The Khipu Database Project
At Harvard's Peabody Museum of Archaeology & Ethnology, researchers manage an important effort to decode the Inca Empire's unique data storage system: the khipu. The Inca, who governed a vast South American territory without a graphic writing system, used these knotted cords to administer their empire. Specialists called khipukamayuqs, or knot-makers, could create and interpret the khipus, which recorded census data, tribute payments, storehouse inventories, and possibly historical narratives. The Khipu Database Project, started in 2002 by Professor Gary Urton, aims to catalog every known khipu from collections worldwide. The database contains detailed records of over 600 of the approximately 1,400 khipus known to exist.
The devices consist of a main primary cord with numerous thinner "pendant" cords attached, which in turn can have their own subsidiaries. Information is encoded through multiple variables. The construction uses materials like cotton and camelid fibers (llama or alpaca). About two-thirds of surviving khipus appear to be quantitative, using a base-10 system. Knots tied at different levels on a cord represent place values for ones, tens, and hundreds. A significant portion of khipus are non-numerical and remain undeciphered. Researchers believe these "narrative" khipus may contain stories, names, or other linguistic information.
A Three-Dimensional Binary Code
The attempt to decipher khipus at Harvard has led to a compelling theory that they represent a three-dimensional binary coding system. Gary Urton proposed that at each stage of a khipu's construction, the maker made a series of binary, either/or choices. This system involves at least seven important variables that can be set in one of two ways. These choices include the material (cotton or wool), the direction of the spin and ply of the yarn (S-ply or Z-ply), the color (from a palette of at least 24 shades), and the direction of the knot itself.
This structure means each knot functions as a "seven-bit binary array". This system could generate 1,536 potential information units, a number greater than the estimated 1,000 to 1,500 signs in Sumerian cuneiform and double the symbols in Egyptian or Mayan hieroglyphics. The database allows researchers to run complex queries, searching for patterns across hundreds of khipus to find matches or repetitions that could unlock this ancient code. By systematically analyzing these structural and symbolic elements, the project moves closer to a unified theory of how the Inca recorded complex information in knotted strings.