s narrative khipus, even if he has a different dimensional, dependent The last was unknown. It was a golden-brown mind-blowing that these people developed idea on how they encoded information. He on touch as well as sight” fibre made from the hair of a vicuna, an this tactile system of writing,” she says. suspects they are semasiographic, a system alpaca-like animal. Hyland realised that She will spend the next two years doing of symbols that convey information without not disappointed. “It was an incredible the term for this hue in the local Quechua more fieldwork in Peru, attempting to being tied to a single language. In other moment,” she says. “But I didn’t have time to language is “paru”. And trying this alongside decipher the Collata khipus and looking for words, they would be akin to road signs, where be awestruck because this was my big chance the other syllables gave, with a little wiggle similar examples elsewhere. we all know what the symbols mean without to study them, and I didn’t have long.” She room, “Yakapar”. That, it turned out, was the Urton too is turning his attention to having to sound anything out. That makes

KIKE CALVO/NATIONAL GEOGRAPHIC CREATIVE GEOGRAPHIC CALVO/NATIONAL KIKE LYNN JOHNSON/NATIONAL GEOGRAPHIC CREATIVE had 48 hours before the man in charge of the name of another of the lineages involved in sense, given that the Inca ran a multi-ethnic, khipus, the village treasurer, had to travel to the revolt that these khipus recorded. multilingual empire, says Urton. Inca inventions a nearby community festival. “We know from the written testimony that There is no solid evidence that any Spaniard Under strict supervision, Hyland set about one of the khipus was made by a member of living at the time learned to read or make a You need only look at the the mountainous areas stories, not just numbers photographing the cords, reviewing the the Yakapar clan and sent to Collata, and we khipu. That suggests that they were more archaeological site of where the Incas lived, they (see main story). manuscripts and taking notes. Each khipu think this is it,” she says. Hyland claims that complicated than conventional writing – Tambomachay to see how also constructed terraces to They certainly went to had hundreds of pendant cords, and they were the Collata khipus show that the cords really or perhaps just conceptually very different. creative the Incas were. The grow crops. It is thought that great lengths to transport the more colourful and complex than anything do hold narratives. “This is a writing system that is inherently site shown (above, bottom they created experimental khipus. Couriers would loop she had ever seen. It was clear the various Yet even if she is right, it is possible these three-dimensional, dependent on touch as left) is near Cusco, once the agricultural stations too, the cords over their shoulders animal fibres used could only be identified by later khipus were influenced by contact well as sight,” says Hyland – and that presents Incas’ capital, and consists of such as the one seen above and run with them across the touch. The villagers told her the khipus were with Spanish writing. “My feeling is that the us with a uniquely tangled mystery. terraced rocks riddled with (bottom right), where they empire. To navigate the terrain, the “language of animals” and insisted that phoneticisation, if it’s there, is a reinvention It also gives us an important insight. If the aqueducts and canals. We tested which crops would a vast network of roads and the different fibres have significance. of khipus,” says Urton. Equally, the Collata Inca used khipus in this way, it might tell us don't know its function, but grow best on terraces at woven grass bridges were built. Her analysis eventually revealed that the khipus might be a regional variation. Possibly something about their world view. With a it may have been a military different altitudes. The last remaining bridge, pendants came in 95 different combinations even a one-off. writing system dependent on touch, says outpost or a spa for the Inca It seems odd that all this known as Queshuachaca (top), of colour, fibre type and direction of ply. That Hyland is the first to admit that we don’t Hyland, “you must have a different way of

political elite. Either way, sophistication arose but straddles a river high in the is within the range of symbols typically found understand the link between these khipus and HYLAND WILLIAM DR being in the world”. ■ it shows how the people writing did not. That is one Andes. Local people band in syllabic writing systems, where a set of signs those dating from before the Spanish arrived. could organise and build. reason to think their knotted together to renew the woven (say, the letters C-A-T) aligns with the sound That doesn’t make them any less interesting Sabine Hyland holds one of Daniel Cossins is a staff feature writer for With little flat ground in cords might record ideas and grass ropes every year. of speech (the word “cat”). “I thought ‘Woah, though. “Even if these later khipus were the incredible Collata khipus New Scientist

36 | NewScientist | 29 September 2018 29 September 2018 | NewScientist | 37