talk SLS 2004

In Knots: emergent knowledge systems and the Inka khipu

Katie King
Women's Studies, University of Maryland, College Park
a talk presented for the panel "Narrative and Emergent Knowledge,"
at the annual meetings of the Society for Literature, Science and the Arts,
Chicago, IL, 11 November 2005

 

The khipu is a string-record recently redated as perhaps as much as 5,000 years old. Yet khipus were in some kinds of active use in the Peruvian highlands as recently as 50 years, or two generations ago. Until the Conquest in the 17th c., the khipu served as the primary record-keeping technology of the Inka Empire.

For the purposes of this very short talk I am going to assume rather than argue that emergent reconfigurations of knowledge production under globalization create, for better and for worse, new cognitive environments. With new angles of vision and new cognitive tools we examine our pasts differently. A case in point is this current reexamination of the Inka khipu, until recently considered "not writing." The work of ethnomathematician Gary Urton and that of anthropologist Frank Salomon force reconsideration of what writing is, and whether the khipu's binary codings are records of poetry and narratives or if the khipu is a quote "simulation device" for planning and dispersing obligations and entitlements in layers of locals and globals in the Inka Empire. How does thinking about the khipu help us also think about emergent knowledge systems?

In order to listen to this talk you may have to access two cognitive skills that might be understood to emerge from products shaped by globalization processes, although perhaps not uniquely. Science journalist Steven Johnson calls these skills respectively "probing" and "telescoping." Imagine you have just received in the mail that longed for piece of computer software which you have been coveting for a while. Excitedly you load it into your computer and immediately set to the task of examining it, not waiting for one minute to read the documentation, but instead engaging its possibilities. You figure out what it does by analogy with your previous knowledge of this or other programs, you experiment and make mistakes testing its and your limits, and you hunt for artifacts like the joking "Easter eggs" sometimes deliberately present in software, or as inadvertent configurations accidentally possible. This is the activity Johnson calls "probing," an engagement that looks for possibilities and tests limits. This is the kind of activity I am engaged in here in this talk and with which you are being invited to also participate.

The second skill often emerges in the course of playing computer video games. "Telescoping" is what you do when you play a game organized in nested hierarchies. To move from level to level you must simultaneously work out what you need to do on a specific focused level, while keeping in mind both the ultimate level toward which you are working, and the next few levels beyond the one you are engaging right now. You are continually simultaneously pulling out and collapsing levels as you dynamically shift focus for varying amounts of time and attention. A variety of material and cognitive structures are organized in such levels, often in something akin to nested hierarchies. I call such levels "layers of locals and globals" to emphasize their relative and relational character, and I use this term to describe both cognitive actions, for example among levels or layers of abstraction and concreteness, or for material organizational structures, such as those essential to social organization in the Inka Empire, and to various local and global regionalisms under globalization today. Such telescoping may be required to analyze what might be called emergent knowledge systems.

I have given you two handouts for this talk, one visual and one textual. The visual one is a cut and paste job from the illustrations from Gary Urton's MacArthur award winning book, Signs of the Inka Khipu, binary coding in the Andean knotted-string records. On one side you see a picture of a piece of a khipu, above it a schematic of possible construction elements in a khipu, below it an illustration of opposite spin-ply relationships in the yarn making up this fiber technology. Andean conceptual systems are radically dualistic: a common person might wear a tunic woven from yarn spun z or clockwise and plied s or counterclockwise, while a pacu or shaman might wear a tunic woven from yarn spun s and plied z. On the right hand of this side of the handout you see at the top a schematic of a bit from a khipu mapping out s- and z-knots, and below that a schematic of the 7 bit binary code Urton theorizes the khipu uses. He calculates that this system could manipulate 1536 unique units, comparable to the sign capacities of early cuneiform, Shang Chinese ideograms, and Egyptian and Mayan hieroglyphs. Seven types of information are coded in binary forms: the material a string is made from, the color of each string and its spin/ply relationship, how it is attached to other cords, what s or z direction the knot is tied in, which of two number classes it belongs to, and which of two kinds of khipu string it might be, either one for recording numbers, or, Urton theorizes, one used to record histories, "Quechua/Inka poetry or other ritual, canonical narrative forms."

The other side of this handout illustrates some of these various kinds of knots, in the upper left side giving a schematic of the binary coding of a single knot. Urton's question is: "How could one 'write' using strings, knots, and colors, rather than pen, paper, and graphemes?" And "why didn't the Inka just invent an iconic, grapheme-based system of signing information? I think the most satisfactory explanation for the question…relates to the status and physical characteristics of the medium3⁄4cloth3⁄4of which the khipu represented one form, or expression. In a seminal article published forty years ago, John V. Murra (1962) instructed us on the central role of cloth as a marker of status, wealth, power, and authority in Inka statecraft. Clearly, cloth was not just any medium among the Inka; it was the medium of choice, and as such, the records of state were, not surprisingly, fabricated of this material.... Thus, the system we're considering here should not be conceived of as non-graphic and non-two-dimensional, as though the khipu can or should be defined by what it is not; rather, the khipu was (positively) three-dimensional and tactile.... it is as though all other ancient civilizations developed graphic scripts in forms similar to the images seen on our computer monitors, whereas the Inka devised a system more closely resembling the binary digital coding sequences of the ASCII code: out of view, inside the workings of the computer."

Urton goes on to point out: "khipu were constructed...as the means of recording information in a society that was typified to a extraordinary degree by dual organization3⁄4for instance, in its moiety systems as well as in other organizational and symbolic forms.... Thus, khipu were dual and binary in relation to their representative, reportative, and information-storage objectives and functions because they existed within a society organized in a dualistic manner…."

What does all this have to do with emergent knowledge systems? For the purposes of this talk I am considering three levels of complexity in relation to ideas of emergence. My quick and dirty definition of complexity here refers to one or more interlocking systems in which, as we say, "the whole is greater than the sum of its parts." I want to bracket the term "the whole" and instead call it "a something" to indicate that it does not have to be a complete totality: instead "a something greater than the sum of its parts." Add to this what we might call "complex behavior": in Steven Johnson's words: "a system with multiple agents dynamically interacting in multiple ways, following local rules...." The kinds of agents here might be those actants described within Latour and others' Actor-Network theory: not only persons, individual and in collectives, but also things, communications, forms of social organization and worldly processes. Johnson points out: "it wouldn't truly be considered emergent until those local interactions resulted in some kind of discernible macrobehavior...a higher-level pattern arising out of parallel complex interactions between local agents." We could refer to these higher-level macrobehaviors as "self-organization." Finally, the really interesting forms of emergence are forms of self-organization that are adaptive, creating more higher-level behaviors over time and adding learning through feedback loops. So we have on the first level complexity, then complex behavior, then complex behavior with self-organization, and finally adaptive self-organization. As you can see, a bit of telescoping is called for here. Try to hold all this in mind as I now address an alternative or perhaps a complementary theory of khipu meaning as advanced by anthropologist Frank Salomon.

Without necessarily challenging Urton's analysis of binary codes within the khipu, Salomon suggests that rather than narratives or poetry, what the khipu codes are simulations and records of the outcomes of simulations. Khipus used within living memory appear to have been worked in pairs. Salomon hypothesizes that one of the pair quote "was a programmatic or planning device…a way to simulate a prospective agenda. As the agenda was realized, execution was recorded" on the other of the pair. "A dialogue between the two sorts of cord artifacts seems, then, to have generated the overall rationality of the community." Salomon goes on: "Cords made by different hands in different materials, as well as knots added and removed, and devices for immobilizing or restraining certain pendants3⁄4all these things strongly suggest that unlike texts, but like simulation devices such as abacuses or molecular modeling kits, quipocamayos were basically operational devices. The Tupicochan evidence suggests that in the planning stage, participants used cords as tokens to try out alternatives and assemble a plan3⁄4subject to further revision3⁄4and thereby rationally deploy labor and other factors of production to meet the complex demands of household/suprahousehold work in polycyclical scheduling. In execution of the plan, too, cords were used to register the completion of obligations as they were fulfilled.... Authority inhered in the nonverbal product. By incrementally adding constraints as assembly progressed, and by forcing attention onto relevant variables while filtering out the 'noise' of rhetoric, rivalry, and so on, the khipu operation would have helped villagers arrive quickly at a collective minimax rationality reconciling the interests of member households and the group.... The quipocamayo process would have compacted social process into an impressive data-dense medium whose clarity did not depend on expansion into words."

"Just as a thought experiment, let us suspend historical caution and imagine the reconstructed Tupicochan system as if it were part of the Inka state…. In this scenario, Inka administrators ...using an all community aggregate khipu, …tender proof of external compliance and entitlement to state authorities. Inka officials consult the village database in formulating their more standardized, more synoptic, less operable record, which will get attention at higher levels. At all levels, ceremony functions to structure modules of recordable action. ...Redistributive celebration using the goods documented is a likely concomitant. Without taking comparison to levels of spurious detail, this outline seems a plausible image of the drama for which both Inka-era ceremonial plazas, …and colonially derived local assembly places, …formed stages."

There are three possible associations I might make here between these various khipu scenarios, and the various levels of complexity surrounding the idea of emergence. The first involves what Salomon calls the "data writing" of the khipu. Data writing is a kind of writing that does not map onto language but instead maps onto features in a social world co-constituted, or "articulated" with such writing: as Salomon puts it "Life had to be lived like a khipu." Such a khipu simulation device operates a range of nested hierarchies, performing layers of locals and globals, cognitive, material, social organizational, at levels of complex behavior and self-organization with opportunities for a range of adaptive possibilities. As the editors of the series Salomon's book The Cord Keepers is in put it: "The Incas were a great mystery—at least according to many Western pundits who could not understand how a complex, highly stratified empire, stretching from southern Columbia to northwest Argentina, with a road system larger than Rome's and a political organization that incorporated millions—could have existed without a 'true' or European-like system of writing and accounting. The Incas' closest instrument was the khipu—a set of knotted cords that served, in ways we hardly understand, as the nerve system of an empire."

The second set of associations with emergence might come from the new technological assemblages that Urton has taken his MacArthur money to create as opportunities for scholarly and alternative self-organizing activity. The new Harvard Khipu Database, created with the help of mathematician, weaver and ethnographer Carrie Brezine, is one element in a complex network of khipu knowledge sites. Such sites are created, shared, and used in many writing technological forms: not only monographs, books, conference talks, but also websites, databases, images, exhibitions, reenactments, television documentaries, tourist and heritage tours, sites and festivals, as well as village and kinship ritual work processes. Gender and nationality, ethnicity and race, indigenous politics and university restructuring, all play roles in such systems animated together with their scholars, curators, collectors, website managers, spinners, conservators, tour creators, television production people in ranging infrastructures. It is out of such necessarily self-organizing local activity that any "code-breaking" of the Inka khipu must come.

Finally, at the level of our own complex behavior encountering such knowledges, from Donna Haraway's recent work her thoughts on what she calls "emergent practices": quote "vulnerable, on-the-ground work that cobbles together non-harmonious agencies and ways of living that are accountable both to their disparate inherited histories and to their barely possible but absolutely necessary joint futures.  For me, that is what significant otherness signifies."

--------------------------------------------------------------------------------------
handout for: Panel on Narrative and Emergent Knowledge (11 Nov 2005);
Conference on "Emergent Systems, Cognitive Environments,"
Annual Meetings of the Society for Literature, Science and the Arts, Chicago IL
In Knots: emergent knowledge systems and the Inka khipu
Katie King, Women's Studies, University of Maryland, College Park / katking@umd.edu
Home Page: http://www.womensstudies.umd.edu/wmstfac/kking/

ABSTRACT: Emergent reconfigurations of knowledge production under globalization create, for better and for worse, new cognitive environments. With new angles of vision and new cognitive tools we examine our pasts differently. A case in point is the current reexamination of the Inka khipu, a string recording device previously considered "not writing." The work of ethnomathematician Gary Urton and anthropologist Frank Salomon forces reconsideration of what writing is, and whether the khipu's binary codings are records of histories and narratives or if the khipu is a "simulation device" for planning and dispersing obligations and entitlements locally and "globally" in the Inka Empire. How does thinking about the khipu help us also think about emergent knowledge systems?

From Steven Johnson (2005):
probing — learning the rules of a simulation by trial and error, while necessarily also checking out its edges, limits and unexpected artifacts or patterns
telescoping — apprehending simultaneously all the structures of nested hierarchy and mobilizing them in various sequences

Gary Urton (2003): "How could one 'write' using strings, knots, and colors, rather than pen, paper, and graphemes?" And "why didn't the Inka just invent an iconic, grapheme-based system of signing information? I think the most satisfactory explanation for the question…relates to the status and physical characteristics of the medium3⁄4cloth3⁄4of which the khipu represented one form, or expression. In a seminal article published forty years ago, John V. Murra (1962) instructed us on the central role of cloth as a marker of status, wealth, power, and authority in Inka statecraft. Clearly, cloth was not just any medium among the Inka; it was the medium of choice, and as such, the records of state were, not surprisingly, fabricated of this material.... Thus, the system we're considering here should not be conceived of as non-graphic and non-two-dimensional, as though the khipu can or should be defined by what it is not; rather, the khipu was (positively) three-dimensional and tactile.... it is as though all other ancient civilizations developed graphic scripts in forms similar to the images seen on our computer monitors, whereas the Inka devised a system more closely resembling the binary digital coding sequences of the ASCII code: out of view, inside the workings of the computer." (37, 40-2)

levels of complexity in relation to ideas of emergence:
Complexity: a something greater than the sum of its parts
Complex behavior: in Steven Johnson's words (2001): "a system with multiple agents dynamically interacting in multiple ways, following local rules...." (19)
Agencies: The kinds of agents here might be those actants described within Latour and others' Actor-Network theory: not only persons, individual and in collectives, but also things, communications, forms of social organization and worldly processes.
Complex behavior with self-organization: Johnson points out: "it wouldn't truly be considered emergent until those local interactions resulted in some kind of discernible macrobehavior...a higher-level pattern arising out of parallel complex interactions between local agents." (19)
Adaptive self-organization: creating more higher-level behaviors over time and adding learning through feedback loops.

Salomon (2004): "was a programmatic or planning device…a way to simulate a prospective agenda. As the agenda was realized, execution was recorded" on the other of the pair. "A dialogue between the two sorts of cord artifacts seems, then, to have generated the overall rationality of the community." Salomon goes on: "Cords made by different hands in different materials, as well as knots added and removed, and devices for immobilizing or restraining certain pendants3⁄4all these things strongly suggest that unlike texts, but like simulation devices such as abacuses or molecular modeling kits, quipocamayos were basically operational devices. The Tupicochan evidence suggests that in the planning stage, participants used cords as tokens to try out alternatives and assemble a plan3⁄4subject to further revision3⁄4and thereby rationally deploy labor and other factors of production to meet the complex demands of household/suprahousehold work in polycyclical scheduling. In execution of the plan, too, cords were used to register the completion of obligations as they were fulfilled.... Authority inhered in the nonverbal product. By incrementally adding constraints as assembly progressed, and by forcing attention onto relevant variables while filtering out the 'noise' of rhetoric, rivalry, and so on, the khipu operation would have helped villagers arrive quickly at a collective minimax rationality reconciling the interests of member households and the group.... The quipocamayo process would have compacted social process into an impressive data-dense medium whose clarity did not depend on expansion into words." (265)

"Just as a thought experiment, let us suspend historical caution and imagine the reconstructed Tupicochan system as if it were part of the Inka state…. In this scenario, Inka administrators ...using an all community aggregate khipu, …tender proof of external compliance and entitlement to state authorities. Inka officials consult the village database in formulating their more standardized, more synoptic, less operable record, which will get attention at higher levels. At all levels, ceremony functions to structure modules of recordable action. ...Redistributive celebration using the goods documented is a likely concomitant. Without taking comparison to levels of spurious detail, this outline seems a plausible image of the drama for which both Inka-era ceremonial plazas, …and colonially derived local assembly places, …formed stages." (274)

"Life had to be lived like a khipu." (278)

Walter D. Mignolo, Irene Silverblatt, Sonia Saldívar-Hull: "The Incas were a great mystery—at least according to many Western pundits who could not understand how a complex, highly stratified empire, stretching from southern Columbia to northwest Argentina, with a road system larger than Rome's and a political organization that incorporated millions—could have existed without a 'true' or European-like system of writing and accounting. The Incas' closest instrument was the khipu—a set of knotted cords that served, in ways we hardly understand, as the nerve system of an empire." (xv)

Haraway (2003): "Answers to these questions can only be put together in emergent practices; i.e., in vulnerable, on-the-ground work that cobbles together non-harmonious agencies and ways of living that are accountable both to their disparate inherited histories and to their barely possible but absolutely necessary joint futures.  For me, that is what significant otherness signifies." (7)

Some materials referred to or possibly of interest:

• Haraway, D. (2003). The companion species manifesto: Dogs, people, and significant otherness. Chicago: Prickly Paradigm.
• Johnson, S. (2001). Emergence: The connected lives of ants, brains, cities, and software. New York: Scribner.
• Johnson, S. (2005). Everything bad is good for you: How's today's popular culture is actually making us smarter. New York: Riverhead.
• King, K. (2004). Pastpresents: Knotted histories under globalization. Unpublished manuscript.
• Quilter, J., & Urton, G. (Eds.). (2002). Narrative threads: Accounting and recounting in Andean khipu. Austin: U Texas.
• Salomon, F. (2004). The cord keepers: Khipus and cultural life in a Peruvian village. Durham: Duke.
• Urton, G. (2003). Signs of the inka khipu: Binary coding in the andean knotted-string records. Austin: U Texas.
• Urton, G., & Brezine, C. (2003-2005). Harvard khipu database project.   Retrieved from http://khipukamayuq.fas.harvard.edu/index.html
• Zerubavel, E. (1997). Social mindscapes: An invitation to cognitive sociology. Cambridge: Harvard.

© November 2005 Katie King. Citation: King, Katie. "In Knots: emergent knowledge systems and the Inka Khipu." Paper presented at the panel on "Narrative and Emergent Knowledge" at the Annual Meetings of the Society for Literature, Science and the Arts Conference: Emergent Systems, Cognitive Environments," Chicago, 11 November 2005. Available online at: http://www.womensstudies.umd.edu/wmstfac/kking/present/KnotsSLSA05.html [last updated 11/8/05; <your visit>]