Agency in Ancient Writing

University Press of Colorado

Contents

The Mediated Image

Reflections on Semasiographic Notation in the Ancient Americas

Margaret A. Jackson

In recent years, the long-standing model of a dichotomous relationship between text and image, a mainstay of the semiotic approach to visual communication, has fallen increasingly under criticism. Our daily interactions with computers and icons of all kinds make it obvious that the world is populated by countless varieties of conventionalized and codified visual signs that operate outside the strictures of language. There are any number of visual codes that are neither purely text nor purely image. This realization is reflected by increased scholarly attention and greatly expanded curiosity about the workings of these alternate forms, which are known as "semasiographies." This shift ultimately makes the present analysis possible.

Scholarship on semasiography ranges from discourse analysis and visual culture studies to linguistics and cognitive science. It is a challenging array of disciplines; the category includes everything from symbolic diagramming and math to transit maps and, of course, pictographies and hieroglyphics — genres of great interest among Americanists. Such multiplicity points to the fact that "semasiography" is actually an umbrella term covering a range of graphic systems and operative structures.

Sometimes called "discourse systems" or "nonphonetic writing," semasiographies are capable of conveying relatively specificinformation in codified or conventionalized form, but they are not designed to directly reflect speech. Scholars such as Brice (1976), Sampson (1985), Boone (2004), Boone and Mignolo (1994), Elkins (1999), Houston (2004a, 2004b), Jackson (2000, 2008), Martin (2006), and others have sought to define "semasiography," which is based on the Greek semasia ("meaning") combined with graph ("drawn/written"). Yet, just as with words like "writing" and "pictures," the concept encompasses a vast, contested territory, making an absolute definition elusive.

Semasiography refers to those conventionalized systems of visual notation arranged primarily around nonphonetic principles of ordering whose overall meanings are derived from the spatial and/or performative relationships among the constituent elements. Semasiographies may incorporate a range of meaningful elements, including phonetic signs, logographs, ideographs, pictures, or direct material relationships. The systems are distinguished from complex iconographies by their capacity to create and record new information (Boone 2004, 317 — 18; Martin 2006, 76). They seem to work best (and occur most often) within strictly limited areas of reference (Sampson 1985, 30; Brice 1976, 41 — 43), even though, in theory, it need not necessarily be so.

Semasiographic systems are often active participants, or mediators, within specific performative contexts, as suggested by the title of this chapter. The idea of mediation is used here in its most traditional sense, that is, as something occupying an intermediate place or position, one that, in some instances, serves to mediate between parties to reconcile or bring about an accord or understanding. It is not used in its most contemporary sense, as a word that connotes a situation where traditional forms, like print or painting, have become "media" through translation into digital or electronic form — that is to say, older forms that become "remediated" (see Bolter and Grusin 1999, 34–50).

Because they focus less on phoneticism, semasiographic models offer a flexible — and less value-laden — way to address visual systems that are not directly tied to spoken language. This does, in fact, represent a new paradigm for studies of Ancient American visual art. The Saussurian dialectic that dominated semiotic discourse for most of the twentieth century, elaborated in what follows, has strongly affected the ways in which indigenous visual cultures in the Americas have been approached, in many cases perpetuating the idea that native people were illiterate (with the attendant pejorative associations). Semasiographic studies represent a means by which to alleviate and (one hopes) eventually eliminate this particular colonialist stigmatization. Apart from the social considerations, the introduction of a revised paradigm allows us to better understand and differentiate among communicatory approaches that may have varying degrees of verbal, visual, or other sensory modal components.

The focus of this chapter, therefore, is to illustrate different semasiographies through drawing attention to several Ancient American examples, summarizing some of the key features distinguishing them, not necessarily decoding the meanings of those examples but rather suggesting the places where semasiographic functions occur. In doing so, it becomes possible to discern variable quantities of pictures, phonetic signs, and operational structures in each visual tradition, which, in turn, bear important correlations to agency and, ultimately, to the kinds of meanings that can be extrapolated from particular object categories.

WORD/IMAGE

The notion that a sign is something that conveys an idea about a thing originates with Charles Sanders Peirce in his now-famous tripartite categorization of signs as icons, indexes, and symbols. Peirce describes an icon as a sign possessing the character of that which makes it significant (the more imagistic of the three), indexes show something about things because of their close connection to them, and symbols are associated with their meanings through conventionalized usage (growing from icons or mixed signs and characterized by varying degrees of codification). Ferdinand de Saussure's subsequent work served to recast Peirce's terminology, linking it in no uncertain terms to Saussure's concept of the linguistic sign, contesting the basic meanings assigned by Peirce in several key regards. In the Saussurian view, signs reflect highly arbitrary relationships between the signifier and that which is signified. (Language as a reflection of ideation serves as his prime example.) Along this line, and generally applied across disciplines, sign has come to mean a signifier retaining no vestige of any "natural" relationship to its source, lying in contrast to symbol, which may retain some likeness to its subject and carry more generalized associations of meaning. In the present chapter, Peirce's fundamental tripartite relationships among icon, index, and symbol are retained. As a designation, sign is used in its limited sense, to mean conventionalized forms whose traces arise from arbitrary or abstract relationships between signifier and signified. When intended in any broader sense, sign is qualified here by the addition of modifying adjectives. Symbol is meant to suggest a form with a relatively generalized base for conventionalization and interpretation, and icon and index retain essentially imagistic, relational connotations.

Writing is often described as a graphic representation of speech — a visual system whose primary focus is on the replication of spoken communication (Coe 1992, 13; Marcus 1992, 29). Although some authors suggest that writing is "any systematized graphic activity that creates sites of interpretation and facilitates communication and sense making" (Boone 2004, 315, citing Rotman 1995, 390), typically and traditionally, written scripts are categorized as those that record speech by phonetic values (as with alphabets) or those that do so according to meaning (logographs). Generally echoing a widely accepted dialectic, articulated at length by Saussure and others (Saussure 1983; Foucault 1993), semioticians have tended to place phonetic writing (spoken sound–based signs) in opposition to pictures (figural imagery). Visual signs more closely related to words have been posited as having a higher degree of discursiveness, and those that are more imagistic have been regarded as more figural (Bryson 1981).

Within a dichotomous model of word and image, semasiographies are conceived as existing somewhere between phonetic writing and pictorial imagery (Martin 2006). Any graphic form not purely phonetic or purely pictorial falls somewhere along a continuum, either more or less conventionalized and glottographic or with more or less iconographic or pictorial component (Figure 1.1).

Early attempts to clarify this intermediate territory include work by Charles Owen (1986), who placed linear sequential writing in opposition to "presentational" picturing (e.g., imagery with no conventionalized or fixed manner of reading). According to Owen, graphic systems residing between linear discourse and presentational images are hybrids, characterized as having configurations of symbols rather than strings of symbols. He noted that, because sequential systems, such as alphabets, are encoded linearly, they must be transmitted, received, and decoded in a strictly defined order; conversely, presentational systems are usually not linear-sequential in their method of communication and therefore may be decoded from more than one starting point. According to Owen, writing is at one end of the discursive scale, followed in descending order by mathematical systems, ideographic languages, and pictographies. These are followed by complex symbolic diagramming systems, such as music notation, Blissymbolics (a transit system notational form), dance notation systems (like Kinetography and Labanotation systems), complex process diagramming, sequence experience notation, and motation. Ultimately, he positions representational drawings, photographs, television, movies, and iconic models as the most presentational (non-sequential) visual forms (Owen 1986, 167–68).

The model of a dialectic continuum becomes problematic at this juncture, however, because certain genres, like mathematical charting systems that involve time, systems diagrams, organization charts, and concept presentations, show little or no sequential or grammatical structure, even though they clearly involve discrete sets of signs and operations. Such systems are almost entirely presentational (configurational) but are not idiosyncratic. It becomes almost impossible to correlate levels of phoneticism with discursive capacity in these systems. This observation creates an impasse of sorts, pointing to a problematic aspect of the binary model and signaling a paradigm whose center is intellectually unwieldy.

In The Domain of Images, James Elkins (1999) challenged the basic premise of the dichotomy, pointing out that there is no such thing as "pure" writing or pictures; all images are both "read" and "seen" simultaneously. They mix words, letters, marks, and iconic images within a larger experience of color, substance, and visual format. One cannot read words on a page without simultaneously seeing the size, color, and materiality of the page itself, nor can one view a picture without also being subjected to canonic structures governing formalist presentation and symbolism. Temporal duration and social function likewise affect discursive structures. To acknowledge this, and to address the enormous variety of visual forms occupying the intermediate territory, Elkins argued that graphic constructions like maps, diagrams, emblems, and schemata actually constitute a third node of visual communication, best referred to as Notation. A modified semiotic model would thus have a tripartite composition (as in Figure 1.2).

No longer would graphic forms fall only along a continuum between iconicity and conventionalization, but they would also have varying degrees of operational and/or spatial correspondences. Within this conception, the overlapping area again constitutes the realm of semasiography, but with greater differentiation. Hieroglyphics continue to reside somewhere between writing and pictures, whereas emblems and schemata, less phonetic but still conventionalized and logographic, situate more toward the Notational node. Mathematical script, with its high degree of conventionalization, finds its place even closer to Notation. In articulating the character of the semasiographic and notational categories, it becomes possible to distinguish the different components of graphic systems that include variable quantities of pictures, phonetic signs, and operational structures.

OPERATIONAL SEMASIOGRAPHIES

If a visual system is not intended to replicate speech per se, then what modalities does it engage? Nonphonetic graphic systems have the capacity to engage almost any cognitive branch, and they communicate across sensory modes. In many cases, their ability to mediate ideas distinguishes them as functional cognitive tools.

Mathematic script, for example, employs specialized signs for numbers, things, and actions to communicate spatial relationships between signs and operatives. Meaning is conveyed through the relative position and size of the symbols, and changes in positional relationships can significantly alter the overall meaning of any given grapheme (e.g., 2a means something quite different from a2 or a/2 or 2/a).

In addition to denoting complex quantitative or spatial relationships, such notations play a dynamic role as functional mediators. When math characters are manipulated in algebraic context, one can "solve" for unknown information. Because the values of the numerals and operatives change according to the spatial configurations in which they are placed, the characters and formulas can actually function to aid in calculation and conceptualization. Certain formal configurations, such as tables, can allow a reader to calculate predictable moments (as in almanacs or train schedules; Figure 1.3). Likewise, diagrams and models can allow viewers to visualize spatial relationships (like chemical or molecular structures).

These systems are obviously not vehicles for the recollection of narrative, and their internal syntaxes do not assume linguistic sequence. They are essentially tools for thought. As notations, they function predominantly as operational semasiographies, where meaning is derived specifically from the operations for which they were developed. In relation to Ancient American visual traditions, many of which are clearly notational, but differ dramatically from one culture to another, several examples of what appear to be operational systems are recognizable in the archaeological and ethnohistoric record.

In a 2004 essay, Elizabeth Boone (2004, 329) articulated an operational concept in relation to Ancient Mesoamerican manuscripts, presenting two excellent examples of Aztec divinatory books, which she likened to tables and charts. Boone noted that the opening pages of the Codex Borgia are schematized into the form of a table (see Figure 1.4; see also Boone 2004, figures 11.22, 11.23). Known as an in extenso almanac, the 260 days of the calendrical cycle are presented within the organizing structure of twenty day signs repeated thirteen times. The grid structure places deity patrons and mantic scenes above and below. The person using the document would count to the appropriate day and, depending on what mantic elements were associated, be able to extract shades of meaning associated with particular days.

Similarly, Boone noted that the famous quadripartite page from the Codex Fejérváry-Mayer is analogous to a diagram (Figure 1.5; see also Boone 2004, figures 11.24, 11.25). Its spatial configuration forms a cosmogram of a quadripartite universe, something its Aztec users would have considered a conceptual truth. It additionally expresses the 260-day calendar, in the form of a small ribbon of dots looping around the perimeter. As with the previous example, the proximity of various parts of the image created conceptual linkages for the person using the document. It is not an idiosyncratic mnemonic device; it contains well-understood graphemes and served as an active tool for the creation of knowledge.

In the Andes, a very different sort of operational device deserves similar consideration. The Andean knotted string khipu also served as a dynamic vehicle for information processing (Figure 1.6). In Inca society, these were used for accounting and, to a less well-understood degree, for communication of narrative elements (Quilter and Urton 2002). Khipus were not counting devices; they did not function as a calculator or an abacus in the sense of moving knots around during the process of mathematical calculations. Neither were they mnemonic structures akin to the medieval European "memory theater," the Art of Memory so eloquently discussed by authors such as Yates (1966) and Carruthers (1990). Nor did khipus function like rosary beads, simply as placeholders for those who would recite memorized materials. By virtue of mutually agreed symbolic values, the knots and strings functioned as readable notations, legible to those persons trained in their workings (Urton 2003).

Khipus recorded information by means of a primary trunk cord, with pendant and subsidiary cords, each with conventionalized knots and colors. Although they potentially included phonetic signs or other logographs, khipus were not designed for the purpose of replicating language. To record information, khipu specialists placed knot-signs in appropriate configurations and in correct relationship to other signs. Reading and operating a khipu depended on these same informed specialists to serve as interlocutors. Information was given voice as a result of readers' interactions with the khipus and other objects like counting stones or yupana,or on occasions of ritualized "telling" events (Harrison 2002, 279; Platt 2002, 259). As with numerical charts, the values associated with any given knot-sign depended on internal relationships of knots, strings, and position within the structuring framework.