When European observers came face to face with the pristine societies of Central America in the late fifteenth and early sixteenth centuries, they began a process of reinterpretation of the exotic and unintelligible features of those societies into more comfortable concepts with which they could govern their new colonies. Native peoples acted as collaborators in this radical reformulation, first of the concepts of their own societies, and then of the institutions and ways of life that shaped those societies. Anyone trying, from the vantage point of the late twentieth century, to recover some of the quality of precolumbian society faces a formidable task. If we take as our goal the aim of making precolumbian Mesoamerica more intelligible, as did the earliest Europeans to enter into a dialogue with its people, then we ensure that we will come to understand these societies as different versions of the same thing, lesser or distorted mirror images of Europe.
Our alternative could rather be to make precolumbian Mesoamerica more distinct, to push at the limits of its strangeness from the Europe that absorbed and reformulated it over half a millennium. I try to do just that in this study, which began as a consideration of the ways that sexual difference affected the formation of political power in prehispanic Mesoamerica. In it I attempt to reexamine materials that have served to characterize Mesoamerican attitudes toward gender and status, guided not by the goal of increasing their transparency but by that of capturing hints they provide of the development of distinctive practices and ideas over more than three millennia.
The difficulties encountered in approaching Mesoamerica illuminate several issues that have emerged as central in recent cross-cultural analyses of gender and power. Among these, the most critical are the paradoxical nature of gender, as presumptively concerned with biological difference but at the same time not determined by such differences; and the erosion of genders as central categories of identity that follows from attention to the unlinking of gender from its "natural" grounding. While other themes—including the relationship between gender, status, and the value of work, and the ways that changing contexts and scales of interaction can affect relationships proposed between gender and status—are considered in the chapters that follow, these two concepts provide the fundamental grounding for the trail I pick though the rich evidence for men's and women's lives and actions in those later sections. It is with an analysis of these concepts that we must begin.
Gender, Nature, and Culture
When a consideration of women's existence and status became an explicit topic of anthropological research, with the publication of works like Woman, Culture, and Society in the early 1970s, the reality of "woman" as a unitary category was taken as a necessary starting point. Women were seen to be, as a category, universally disadvantaged with respect to men, a proposition stated as axiomatic in at least one influential introduction to kinship studies. Some of the most important early anthropological analysis of gender was devoted to explaining this presumed reality, and such explanation quickly focused on the symbolic dimension of femininity, as determined by women's biological role. Woman was to man as nature was to culture; and unlike man, who could remake himself socially, woman was irrevocably marked as of nature through the biological processes of menstruation and childbirth. In the most self-empowering versions of this logic, being the embodiment of the forces of nature was presented as a source of power for women; a power that—threatening as it was to men, who lacked this resource—was denied by men as a means to assert control over women. In its most blunt form, this thread led directly to the development of a mythology of an earlier stage of human society when women, representing a generalized Goddess, exercised the power rightfully theirs as the unique vessels of reproduction.
Although anthropological analyses of gender have long since abandoned this original schematic, and historical research has demonstrated its roots in conventional historical speculation of the nineteenth century, a residue of the original formulation of gender as a problem has remained and continually arises within even the most sophisticated analyses. While the historical narrative of women's natural power and fall from grace may have been thoroughly debunked, the naturalness of woman as a category of study, or even as the primary category of social analysis, has proven more tenacious. The major spurs to recognizing problems with the initial conceptualization of women's status in anthropological analysis have undoubtedly been analyses that introduce crosscutting dimensions of seniority, race, class, and sexuality into the presumed unity of the category woman. These studies, which by no means speak with one coherent voice, nonetheless share a beginning commitment to the ideas that gender cannot be taken as the original and natural ground for identity, that sharing specific biological features does not imply a unity of purpose or viewpoint, and that perhaps the presumption that the question to be studied is how given biological features are valued in a society is not, in fact, a particularly productive question.
Under the assumptions of a biologically based analysis of women's status, the goal of this book would be to trace the ways that women's association with fertility served as fodder for discourses that associated women's bodies and their potential with different social values. My procedure would be first to look for women in precolumbian societies—identifying them ideally by their biological characteristics in burials, by the depiction of such characteristics in representational media, and by the sex-specific behaviors that might be documented by the distributions of tools, ornaments, or vocabulary in written documents. Once having ascertained the presence of women, we could summarize the ways their experience differed from that of men, where they were and were not allowed to act, what they were allowed or required to do, whether and how they could claim credit for their actions. I could ultimately construct a history of the gradual tightening of opportunities for women as Mesoamerican states became increasingly centralized through time, and control of women's production and reproductive potential became more critical to the men in power. In fact, I have engaged in a series of studies that could easily be characterized in this fashion, producing in the end a narrative strangely similar to the accounts of European Goddess literature, indulging in what Margaret Conkey and Sarah Williams call a "narrative of closure" typical of the slide of archaeology into "origins research."
But in the process, in order to sustain this vision, I (or those summarizing my work) would have to ignore a number of uncomfortable points of severe ambiguity. The easily detectable points of discomfort—in the face of a procedure that should be guaranteed not to see ambiguity—include female-sexed skeletons whose associations are more like those of male-sexed skeletons than of other females in their own community; images in which identifiable primary or secondary sexual characteristics are absent, even when the body is nude; distributions of "women's" tools in settings where men are the expected participants (and the reverse); and texts in which gender is unmarked linguistically or apparent gender appears to be violated. These difficulties are perhaps epitomized by the ambiguity of a group of Classic Maya figures who wear a well-defined costume, and have serially been identified as women, male priests, women again, royal men dressed as women, royal men or women dressed as a bigendered deity, and royal women dressed as men dressed as a deity. How can the gender status of the individuals depicted beso fluid? How, analytically, can we even begin to talk about the relationships between gender and power if we cannot even identify the gender identity of our subjects?
Gender Identity and Gendered Performance
One way out of the dilemma may be to admit that because we must rely on secondary evidence when addressing precolumbian societies, we cannot ever really know the gender identity of our subjects. Perhaps we should be satisfied with analyses at a scale coarser than the individual, taking as our unit of study the household (or houseful), the site, the polity, the society, the culture.
Such a strategy suffers from two flawed assumptions. The first is the notion that a concern with the way gender works is ultimately a question about individuals. As numerous examples from Mesoamerica demonstrate, realistic models at any of these scales require assumptions about the way gender works, leaving us not with ungendered models, but with models that simply project certain gender arrangements, often with no selfconsciousness, into these societies. While gender is certainly a vital issue for real individuals in society, it also permeates models about society that rest on categorization and generalization about the behavior of people in groups.
The second problematic assumption embodied in the illusion of escape from the ambiguity of gender is the idea that gender is clearly an identity. This is the idea that underwrites the legacy of the original formulation of gendered research questions in anthropology, the concept that woman is a category with a privileged unitary status deriving from the common experience within society of persons with particular sexual organs. Here, confronting again the natural grounding of gender in biology, I find my thread out of this maze, the thread I will follow in the succeeding pages. What if gender is not simply the particularly contextualized socially valued construal of the body one is born with?
This alternative formulation of the issue has engaged a number of contemporary theorists in diverse fields. In biology itself, Ruth Hubbard has pointed out how the assumption that people are supposed to be either male or female guides action surrounding birth in North American culture, including the treatment of babies whose sexual characteristics do not conform to this expectation. Thomas Laqueur has shown how European concepts of gender difference structured the way sexual organs were represented and ideas about the development of the body were fostered, from the Classical Greek tradition through the twentieth century. His analysis demonstrates that what was seen depended on the prior existence of a theory about what there was to be seen. He specifically identifies the idea of immutable biological gender identity stemming from extreme difference (to the most basic biochemical level) as a relatively late idea in European history, replacing an ideology that saw sexual difference as a matter of degree of development of shared potential. As these scholars and others have demonstrated, the assumption of fixed dichotomous genders grounded in absolute distinctions in biology is a commonplace of the modern western European intellectual tradition. There is no reason to assume that this was a feature of other societies and other times, any more than it serves for the earlier history of European society. With its presumed grounding in the body gone, an analysis of gender must go beyond the characterization of the presence, recognition, and value of biological categories, and engage a theory of gender itself.
Judith Butler has argued strongly for the decoupling of gender from the "natural" body, and indeed, from any fixity that might allow characterization of gender as an aspect of "identity," something one is. Butler offers instead a vision of gender as activity, something one does, a kind of performance. In Butler's view, people perform gender, and their performances are the fluid medium through which gender is reflexively shaped within specific social settings. As performance, gender is a way of being in the world, a way of dressing, of using the body, of revealing, concealing, modifying, and presenting the physical self. Gendered performances are learned and practiced, and they gain their intelligibility through social acts of interpretation, that is, when others understand a performing body's gender. Gendered performance, in Butler's analysis, is particularly centered on sexuality, and the readings of bodies that are interpretations of gender are always concerned with sexual possibility?
Throughout the following chapters, I will pursue the argument that certain archaeological materials and settings were the media and stages for gendered performances. With Butler's insight as a guide, the confusion surrounding apparent cross-dressing in the debate about a particular Classic Maya costume becomes evidence for a gendered performance in which ambiguity of a particular kind is displayed and the lack of clarity becomes a deliberate aspect of the political manipulation that is implicit in all these images. If gender is not a fixed and biologically grounded identity, then in fact the fixing of specific gendered performance in everyday life and in representation takes on a slightly different significance and raises other questions. What were the reasons for the creation of the images, settings, and objects that appear in relation to sexual status in precolumbian Mesoamerica?
Gender, Power, and Material Culture
Material evidence is abundant for the presence of differently gendered individuals in Mesoamerican societies. Representations depicting the human form are among the earliest enduring media of Formative Period societies. A single low-fired anthropomorphic pottery figurine from Tlapacoya, at ca. 2300 B.C., actually precedes the adoption of clay as a medium for pots in the Basin of Mexico. Figurines accompany the earliest sequences of pottery known from villages on the Pacific Coast from Mexico to El Salvador, in the Mexican and Guatemalan Highlands, and along the Gulf Coast and in the Honduran lowlands, between 1500 and 800 B.C. Soon after the establishment of permanent villages using pottery as a medium for representation, societies in several regions employ monumental media, especially stone, to depict anthropomorphic images. Both small-scale and large-scale media of human representation continue in use up to the epoch of Spanish contact.
The earliest Mesoamerican settlements also yield evidence of body ornaments in a variety of more or less durable materials: greenstone, shell, iron ore, and other minerals. The testimony of figurines adds an abundance of other forms of body modification, especially elaborate treatments of hair. Along with meager but suggestive evidence from Formative Period burials, discussed in the next chapter, representational images suggest that differences in costuming were already developing along lines well illustrated in later Mesoamerican societies, featuring distinctive items of male and female costume combined with other elements related to age, status, and role. Like large- and small-scale images, distinctive costume ornaments are a part of all later archaeological assemblages in Mesoamerica.
The Formative Period villages of Mesoamerica themselves provided a third material medium for the performance of difference, including gender, in the form of the spaces created through the use of architecture. The placement of residences, houses and their surrounding spaces, and, in some settlements, of larger structures, with respect to each other and in relation to natural features, created different arenas for social interaction at varied scales. It was in and among these spatial arenas that everyday performance of gender took place. It is the existence of such different settings in Maya and Mexica communities of the sixteenth century that allowed the articulation of rules of circulation dictating where women of particular ages and statuses might and might not go.
The permanence of architectural settings and anthropomorphic representations contrasts with the fleeting and fluid nature of embodied gender performance. It is as if in Formative Mesoamerican villages the impermanent were being made permanent, precisely when villages were being transformed into parts of more or less stratified and centralized polities. I suggest that the flourishing of permanent media of gender during the Formative Period, including the costume ornaments cast in pottery, stone, and shell, is not an accident. I argue that rather than acting simply as reflections of a given biological identity, male or female, the media of gender—representations, costume, and action in space—were the means through which people sought to constrain and control the behavior of others in their communities. Once involved in attempts at constraint and control, these media constituted sites of discourses about what it was to enact gender within Mesoamerican societies: sites of formation of gender ideologies, but also of resistance to and transformation of those ideologies. The materiality of these media was central to their use as enduring propositions that transcended temporal limitations on the individual.
I draw here on the work of Paul Connerton about the differences that exist between what he called embodied and inscribed practices. Embodied practices—which I argue would include Butler's performances of gender—are personal, experiential, and yet inherently fleeting. While they are preserved in memory and may be the subject of social comment, accumulating reputation and surviving through oral performance, their primary power stems from their momentary sensory impact. Inscribed practices, in contrast, are social, shared among many viewers, and potentially transcend the spatial and temporal limitations of any individual. While they also have sensory impact, it is distinct from the engaged perspective of embodied practices. Connerton stresses the fact that inscribed practices are subject to conscious recognition to a degree not automatic for embodied practices, which can exist on a level of unselfconscious habit.
Michael Herzfeld makes similar points in his discussion of the conversion of time into history through architectural style in a Cretan village. He distinguishes between monumental time and social time, the former operating at a scale beyond the individual and the latter being measured by individual action, the embodied practice of Connerton. Herzfeld is particularly concerned with the political process through which local antiquities authorities impose on the inhabitants of particular buildings a set of requirements fixing the architectural form of the building in a particular monumental time. He illuminates how the control of the physical layout and ornamentation of space overrides the meaning with which residents imbue their house in everyday life. Herzfeld also shows that the imposition of control is not unchallenged, with the physical structure of buildings becoming the site of political cotiflict unavoidably publicly displayed. Houses demolished, altered, or transformed into the required form are large-scale and permanent witnesses to political gains and losses.
What Connerton and Herzfeld's analyses demonstrate is the importance of forms of material culture as media for struggles to determine what will endure within a society. Richard Parmentier discusses the way specific material features in the landscape of Belau are recognized as signs of and in history, "the permanent signs which function as present evidence of a significant past (p. 308)." He argues that consideration of "the different ways history is connected to notions of time" requires attention to "the distribution within a given society of power to control the significance of events by creating or destroying historical evidence and by constructing historical discourse for specific ideological ends" (p. 5). Parmentier suggests that the selection of specific media as signs of/in history is motivated by their "sedimenting" status as "concrete embodiments" "endowed with the essentialized or reified property of historicity" as a result of their original association with the events for which they stand (pp. 11—13). Connerton, Herzfeld, and Parmentier are all explicitly concerned with the ways in which nontextual media carry extraordinary power in the creation of histories that transcend individual lives.
The media of gendered performance in precolumbian Mesoamerica served to inscribe bodily practices, to transform social time into monumental time. The efforts that were invested in creating these media throughout Mesoamerican history were substantial, including the provision of raw materials and the support of craftworkers in the exploration, mastery, and exercise of their skills. These investments of time, people, and materials become more intelligible when we take the products of this work seriously as assertions of social control. To fully explore the use of such media in dynamics related to the definition and control of sexuality and gender, however, we need to do more than recognize that they have unique kinds of power. We need to ask how they worked, in what ways they may have affected their makers and users, and in what ways they could have been creatively reworked to support alternative interpretations. These questions lead us to a consideration of the semiotics of gendered performances and representations.
Gendered Performance and Representation
Gendered performances in Mesoamerica included both everyday activities and the marked events in which concepts of gender themselves were explicitly being shaped: life-cycle rituals and ceremonies tied to the shared agricultural and divinatory calendars. While the performances themselves were fleeting, their settings and regalia could survive and provide a means to approach gendered performance archaeologically. At the same time, the inscription of bodily practices in the form of more permanent representational media provides material for examination of deliberate attempts to shape gender.
Gendered performance in everyday life left behind the residues of activities ranging from food preparation and consumption though specialized craft working. The crucial questions to ask about these performances are not the obvious ones that call for assigning tasks to fixed genders and delimiting spatial arenas for men and women from the distribution of these residues. Instead, in the pages that follow, I will examine how and why specific kinds of action came to be representative of certain kinds of gender, so that among the Aztecs, to weave was to be female, but not only female: young, marriageable, and perhaps primarily of the nobility. Examining those kinds of actions that were stereotyped by gender in Mesoamerica repeatedly exposes logics of everyday practice stressing interdependencies between social actors rather than hierarchies among them.
Marked gendered performance in ritual or public ceremony illuminates the same logic, but here the playing out of gender is more selfconscious. In ceremonies, costume and forms of movement (athletic contests, dance, and procession) were prescribed for participants, but within the shared framework of the event, individual performances could distinguish one person from another. I suggest that Nancy Munn's concept of "beautification" as a social means to make a person more "persuasive" to others can be applied to Mesoamerican performances.
Munn describes specific social practices in Gawa, an Oceanian society, through which the sensory attraction of the body is deliberately increased by the addition of skillfully worked materials that distinguish one person from others, practices Munn labels "beautification." Beautification draws specific attention to an individual and contributes to the relative social assessment of that person, resulting in a quality of renown of great significance in this small-scale society. The identity built up though beautification and, in particular, the evaluations other people make of it become grounds for differential perceptions of effectiveness and hence of different patterns of action toward people perceived as more "persuasive" or "potent." Similar statements are embedded in other ethnographies. The Amazonian Cashinahua describe feather ornamentation as promoting dua, a quality of health and attractiveness, by acting as "medicine" (dau) and by making externally visible the inherent aura of each individual. The Amazonian Waiwai regard the beautification produced by wearing feather costume as "a profoundly political matter" in which individual attractiveness provides grounds for interpersonal competition: "the efforts put into beautification are one of several means of displaying one's suitability as a good candidate for marriage, a kind of visual persuasiveness...." Annette Weiner's discussion of moments of beautified gendered performance in the Trobriand Islands shows that some of the most important social settings for such bodily distinction are ceremonies when newly sexually active girls and boys perform in public dances and displays, leading to the identification of desired sexual or marital partners, as is also the case among the Waiwai, the Papuan Gawa, and the Cashinahua.
Western observers tend to see shell ornaments, feather costume, and the like as inherently beautiful, but analysis of practices of beautification does not require agreement between the aesthetic standards of the analyst and the subject of study. Evidence of efforts expended to distinguish personal appearance though modification of the body can be grounds to consider beautification and its relative social evaluation. For example, while the wearing of palm-leaf skirts in the Trobriands is described by Annette Weiner in terms similar to Nancy Munn's discussion of beautification on Gawa, the aesthetic dimensions of differences in these skirts are not as immediately evident to Western sensibilities as those of shell valuables, but they are clear to Trobriand observers. Weiner compares gender performances in the Trobriands, where male renown was created though participation in kula exchange and women's value was asserted though display of the products of agricultural labor during mortuary rituals. Both performances are subject to social evaluation by others.
The recognition of shared standards for adornment and the measurement of each attempt at beautification against those standards demarcates a community that argues about what is beautiful. The anthropological subdiscipline of ethnoaesthetics is in fact based on this understanding. Ethnoaesthetic studies use a methodology of asking members of a production community to rate or rank craft products, arriving through this means at an understanding both of the general standards shared, and also of the points of contestation, where different makers disagree. Failure of agreement is as much a part of the establishment of a common community as is agreement. Norman Whitten presents an ethnographic example from the Ecuadorian Amazon in which the skillful creation of pottery and its social evaluation are explicit bases for the formation of individual power. His ethnographic account shows that the existence of variation in the understanding of pottery decoration is intrinsic to dynamics within the community, establishing internal differences.
In performance in ancient Mesoamerica, the incorporation of common practices of body Ornamentation and movement was linked with the expression of individual distinction in a tension that at times was related specifically to sexuality (see Chapter 5). The identification of marked gendered performance in precolumbian societies is possible because the regalia and settings of action were executed in permanent media, and because representational imagery further inscribed these performances. In Mesoamerica, costume was a medium for creation of specific social identity through a cumulative layering of age-, gender-, and status-appropriate ornaments and items of clothing. In images from sites such as Tikal and Calakmul, signs for individual names and titles of personages from the textual systems of the Maya were placed directly in costume, along the fringe of a robe or inserted in a feathered headdress. The incorporation of names in costume transformed dress into a literal text, an extreme version of the particularistic historical character of costume.
While costume documented personal history, it also served, perhaps primarily, to beautify the body. I suggest that beautification, modification of appearance by addition of materials that enhanced the body or were themselves objects of admiration, was particularly important to newly sexually mature young adults in Mesoamerica. Because beautified persons engaged in performance were members of larger social groups, their success was both individual and social, meriting the support provided for them by the group through the creation of distinguished costumes. In Mesoamerica, as was also true in Indonesia, Oceania, and the Northwest Coast of North America, costumes prepared for the use of younger members of social groups often employed heirloom objects, retained by the group and passed on though generations along with the histories they accumulated though their use in individual performances. It is no accident that such heirloom ornaments were conserved by Mesoamerican societies, including the Classic Maya, and that costume items were one of the major classes of valuables exchanged in marriage negotiations. Imperishable costume ornaments survive archaeologically and provide us the permanent, lnscrlbed evidence of fleeting performances, and testify to the social importance of performance to larger social groups.
Examples of heirloom valuables retained by Mesoamerican peoples include "Olmec" jade objects found in Postclassic settings at the Aztec Templo Mayor of Tenochtitlan and on Yucatec Maya Cozumel; jades in a Late to Terminal Classic context in northeast Yucatan similar to Middle Formative examples from La Venta; objects deposited in Chichen Itza's Sacred Cenote during Terminal Classic to Early Postclassic times that range from Middle and Late Formative styles typical of Belize to Late Classic jades carved with inscriptions referring to Palenque and Piedras Negras; and the incised jade celt-shaped belt pendant known as the Leyden Plaque, whose inscription records Early Classic events at Tikal, but which was found near the mouth of the Motagua River in a Terminal Classic or Postclassic setting.
That heirloomed costume ornaments were displayed, not simply reburied, by later peoples is supported by the image on a stela from Late Classic Yaxchilan, described by Carolyn Tate as wearing an Olmec-style figural pendant. Through their re-use, heirloomed costume ornaments accrued histories. Literate societies added notations of specific events to heirloomed objects. These could be added long after the original creation of the ornament itself: a pendant from Dumbarton Oaks, stylistically Olmec, bears on the reverse a low-relief incised figure and inscription in Early Classic Maya style.
The common form of the brief inscriptions on Classic Maya costume ornaments is a possessive statement: "his or her [object typel, name and titles of a person." More extensive texts recording political ceremonies were incised on jade and shell ornaments from sites such as Palenque, Piedras Negras, Tikal, and Altun Ha. Both kinds of texts served to fix the use of the ornament at a particular point in time, and in this sense, they are equivalent: they inscribe a specific history for and on the object. This history is necessarily carried along as it is transmitted from person to person.
Many of the texts on Classic Maya inscribed heirlooms are selfreferential, reinforcing the role of text in creating histories for objects. "Name-tag" texts on costume ornaments reiterate the type of object: "his ear spools," for example. The Dumbarton Oaks pendant's Early Classic Maya text and image employ the same visual form, an image of the profile view of a seated person, to inscribe the "seating" of the named noble, in political office on what may be an item of regalia of that individual. The Leyden Plaque, which also records a "seating" event, depicts its standing subject wearing a belt ornamented with a set of three pendant celts like the Leyden Plaque itself, confirming the use in this ceremony of the same or, similar objects.
This layering of reference reaches a peak in a pendant recovered from the Cenote at Chichen Itza that bears a text referring to events at Palenque. The bead itself was oriented horizontally when worn. Its two opposing long faces are inscribed with a text and an image, both oriented perpendicular to the orientation of the ornament when in use. The text refers to events in the heir-designation of a young Palenque lord, literally described as "taking a step into" or "entering" the succession of rulers; the image depicts a standing male figure with its heel raised in the Classic Maya stylization of movement. This figure wears on his chest a horizontally oriented long bead pendant, like the object on which the image is incised.
The self-referentiality of ornaments inscribed with texts is more obvious, but other Maya heirlooms subject to recirculation and ultimate deposition in the Cenote at Chichen Itza also historicized their own presentation on the bodies of powerful nobles. The jade pendants in what Tatiana Proskouriakoff called Nebaj style show seated rulers in elaborate costume, wearing pendants of the same general size and shape. With or without written exegesis incorporated on them, these Classic Maya ornaments (and presumably similar heirloomed costume ornaments from less literate societies) materialized the occasions of their own use and served as permanent records of specific historical events. The prior ceremonies they recorded were cited as precedents in later performances through the wearing of the same costume ornament.
Costume ornaments, the lasting media of fleeting embodied performance, are material testimony both to social investment of raw materials and craft skills in marked performances, and to the social evaluation of these performances as aesthetically realized repetitions of earlier precedents. Permanent representations of performances transformed the spatial and temporal reach of these social investments. Marked in monumental time and executed at monumental scale, they served to transform the spatial settings in which they were located. These transformations were effected through a double movement inherent in representations, the focus of much of my analysis in this study.
This analysis is what Dan Sperber called "an epidemiology of representations" that "is not about representations but about the process of their distribution." Sperber argues that representations begin as private mental constructs. When given material form as public representations (whether performances or inscribed media) they may be transformed into new mental representations by others who witness them. Sperber's discussion views the spread of representations and the responses to them as a chain of transmission that the use of the term epidemiology explicitly likens to the spread of a disease: outside conscious control, moving inexorably from host to host. The beginning point of my analysis, then, is the stipulation that the anthropomorphic images and the performances that took place in the settings and using the costumes I discuss were public in this sense, inspiring new mental representations in those who viewed them.
Because these images suggest aspects of human appearance and behavior through resemblance, they are iconic, and of particular interest in the production of inequality. As Michael Herzfeld puts it, "iconic relations...because they either 'look natural' or can be 'naturalized,' are a good deal more labile [than symbols, "arbitrary" signs], and lend themselves with particular ease to totalizing cultural ideologies." Because they appear to be chosen merely to establish resemblance, the ideological dimension in selection of features for iconic representations is less obvious and constitutes a subtle means to spread stereotypes of natural or essential human behavior. I assume that no detail of Mesoamerican human representations is simply natural, but rather is constrained by and constrains assumptions about what possibilities existed for living people.
Each image can then be approached as what Roland Barthes called a "pregnant moment": "In order to tell a story, the painter possesses only one moment... [the image] will be a hieroglyph in which can be read at a glance.., the present, the past, and the future, i.e., the historical meaning of the represented gesture." Barthes suggested that even a single still image implies its history, actions before and after the moment. But since there are many possible sequences of action that could precede or follow from a given moment, "all images are polysemous; they imply, underlying their signifiers, a 'floating chain' of signifleds, the reader able to choose some and ignore others." While the patrons of Mesoamerican artists may have had in mind particular narratives, they could not completely determine how the viewers of those images would interpret and react to them.
Further, images do not exist in isolation, but in active contact with others. Contrasts between images provide one avenue to identify points of what Michael Herzfeld calls disemia: "At each level of social organization, the relations between insiders and outsiders are ordered according to topically distinctive principles, but they always remain predicated on the distinction between the inside and the outside of whatever social group is in question. This is disemia, a mode of organizing social knowledge through cultural form." Just as the evaluation of performances marks the difference between those who share standards and those who do not, so the differences that exist between images rest on differences between Herzfeld's insiders and outsiders. The profusion of detail in Mesoamerican images, often requiring extended exegesis for intelligibility, is one sign of the importance of images as sites of differentiation between groups and individuals whose narrative understandings of an image were distinct.
Barthes points to two simultaneous forms of signification that are present in any image. Each feature selected excludes a range of other possibilities, and creates more-or-less arbitrary associations with other instances of the same feature. Thus Classic Maya images depicting figures dressed in a netted skirt with a shark's head at the belt recall other examples of the same costume, and suggest that there should be some uniform quality shared by these figures. But at the same time, the features selected suggest connections with earlier and later actions, giving each image a unique character: at Classic Maya Piedras Negras, images of individuals seated at the top of a ladder invoke not only the ascent up this device, but an act of human sacrifice, by the placement of a dead body at the base of the ladder and a cloth marked by dark footprints along its length. The crowding of such elements in Mesoamerican images can be seen as a visual strategy through which an attempt is made to constrain the interpretive freedom inherent in images. The signs united in such implied narratives are bound in a relationship of "double implication: two terms presuppose one another," transforming chronological order to a logical binding "capable of integrating backwards and forwards movements" through the narrative. What is shown as a sequence of action is transformed into a causal chain, where each step leads inevitably to what follows. Narrative "seems to found in nature the signs of culture."
By placing politically charged images in particular spatial arrangements, Mesoamerican patrons provided a preferred body of images for this process of narrativization, and these arrangements can be used not only as guides to our understanding of preferred interpretations, but also as evidence of those topics around which the greatest tensions, and consequent desire for control, existed. The addition of text to images is yet another means adopted to further channel interpretation, a means widely used in Mesoamerica beginning in the late Formative Period. Treating texts as evidence of the desire to limit interpretation provides another indication of the sites of tension surrounding representational images, as texts selectively emphasize certain possible narratives and downplay others.
The remainder of this study, then, will trace an epidemiology of images, with particular attention to the fissures that open up in any proposed unified interpretation. I take images as a starting point because throughout prehispanic Mesoamerican history they were media for the formalization of ever more limited stereotypes of everyday and ritual performance, offering sanctioned precedents for citation. Following threads suggested by disjunctions between images, I also trace other less tractable, more diverse, less controlled media for gendered performances: the regalia and costumes through which individual members of Mesoamerican societies negotiated their own standing within social groups of varying scale. Along the way we will encounter practices and representations that I argue demonstrate that gender in precolumbian Mesoamerica was more fluid and negotiable than was suggested in the accounts left for us by the first European observers.