Beruflich Dokumente
Kultur Dokumente
Most scholars agree that the urban states of Classic Mexico developed from Formative
chiefdoms which preceded them. They disagree over whether that development (1) took place
over the whole area from the Basin of Mexico to Chiapas, or (2) emanated entirely from one
unique culture on the Gulf Coast. Recently Diehl and Coe (1996) put forth 11 assertions in
defense of the second scenario, which assumes an Olmec “Mother Culture.” This paper
disputes those assertions. It suggests that a model for rapid evolution, originally presented by
biologist Sewall Wright, provides a better explanation for the explosive development of For-
mative Mexican society. © 2000 Academic Press
1
0278-4165/00 $35.00
Copyright © 2000 by Academic Press
All rights of reproduction in any form reserved.
2 FLANNERY AND MARCUS
ture and contributed “little if anything to communities. Chiefdoms are not a mono-
later [Mesoamerican] civilization.” lithic category; they come in many differ-
Our school would be happy to chal- ent types. Some, like those of Panama’s
lenge the Olmec-centrists to a tug-of-war, Azuero Peninsula, were sedentary and
since half the members of their team are flamboyant (Lothrop 1937; Linares 1977;
dead. However, their portrayal of our po- Helms 1979). Others, like those of Iran’s
sition is not accurate—a familiar problem Zagros Mountains, were pastoral and
when one is being used as a straw man. non-flamboyant (Barth 1964; Flannery in
We would not describe the Olmec as “no press). Within Polynesia alone, Goldman
more advanced” or “contributing little.” (1970) has classified some chiefdoms as
Their contribution has simply been exag- “traditional” (based more on sacred au-
gerated by Olmec-centrists, who credit thority), others as “open” (based more on
the Olmec with many things their neigh- secular power), and still others as “strati-
bors did earlier or better. fied” (large, with a combination of sacred
OA presents 11 “traits” which allegedly authority and secular power). Nowadays
show the Olmecs’ maternal role in Me- the term “paramount” often substitutes
soamerica’s genealogy (Diehl and Coe for Goldman’s “stratified.” While rank in
1996:23). We find those traits unconvinc- traditional chiefdoms usually took the
ing and suggest that there are better form of a continuum from higher to lower
frameworks than the Mother Culture status, a few paramount chiefdoms—like
model, which we do not find appropriate those in Hawai’i (Kirch 1984: Fig. 85)—
for any world region. One alternative is a achieved stratification by cutting lower-
model for the conditions leading to rapid status families out of the genealogy, re-
evolution, presented by the distinguished ducing them to a commoner class.
biologist Sewall Wright (1939). Even be- In some parts of the ancient world,
fore refuting the 11 traits, however, we chiefdoms persisted for centuries. Re-
must modify the authors’ caricature of our search in such regions has defined a long-
position. term process called “chiefly cycling” (H.
Wright 1984; Anderson 1994). In this pro-
PRIMUS INTER PARES: A cess, paramount chiefdoms rose, peaked,
CLARIFICATION then collapsed amid a regional landscape
of smaller traditional or open chiefdoms.
Any model for the Olmec and their It is increasingly clear that paramount
neighbors must be based on our current chiefdoms formed by taking over their
understanding of complex societies, weaker neighbors (Carneiro 1981, 1991).
which is far greater now than in Vaillant’s Their collapses resulted from such factors
or Covarrubias’s day (Anderson 1994; as competition between chiefly families or
Carneiro 1981, 1991; Drennan and Uribe factions, endemic raiding, agricultural
1987; Earle 1991a, b, 1997; Flannery 1995, failure, or demographic imbalance, and
1999; Goldman 1970; Johnson 1987; Kirch usually took the form of fragmentation
1984; Kirch and Green 1987; Marcus 1989, back into the smaller units from which
1992; Marcus and Flannery 1996; Spencer they had been created. We view the
1993, 1998; H. Wright 1984, 1986). Olmec as one more set of paramount
Among the most interesting societies in chiefdoms that rose, peaked, and eventu-
the ethnographic and archaeological ally collapsed in a landscape of traditional
records are chiefdoms—societies based on and open chiefdoms.
hereditary differences in rank, in which A rare paramount chiefdom might suc-
the chief’s authority extends to satellite ceed in subduing and incorporating other
MEXICAN CHIEFDOMS AND MOTHER CULTURE 3
large chiefdoms, creating a polity so great shell, iron ore mirrors, and the like (Pires-
that it could no longer be administered as Ferreira 1975). Tlapacoya in the Basin of
a chiefdom (Spencer 1998). This is how Mexico sent Paloma Negative and Cesto
indigenous states formed in Madagascar White pottery to San José Mogote in the
(Dewar and Wright 1993) and among the Valley of Oaxaca; Oaxaca sent Leandro
Zulu, Ashanti, Hunza, and Hawai’ians Gray and Delfina Fine Gray pottery to
(Flannery 1999). It is becoming increas- Tlapacoya and to San Lorenzo, Veracruz
ingly clear that the first states in south- (Marcus 1989:192; Flannery and Marcus
west Iran, Egypt, Peru, Oaxaca, and the 1994:259 –263, 286). San José Mogote re-
Maya region also formed this way (H. ceived turtle shell drums and pearly
Wright 1986; Flannery 1995; Marcus 1992, freshwater mussels from the San Lorenzo
1993, 1998a; Marcus and Flannery 1996). It region; it also received Guamuchal
makes the study of chiefdoms all the more Brushed pottery from Chiapas (Flannery
interesting to discover that, on at least and Marcus 1994:286). Magnetite from
some occasions, they became the “precur- Oaxaca reached San Pablo in Morelos and
sors” of states (Carneiro 1981; H. Wright San Lorenzo in Veracruz (Pires-Ferreira
1984). 1975).
It took more than 1000 years for Mexi- There are two reasons why such ex-
co’s Formative societies to become com- changes of goods should not surprise us.
plex enough to serve as precursors for The first is that intersite distances were
states. By the middle of the second mil- not great. Given foot travel estimates of
lennium b.c., agricultural villages were 4.5 km per h (Morley 1938:234) or 32 km
spread over the whole area from the Basin per day (Hammond 1978), even a trip from
of Mexico to the Pacific Coast of Chiapas. the Basin of Mexico to the Chiapas Coast
Some, but not all, of these village societies would take less than a month. The second
had been reorganized into states by the reason is that chiefly elites are always ea-
beginning of the Christian era. ger for prestigious gifts from other chiefly
We know less about this transitional pe- elites.
riod than we should, since some archae-
ologists assume that their sites belong to THE OLMEC IN WIDER CONTEXT
chiefdoms without producing evidence of
the requisite sociopolitical institutions. Let us now look at the Olmec in the
Elsewhere we have suggested that as context of chiefdoms worldwide. The apo-
many as ten lines of evidence may be nec- gee of this flamboyant society took place
essary to confirm a chiefdom (Marcus and between 1150 and 300 b.c. on Mexico’s
Flannery 1996:110). At this writing, we are Gulf Coast (Grove 1997). What we know of
confident that the Valleys of Mexico, its demographic history suggests typical
Puebla, Morelos, and Oaxaca, and various chiefly cycling. San Lorenzo, perhaps the
parts of Guerrero, Chiapas, and southern earliest Olmec center, peaked between
Veracruz-Tabasco had chiefly societies by 1150 and 900 b.c.; it then suffered a loss of
1150 b.c. We are less confident about areas population and many of its stone monu-
such as the Tehuacán Valley and the Cañ- ments were defaced, most likely by a rival
ada de Cuicatlán, but they show evidence chiefdom (Coe and Diehl 1980a, b;
of modest chiefdoms by 600 – 450 b.c. Cyphers 1997). San Lorenzo’s population
(Spencer 1993). was partially restored between 600 and
Most chiefly centers of 1150 – 450 b.c. 400 b.c., after which it collapsed again and
were in frequent contact with each other, lay abandoned for centuries rather than
exchanging goods like obsidian, marine becoming part of a state.
4 FLANNERY AND MARCUS
La Venta, a second Olmec center some mez 1995). 3 Surveys of the American Bot-
90 km to the northeast, rose to promi- tom, the alluvial valley surrounding
nence between 900 and 600 b.c. (Drucker Cahokia, suggest that the site’s immediate
et al. 1959; González Lauck 1996). It is “sustaining area” may have covered 3000
probably no accident that La Venta’s rise km 2. By A.D. 1400 it had collapsed without
coincided with San Lorenzo’s 900 – 600 b.c. becoming part of a state.
hiatus. Whether La Venta or a third Like the Olmec, Cahokia was once con-
chiefly center was responsible for defac- ceived of as a “mother culture.” Forty
ing San Lorenzo’s monuments, this cycle years ago, when we had much less infor-
of synchronized rises and collapses is typ- mation than we do now, the American
ical of chiefdoms competing for labor and Bottom was considered “something of a
resources (H. Wright 1984; Anderson font from which all Mississippian [cul-
1994). ture] arose, even the source of invading
Indeed, the Olmec resembled many waves of population” for other parts of
other chiefdoms worldwide. Some of their the eastern United States (Anderson 1994:
chiefly centers covered hundreds of hect- 144). Over the past four decades, that
ares, like the largest Mississippian centers model of Cahokian mother culture has
of North America. The Olmec built been replaced by a multiple-center model.
earthen mounds like some Polynesian The Mississippian is now seen as “emerg-
chiefdoms. They set up huge stone sculp- ing” (Smith 1990) simultaneously from
local Woodland cultures all over the
tures like chiefdoms on Easter Island, and
Southeast, and “any recourse to popula-
carved wooden statues and jade sumptu-
tion movement is suspect” (Anderson
ary goods like the Maori. While they were
1994:144).
not identical to any of those other chief-
Even within the 3000 km 2 American
doms, the difference was more of degree
Bottom, Milner (1990:29) would now see
than kind.
Cahokia as primus inter pares, the domi-
Many chiefly centers sprawled over ar-
nant political entity among a number of
eas larger than that of a typical Bronze organizationally similar (if less complex)
Age city. This results from the fact that semiautonomous chiefdoms which exer-
chiefs cannot control people at a distance, cised considerable control over their own
as states can; many chiefs therefore con- territories—something analogous, in
centrated thousands of farmers, warriors, other words, to Powhatan’s confederacy of
and craftsmen as close to their residences 200 villages (Rountree 1989). Anderson
as possible. Conversely, when a chiefdom (1994:141) points to a significant difference
“cycled down,” its loss of population between Cahokia and most early states:
could be as spectacular as that recorded “the complete absence [at Cahokia] of ev-
from San Lorenzo at 900 b.c. by Symonds idence for formal, differentiated adminis-
and Lunagómez (1997:135). trative structures.”
Even at their peaks, San Lorenzo and La While it lacked administrative struc-
Venta were smaller than Cahokia, a Mis- tures, Cahokia did build earthen mounds.
sissippian chiefly center in Illinois. At its One of these, Monks Mound, stands 30 m
apogee in A.D. 1250, Cahokia is estimated high and covers an area 300 ⫻ 212 m
to have covered 13 km 2 (Milner 1998:109; (Anderson 1994:138). It is the largest
Pauketat 1994). This is six times the cur- 3
The largest estimates for San Lorenzo would in-
rent estimate for La Venta (González clude, within the boundaries of that one site, locali-
Lauck 1996:75) and twice the most hyper- ties which other reports consider separate sites in the
bolic estimate for San Lorenzo (Lunagó- settlement hierarchy below San Lorenzo.
MEXICAN CHIEFDOMS AND MOTHER CULTURE 5
major chiefly center of the period 1150 – (Flannery and Marcus 1994:31–33; Marcus
450 b.c. whose hinterland has been sys- and Flannery 1996:87). Complex A at La
tematically surveyed had villages and Venta had a similar orientation (Drucker
hamlets hierarchically below it. The Basin et al. 1959), but since the Oaxaca Men’s
of Mexico (Sanders, Parsons, and Santley Houses antedate Complex A by 500 years,
1979; Niederberger 1996: Map 1), the Val- one can hardly credit the Olmec with
ley of Morelos (Hirth 1980, Grove 1987), Mexico’s first solar or astral alignments.
the Valley of Oaxaca (Kowalewski et al. Trait 3. Although admitting that “we
1989; Marcus and Flannery 1996), the lack precise data on the size of Olmec
Chiapas coast (Clark and Blake 1994), and polities,” the authors of OA argue that the
northern Belize (Hammond 1991) all had territories controlled by Olmec centers
hierarchies of villages and hamlets below may have been “significantly larger than
major centers. As for “craft workshops,” those of their contemporaries.” The truth
examples include the Matadamas chert is that we also lack precise data on the size
quarries (Whalen 1986) and Fábrica San of their contemporaries’ polities, making
José saltworks (Drennan 1976) in the Val- the whole topic speculative.
ley of Oaxaca. “Special ritual locales” are Trait 4. The Olmec, OA asserts, “had a
also widespread; consider the painted highly sophisticated symbol system ex-
cliffs and caves above the site of Tlapa- pressed in a coherent art style.” We defer
coya in the Basin of Mexico which, ori- our discussion of this trait to a later sec-
ented east toward the volcanoes Ixtacci-
tion, where we show that San Lorenzo had
huatl and Popocatepetl, receive the early
only a subset of the repertoire of symbols
light of sunrise and “may have constituted
used throughout early Mexico.
a significant component of sacred space”
Trait 5. The Olmec invented monumen-
(Niederberger 1996:87). The painted cave
tal stone carving, which was “a defining
of Oxtotitlán in Guerrero (Grove 1970)
characteristic of every Mesoamerican civ-
would be a second example.
ilization.” We agree that monumental
Trait 2. Although we “cannot yet deci-
pher the meanings,” San Lorenzo and La sculpture was a defining characteristic of
Venta “were laid out as cosmograms.” the Olmec; the question is, how accurate
This is sheer speculation, based on Coe’s an indicator of sociopolitical complexity is
belief that San Lorenzo was laid out to it? We have already shown that Easter
resemble a “gigantic bird flying east” (Coe Island, a modest chiefdom by Polynesian
and Diehl 1980a:387). This notion is re- standards, produced 100 times as many
futed by geological studies which show colossal heads as are known from San
that, although modified by architectural Lorenzo.
terracing, the overall shape of the San Trait 6. Predictably, the authors of OA
Lorenzo plateau is largely the result of use the colossal heads for a second trait.
natural erosion (Cyphers 1997:102–105). Both the heads, and the wooden busts
While true cosmograms have not been found in the spring at El Manatı́, are
found, many early Mesoamerican cultures thought by them to be “portraits of rul-
used solar or astral principles in orienting ers.” Again, this is pure speculation. Like
important buildings. As early as 1350 b.c., the statues of Easter Island, the Olmec
the occupants of the Valley of Oaxaca colossal heads might represent chiefly an-
were apparently aligning their Men’s cestors. As for the busts of El Manatı́, they
Houses to the sun’s path during the equi- might be (1) ancestors, like some Maori
nox. This resulted in an orientation 8° N of woodcarvings, or (2) surrogate sacrificial
east, or as it is often given, 8° W of north victims tossed into a spring.
8 FLANNERY AND MARCUS
Trait 7. This trait, “special ritual locales,” the arid Tehuacán Valley used a dry cave
has already been discussed under Trait 1. for such sacrifices, while the occupants of
Trait 8. The ballgame, OA claims, finds the humid Gulf Coast used a spring,
its “oldest known evidence” in Olmec de- hardly seems earth-shaking.
posits; San Lorenzo’s Palangana mound Trait 11. The Olmec had “extensive
complex (600 – 400 b.c.) “is the first known, trade networks.” While they admit that
purposefully constructed ballcourt.” This most Formative cultures had extensive
assertion is contradicted by Hill et al. networks, the authors of OA insist that the
(1998), who claim to have found a 1400 b.c. Olmec “moved a greater quantity and
ballcourt at Paso de la Amada, Chiapas. more different kinds of goods” than their
The game itself is surely older than the
contemporaries (they then pad the list
Olmec; we even have one preceramic
with “probable exports” for which we
camp site with a boulder-lined area that
have no physical evidence.) The fact is
could be for ballgames (Marcus and Flan-
that we currently have no objective, quan-
nery 1996:58 –59).
The most convincing evidence for an tified measure of goods moved by any
early Mexican ballgame comes from rub- Formative society, especially in the case of
ber balls preserved by waterlogging in the perishables.
spring at El Manatı́ (Ortı́z and Rodrı́guez We cannot resist pointing out the irony
1989, 1999). The discovery of these balls, of the OA authors’ position on Trait 11: All
however, is an accident of good preserva- Formative cultures had trade, but the
tion. We cannot assume that similar ball- Olmec had the most trade. Doesn’t that
games were unknown in the highlands; make the Olmec primus inter pares?
after all, there are very early figurines of
ballplayers at El Opeño, Michoacán (Oliv-
eros 1974). TRAITS CONSPICUOUS BY THEIR
Trait 9. The authors of OA use El Manatı́ ABSENCE
for a second trait: the first “ritual use of
rubber.” It makes sense that the first ritual As interesting as the 11 traits given in
use of rubber might occur on the Gulf OA are the “firsts” the authors do not list
Coast, where rubber trees are native—just for the Olmec. These include the first use
as it makes sense that the first ritual use of of lime plaster, adobe brick, and stone
obsidian and magnetite might occur in the
masonry, three materials emblematic of
highlands, where those raw materials are
Classic Mesoamerican civilization. OA
native. The point is, every region has
cannot list these as Olmec innovations be-
something it did “first.”
cause their first use occurred in the Mex-
Trait 10. But wait; El Manatı́ gets used
for a third trait. It provides the Olmec with ican highlands. In the Valley of Oaxaca,
the oldest evidence for “infant sacrifice in for example, lime plaster was used in
water-related rituals.” Men’s Houses as early as 1350 b.c.; adobes
The truth is that by the time El Manatı́ were used in public buildings by 1000 b.c.;
was occupied, infant sacrifice had existed and stone masonry platforms up to 2.5 m
in Mexico for thousands of years. Several in height were in use by 1000 b.c. (Marcus
infants were sacrificed (perhaps even can- and Flannery 1996:87, 109 –110). By the time
nibalized) in Level XIV of Coxcatlán Cave such construction techniques reached Com-
in the Tehuacán Valley, an occupation plex A, La Venta (Drucker et al. 1959), they
dating to 5000 b.c. (MacNeish et al. 1972: had been used in the highlands for cen-
266 –270). The fact that the occupants of turies.
MEXICAN CHIEFDOMS AND MOTHER CULTURE 9
FIG. 3. Map of Formative Mexico, showing style provinces and places mentioned in the text.
Hachured area, highland province. Shaded area, lowland province. The style boundary emerged at
1400 –1150 b.c. (Clark 1991: Fig. 8) and remained intact through 1150 – 850 b.c. 1, Tlatilco; 2, Tlapa-
coya; 3, Coapexco; 4, Gualupita and Atlihuayan; 5, Nexpa and San Pablo; 6, Chalcatzingo; 7, Las
Bocas; 8, Ajalpan and Coxcatlán Cave; 9, Teopantecuanitlán; 10, Oxtotitlán Cave; 11, Nochixtlán and
Etlatongo; 12, Cuicatlán; 13, San José Mogote and Tierras Largas; 14, La Venta; 15, San Lorenzo; 16,
El Manatı́; 17, Las Limas; 18, Mirador-Plumajillo; 19, Chiapa de Corzo; 20, Paso de la Amada.
FIG. 4. A complex of red-on-buff vessels characterized the valleys of the highland style province
at 1400 –1150 b.c. (a) Jar from Burial 1 of Nexpa. (b) Jar from Tierras Largas. (c) Jar from Tlapacoya.
(i,j) Jar and bottle from Ajalpan. (d, e) Hemispherical bowls from Tierras Largas. (f, g) Hemispherical
bowls from Tlapacoya. (h, k) Hemispherical bowls from Ajalpan. (Redrawn from Grove 1974;
Niederberger 1976; Flannery and Marcus 1994; MacNeish et al. 1970.)
The Basin of Mexico, Morelos, Puebla, the Despite these regional differences, a
Tehuacán Valley, the Valleys of Oaxaca few pottery types were present on both
and Nochixtlán, and the Cuicatlán Cañ- sides of the style boundary. One of these
ada all shared red-on-buff bowls, bottles, was a pure white product called “kaolin
and jars (Fig. 4). East of Tehuacán and ware,” believed on the basis of petro-
Oaxaca, this red-on-buff complex gradu- graphic analysis to have been made in two
ally gave way to one linking southern Ve- to three different regions (Fig. 6). Also
racruz, Tabasco, and Chiapas. This low- found on both sides of the boundary were
land complex featured tecomates or tecomates decorated with rocker stamping
neckless jars with bichrome slips, fluting, in zones (Fig. 7). Such vessels make the
or crosshatching (Fig. 5). point that plastic decoration was already
MEXICAN CHIEFDOMS AND MOTHER CULTURE 11
FIG. 5. Southern Veracruz, Tabasco, and Chiapas were part of a lowland style province at
1400 –1150 b.c. (a) Chilpate Red-on-Cream tecomate, San Lorenzo. (b) Tepa Red-and-White teco-
mate, coastal Chiapas. (c) Centavito Red fluted tecomate, San Lorenzo. (d) Cotán Red fluted
tecomate, coastal Chiapas. (e) Tusta Red fluted tecomate, coastal Chiapas. (f) Achiotal Gray
tecomate with zoned crosshatch, San Lorenzo. (g) Salta Orange tecomate with zoned crosshatch,
coastal Chiapas. (Redrawn from Coe and Diehl 1980a; Blake et al. 1995.)
popular at 1400 –1150 b.c., in what we as- vored demographic growth, craft special-
sume the Olmec-centrists would have to ization, increased interregional exchange,
consider “Grandmother Cultures.” greater disparities in social rank, and
more elaborate ceremonialism.” Certain
The So-Called “Early Horizon” communities (often the largest in each re-
gion) seem to display these characteristics
Sometime around 1200 –1150 b.c., in the more than others. The increased interre-
words of Tolstoy (1989:275), “conditions gional exchange mentioned by Tolstoy in-
over much of Mesoamerica evidently fa- volved obsidian, marine shell, iron ores
12 FLANNERY AND MARCUS
FIG. 6. Despite Mexico’s division into style provinces at 1400 –1150 b.c., some luxury pottery
types showed up everywhere. Above, collared tecomates in kaolin ware from San José Mogote (a,b)
and San Lorenzo (c). Below, kaolin bottles from San José Mogote (d) and San Lorenzo (e). (Redrawn
from Flannery and Marcus 1994; Coe and Diehl 1980a.)
and pigments, jade, mica, stingray spines, blages of the lowlands. Assemblages from
turtle shell drums, and pottery. Often southern Veracruz, Tabasco, and Chia-
flamboyant, the pottery came in white, pas— components of the old lowland
black, gray, red, red-and-white, and province—still found their strongest ties
black-and-white. Its plastic decoration, with each other. Another reason the
while still including rocker stamping, now Olmec-centrists’ model will not work is
featured delicate fine-line incising, deep that many of the ceramic features they
excising or carving, and combinations of attribute to the Olmec appear earlier, are
these. Many of the carved and incised mo- more abundant, and/or are better made at
tifs of 1150 b.c. were so stereotyped and Tlapacoya, Tlatilco, Las Bocas, and San
pan-Mesoamerican that some scholars as- José Mogote than at San Lorenzo or La
sign them to an “Early Horizon” (see Venta (Grove 1989).
Grove 1989 for discussion). Almost 30 years ago, Joralemon (1971)
Olmec-centrists want us to believe that assembled an inventory of 176 allegedly
this style was created by the Olmec and “Olmec” motifs. While widely cited by
imposed on the rest of Mexico. There are Olmec-centrists (e.g. Coe and Diehl 1980a,
several reasons why that is unconvincing. b), this study has two flawed assumptions:
One reason is that Mexico did not, in fact, (1) a belief that every motif was Olmec no
become one uniform style province be- matter what region it came from, and (2)
tween 1150 and 850 b.c. Ceramic assem- the notion that every motif was a deity.
blages from the Basin of Mexico, Puebla, Joralemon created a “pantheon” of al-
Morelos, and Oaxaca— components of the leged “Olmec gods,” but he did so relying
old red-on-buff province—still resembled heavily on decorated wares from Tlatilco,
each other more than they did the assem- Tlapacoya, Las Bocas, and other sites in
MEXICAN CHIEFDOMS AND MOTHER CULTURE 13
stylistic provinces seen at 1400 –1150 b.c. drilled cubes (Fig. 14). These unusual ar-
(Fig. 3). Ties between the Basin of Mexico, tifacts—present also at Las Limas, Vera-
Morelos, Puebla, and Oaxaca remained cruz (Agrinier 1989:21)—are virtually ab-
strong, with San José Mogote and Tlatilco/ sent to the west of the style boundary.
Tlapacoya using similar distinctive arti-
facts (Figs. 11, 12) and displaying similar EVALUATING THE CLAIM OF
motifs on similar vessels (Fig. 13). Ties “INTRUSIVENESS”
between Veracruz/Tabasco and Chiapas
also remained close; for example, a Brain- Having shown that the major stylistic
erd-Robinson matrix calculated by provinces of early Mexico were un-
Agrinier (1989) shows strong similarity in changed by the rise of the Olmec, let us
ceramic assemblages between San look at the claim in OA that “monumental
Lorenzo (Veracruz) and Mirador-Pluma- three-dimensional stone sculpture; hol-
jillo (Chiapas). Further artifact similarities low whiteware figurines depicting babies;
between those two sites include thou- and Calzadas Carved pottery” were “in-
sands of iron ore “lug nuts” or multi- trusive elements” at highland centers
MEXICAN CHIEFDOMS AND MOTHER CULTURE 15
FIG. 11. Diagnostic of the highland style province at 1150 – 850 b.c. were white-slipped spouted
trays, used to mix and pour pigments. From L to R, these examples come from Tlatilco (Porter 1953),
Gualupita (Vaillant and Vaillant 1934), and Tierras Largas (Flannery and Marcus 1994). Diameter of
a, 15.5 cm. Such trays were not a significant part of the Gulf Coast inventory; Coe and Diehl (1980a)
apparently did not find a single one at San Lorenzo. We wonder why Olmec-centrists continue to
feature spouted trays from the highlands in their exhibits of “Olmec art” (Art Museum of Princeton
University 1996:325). Such indiscriminate application of the term “Olmec” waters down whatever
regional and cultural significance it might have had, reducing it to a synonym for “pretty.”
dance and variety. San Lorenzo alone has for their “intrusion” into the Mexican
produced more than 70 stone monuments, highlands.
including 10 colossal heads (Cyphers
1997). To be sure, since many heads were Hollow Whiteware Figurines Depicting
found reused, rededicated, defaced, re- Babies
worked, or out of context, we cannot be
sure how many actually date to the Early Hollow white-slipped “baby dolls” ap-
Horizon. Many similar monuments from pear to have been present at every major
La Venta are thought to be Middle Forma- Mexican site of 1150 –500 b.c. Tlatilco, Tla-
tive (850 –500 b.c.) in date (Drucker et al. pacoya, Gualupita, Las Bocas, Teopante-
1959; Hammond 1988; Graham 1989; cuanitlán, San José Mogote, Etlatongo,
Grove 1997). San Lorenzo, La Venta, and Paso de la
The real question is, how often does Amada have all produced fragments or
such sculpture appear as an “intrusive el- complete specimens.
ement” in the Mexican highlands? Teo- For hollow white dolls, we lack detailed
pantecuanitlán (Guerrero) has some statistics comparable to those for the
three-dimensional monuments (Martı́nez carved pottery discussed below. It is in-
Donjuán 1985, 1994), but most of these are structive, however, to examine examples
Middle Formative and might have been for which proveniences are known or al-
influenced by the much nearer highland leged. Consider the catalogues for two re-
site of Chalcatzingo (Grove 1987). Oax- cent exhibits of supposedly “Olmec” art:
aca’s Early Horizon sculptures, such as (1) one held by the National Gallery of Art
Monuments 1 and 2 of San José Mogote, in Washington, D.C. (Benson and de la
are not Olmec in style (Marcus 1989:165; Fuente 1996) and (2) one held by The Art
Flannery and Marcus 1994: Fig. 18.9). In Museum of Princeton University (1996).
the Basin of Mexico neither Tlatilco, Co- The National Gallery catalog illustrates
apexco, nor Tlapacoya has produced stone seven hollow white dolls of young indi-
monuments imitating those of the Olmec. viduals. All are masterpieces; none are
Thus, while conceding a Gulf Coast origin from the Gulf Coast. Two are from Tla-
for colossal heads, we find little evidence tilco, two are from Tlapacoya, one is from
MEXICAN CHIEFDOMS AND MOTHER CULTURE 17
FIG. 13. Between 1150 and 850 b.c. pottery as- FIG. 14. Just as highland Mexican sites of 1150 –
semblages of the highland style province shared 850 b.c. shared ground-stone yuguitos, many lowland
similar combinations of vessel shape and motif. sites shared multidrilled iron ore cubes or “lug
Here we see dark bottles with crosshatched sun- nuts.” These examples, averaging 3.1 cm thick, come
burst motifs from Tlatilco (a) and San José Mogote from Mirador-Plumajillo (a– c) and San Lorenzo (d–f).
(b). (Redrawn from Porter 1953:Pl. 6I; Flannery and (Drawn from photographs in Agrinier 1989:25; Coe
Marcus 1994:99). Height of (a) 16.2 cm. and Diehl 1980a:242.)
MEXICAN CHIEFDOMS AND MOTHER CULTURE 19
FIG. 16. Frequencies of classified sherds in Levels 13– 6 at Zohapilco-Tlapacoya IV, Basin of Mexico. Pottery types bearing pan-Mesoamerican
motifs are printed in capital letters. (Based on Niederberger 1976: Pl. 32.)
MEXICAN CHIEFDOMS AND MOTHER CULTURE 21
TABLE 1
Zohapilco-Tlapacoya IV. Total Sherd Counts of the 6 Pottery Types Bearing Pan-Mesoamerican Motifs,
Levels 13-6, Phases Nevada Through Manantial. (Source: Niederberger 1976: 164.)
Type 13 12 10–11 9 8 7 6
All classified sherds 1954 971 15637 20418 30914 11828 2278
wares were also used for pan-Mesoameri- were common by 1250 b.c. In Level 8,
can motifs. Pilli White appeared in nu- whose volume was 12 m 3, there were 7728
merous bowl forms, some carved with sherds of Tortuga Polished; in Level 7,
Pyne’s Motif 1 and others showing her amounting to only 4.4 m 3, there were 2569
Motifs 10 and 15 in fine-line hachure sherds of that type. Moreover, pan-
(Niederberger 1976: Pl. 42). Paloma Nega-
tive, a related ware of the Ayotla phase,
was used for one of the most elegant ver-
sions of Earth ever found, a vessel eclips-
ing any found by Coe and Diehl at San
Lorenzo (Fig. 17).
Finally we come to Atoyac Fine Gray, an
imported ware decorated with Pyne’s Mo-
tifs 1, 2, and 7 (Niederberger 1976: Pl. 46).
Some vessels of this type (under the ear-
lier name “Tlapacoya Gray”) have been
studied by geologists Howel Williams and
Wayne Lambert, who consider them to
have been made in Oaxaca (Weaver 1967:
30; Lambert 1972; Niederberger 1987:564;
Flannery and Marcus 1994:259 –262). We
suspect that many of these vessels belong
to a Oaxaca type called Delfina Fine Gray. FIG. 17. Four angry versions of Earth/Earth-
However, other Atoyac Fine Gray vessels quake– one for each of the four great Mesoamerican
world directions– circle this bowl from Tlapacoya.
illustrated by Niederberger (1976: Pl. 46) The type, Paloma Negative, combines (1) white slip
were probably made locally. and (2) resist white over pale brown. Locally made at
To summarize: carved and incised pan- Tlapacoya, Paloma Negative was traded as far as
Mesoamerican motifs were neither rare Oaxaca. Highland vessels like this should not be
called “Olmec.” Coe and Diehl (1980a) report no
nor “intrusive” at Tlapacoya. The dark
sherds of this ware from San Lorenzo and illustrate
gray wares on which they occurred had no vessel approaching it in sophistication. (Drawn
been among the most common local types from a photograph in Benson and de la Fuente 1996:
at 1350 b.c., and the motifs themselves 202.)
22 FLANNERY AND MARCUS
Mesoamerican motifs (sometimes bril- San José phase transition, simply required
liantly executed) also occurred on Tlapa- burnishing the ware twice instead of once,
coya’s white-rimmed black, white- then firing it in a reducing atmosphere
slipped, and resist white wares. (Flannery and Marcus 1994:157–165). Le-
andro Gray went on to constitute 23% of
San José Mogote all sherds in middle San José times, a per-
centage comparable to that of Tortuga
The Valley of Oaxaca lies 330 km from Polished in Tlapacoya’s early Manantial
Tlapacoya, but only 210 km from San phase.
Lorenzo. If the Olmec were truly the Table 2 gives the actual counts of Lean-
source of inspiration alleged in OA, Oa- dro Gray, Delfina Fine Gray, San José
xaca’s Early Formative ceramics should Black-and-White, and Atoyac Yellow–
resemble San Lorenzo’s more than Tlapa- white sherds from an excavation in Area A
coya’s. In fact, the reverse is true (Flan- of San José Mogote (Flannery and Marcus
nery and Marcus 1994). 1994: Figs. 14.1, 14.4). We have chosen to
The relevant periods in Oaxaca are the feature this excavation because it covered
Tierras Largas phase (1400 –1150 b.c.) and 12 m 2, virtually the same area as an im-
San José phase (1150 – 850 b.c.). Fig. 18 portant excavation at San Lorenzo which
shows the changing frequencies of pottery we will discuss below. The stratigraphic
types during the course of these periods, levels consist of a midden (Zone D) and
including the crucial Tierras Largas/San the remains of four superimposed house-
José transition. All proveniences used in hold units (Units C4 –C1). The details can
Fig. 18 come from San José Mogote and be found in Flannery and Marcus (1994:
Tierras Largas, two sites excavated by nat- Table 14.1).
ural stratigraphic units. Complete sherd The Zone D midden was roughly 40 cm
counts can be found in Flannery and Mar- thick. The volume excavated was 4.8 –5.0
cus (1994). m 3, slightly greater than that of Level 7 at
Four pottery types of the San José phase Tlapacoya. The number of Leandro Gray
were used as the medium for pan- sherds from Zone D (2332) is similar to the
Mesoamerican motifs. One, Leandro number of Tortuga Polished sherds from
Gray, resembles Tlapacoya’s Tortuga Pol- Tlapacoya’s Level 7 (2569). On the other
ished and Volcán Polished. Another, San hand, the number of Delfina Fine Gray
José Black-and-White, resembles Tlapa- sherds from Zone D (106) is greater than
coya’s Valle White-rim Black. Still an- the number of Atoyac Fine Gray sherds
other, Atoyac Yellow-white, resembles from Tlapacoya’s Level 7 (14). This is rea-
Tlapacoya’s Pilli White. Finally we come sonable, since petrographic evidence sug-
to Delfina Fine Gray, an export ware gests that such gray ware is native to Oa-
which—as we saw above—was traded to xaca.
(and imitated by) Tlapacoya. Household Units C4-C1 each produced
Leandro Gray was one of the most com- fewer sherds than Zone D, since the vol-
mon pottery types of the San José phase, ume of earth removed from each was on
usually exceeded in frequency only by the order of 2.4 m 3. Nevertheless, each
utilitarian cooking jars (Fidencio Coarse). household produced 674 to 1667 sherds of
Leandro Gray grew out of Tierras Largas Leandro Gray, and 16 to 43 sherds of
Burnished Plain, the most common utili- Delfina Fine Gray. Such quantities of
tarian ware of the Tierras Largas phase. sherds are consistent with what might be
The changes producing Leandro Gray, expected from volumes of earth half that
which emerged during the Tierras Largas/ of Tlapacoya’s Level 7. Like Tortuga Pol-
MEXICAN CHIEFDOMS AND MOTHER CULTURE
FIG. 18. Frequencies of classified sherds from eight proveniences at San José Mogote (SJM) and Tierras Largas (TL), Valley of Oaxaca. Pottery
types bearing pan-Mesoamerican motifs are printed in capital letters. H.16, House 16; H.C3, Household Unit C3; C/D2, Area C, Level D2; C/E, Area
C, Level E; LTL-3, House LTL-3; C/F, Area C, Level F; C/G, Area C, Level G; C/G2, Area C, Level G2. (Raw data from Flannery and Marcus 1994.)
23
24 FLANNERY AND MARCUS
TABLE 2
San José Mogote, Area A. Total Sherd Counts of the 4 Pottery Types Bearing Pan-Mesoamerican Motifs.
The Stratigraphic Units (All Belonging to the San José Phase) are the Zone D Midden (D) and Household
Units C4 Through C1. (Source: Flannery and Marcus 1994: Table 14.1.)
Type D C4 C3 C2 C1
ished, Leandro Gray was produced in a allel lines turn up or down at intervals
wide variety of vessel shapes: cylinders, (Flannery and Marcus 1994: Figs. 12.19 –
outleaned-wall bowls, tecomates, bol- 12.22). Yellow-white sherds of the San José
stered-rim bowls, spouted trays, vertical- phase suggest that the double-line break
necked jars, and many others. The variety originated as a simplified version of Earth,
of pan-Mesoamerican motifs was also with its cleft head and associated “music
great, including both carved examples brackets” (Fig. 19). This is significant for
(Pyne’s Motifs 1–7) and fine-line incised
examples (Pyne’s Motifs 8 –11 and 15–18).
Atoyac Yellow–white, a ware almost as
popular as Leandro Gray, was also used as
a medium for pan-Mesoamerican motifs
(Marcus 1989). Like Leandro Gray, it first
appeared during the Tierras Largas/San
José phase transition and grew out of Tier-
ras Largas Burnished Plain. (In this case,
the new ware was created simply by giv-
ing Tierras Largas Burnished Plain a white
slip.) In contrast to Leandro Gray—which FIG. 19. As early as 1150 b.c., abstract versions of
Earth/Earthquake were incised on white-slipped
was most often used for depictions of Sky/ pottery in the highland style province. This sherd of
Lightning—Atoyac Yellow–white was most Atoyac Yellow–white from the Valley of Oaxaca
often used for depictions of Earth/Earth- shows the cranial fissure (Pyne’s Motif 13) and “mu-
quake (Pyne’s Motifs 8 –10, 12, and 14). sic brackets” often associated with depictions of
Earth (see Fig. 8). No comparable white ware with
The Valley of Oaxaca was one of the incised Earth/Earthquake motifs has been found in
earliest regions to feature the “double- 1150 – 850 b.c. levels at San Lorenzo. (Drawn from a
line-break,” an incised motif in which par- photograph in Flannery and Marcus 1994:147.)
MEXICAN CHIEFDOMS AND MOTHER CULTURE 25
three reasons. First, it reinforces the asso- the start of the San Lorenzo phase (Coe
ciation of white ware with Earth/Earth- and Diehl 1980a: Fig. 97), rather than hav-
quake motifs. Second, it emphasizes the ing a long previous history like Tlapa-
ties between Oaxaca and the Basin of coya’s Tortuga Polished. Equally surpris-
Mexico, where similar motifs occur on ing is the fact that Calzadas Carved seems
Pilli White (compare Niederberger 1987: to be relatively rare, not exceeding 4% of
Figs. 475– 476 with Flannery and Marcus the classified sherds. Having been shown
1994: Fig. 19.1). Third, it suggests that the Cyphers’ new collections from San
double-line break variant of the Earth mo- Lorenzo, we have no doubt that she will
tif originated in the Mexican highlands one day be able to divide Calzadas Carved
around 1150 b.c. Not until three hundred into (1) a softer and darker gray ware like
years later, in the Nacaste phase, did a com- Leandro Gray/Tortuga Polished, and (2) a
parable incised white ware show up at San harder and lighter gray ware like Delfina
Lorenzo (Coe and Diehl 1980a:194). Fine Gray/Atoyac Fine Gray. At this writ-
Finally we come to San José Black-and- ing, however, we are limited to Coe and
White, Oaxaca’s version of Tlapacoya’s Diehl’s types. Let us therefore look at their
Valle White-rim Black. Such ware was not four published stratigraphic units.
present in Oaxaca until 1150 b.c.; once SL-PNW-St. II, a major stratigraphic
present, however, it was carved with unit for which Coe and Diehl present both
Pyne’s Motifs 7 and 11. a frequency graph and a sherd count, be-
gan as a 12 m 2 excavation (Coe and Diehl
San Lorenzo 1980a: Fig. 51, Fig. 97, Table 4-1). In its
lower levels the excavated area was twice
We turn now to San Lorenzo, the al- reduced, but its upper levels are compa-
leged wellspring of pan-Mesoamerican rable in volume to the 12 m 2 excavation in
motifs. In their report on the Yale project, Area A at San José Mogote. Levels O-K1
Coe and Diehl (1980a: Tables 4-1 to 4-4) are attributed to the “pre-Olmec” Bajı́o
publish the sherd counts from four strati- and Chicharras phases (1300 –1150 b.c.);
graphic excavations at San Lorenzo. We K2 is mixed; and Levels J–F are assigned
assume that these were their best strati- to the San Lorenzo phase, 1150 – 850 b.c.
graphic units, since they chose to publish (Fig. 20, Table 3).
them in detail. Calzadas Carved, regarded by Coe and
Our first surprise is that Coe and Diehl Diehl (1980a:159) as “100 percent Olmec,”
define only one pottery type—Calzadas occurred in Levels K2-F. What stands out
Carved—which bears pan-Mesoamerican is the small number of sherds— only 29 in
motifs. Their white-rimmed black ware all of SL-PNW-St. II. Level K2, whose vol-
does not bear such motifs, and even more ume was somewhere between 3 and 6 m 3
significantly, the San Lorenzo phase has no (3 ⫻ 2 ⫻ 0.5–1.0 m), produced 1617 classi-
incised white ware analogous to Atoyac Yel- fiable sherds, of which only 19 were Cal-
low–white or Pilli White. This fact has been zadas Carved. Level F, whose volume was
confirmed by Ann Cyphers (personal roughly 6 –9 m 3 (4 ⫻ 3 ⫻ 0.5– 0.75 m), pro-
communication, 1996) following her re- duced 133 classifiable sherds, of which
cent excavations at San Lorenzo. Owing to only 5 were Calzadas Carved. Nor do the
this lack of incised white wares, the San surprises end there: the total number of
Lorenzo phase has surprisingly few pan- Calzadas Carved sherds produced by the
Mesoamerican motifs featuring Earth/ Yale project’s four published stratigraphic
Earthquake. cuts was only 38 (Coe and Diehl 1980a:
Calzadas Carved appears abruptly at Tables 4-1 to 4-4).
26
FLANNERY AND MARCUS
FIG. 20. Frequencies of classified sherds in Levels O–D of Stratigraphic Unit SL-PNW-St. II at San Lorenzo, Veracruz. The lone pottery type
bearing pan-Mesoamerican motifs is printed in capital letters. (Based on Coe and Diehl 1980a: Fig. 97.)
MEXICAN CHIEFDOMS AND MOTHER CULTURE 27
TABLE 3
San Lorenzo, Unit SL-PNW-St.II. Total Sherd Counts of Calzadas Carved, the Only Pottery Type Bearing
Pan-Mesoamerican Motifs, in Levels O-B, Phases Bajı́o Through Nacaste. (Source: Coe and Diehl 1980a:
Table 4-1.)
Type O L K1 K2 J H F D C B
Calzadas Carved — — — 19 2 3 5 — — —
All classified sherds 288 155 386 1617 87 214 133 2137 689 4399
We know, of course, how Coe and Diehl White-rimmed black ware began at San
will rationalize these low counts; they will Lorenzo in the Chicharras phase (1200
argue that poor preservation of sherd sur- b.c.). But Perdida Black-and-White, the
faces at San Lorenzo made it impossible to relevant San Lorenzo type, was neither
identify those sherds of Calzadas Carved carved nor incised (Coe and Diehl 1980a:
vessels that did not bear the actual excis- Fig. 156). It most closely resembles Coate-
ing (Coe and Diehl 1980a:131). We re- pec White-rimmed Black from the Te-
spond to this apologia by giving the counts huacán and Oaxaca Valleys (MacNeish,
of excised Leandro Gray sherds from Area Peterson, and Flannery 1970:108; Flannery
A of San José Mogote in Table 2. The Zone and Marcus 1994:274), and may in fact be
D midden alone, with a volume of no the same ware. Coatepec White-rimmed
more than 5 m 3, produced 282 excised Black, while extremely well-made, was
sherds of Leandro Gray and 9 more of not carved. This lack of carved white-
Delfina Fine Gray. Household Unit C4, rimmed black ware contributes to the
with a volume of only 2.4 m 3, produced smaller repertoire of pan-Mesoamerican
141 excised sherds of Leandro Gray and 5 motifs at San Lorenzo.
more of Delfina Fine Gray. Even if we That’s right: San Lorenzo displays fewer
count only those gray sherds bearing ac- pan-Mesoamerican motifs than either Tlapa-
tual excising, Area A produced 678. coya or San José Mogote. We should not be
To be sure, since Cyphers has opened surprised, since Tlapacoya had 6 pottery
up larger areas of San Lorenzo, her sam- types bearing such motifs, San José Mo-
ple of Calzadas Carved is now larger than gote had 4, and San Lorenzo had only one.
Coe and Diehl’s. No amount of earth We have already stressed San Lorenzo’s
moved, however, will make up for the lack of Earth/Earthquake motifs on white-
aforementioned lack of white-slipped slipped ware; even Calzadas Carved,
ware comparable to Pilli White and however, shows an impoverishment of
Atoyac Yellow–white. At Tlapacoya and motifs. Aside from a sunburst motif (con-
San José Mogote, such white wares bear fidently called “God III, an eagle” by Coe
fully half the pan-Mesoamerican motifs; and Diehl 1980a:166), most illustrated mo-
take away the white ware and one loses tifs on Calzadas Carved are versions of
most of the depictions of Earth/Earth- Pyne’s Motif 1 (e.g. Coe and Diehl 1980a:
quake. Area A of San José Mogote had Figs. 138, 143).
more than 300 white-slipped sherds with When one focuses in detail on the use of
variants of pan-Mesoamerican motifs; San Pyne’s Motif 1 by the makers of Calzadas
Lorenzo phase levels in SL-PNW-St. II Carved, one sees another difference be-
had none. tween the highland and lowland style
28 FLANNERY AND MARCUS
FIG. 21. In the highland province, Pyne’s Motif 1 was usually carved onto vessels at a 45° angle,
as shown in (a) and (b). In the lowland province, Motif 1 was usually carved horizontally, as shown
in (c) and (d). (a) Volcán Polished bolstered-rim bowl from Tlapacoya (Niederberger 1976:170). (b)
Leandro Gray cylindrical bowl from the Valley of Oaxaca (Flannery and Marcus 1994:181). (c)
Calzadas Carved bolstered-rim bowl from San Lorenzo (Coe and Diehl 1980a:163). (d) Cuadros
phase bolstered-rim bowl from the Chiapas coast (Blake et al. 1995:178). (To be sure, gifts and
visitors crossed the style boundary often enough to provide some exceptions. For example, Blake
et al. [1995: Fig. 17a] illustrate a gray jar neck from Chiapas with Motif 1 at a 45° angle, done in the
way typical of jar necks at San José Mogote.)
provinces. At Tlapacoya and San José on cylindrical bowls. Pyne was allowed to
Mogote, Motif 1 was usually placed on take small pieces off these sherds so that
bowls at a 45° angle (Figs. 21a– b). At San William O. Payne, the Oaxaca project ce-
Lorenzo and various Chiapas sites, on the ramicist, could examine them under the
other hand, the same motif was usually microscope. Four of the eight fragments
placed horizontally (Figs. 21c– d). appear to be Leandro Gray, one resem-
The occasional exceptions to this pat- bled Delfina Fine Gray, and two others
tern are interesting. In 1972 Pyne, after contained decomposed gneiss or altered
studying hundreds of carved sherds from pegmatite like that present in Formative
Oaxaca, examined the Yale collection of Oaxaca clays (Flannery and Marcus 1994:
Calzadas Carved from San Lorenzo. 262–263). This contrasts with locally made
While most of the carved motifs were set Calzadas Carved, which is tempered with
horizontally on bolstered-rim bowls, Pyne “fine, quartzite sand” (Coe and Diehl
noticed eight which were set at a 45° angle 1980a:162). Thus at least 7 examples of this
MEXICAN CHIEFDOMS AND MOTHER CULTURE 29
allegedly “100 percent Olmec” ware may white (or resist white) wares unknown in
be from Oaxaca. the San Lorenzo phase.
It is significant that Pyne was successful At San José Mogote, pan-Mesoamerican
at picking intrusive Oaxaca sherds out of motifs occur on four different pottery
the Yale collections based solely on vessel types, ranging from dark gray to fine gray,
shape and the 45° placement of motifs; it white, and white-rim black. Because
confirms the relationship between stylistic white-slipped ware is so common in Oa-
preference and region of origin. (Paren- xaca, these four types represent more than
thetically, we did not notice any additional a third of the pottery assemblage, occur-
“Oaxaca-like” sherds in Cyphers’ collec- ring at densities of up to 1180 sherds per
tions of Calzadas Carved, which came cubic meter. Sky/Lightning motifs (nos.
from different proveniences at San 1– 6) were more common on gray ware,
Lorenzo.) while Earth/Earthquake motifs (nos. 8 –10,
12, 14 –15) were more common on white
THE NEED FOR A MORE ware.
RESTRICTED DEFINITION OF At San Lorenzo, pan-Mesoamerican
“OLMEC STYLE” motifs occur only on Calzadas Carved.
Carved sherds represent less than 4% of
the pottery assemblage, occurring at den-
In sum, only for colossal sculpture can a
sities of 3– 6/m 3 in Coe and Diehl’s four
case be made that it is “indigenous” to the
best stratigraphic proveniences. The de-
Gulf Coast. Even Olmec-centrists turn to
sign repertoire is essentially limited to
the central highlands of Mexico when
Pyne’s Motifs 1, 7, 11, and a sunburst. In
their art exhibit requires lots of complete,
well-made examples of hollow white baby part because the San Lorenzo phase lacks
dolls. We should stop calling these dolls a white ware equivalent to Pilli White or
“Olmec,” since to do so results in the par- Atoyac Yellow-white, it also lacks most of
adox pointed out by Serra Puche et al. the Earth/Earthquake motifs so wide-
(1996:39): it leaves us with “more objects spread in the highlands. Motif 1, when
of [alleged] ‘Olmec’ style in the highlands present, is usually set horizontally; a few
of Mesoamerica than on the coasts of Ta- sherds with motifs set at a 45° angle
basco or Veracruz.” turned out to be made of clays like those
In the case of pottery carved with Earth used in Oaxaca.
and Sky motifs, the notion that it is “in- What would an impartial observer con-
trusive” in the Mexican highlands is non- clude from this? That the Basin of Mexico
sense. At Tlapacoya, pan-Mesoamerican has so far produced the most abundant,
motifs occur on six different pottery types varied, and skillfully produced assem-
ranging from dark gray to fine gray, white, blage of vessels with pan-Mesoamerican
white-rim black, and resist white. Those motifs, and that the farther away you
types represent more than a quarter of the travel, the more impoverished the assem-
pottery assemblage, occurring at densities blages are in surface color, vessel shape,
of up to 769 sherds per cubic meter. Vessel range of motifs, and quality of execution.
shapes are diverse, and at least 6 of Pyne’s Many regions contributed to the richness
motifs (nos. 1, 2, 7, 10, 12, 15) were com- and diversity of Early Horizon ceramics,
mon. Earth/Earthquake motifs from Tla- and San Lorenzo never had more than a
pacoya include some of the most bril- subset of the shapes and motifs. We
liantly executed masterpieces of Early should stop calling such pottery “Olmec,”
Horizon art, and many of them are on and restrict that term to the chiefdoms of
30 FLANNERY AND MARCUS
southern Veracruz/Tabasco and the ob- mound groups, but no more impressive
jects most diagnostic of that region. than Tonga’s; they set up colossal heads,
but not as many as Easter Island; they
DOWNSIZING THE “OLMEC carved jade and wood, but no more skill-
PANTHEON” fully than the Maori. The Olmec were im-
pressive enough not to need the hyper-
Our data on carved pottery do more bolic claims of Olmec-centrists. Yet like all
than refute the notion that it was intrusive chiefdoms, they were a product of their
in the highlands. They spell Götterdäm- time, their place, and their interactions
merung for the pantheon of Olmec “gods.” with their neighbors.
Recall that in his study of Formative It is no accident, we believe, that 1150 –
iconography, Joralemon (1971) drew 850 b.c. was a period of rapid social evo-
heavily on the Basin of Mexico, Morelos, lution in Mesoamerican prehistory. It was
Puebla, and Guerrero in assembling his a period during which many competing
inventory of motifs. Each motif was then chiefly centers were concentrating man-
assumed to be a “god.” Expecting to find power, intensifying agriculture, exchang-
all these “gods” at San Lorenzo, Coe and ing sumptuary goods, and borrowing
Diehl (1980a:166) expressed disappoint- ideas from each other. We believe it was
ment when they found only two. The rea- the intensity of this competitive interac-
son for their disappointment should now tion, rather than the supremacy of any one
be clear: it was highland Mexico that had culture, that made social evolution so
the greatest repertoire of pan-Mesoameri- rapid. The social landscape of Mexico was
can motifs. Far from being the source of all one in which dozens of emerging chiefly
Early Horizon iconography, the Olmec centers were (1) sufficiently isolated to
were “out of the loop” relative to the Basin find the best adaptations for their respec-
of Mexico, Morelos, Puebla, and Oaxaca. tive regions, yet (2) sufficiently in contact
Had Joralemon been restricted to motifs to borrow relevant innovations from other
actually found on pottery at San Lorenzo, he regions as they arose. This is analogous to
would have had to conclude that the a situation the biologist Sewall Wright
Olmec were nearly monotheistic. (1939) once identified as favoring rapid
evolution.
SEWALL WRIGHT’S MODEL: Wright modeled a hypothetical “adap-
AN ALTERNATIVE TO THE tive landscape” of “very rugged charac-
OLMEC-CENTRIC VIEW ter” on which a genetically flexible popu-
lation might live (Fig. 22). The rises in this
It should now be clear why our position landscape were not hills, but peaks of pos-
cannot be described as primus inter pares. itive selective values—that is, gene com-
The Olmec may have been “first among binations that would be selected for.
equals” in sculpture; some Olmec chief- Those peaks were in turn separated by
doms may even have been “first” in pop- “saddles” of low selective values.
ulation size. But they were not the first to Let us imagine, Wright (1939:42) said,
use adobes, stone masonry, and lime plas- “an indefinitely large but freely inter-
ter, nor to lay out buildings 8° N of east. breeding species living under conditions
Nor were they “first among equals” in the which have not changed . . . for a long
production of white-slipped baby dolls or time.” As the result of natural selection,
carved pottery with pan-Mesoamerican that species has come to occupy a certain
motifs. Olmec chiefly centers were big, field of variation around one of the adap-
but not as big as Cahokia; they built tive peaks—in the case of Fig. 22a, the
MEXICAN CHIEFDOMS AND MOTHER CULTURE 31
FIG. 22. Six possible models for evolutionary change, redrawn from Sewall Wright (1939:Fig. 9).
Contour lines indicate a rugged adaptive landscape, with genetic combinations of high selective
value shown as peaks (⫹). Model f has the greatest potential for rapid evolutionary advance (see
text).
peak in the upper left corner of the map. then becomes: how may the species “con-
There is, however, an even higher peak of tinually find its way from lower to higher
selective values available in the lower peaks in such a system” (ibid.)?
right corner. The evolutionary problem Wright considered six possible scenar-
32 FLANNERY AND MARCUS
ios for change in this system. If mutation slow” and depends largely on “trial and
rates were to increase, or selection pres- error” (S. Wright 1939:45).
sure were to be reduced (Fig. 22a), the Finally we come to the scenario of most
population might spread downhill from relevance to this article, Fig. 22f. This is
its peak and, by further spreading across a the case of a large species that is subdi-
saddle, fortuitously reach a higher peak. vided into smaller local races, each breed-
The cost of such a spread, however, would ing largely within itself but occasionally
be an average lowering of the popula- cross-breeding (S. Wright 1939:46). With
tion’s adaptive level. Wright then consid- many local races spread out over the rug-
ered the opposite scenario, one in which ged landscape, there is a good chance that
the mutation rate decreased or selection at least one will come under the influence
pressure increased (Fig. 22b). In this case, of another peak, acquiring a preadapta-
the population might shrink to cover only tion useful to the species as a whole. Bet-
the top of the peak, strengthening its ad- ter still, several of the local races may ac-
aptation but decreasing its chance of cap- quire preadaptations. Those races will
turing a neighboring pinnacle. expand in number and, by cross breeding
Were the environment to change (Fig. with the others, make useful preadapta-
22c), so would the adaptive landscape; the tions available to all. This in turn pulls the
species might now find itself in a saddle. whole species to a higher position. “Fine
Under conditions of severe selection pres- division of a species into partially isolated
sure, the species would “merely be kept local populations,” said Wright (1939:46),
continually on the move” (S. Wright 1939: “provides a most effective mechanism for
44). It would, over time, be shuffled out of trial and error in the field of gene combi-
low peaks more easily than high ones, nations and thus for evolutionary advance
“and thus should gradually work its way by intergroup selection.”
to the higher general regions of the field To be sure, Wright’s conclusions
as a whole” (ibid.). Wright called this “an emerged from research on genetics, and
evolutionary process of major impor- are most relevant to biological evolution.
tance,” but one which requires a long pe- We believe, however, that Wright had dis-
riod of time and continual environmental covered a deeper underlying principle,
change. one relevant to sociocultural evolution as
Two of Wright’s scenarios, shown in well. That principle can be stated as fol-
Figs. 22d and e, relate to inbreeding. Un- lows: One of the most favorable scenarios
der conditions of close inbreeding (d) cer- for rapid evolutionary change is the divi-
tain alleles gradually become fixed, re- sion of a large population into numerous
gardless of whether or not they are smaller units, all adaptively autonomous
selectively advantageous. The result is a but still periodically in contact. Their au-
population low in variability, one which tonomy increases the likelihood that each
soon moves erratically downward from its will adapt successfully to its own social
peak, and finally becomes so homoge- and environmental setting; their periodic
neous that its change is slow and largely contact increases the likelihood that any
non-adaptive. Under conditions of only beneficial innovation will eventually be
slight inbreeding (e), the species tends to picked up by the entire population. We
wander continuously around its peak believe that Wright’s principle underlies
without leaving it entirely. It may find a several successful middle-range theories,
higher adaptive peak during this wander- including Renfrew and Cherry’s (1986)
ing, but its rate of progress is “extremely “peer-polity interaction.”
MEXICAN CHIEFDOMS AND MOTHER CULTURE 33
McKern, W. Niederberger, C.
1929 Archaeology of Tonga. Bernice P. Bishop Mu- 1976 Zohapilco: Cinco milenios de ocupación hu-
seum Bulletin 60. Hawaii. mana en un sitio lacustre de la Cuenca de
Marcus, J. México. Colección Cientı́fica, Arqueologı́a 30.
1989 Zapotec chiefdoms and the nature of forma- Instituto Nacional de Antropologı́a e Histo-
tive religions. In Regional perspectives on the ria, Mexico.
Olmec, edited by R. J. Sharer and D. C. 1987 Paléopaysages et archéologie pré-urbaine
Grove, pp. 148 –197. Cambridge Univ. Press, du Bassin de Mexique. Etudes Mésoameric-
Cambridge. aines 11, Mexico.
1992 Dynamic cycles of Mesoamerican states. Na- 1996 The basin of Mexico: A multimillennial de-
tional Geographic Research & Exploration velopment toward cultural complexity. In
8:392– 411. Olmec art of ancient Mexico, edited by E. P.
1993 Ancient Maya political organization. In Low- Benson and B. de la Fuente, pp. 83–93. Na-
land Maya civilization in the eighth century tional Gallery of Art, Washington, DC.
A.D., edited by J. A. Sabloff and J. S. Hen- Oliveros, J. A.
derson, pp. 111–183. Dumbarton Oaks, 1974 Nuevas exploraciones en El Opeño, Mi-
Washingon, DC. choacán. In The archaeology of west Mexico,
1998a The peaks and valleys of ancient states: An pp. 182–201. Sociedad de Estudios Avanza-
extension of the dynamic model. In Archaic dos del Occidente de México, Ajijic, Jalisco.
States, edited by G. M. Feinman and J. Mar- Ortı́z, P., and M. Rodrı́guez
cus, pp. 59 –94. SAR Press, Santa Fe, NM. 1989 Proyecto Manatı́ 1989. Arqueologı́a 1:23–52.
1998b Women’s ritual in formative Oaxaca: Figu- Mexico.
rine-making, divination, death and the an- 1999 Olmec ritual behavior at El Manatı́: A sacred
cestors. Memoirs of the University of Michigan space. In Social Patterns in Pre-Classic Me-
Museum of Anthropology 33. Ann Arbor. soamerica, edited by D. C. Grove and R. A.
1999 Men’s and women’s ritual in formative Oa- Joyce, pp. 225–254. Dumbarton Oaks, Wash-
xaca. In Social patterns in Pre-Classic Me- ington, DC.
soamerica, edited by D. C. Grove and R. A. Pauketat, T.
Joyce, pp. 67–96. Dumbarton Oaks, Wash- 1994 The ascent of chiefs. University of Alabama,
ington, DC. Tuscaloosa.
Marcus, J., and K. V. Flannery Pires-Ferreira, J.
1996 Zapotec civilization: How urban society evolved 1975 Formative Mesoamerican exchange net-
in Mexico’s Oaxaca Valley. Thames and Hud- works, with special reference to the Valley
son, London. of Oaxaca. Memoirs of the University of Mich-
Martı́nez Donjuán, G. igan Museum of Anthropology 7. Ann Arbor.
1985 El sitio olmeca de Teopantecuanitlán en Porter, M.
Guerrero. Anales de Antropologı́a 22:215–226. 1953 Tlatilco and the Pre-Classic Cultures of the
1994 Los olmecas en el estado de Guerrero. In Los New World. Viking Fund Publications in An-
Olmecas en Mesoamérica, edited by J. E. thropology, 19. New York: Wenner-Gren
Clark, pp. 143–163. Citibank, Mexico. Foundation for Anthropological Research.
Mead, S. M., A. Sullivan, D. R. Simmons, A. Sal- Pyne, N.
mond, B. Kernot, and P. Sciascia 1976 The fire-serpent and were-jaguar in Forma-
1985 Te Maori: Art from New Zealand collections. tive Oaxaca: A contingency table analysis. In
Harry N. Abrams, New York. The Early Mesoamerican Village, edited by
Milner, G. R. K. V. Flannery, pp. 272–280. Academic Press,
1990 The Late Prehistoric Cahokia cultural sys- New York.
tem of the Mississippi River Valley: Foun- Renfrew, C., and J. Cherry (Editors)
dations, florescence, and fragmentation. 1986 Peer–polity interaction and socio-political
Journal of World Prehistory 4:1– 43. change. Cambridge Univ. Press, Cambridge.
1998 The Cahokia Chiefdom: The archaeology of a Rountree, H. C.
Mississippian society. Smithsonian Institution 1989 The Powhatan Indians of Virginia. University
Press, Washington, DC. of Oklahoma Press, Norman.
Morley, S. G. Sanders, W. T., J. R. Parsons, and R. S. Santley
1938 The inscriptions of Peten, Volume 2. Carne- 1979 The basin of Mexico: Ecological processes in the
gie Institution of Washington Publication 437. evolution of a civilization. Academic Press,
Washington, DC. New York.
MEXICAN CHIEFDOMS AND MOTHER CULTURE 37
Serra Puche, M. C., with F. González de la Vara and Vaillant, Suzannah B., and George C. Vaillant
K. R. Durand V. 1934 Excavations at Gualupita. Anthropological
1996 Daily life in Olmec times. In Olmec art of Papers of the American Museum of Natural His-
Ancient Mexico, edited by E. P. Benson and B. tory, Vol. 35, Part 1. New York.
de la Fuente, pp. 35–39. National Gallery of
Van Tilburg, J.
Art, Washington, DC.
1994 Easter Island. Smithsonian Institution Press,
Smith, B. D. (Editor)
Washington, DC.
1986 The Mississippian emergence. Smithsonian In-
stitution Press, Washington, DC. Weaver, M. P.
Spencer, C. S. 1967 Tlapacoya pottery in the museum collection.
1993 Human agency, biased transmission, and Indian Notes and Monographs, Miscellaneous
the cultural evolution of chiefly authority. Series 56. Museum of the American Indian,
Journal of Anthropological Archaeology 12:41– Heye Foundation, New York.
74. Whalen, M.
1998 A mathematical model of primary state for- 1986 Sources of the Guilá Naquitz chipped stone.
mation. Cultural Dynamics 10:5–20. In Guilá Naquitz, edited by K. V. Flannery,
Symonds, S., and R. Lunagómez pp. 141–146. Academic Press, New York.
1997 El sistema de asentamientos y el desarrollo
de poblaciones en San Lorenzo Tenochtit- Wright, H. T.
lán, Veracruz. In Población, Subsistencia y Me- 1984 Prestate political formations. In On the evo-
dio Ambiente en San Lorenzo Tenochtitlán, lution of complex societies: Essays in honor of
México, edited by A. Cyphers, pp. 119 –152. Harry Hoijer, pp. 41–77. Undena Press,
Universidad Nacional Autónoma de Méx- Malibu.
ico, Mexico. 1986 The evolution of civilizations. In American
archaeology past and future, edited by D. J.
Tolstoy, P.
Meltzer, D. Fowler, and J. A. Sabloff, pp.
1989 Western Mesoamerica and the Olmec. In
Regional perspectives on the Olmec, edited by 323–365. Smithsonian Institution Press,
R. J. Sharer and D. C. Grove, pp. 275–302. Washington, DC.
Cambridge Univ. Press, Cambridge. Wright, S.
Tolstoy, P., and L. Paradis 1939 Statistical genetics in relation to evolution.
1970 Early and middle preclassic culture in the Actualités Scientifiques et Industrielles 802:37–
Basin of Mexico. Science 167:344 –351. 60. Hermann, Paris.