How the Inca Used Intricately-Knotted Cords, Called Khipu, to Write Their Histories, Send Messages & Keep Records

Those of us who learned to write in a (most­ly) pho­net­ic lan­guage learned to take it for grant­ed that writ­ing should cor­re­spond (rough­ly) to sound. Then we learned of the pic­tographs, ideo­graphs, and logograms of the Chi­nese alpha­bet, or of Ancient Egypt­ian or Mayan, or of oth­er non-phone­mic orthogra­phies, and we were forced to revise ear­li­er assump­tions. Those who pur­sue the study of sym­bol­ic sys­tems even fur­ther will even­tu­al­ly come to meet khipu, the Incan sys­tem of record-keep­ing that uses intri­cate­ly knot­ted rope.

Khipu, long thought an aba­cus-like means of book­keep­ing, has recent­ly been acknowl­edged as much more than that, coun­ter­ing a schol­ar­ly view Daniel Cossins sum­ma­rizes at New Sci­en­tist as the belief that the Incas, despite their tech­no­log­i­cal and polit­i­cal “sophis­ti­ca­tion… nev­er learned to write.” This Euro­pean logo­cen­trism (in the Der­ridean sense), per­sist­ed for cen­turies despite some evi­dence to the con­trary four hun­dred years ago.

For exam­ple, the poet Gar­cila­so de la Vega, son of an Incan princess and Span­ish con­quis­ta­dor, wrote in 1609 that the Incas “record­ed on knots every­thing that could be count­ed, even men­tion­ing bat­tles and fights, all the embassies that had come to vis­it the Inca, and all the speech­es and argu­ments they had uttered.” There may be some hyper­bole here. In any case, the point “was moot,” notes Cossins, “because no one could read any of them.”

Like most­ly illit­er­ate cul­tures in the West and East that relied on scribes for record-keep­ing, Incan civ­i­liza­tion relied on khipumayuq, “or the keep­ers of the khi­pus, a spe­cial­ly trained caste who could tie and read the cords.” As explor­er Ale­jan­dro Chu and Patri­cia Lan­da, Con­ser­va­tor of the Inc­ahuasi Arche­o­log­i­cal Project, explain in the Nation­al Geo­graph­ic video at the top, these spe­cial­ists died, or were killed off, before they could pass their knowl­edge to the next gen­er­a­tions.

But the lin­guis­tic code, it seems, may have been cracked—by an under­grad­u­ate fresh­man eco­nom­ics major at Har­vard named Man­ny Medra­no. As Atlas Obscu­ra report­ed last year, Medra­no, work­ing under his pro­fes­sor of Pre-Columbian stud­ies, Gary Urton, spent his spring break match­ing a set of six khipu against a colo­nial-era Span­ish cen­sus doc­u­ment. He was able to con­firm what schol­ars had long assumed, that khipu kept track of cen­sus and oth­er admin­is­tra­tive data.

More­over, though, Medra­no “noticed that the way each cord was tied onto the khipu seemed to cor­re­spond to the social sta­tus of the 132 peo­ple record­ed in the cen­sus doc­u­ment. The col­ors of the strings also appeared to be relat­ed to the people’s first names.” (Now a senior, Medrano’s find­ings have been pub­lished in the jour­nal Eth­no­his­to­ry; he is first author on the paper, “indi­cat­ing that he con­tributed the bulk of the research”).

This research shows how khipu can tell sto­ries as well as record data sets. Medra­no built upon decades of work done by Urton and oth­er schol­ars, which Cossins sum­ma­rizes in more detail. Oth­er ethno­g­ra­phers like St. Andrews’ Sabine Hyland have had sim­i­lar epipha­nies. Hyland chanced upon a woman in Lima who point­ed her to khi­pus in the vil­lage of San Juan de Col­la­ta. The vil­lagers “believe them to be nar­ra­tive epis­tles,” writes Cossins, “cre­at­ed by local chiefs dur­ing a rebel­lion against the Span­ish in the late 18th cen­tu­ry.”

After care­ful analy­sis, Hyland found that the khi­pus’ pen­dant cords “came in 95 dif­fer­ent com­bi­na­tions of colour, fibre type and direc­tion of ply. That is with­in the range of sym­bols typ­i­cal­ly found in syl­lab­ic writ­ing sys­tems.” She has since hypoth­e­sized that khipu “con­tain a com­bi­na­tion of pho­net­ic sym­bols and ideo­graph­ic ones, where a sym­bol rep­re­sents a whole word.”

Hyland grants it’s pos­si­ble that lat­er khi­pus made after con­tact with the Span­ish may have absorbed an alpha­bet from Span­ish writ­ing. Nev­er­the­less, these find­ings should make us won­der what oth­er arti­facts from around the world pre­serve a lan­guage West­ern schol­ars have nev­er learned how to read.

Attempts to deci­pher khi­pus use all sorts of com­par­a­tive meth­ods, from com­par­ing them with each oth­er to com­par­ing them with con­tem­po­rary Span­ish doc­u­ments. But one inno­v­a­tive method at MIT began by com­par­ing Incan khipu with stu­dent attempts to cre­ate their own rope lan­guage, in a 2007 course led by the “Khipu Research Group,” a col­lec­tion of schol­ars, includ­ing Urton, from arche­ol­o­gy, elec­tri­cal engi­neer­ing, and com­put­er sci­ence.

“To gain insight into this ques­tion” of how the code might work, the syl­labus notes, “this class will explore how you would record lan­guage with knots in rope.” Maybe you’d rather skip the guess­work and learn how to make a khipu the way the Inca may have done? If so, see the series of six videos above by Har­vard Ph.D. stu­dent in arche­ol­o­gy, Jon Clin­daniel. And to learn as much about khipu as you might ever hope to know, check out the Khipu Data­base Project at Har­vard, whose goal is to col­lect “all known infor­ma­tion about khipu into one cen­tral­ized repos­i­to­ry.”

Relat­ed Con­tent:

Take a Vir­tu­al Tour of Machu Pic­chu, One of the New 7 Won­ders of the World

How to Write in Cuneiform, the Old­est Writ­ing Sys­tem in the World: A Short, Charm­ing Intro­duc­tion

Trigonom­e­try Dis­cov­ered on a 3700-Year-Old Ancient Baby­lon­ian Tablet

Josh Jones is a writer and musi­cian based in Durham, NC. Fol­low him at @jdmagness