Abstract
Despite ongoing efforts to compile both Andean khipus and their written colonial references, initiatives in this domain have emphasized the benefits of aggregation vis-à-vis preservation and diffusion, largely forgoing opportunities to analyze khipu data in aggregate. This article introduces multivariate statistical analysis to colonial khipu texts, enlisting the aid of a heretofore little-studied source: the Textos Andinos, a compilation of sixteenth-century Spanish transcriptions of Indigenous khipu “readings.” The largest syntactically annotated corpus of khipu transcriptions to date is compiled. Textual interpretation informs an exegetical typology of “paper khipus”—a division of the texts into distinguishable categories. The initial typology is expanded using the outcome of its statistical evaluation. Pre- versus postconquest content and the incorporation of currency emerge as the primary distinguishing attributes of khipu transcriptions. The expanded typology in turn enables the assessment of previous hypotheses in the study of paper khipus, responding to criticisms of their generalizability; suggestions of a diminishment in khipu complexity following the Spanish conquest are revisited and corroborated to this effect. A corpus-based study of khipu transcriptions offers a promising inroad to negotiating the highly mediated conditions of their original creation while expanding the study of khipus in the early colonial Andes. The aggregative methodology is proposed to ethnohistorians as an additional strategy for complementing and enriching historical interpretation.
