With National Science Foundation support, Dr. Gary Urton, will address the problem of how to interpret information stored in 500+ year-old knotted cord records - the Inka khipu. Khipus were the knotted-string devices used for record keeping in the Inka Empire of Pre-Columbian South America. Beginning in 2002, Dr. Urton and colleagues began collecting and recording detailed observations on khipu samples in museum collections in Europe and North and South America. With observations on hundreds of khipus in hand, Urton set about building a searchable database -- the Khipu Database (KDB) -- which currently holds information on some 520 of the 850+/- extant khipus in museums around the world. Samples in the KDB pertain to at least a dozen "archives" that were retained at different locales within Inka territory. This is the first time an entirely searchable khipu database has existed in the history of the study of this enigmatic recording system. The khipu was said by Spanish chroniclers of the 16th century to have been used for recording information on state administration, histories, songs and other forms of narratives that were performed by cord-keepers. While students of the khipu have succeeded in determining how numerical data were recorded by state administrators (for census, tribute records, etc.), we lack a basic understanding of how personal names, place names, identifiers of events, etc., were recorded in the cords, knots and colors of the khipu. Decipherment of the latter is the principal objective of this project.

The intellectual merit of this research will be realized in the investigation of what was the unique, alternative communication technology and intellectual traditions whereby peoples living in the high Andes before the time of European contact communicated across frontiers separating speakers of multiple different languages as they succeeded in building the largest empire of the ancient New World in what was one of the most challenging and inhospitable environments on earth. The research will advance our understanding of human sign production and manipulation in early, pre-modern societies and should, if our efforts prove successful, result in the decipherment of the khipu. This project will hopefully allow us to interpret, or "read" the knotted-cord records left by the Inkas - the lords of the last great empire of Pre-Columbian South America.

The method of analysis to be followed is based on the hypothesis that khipu recording signs were registered in the form of semasiographic (meaning sign) units of information stored on the cords and arrayed in the form of cord structures, knots and patterned arrays of colors according to standardized formatting and encoding principles constituting conventionalized sign values. Focus will be on the study of what are termed "cord types" -- standardized configurations of cord structure, colors and numbers -- found within individual samples and across the entire khipu corpus. "Translation" values for cord patterns will be sought in Inka political, social, economic, ritual, and calendrical institutions and practices that are structured and organized in patterns similar to those identified in the khipus.

Project Report

This project addressed the problem of how the Incas of Pre-Columbian South America recorded information on their unique three-dimensional, knotted-string recording device, known as khipu (or quipu). We have long known that quantitative information (e.g., on censuses, tribute, etc.) was recorded in the base-10 numbering system of Quechua, the language of administration in the Inca Empire. However, we have to date not determined how identities (e.g., names of people, places, products, etc.) were recorded on these devices. This matter is important because with the khipu, the Incas were able to communicate information in a manner completely different from every other ancient civilization (the Mesopotamians, Egyptians, Chinese, and Mayas). If we can determine how they encoded information in this unique recording system, we may gain significant insights into alternatives to information recording that are relevant today. The khipu is particularly relevant in this regard as it depended on the binary coding of sign elements, on a principle analogous to the binary coding used in computers. The project analyzed data, using a database developed at Harvard University, from study of some 620 of the known 870 samples in museums around the world. Significant advances were made in a variety of areas of khipu recording, most notably, in administrative accounting. The principal advances made in this regard came from study of a recently excavated sample of 28 khipus excavated at an Inca storage facility (Inkawasi) on the south coast of Peru. Close study of these khipus shows that the khipu accountants employed a number of different "checks-and-balances" accounting techniques, some quite similar to the Western system of double-entry bookkeeping. The record-keepers at Inkawasi also devised a way of establishing a standardized unit of measure by marking off the floors of their storage bins in a precise grid composed of squares measuring 23cm. x 23 cm. This is the first time we have had evidence of the establishment of such standardized measures, a development which promises to give deeper insights into the quantitative and semantic values used in khipu record-keeping. Finally, considerable time was spent on the study of a set of 6 khipus from the Santa River Valley, on the north-central coast of Peru. The numerical data recorded on these 6 khipus matches quite closely data recorded (in writing) in a 17th century Spanish colonial tribute document. The document contains the names of 134 tribute payers; these identities are possibly recorded (most likely in color coding) on the 6 khipus. It is possible that this convergence between a colonial document and 6 khipus represents the long sought-after "Rosetta khipu!" In sum, significant advances were made with this project in terms of deepening and broadening our understanding of the recording of information -- both administrative and narrative -- in the semasiographic recording system of the khipus of the Inca Empire of ancient South America. These advances will contribute to on-going, comparative studies in information sciences and writing systems, worldwide.

National Science Foundation (NSF)
Division of Behavioral and Cognitive Sciences (BCS)
Standard Grant (Standard)
Application #
Program Officer
John E. Yellen
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Harvard University
United States
Zip Code