Testimony from knotted strings: An archival reconstruction of early colonial Andean khipu readings

ABSTRACT

Despite ongoing efforts to document the use of khipus (Andean knotted-string recording devices) in the decades following the Spanish conquest, the scenes in which former Inka-era khipukamayuqs (cord keepers) rendered their cords for colonial observers remain unclear. This study confronts this historical blind spot, endeavouring to reconstruct early colonial khipu ‘readings’ by way of a heretofore unexplored occurrence: scribal modifications entered in transcriptions of khipu-based testimonies before the colonial high courts. Following an overview of evidentiary cord readings and previous treatments of scribal emendations, the prevalence of visible markings is assessed in the largest compilation of khipu transcriptions. Legal proceedings between 1558–1568 before the Audiencias of Lima and La Plata form two case studies. The documents are representative of other khipu transcriptions in lacking an exhaustive description of their own preparation; as such, the possibility that one or more is a secondary copy cannot be entirely excluded. Scribal modifications in three testimonies from the proceedings are deployed as heuristic devices to probe the presentation of ethnocategories and chronology by the cord keepers within this fraught documentary space. Despite the possibility of unattested documentary interference, the case studies attest to slippages in khipu reading that, in their sum, enable the proposal of punctuated narrative cadence, reversions to verbal shorthand, and relational discourse as recurring qualities of early colonial khipu readings. It is argued that this interpretive exercise enables closer study of the relationship between khipus and their written colonial references, serving other inquiries into ‘lost’ historical scenes.

KEYWORDS:

• Khipu (quipu)
• transcription
• colonial Andes
• legal archives

Acknowledgments

I would like to thank Sabine Hyland, José Carlos de la Puente Luna, and the anonymous reviewers for History and Anthropology, who provided constructive comments on earlier versions of this article. I am grateful for the generous support of the Marshall Aid Commemoration Commission. All translations are my own unless otherwise noted.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 I use the terms ‘rendering’ and ‘reading’ to refer to the vocalization of information recorded in khipus. I follow Brokaw (2003) and Puente Luna (2015) in this regard (among others), acknowledging that the latter term may evoke the longstanding debate – outside of the scope of this article – over whether khipus constitute a writing system (see Ascher 2002a for an overview; Bauer 2014, 349 calls for moving beyond this question).

2 Murra (1982, 245), at the time of his writing, pointed to the scarcity of paper khipus as a barrier to their analysis. Brokaw (2003, 142) has also identified corpus size as a relevant consideration.

3 Pärssinen and Kiviharju (2004, 84) claim that memorias were generally understood by Spaniards to be khipu transcriptions, though this overlooks earlier memorias of Peninsular origin (Huamanchumo de la Cuba 2019, 152n9).

4 I limit this comparison, acknowledging the rather totalizing dichotomy drawn by Lienhard and Cornejo Polar between orality (native) and literacy/writing (European). Quispe-Agnoli (2019), Fossa (2001), and Rowe (2003), among others, have offered critiques.

5 This document is of disputed provenience. A comprehensive overview is offered by Domínguez Faura (2008).

6 Durston (2008, 63, 63n71) points to the reversal of nouns and numerals in a Quechua-language petition (‘e.g., vallpa socta rather than socta vallpa for ‘six hens’’) as evidence that it may have been transcribed from a khipu, preserving the structure of the latter in moving from general categories (hens) to more specific information (counts of hens).

7 This is distinct from the idea that specific texts exhibit homologies with khipus. The latter, important in its own right, has been thoroughly developed elsewhere (Rowe 1985, 197–98; Brokaw 2001; 2002; 2003; Romiti Vinelli 2008; Rasmussen 2012, chap. 3; Bauer 2014). Different still is Ascher’s (2002a, 108–11) encipherment methodology, which constructs hypothetical khipus from surviving information (e.g. Pärssinen 1992, 34–50; Mazzotti 2000, 161).

8 This translation is chosen in accordance with the revision and English translation of Mazzotti’s original text (Mazzotti [1996] 2008).

9 A corpus linguistic analysis of the two volumes is offered in Medrano forthcoming.

10 ‘/₁₂₄₀ Yten le dieron nueue hanegas de quinua. /₁₂₄₁ Yten le dieron [nueve ha] çiento e treynta e seis hane- /₁₂₄₂ gas de papas.’ The palaeographic symbols (see table 1) correspond to those of Pärssinen and Kiviharju (2004, 77–78), except for the strikethrough notation that I use to denote crossed-out, but still-legible text. Several scribal modifications in the 1561 document (ii-a, iii-b, iv) are noted at the bottom of their respective folios with ‘va testado’ (glossed as ‘it is corrected’), though others (excerpts i, ii-b, iii-a) are not accompanied by this statement.

11 ‘Yten dieron al dicho Salazar çinquenta çestillos de pescado. /₆₂₉ [Yten] Después d'esto. [que] e muerto el marqués don /₆₃₀ Françisco Piçarro vinieron a estos tambos un fulano Salcedo.’

12 ‘/₇₇₉ Yten le dieron quinientos e quarenta e ocho corderos. /₇₈₀ [Yte] Después d'esto. subió desde los Reyes don Diego de /₇₈₁ Almagro el Moço a este repartimiento.’

13 ‘/₈₁ Yten ranchearon los dichos soldados [m] dos mil e trezientas /₈₂ e treynta e seis hanegas de quinoa.’

14 ‘/₁₅₁ Yten dieron al dicho tesorero y soldados [mill] siete mil /₁₅₂ e quinientas e ochenta e dos fanegas de maýz /₁₅₃ para su comida.’

15 ‘/₁₃₉₆ Yten les dieron mil e noveçientas e dos ycangas de /₁₃₉₇ [—]/pes\\cado/ [fruta].’

16 ‘/₃₆₃ Más le dimos de obejas 592 y de /₃₆₄ corderos 24 y le dimos 200 \puercos/ y le /₃₆₅ dimos 550 \gallinas/ y le dimos 4232 guebos … /₃₈₆ Más le dimos 601 anegas \de mayz/ y de quin- /₃₈₇ gua quatro anegas.’

17 Chirinos Rivera (2010, 65, 310–11) has argued that the Huanca cord keepers employed ‘coarse’ classifications, under which fruit, salt, and fish would all have been associated as items measured in baskets.

18 I thank the anonymous reviewer who brought this point to my attention.

19 Given the location of the chinu within the Andean cord tradition, I use the term ‘khipu’ throughout (Brokaw 2010, 2).

20 Lorenzo de Estupiñan, corregidor of La Paz and visitador general of the Charcas repartimientos, relayed Gasca’s tax to the area in 1551 (Platt, Bouysse-Cassagne, and Harris [2006] 2011, 322).

21 ‘/₅₀ Declararon que los quatro años pri- /₅₁ meros no aver dado ninguna ropa al dicho /₅₂ Martín de Robles. y desde que les fue /₅₃ dada la tasa por el dicho Estopinán /₅₄ dizen averle dado \en el quinto año/ veinte e dos pieças /₅₅ e media de ropa y en el sesto /₅₆ diez y seis pieças.’

22 ‘/₃₆₂ Yten declararon qu’el sesto año /₃₆₃ ansimismo sembraron para /₃₆₄ el dicho Martín de Robles otras diez /₃₆₅ e siete fanegas de maíz … /₃₇₈ El sétimo año lo mismo qu'el ses- /₃₇₉ to de la misma manera. /₃₈₀ El otabo lo mismo y de la /₃₈₁ misma forma y manera. /₃₈₂ digo qu'el [ses] otabo año senbraron solas /₃₈₃ quinze fanegas.’ ‘Digo’ is glossed here as ‘I clarify.’

23 The trial is perhaps the best-known of early colonial Andean legal proceedings to khipu scholars (Assadourian 1998; Platt 2002; Fossa 2000b; 2006, 366–77; Loza 2001, 82–87; Brokaw 2010, 201–13; Curatola Petrocchi and Puente Luna 2013, 194–205; Urton 1998).

24 Brokaw (2010, 50; see also 2003) refers similarly to ‘paradigmatic information structures,’ relational orders transposed into khipus from underlying sociocultural norms.