Table Of ContentTWIN TOLLANS
Chichén Itzá, Tula, and the Epiclassic to Early
Postclassic Mesoamerican World
JEFF KARL KOWALSKI &
CYNTHIA KRISTAN-GRAHAM
Editors
Dumbarton Oaks Research Library and Collection | Washington, D.C.
© 2011 Dumbarton Oaks
Trustees for Harvard University, Washington, D.C.
All rights reserved.
Printed in the United States of America
17 16 15 14 13 12 11 1 2 3 4 5 6 7
Library of Congress Cataloging-in-Publication Data
Twin Tollans : Chichén Itzá, Tula, and the Epiclassic to Early Postclassic
Mesoamerican world / Jeff Karl Kowalski and Cynthia Kristan-Graham, editors.
— revised ed.
p. cm.
Includes index.
Based on papers presented at the two-day colloquium “Rethinking Chichén
Itzá, Tula, and Tollan,” held at Dumbarton Oaks, Washington, D.C., on February
19–20, 2000.
ISBN 978-0-88402-323-4 (cloth : alk. paper)
ISBN 978-0-88402-372-2 (pbk. : alk. paper)
1. Chichen Itza site (Mexico)—Congresses.
2. Tula site (Tula de Allende, Mexico)—Congresses.
3. Indians of Mexico—Antiquities—Congresses.
4. Indian architecture—Mexico—Congresses.
5. Excavations (Archaeology)—Mexico—Congresses.
6. Mexico—Antiquities—Congresses.
I. Kowalski, Jeff Karl. II. Kristan-Graham, Cynthia.
F1435.1.C5T95 2011
972'.01—dc23
2011017155
Chichén Itzá, Tula, and Tollan
Changing Perspectives on a Recurring Problem
in Mesoamerican Archaeology and Art History
Cynthia Kristan-Graham Jeff Karl Kowalski
Reading Chichén Itzá, Tula, and the Toltecs1 These interpretations have moved through
several critical phases. Beginning in the nine-
Like the doubled feathered serpent columns that teenth century, the similarities between the two
adorn the entrances to their principal temples sites were noticed and ascribed to the spread of a
(Figures 1–2), the two sites of Chichén Itzá and near-mythical Toltec civilization. Then, during
Tula have been considered to represent “twin cit- the first half of the twentieth century, more care-
ies,” paired political capitals that share so many ful archaeological investigations at the two sites
aspects of architectural plan, sculptural reper- documented parallels in building plan, sculp-
tory, and iconographic motifs that they represent a tural forms, and iconographic motifs. Scholars
unique case of cultural contact and artistic conver- endeavored to explain them by correlating cen-
gence in ancient Mesoamerica (Jones 1995). Along tral Mexican and Yucatec Maya historical sources
with such effigy serpent columns, other important so that the departure of the priest-ruler Ce Acatl
visual parallels between these two sites include Topiltzin Quetzalcoatl from Tula would coincide
galleries and halls whose roofs were supported by with the arrival of K’uk’ulkan at Chichén Itzá,
rows of columns or pillars; pillars bearing images which was considered to have coincided with a
of soldiers wearing related butterfly pectorals Tula Toltec conquest of Yucatan. Finally, during
and “pillbox” helmets and carrying atlatl spear- the second half of the twentieth century—with the
throwers; reclining chacmool sculptures (Figure rise of processual and post-processual archaeol-
3); small atlanteans and standard-bearers; and ogy, more critical appraisal of historical sources,
relief sculptures featuring “jaguar-serpent-bird” and more detailed archaeological investigations—
(or “Tlahuizcalpantecuhtli”) icons and images the similarities between Tula and Chichén Itzá
of predatory animals and raptorial birds holding were increasingly explained as the result of con-
human hearts. That these two centers share an tact and collaboration between their rulers and
important ensemble of traits has generally been those of other Epiclassic and Early Postclassic
recognized, but how to interpret the significance of Mesoamerican sites and regions, involving efforts
this fact has been a perennial and often hotly con- to establish political legitimacy and long-dis-
tested problem. tance networks of commercial exchange and elite
11
figure 1 Feathered serpent columns, Temple of the figure 2 Feathered serpent column drums and
Warriors, Chichén Itzá. Photograph by Cynthia atlantean sculptures, Pyramid B, Tula. Photograph by
Kristan-Graham. Mark Miller Graham.
prestige in the wake of Teotihuacan’s demise and The close similarities between Chichén Itzá
the collapse of the Classic Maya cities of the south- and Tula, Hidalgo, were first pointed out and com-
ern lowlands (Figure 4). mented on in detail by the explorer and archaeol-
The Chichén Itzá–Tula paradigm that took ogist Désiré Charnay (1885, 1887). Writing about
shape at the end of the nineteenth century was the distinctive appearance of the feathered ser-
rooted as much in ethnohistory as in archaeology pent doorway columns of the Castillo at Chichén
and art history. While also directing the reader to Itzá (Figures 5–6), he noted that: “These shafts
Susan Gillespie’s wide-ranging and theoretically are almost an exact reproduction of a Toltec col-
sophisticated discussion about these issues and the umn we unearthed at Tula . . . The two columns
intellectual problems that the Tula and Chichén are found three hundred leagues from each other,
Itzá question has engendered, we will briefly exam- separated by an interval of several centuries; but if,
ine how the Chichén Itzá–Tula paradigm has been as we firmly believe, the Tula column is Toltec, the
framed over time, in tandem with diverse strategies other must be so too, for it could not be the result
of reading and interpretation. of mere accident” (Charnay 1887:341–343).
2 kristan-graham and kowalski
figure 3 Reclining chacmool sculpture in front of the feathered serpent columns in the portal of the Temple of the
Warriors, Chichén Itzá. Photograph by Jeff Kowalski.
Charnay’s recognition of the parallels between overview, correctly notes the “eye of the beholder”
Chichén Itzá and Tula, and his identification of nature of this problem, pointing out that for some
such shared features as “Toltec,” along with the Tula and Chichén Itzá “are practically mirror
presumption that the forms at Chichén Itzá are images of one another” (McCafferty, this volume),
reproductions of prototypes at Tula, has had a pro- whereas for others their site plans are quite differ-
found and continuing impact on Mesoamerican ent (Cobos, this volume; Lincoln 1986). As we hope
studies. Archaeological knowledge of both Tula this volume makes clear, there are significant dif-
and Chichén Itzá continued to grow through- ferences between the two sites, but they also share
out the later nineteenth and the earlier twentieth quite specific building forms, sculptural images,
centuries, and during that time several scholars and politico-religious symbolism. In some cases,
(e.g., Batres 1906; Spinden 1975 [1913]:205–208; see as, for example, in the parallels in plan, general
Gillespie, this volume) added details consistent sculptural program, and specific iconographic ref-
with the idea that a special relationship existed erences shared by the Temple of the Warriors at
between Chichén Itzá and Tula and that the dis- Chichén Itzá and Pyramid B at Tula (see Figures
tinctive traits they shared were evidence of their 9–10 in Kowalski, this volume, and Figures 3 and
Toltec cultural identity. 6 in Kristan-Graham, this volume), there are no
Charnay’s recognition of some important clear counterparts of such closely comparable
basic similarities between the two sites is diffi- buildings at any other sites occupying the some
cult to deny. Michael E. Smith, in his summary eight-hundred-mile distance separating the two
Chichén Itzá, Tula, and Tollan 3
figure 4 Map of Mesoamerica, showing the location of the principal sites referred to in this volume
(after Kowalski 1999:4).
capitals (although possible parallels and models which the rise of centers of high culture through-
occur at sites such as Alta Vista and La Quemada out Mexico and Central America was explained by
to the northwest of Tula; see Kristan-Graham, this attributing them to the impact of migrations of the
volume). Toltecs following the fall of Tula. This interpreta-
Although he perceptively noted the real corre- tion focused more on the impact of the Toltecs as
spondences between the two sites, Charnay’s inter- a super-“race” than on the accounts of Topiltzin
pretation was overly simplistic. Charnay shared Quetzalcoatl.
with other late nineteenth-century scholars an Although most of Charnay’s interpretations
interest in explaining the rise of “civilization” as of Mesoamerican prehistory, particularly his
a process in which superior knowledge and tech- “conviction that Toltec civilization was the foun-
nical advances were perfected by particular peo- tainhead of all the high cultures of Mexico and
ples or racial stocks, who then disseminated them Central America” (Jones 1995:7, 27–31), have long
to other regions, primarily through diffusionis- since been abandoned, his claims that the resem-
tic mechanisms such as large-scale migrations or blances between the architecture and art of Tula
conquests (see Gillespie, this volume, and Smith, and Chichén Itzá indicate that the two cities had
this volume, for further discussion). Thus Charnay a special relationship with one another, reflecting
adopted a positivistic attitude that relied on eth- their common identity as expressions of a “Toltec”
nohistoric sources (as well as artifacts) to present civilization, have exerted a continuing power in
a straightforward account of prehistoric events in Mesoamerican archaeology (see Tozzer 1957:188).
4 kristan-graham and kowalski
figure 5 The Castillo (Structure 2D5), a radially symmetrical pyramid-temple located at the conceptual center of
the North Terrace at Chichén Itzá. View from the northeast; photograph taken prior to the INAH restoration of the
1980s. Photograph by Jeff Kowalski.
Charnay’s characterization of the Toltecs as was the popular author of the day on Tula and the
an omnipresent people in Mesoamerica and his Toltecs, while Brinton’s works were read primarily
unilineal model of a highland Mexico-to-Maya by scholars. The fact that Brinton’s essay appeared
movement of people and cultural traits was both shortly after Charnay’s books were published indi-
too general and too myopic to explain the complex cates that there could scarcely be another writer
social processes that developed in the Epiclassic whom Brinton was so vehemently criticizing.
and Terminal Classic periods. Yet Charnay did Brinton characterized the Toltecs as a branch
contribute a new vision of Mesoamerica, because of the Nahua people whom the Aztecs elevated to
“for the first time the Maya and the Mexica of cen- “superhuman” status in order to legitimate their
tral Mexico are seen as a unity, and their joint ter- own claim to rulership in central Mexico (although
ritories as a super-zone” (Bernal 1980:127). he is probably guilty of overstatement when he
Charnay’s campaign to extol the exemplary argues that the Toltecs were simply invented as
virtues of the Toltecs prompted a critical response noble ancestors). Although Brinton was given to
from Daniel Garrison Brinton (1887) in the now hyperbole and factual errors, he did make some
famous essay “Were the Toltecs an Historic lasting contributions to the discourse on the
Nationality?” Brinton does not mention Charnay Toltecs. He made a convincing case that we know
by name, and he was certainly aware that the the Toltecs from written sources more than from
notion of a Toltec foundational culture was com- archaeology. Brinton observed that the Toltecs are
mon in other accounts of pre-Hispanic American credited with so many feats and cultural advance-
history (e.g., Valentini 1883). Nevertheless, Charnay ments that they seem more mythic than real.
Chichén Itzá, Tula, and Tollan 5
figure 6 Feathered serpent
columns in the northern
entrance to the Castillo
temple. Photograph by
Jeff Kowalski.
Overall, Brinton demonstrated that, above any- most Classic- and Postclassic-period large centers
thing else, the Toltecs are a textual phenomenon and impressive works of art and architecture in
(Kristan-Graham 1996). From his close reading Mesoamerica, while the other held that the Toltecs
of the ethnohistorical literature, Eduard Seler fol- were not even a real people. Brinton’s essay seems
lowed Brinton in asserting that Quetzalcoatl and to have, at first, been little read and cited, yet his
Tollan, and by extension the Toltecs, were fabri- ideas about the power of the text and the Toltecs
cations of myth and that the mythic Tollan was being primarily characters from the pages of eth-
not to be identified with Tula, which he called the nohistory are more in line with modern theories of
historical Tollan (see Gillespie, this volume, for a the text and of reading (see the “Mapping Tollan”
more detailed discussion of Seler’s ideas). section below, and the chapters by Gillespie and
As the twentieth century began, there were Smith, this volume).
two divergent positions on the “Toltec question.” After Charnay and Brinton, the work of Paul
One held that the Toltecs were responsible for Kirchhoff (1955), Wigberto Jiménez Moreno (1941),
6 kristan-graham and kowalski
Nigel Davies (1977), and others belied the tension they are the describer’s descriptions, not those of
inherent in dealing with sources that are ostensibly the described.”
about the same historic events but which contain Surely the same is true of ethnohistory. The
contradictory details of character, plot, and loca- century after the conquest in which the Spaniards
tion. Scholars dealt with such challenges in a num- collected and wrote the primary ethnohistoric
ber of ways, from opting to use the sources that sources was an era of acculturation, defined by the
seem to be the most reliable (Kirchhoff and Jiménez Counter Reformation and the Spanish Inquisition.
Moreno), to fitting (or forcing) the sources into a Consequently, the sources embody the radical
more or less unified narrative (Jiménez Moreno), changes that the indigenous cultures were sub-
to explaining discrepancies as the result of “mis- jected to as well as European attitudes of a chang-
takes” or regional or cultural biases (Davies). ing world. This does not negate the importance or
As we know, the ethnohistoric record is rife use of the data, but it does indicate that ethnohis-
with allusions to Topiltzin Quetzalcoatl and tory is best viewed as documents that mediate the
Tollan that may or may not refer to Tula, Hidalgo. observed and the observer, the vanquished and the
Differing reading strategies are always in play, victor. Such an approach views ethnohistoric doc-
from literal to symbolic to private understandings uments as part of a literary tradition that may ulti-
(see Manguel 1996 for an analysis of the various mately tell us more about literary genres, symbols,
modes of reading and interpretation in the West). and rhetorical uses of language than about “real
Are we to read ethnohistorical documents as “real history.” In our own time, the discipline of history
history” that tell us “what really happened”? Or are itself has become introspective to the point where
the sources to be understood as vessels of allegori- the idea of an empirical sort of history that is
cal language and cultural attitudes? knowable and retrievable has come into question.
Even H. B. Nicholson (1979), who quite liter- Indeed, Benjamin Keen (1971), Jacques Lafaye
ally wrote the book about the books that describe (1976), David Carrasco (1982), Jorge Klor de Alva
Topiltzin Quetzalcoatl and Tollan, has stated that et al. (1988), Susan Gillespie (1989), and Enrique
the ethnohistoric literature may well contain a ker- Florescano (1994) have shown that both reading
nel of truth, but that it is not always easy or pos- ethnohistory and using ethnohistory to construct
sible to identify what that kernel might be. It is a vision of the past are complex, ongoing processes
Nicholson’s (1957, 2001) painstaking analysis of the wherein concerns of colonialism and Mexican
ethnohistoric literature, revealing minor to signifi- nationalism in part temper how the pre-Hispanic
cant differences among plot lines and characters, past is read and rewritten. These approaches
that suggests the Topiltzin Quetzalcoatl and Tollan intersect with the ideas of the critic Edward Said
sagas are not reliable guides to understanding pre- (1994:17), who has observed of the highly charged
Hispanic events. political nature of reading and writing cultural
There is a growing trend to read ethnohis- history that “more important than the past itself
tory as a compilation of poetic statements about . . . is its bearing upon cultural attitudes in the
the mythic and lived past that were created within present.” Moreover, Carrasco and Gillespie dem-
the contours of the post-contact, colonial, and even onstrate that the Topiltzin Quetzalcoatl of Tollan
modern moments, seasoned with varying degrees tales, as we know them today, are contact period
of cultural subjectivity. Ethnohistory and ethnog- and later in date and cannot be used as reliable
raphy are now more properly viewed as part of a guides to pre-Hispanic ideas about Tollan.
broad network of cultural codes, literary genres, However we choose to use the ethnohis-
and allegorical language that intersects with social toric literature, our readings are open to uses and
and political concerns (Clifford and Marcus 1986). abuses. A case in point is the oft-cited Toltec or
Clifford Geertz (1988:144–145) has observed that central Mexican conquest of Chichén Itzá, clear
“all ethnographic descriptions are homemade . . . references to which neither Ralph Roys (1966) nor
Chichén Itzá, Tula, and Tollan 7