Textile Society of America


Date of this Version



Textile Narratives & Conversions: Proceedings of the 10th Biennial Symposium of the Textile Society of America, October 11–14, Toronto, Ontario


Copyright 2006 by the author.


Considering the various mathematical ideas we’ve been exploring briefly, how might these differ in textiles from ancient Peru?

In the ancient textiles that I study, Andean people seem to have made conscious use of the math embedded in textile processes for generating several kinds of graphic codes. You may have heard about a device called the khipu: it is a bunch of colored and knotted cords that looks like a string mop, but it is really an ancient record-keeping device. The knots stand for numbers in the base 10 system. Long knots were used for numbers from 2 to 9, while clusters of single knots stood for the 10’s, 100’s, and 1000’s. The colors of cords were likely used for categories of things, and the direction of the twist for another kind of distinction. Although no one can read a khipu, we have a pretty good idea of how the fiber qualities and structures mathematically encoded information such as census figures or quantities of goods collected through taxation during the Inka period, c. 1500 A.D.

Two thousand years ago, people on the South Coast of Peru were embroidering mantles that look something like khipus. Some Linear-style mantles from the site of Paracas Necropolis have many parallel bands with reverse-curve snakes repeated within them. The snake-like patterns usually alternate three colors, and look like three-ply cords. In an example from the Museum of Fine Arts, Boston (1972.353), all of the band patterns twist in the same direction, except for one. The number of colors and bands, and the twist direction, are no doubt significant, as they vary from mantle to mantle in this style.

It is fairly clear that fiber structures were used as a model for some visual codes that we see on Andean textiles in different periods – particularly those patterns that have a geometric appearance. Headbands from the mummy bundles excavated at the site of Paracas Necropolis sometimes look like cords made of five plied elements. Other headbands from different bundles show from two to eight plied elements in their patterns.

Cord patterns are not the only fiber structures that occur. Braid structures made with three, four, or eight ends also occur as patterns on the headbands.