El Mito de La Cultura Madre FLANNERY

8/9/2019 El Mito de La Cultura Madre FLANNERY

1/37

Formative Mexican Chiefdoms and the Myth of the Mother Culture

Kent V. Flannery and Joyce Marcus

Museum of Anthropology, University of Michigan, Ann Arbor, Michigan 48109-1079

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

INTRODUCTION

On occasion, archaeologists revive ideasso anachronistic as to have been declareddead. The most recent attempt came w henRichard Diehl and Michael Coe (1996)parted the icy lips of the Olmec Mother

Culture and gave it mouth-to-mouth re-suscitation.1

The notion that the Olmec of the GulfCoast w ere the mother of all Mesoameri-can civilizations goes back more than halfa century (Covarrubias 1944), to a timewhen Formative archaeology was in itsinfancy. Scholars of the 1940s saw generalstylistic similarities between the Gulf

Coast and the Mexican highlands; sinceOlmec centers had stone m onuments andthe highlands generally did not, it wasassumed that the Gulf Coast was in theforefront and the highlands were begging

to be civilized. Five de cades of subse quen texcavation have shown the situation to bemore complex than that, but old ideas diehard.

In Olmec Archaeology (hereafter ab-breviated OA), Diehl and Coe (1996:11)propose that there are two contrastingschools of thought on the relationshipbetween the Olmec and the rest of Me-soamerica. In the Olmec-centric school theyplace themselves, John Clark, Beatriz de laFuente, Paul Tolstoy, and the late AlfonsoCaso, Ignacio Bernal, Miguel Covarrubias,Matthew Stirling, and George Vaillant.This group, they allege, agrees with themthat the Olmec were different from theircontemporaries in kind rather than de-gree, creating the entire symbolic systemof 1150500 b.c.2 and becoming theMother Culture of later Mesoamericancivilization. In the primus inter pares schoolthey place William R. Coe, Arthur De mar-est, John Graham, David Grove, NormanHammond, Flannery and Marcus, RobertStuckenrath, Jr., and the late Sir EricThompson. They describe this school asbelieving that the Olmec were no moreadvanced than any other Formative cul-

1 While Diehl is given as the co-author of the 1996

resuscitation, he and Coe are not always in full

agreement. For example, Diehl believes (as do we)

that the Olmec were a set of chiefdoms;Coe does not(Coe and Diehl 1980b:147). Coe believes that the

Olmec site of San Lorenzo is a gigantic bird effigy;

Diehl (personal communication, 1990) does not. We

thus feel uncomfortable including Diehl in our re-

buttal of what are largely Coes views. Our compro-

mise is simply to refer to the Diehl and Coe (1996)

paper by its title, Olmec Archaeology.

2 In this pape r, lowe rcase b.c. is use d for uncali-

brated radiocarbon years before the Christian era.

Journal of Anthropological Archaeology 19, 137 (2000)

doi:10.1006/jaar.1999.0359, available online at http://www.idealibrary.com on

10278-4165/00 $35.00

Copyright 2000 by Academic PressAll rights of reproduction in any form reserved.


2/37

ture and contributed little if anything tolater [Mesoamerican] civilization.

Our school would be happy to chal-lenge the Olmec-centrists to a tug-of-war,since half the members of their team are

dead. However, their portrayal of our po-sition is not accuratea familiar problemwhen one is being used as a straw man.We would not describe the Olmec as nomore advanced or contributing little.Their contribution has simply been exag-gerated by Olmec-centrists, who creditthe Olmec with many things their neigh-bors did earlier or better.

OA presents 11 traits which allegedlyshow the Olmecs maternal role in Me-soamericas genealogy (Diehl and Coe1996:23). We find those traits unconvinc-ing and suggest that there are betterframeworks than the Mother Culturemodel, which we do not find appropriatefor any world region. One alternative is amodel for the conditions leading to rapid

evolution, presented by the distinguishedbiologist Sewall Wright (1939). Even be-fore refuting the 11 traits, how ever, wemust modify the authors caricature of ourposition.

PRIMUS INTER PARES: A

CLARIFICATION

Any model for the Olmec and theirneighbors must be based on our currentunderstanding of complex societies,which is far greater now than in Vaillantsor Covarrubiass day (Anderson 1994;Carneiro 1981, 1991; Drennan and Uribe1987; Earle 1991a, b, 1997; Flannery 1995,1999; Goldman 1970; Johnson 1987; Kirch1984; Kirch and Green 1987; Marcus 1989,

1992; Marcus and Flannery 1996; Spencer1993, 1998; H. Wright 1984, 1986).

Among the most interesting societies inthe ethnographic and archaeologicalrecords are chiefdomssocieties based onhereditary differences in rank, in whichthe chiefs authority extends to satellite

communities. Chiefdoms are not a mono-lithic category; they come in many differ-ent types. Some, like those of PanamasAzuero Peninsula, were sedentary andflamboyant (Lothrop 1937; Linares 1977;

Helms 1979). Others, like those of IransZagros Mountains, were pastoral andnon-flamboyant (Barth 1964; Flannery inpress). Within Polynesia alone, Goldman(1970) has classified some chiefdoms astraditional (based more on sacred au-thority), others as open (based more onsecular power), and still others as strati-fied (large, with a combination of sacred

authority and secular power). Nowadaysthe term paramount often substitutesfor Goldmans stratified. While rank intraditional chiefdoms usually took theform of a continuum from higher to lowe rstatus, a few paramount chiefdomslikethose in Hawaii (Kirch 1984: Fig. 85)achieved stratification by cutting lower-status families out of the genealogy, re-

ducing them to a commoner class.In some parts of the ancient world,

chiefdoms persisted for centuries. Re-search in such regions has defined a long-term process called chiefly cycling (H.Wright 1984; Anderson 1994). In this pro-cess, paramount chiefdoms rose, peaked,then collapsed amid a regional landscapeof smaller traditional or open chiefdoms.

It is increasingly clear that paramountchiefdoms formed by taking over theirwe aker neighbors (Carneiro 1981, 1991).Their collapses resulted from such factorsas competition between chiefly families orfactions, endem ic raiding, agriculturalfailure, or demographic imbalance, andusually took the form of fragmentationback into the smaller units from which

th ey had be e n cre ate d. We vie w th eOlmec as one more set of paramountchiefdoms that rose, peaked, and eventu-ally collapsed in a landscape of traditionaland open chiefdoms.

A rare paramount chiefdom might suc-ceed in subduing and incorporating other

2 FLAN NERY AND MARCUS


3/37

large chiefdoms, creating a polity so greatthat it could no longer be administered asa chiefdom (Spencer 1998). This is howindigenous states formed in Madagascar(Dewar and Wright 1993) and among the

Zulu, Ashanti, Hunza, and Hawaiians(Flannery 1999). It is becoming increas-ingly clear that the first states in south-west Iran, Egypt, Peru, Oaxaca, and theMaya region also formed this way (H.Wright 1986; Flannery 1995; Marcus 1992,1993, 1998a; Marcus and Flanne ry 1996). Itmakes the study of chiefdoms all the moreinteresting to discover that, on at least

some occasions, they be came the precur-sors of states (Carneiro 1981; H. Wright1984).

It took more than 1000 years for Mexi-cos Formative societies to become com-plex enough to serve as precursors forstates. By the middle of the second mil-lennium b.c., agricultural villages werespread over the whole area from the Basin

of Mexico to the Pacific Coast of Chiapas.Some , but not all, of these village societieshad been reorganized into states by thebeginning of the Christian era.

We know less about this transitional pe-riod than we should, since some archae-ologists assume that their sites belong tochiefdoms without producing evidence ofthe requisite sociopolitical institutions.

Elsewhere we have suggested that asmany as ten lines of evidence may be nec-essary to confirm a chiefdom (Marcus andFlannery 1996:110). At this writing, we areconfident that the Valleys of Mexico,Puebla, Morelos, and Oaxaca, and variousparts of Guerrero, Chiapas, and southernVeracruz-Tabasco had chiefly societies by1150 b.c. We are less confident about areas

su ch as the Tehu acan Valle y and the Can-ada de Cuicatlan, but the y show evidenceof modest chiefdoms by 600 450 b.c.(Spencer 1993).

Most chiefly centers of 1150 450 b.c.were in frequent contact with each other,exchanging goods like obsidian, m arine

shell, iron ore mirrors, and the like (Pires-Ferreira 1975). Tlapacoya in the Basin ofMexico sent Paloma Negative and CestoWhite pottery to San Jose Mogote in theValley of Oaxaca; Oaxaca sent Leandro

Gray and Delfina Fine Gray pottery toTlapacoya and to San Lorenzo, Veracruz(Marcus 1989:192; Flanne ry and Marcus1994:259263, 286). San Jose Mogote re-ceived turtle shell drums and pearlyfresh water musse ls from the San Lorenzoregion; it also received GuamuchalBrushed pottery from Chiapas (Flanneryand Marcus 1994:286). Magnetite from

Oaxaca reached San Pablo in Morelos andSan Lorenzo in Veracruz (Pires-Ferreira1975).

There are two reasons why such ex-changes of goods should not surprise us.The first is that intersite distances werenot great. Given foot travel estimates of4.5 km per h (Morley 1938:234) or 32 kmper day (Hammond 1978), even a trip from

the Basin of Mexico to the Chiapas Coastwould take less than a month. The secondreason is that chiefly elites are always ea-ger for prestigious gifts from other chieflyelites.

THE O LMEC IN WID ER CONTEXT

Let us now look at the Olmec in the

context of chiefdoms worldwide . The apo-gee of this flamboyant society took placebetween 1150 and 300 b.c. on MexicosGulf Coast (Grove 1997). What we know ofits demographic history suggests typicalchiefly cycling. San Lorenzo, perhaps theearliest Olmec center, peaked between1150 and 900 b.c.; it then s uffered a loss ofpopulation and many of its stone monu-

me nts were de faced, mo st likely by a rivalch ie fd om (Co e an d Die hl 1980a, b;Cyphers 1997). San Lorenzos populationwas partially restored between 600 and400 b.c., after wh ich it collapse d again andlay abandoned for centuries rather thanbecoming part of a state.

3MEXICAN CHIEFDOMS AN D MOTHER CULTURE


4/37

La Venta, a second Olmec center some

90 km to the northeast, rose to promi-

nence between 900 and 600 b.c. (Drucker

et al. 1959; Gonzalez Lauck 1996). It is

probably no accident that La Ventas rise

coincided with San Lorenzos 900600 b.c.hiatus. Whether La Venta or a third

chiefly center was responsible for defac-

ing San Lorenzos monuments, this cycle

of synchronized rises and collapse s is typ-

ical o f chiefdoms compe ting for labor and

resources (H. Wright 1984; Anderson

1994).

Indeed, the Olmec resembled many

other chiefdoms worldwide. Some of theirchiefly centers covered hundreds of hect-

ares, like the largest Mississippian centers

o f No rth America. The Olmec bu ilt

earthen mounds like some Polynesian

chiefdoms. They set up huge stone sculp-

tures like chiefdoms on Easter Island, and

carved wooden statues and jade sumptu-

ary goods like the Maori. While they were

not identical to any of those other chief-

doms, the difference was more of degree

than kind.

Many chiefly centers sprawled over ar-

eas larger than that of a typical Bronze

Age city. This results from the fact that

chiefs cannot control pe ople at a distance,

as states can; many chiefs therefore con-

centrated thousands of farmers, warriors,and craftsmen as close to their residences

as possible. Conversely, when a chiefdom

cycled down, its loss of population

could be as spectacular as that recorded

from San Lorenzo at 900 b.c. by Symonds

and Lun agomez (1997:135).

Even at their peaks, San Lorenzo and La

Venta were smaller than Cahokia, a Mis-

sissippian chiefly center in Illinois. At itsapogee in A.D. 1250, Cahokia is estimated

to have covered 13 km 2 (Milner 1998:109;

Pauketat 1994). This is six times the cur-

rent estimate for La Venta (Gonzalez

Lauck 1996:75) and twice the most hyper-

bolic estimate for San Lorenzo (Lunago-

mez 1995).3 Surveys of the American Bot-tom, the alluvial valley surroundingCahokia, suggest that the sites immediatesustaining area may have covered 3000km 2. By A.D. 1400 it had collapsed w ithout

becoming part of a state.Like the Olmec, Cahokia was once con-

ceived of as a mother culture. Fortyyears ago, when we had much less infor-mation than we do now, the AmericanBottom was considered something of afont from which all Mississippian [cul-ture] arose, even the source of invadingwaves of population for other parts of

the eastern United States (Anderson 1994:144). Over the past four decades, thatmodel of Cahokian mother culture hasbee n replaced by a multiple-ce nter mode l.The Mississippian is now seen as emerg-ing (Smith 1990) simultaneously fromlocal Woodland cultures all over theSoutheast, and any recourse to popula-tion movement is suspect (Anderson

1994:144).Even within the 3000 km 2 American

Bottom, Milner (1990:29) would now seeCahokia as primus inter pares, the domi-nant political entity among a number oforganizationally similar (if less complex)semiautonomous chiefdoms which exer-cised considerable control over their ownterritoriessomething analogous, in

other wo rds, to Powhatans confede racy of200 villages (Rountree 1989). Anderson(1994:141) points to a significant differencebetween Cahokia and most early states:the complete absence [at Cahokia] of ev-idence for formal, differentiated adminis-trative structures.

While it lacked administrative struc-tures, Cahokia did build e arthen moun ds.

One of these, Monks Mound, stands 30 mhigh and covers an area 300 212 m(Anderson 1994:138). It is the largest

3 The largest estimates for San Lorenzo would in-

clude, within the boundaries of that one site, locali-

ties which other reports consider separate sites in the

settlement hierarchy below San Lorenzo.



5/37

earthen structure in the New World,dwarfing even the largest pyramid at LaVenta (Drucker et al. 1959:11).

Some Polynesian chiefdoms also builtmound groups rivaling La Venta. The Tui

Tonga, or hereditary chief, of Tonga ruledfrom a fortified ceremonial center calledLapaha on the island of Tongatapu(McKern 1929; Kirch 1984:228). Stretchingalong the shore of a lagoon for 1.5 km,Lapaha featured a series of plazas and2030 earthen mounds, both circular andrectangular. Historically known chiefs ortheir brothers and sisters lie buried under

specific mounds.Several Polynesian chiefdoms , most no-

tably Easter Island, erected monumentalstone statues analogous to the colossalheads of San Lorenzo and La Venta (Fig.1). Based on ethnohistoric records, Bahnand Flenley (1992) suggest that the hugestatues or moai of Easter Island representimportant ancestors. They were set on

slopes above villages to stand guard dur-ing the ceremonies of their descendants,but were sometimes toppled by victoriousenemies. Easter Island was an open chief-dom in Goldmans (1970:21) terms, notnearly as powerful as the paramountchiefdoms of Tonga and Hawaii. Despitethis, Easter Island carved 900 to 1000 moai(Van Tilburg 1994), roughly 100 times as

many colossal heads as are known fromSan Lorenzo (Diehl and Coe 1996:15).

Many chiefdoms carved jade, anotheractivity for which the Olmec are known.Among the most spe ctacular we re the tikisof jadeite or nephrite carved by the Maori(Mead et al. 1985), a traditional Polynesianchiefdom. Some carvers were renowne dthroughout New Zealand, and the best

jades were given names and became heir-looms for the elite.

Maori chiefs also supported carvers ofwooden house posts and statues (e.g.Mead et al. 1985: Pl. 37, 39, 47, 58). Thiscraft is another with which the Olmec arenow credited, based on the discovery of

waterlogged wooden busts in a spring atEl Manat , Veracruz (Ortz an d Rodrguez1989, 1999). As Fig. 2 should make clear,however, the best Maori carvers took abackseat to no one, including the Olmec.

Maori chiefs house s had carved lintels,thresholds, side posts, roof beams, andsupport posts. Each house was considereda work of art and given a name. Similar

craftsmanship was lavished even on stor-age houses, which might be given nameslike Te Oha, The Abundance. Maorichiefs also commissioned Meeting Housesfor their followers. The roofs of theseMeeting Houses were supported by hugeupright posts, sometimes carved to re-

FIG. 1. Many chiefdoms erected stone sculptures

of chiefly ancestors. (a) Moai 27 from Easter Island

(height 5.45 m). (b) Colossal Head 4 from San

Lorenzo (height 1.78 m). Redrawn from Van Tilburg

(1994) and de la Fuente (1975).



6/37

sem ble warriors; the door posts we re also

carved, often depicting legendary ances-tors.

San Lorenzo has produced one featurewhich (although enigmatic) hints that theOlmec may have had comparable publicbuildings, but with roofs supported by ba-salt columns instead of woo den posts. The

evidence consists of a carved basalt col-umn (now broken in half), its upright baseset in a patch of red clay floor with asso-ciated steps and a stone-lined drain(Cyphe rs 1997: Fig. 7.15).4

In sum, the Olmec fit comfortablywithin the parameters of chiefdomsworldwide . They built mounds and plazaslike Tongan chiefdoms, carved jades andwooden statues like the Maori, erected co-lossal heads like Easter Island, and con-centrated thousands of farmers, warriors,and artisans in sprawling settlements asthe chiefs o f Cahokia did. The Olmec look

impressive relative to their contemporar-ies, but not in comparison to later societieslike those centered at Teotihuacan andMonte Alban.

THE 11 OLMEC-CEN TRIC TRAITS

PROPOSED IN OA

In spite of the Olmecs resemblance to

other chiefdoms, Coe has always imag-ined them to be a colonizing empire, Me-soamericas first true civilization. Let uslook at the 11 traits which he and hisco-author believe support the Olmec-cen-tric view (Diehl and Coe 1996:11).

Trait 1. San Lorenzo and La Venta hadmultitiered, hierarchical settlement sys-tems that integrated towns, smaller vil-

lages, tiny hamlets, craft workshops andspecial ritual localessystems that oc-curred nowhere else in Mesoamerica untilcenturies later. (Convinced that specialritual locales were unique to the Olmec,the authors of OA use them again asTrait 7.)

Someone evidently hasnt been readingthe settlement pattern literature. Every

4 The basalt column is called Monument 57. Un-

fortunately, the patch of red clay floor with s teps has

been nicknamed El Palacio Rojo, an easily remem-

bered but misleading label since we lack a plan of

the building, and what we do have looks nothing like

a Mesoamerican palace (see Flannery 1998 for exam-

ples).

FIG. 2. Many chiefdoms were noted for elaborate

wood carving. (a) Maori carved post (height 175 cm).

(b) Carved bust from El Manat (height 55 cm).

Drawn from photos in Mead et al. (1985) and Ort z

and Rodrgue z (1989).



7/37

major chiefly center of the period 1150450 b.c. whose hinterland has been sys-tematically surveyed had villages andhamlets hierarchically below it. The Basinof Mexico (Sanders, Parsons, and Santley

1979; Niede rberger 1996: Map 1), the Val-ley of Morelos (Hirth 1980, Grove 1987),the Valley of Oaxaca (Kowalewski e t al.1989; Marcus and Flannery 1996), theChiapas coast (Clark and Blake 1994), andnorthern Belize (Hammond 1991) all hadhierarchies of villages and hamlets belowmajor centers. As for craft workshops,examples include the Matadamas chert

quarries (Whalen 1986) and Fabrica SanJose saltworks (Drenn an 1976) in the Val-ley of Oaxaca. Special ritual locales arealso widespread; consider the paintedcliffs and caves above the site of Tlapa-coya in the Basin of Mexico which, ori-ented east toward the volcanoes Ixtacci-huatl and Popocatepetl, receive the earlylight of sunrise and may have constituted

a significant component of sacred space(Niederberger 1996:87). The painted caveof Oxtotitlan in Guerrero (Grove 1970)would be a second example.

Trait 2. Although we cannot yet deci-pher the meanings, San Lorenzo and LaVenta were laid out as cosmograms.This is sheer speculation, based on Coesbelief that San Lorenzo was laid out to

resem ble a gigantic bird flying e ast (Coeand Diehl 1980a:387). This notion is re-futed by geological studies which showthat, although modifie d by architecturalterracing, the overall shape of the SanLorenzo plateau is largely the result ofnatural erosion (Cyphers 1997:102105).

While true cosmograms have not beenfound, many early Mesoamerican cultures

used solar or astral principles in orientingimportant buildings. As early as 1350 b.c.,the occupants of the Valley of Oaxacawere apparently aligning their MensHouses to the suns path during the equi-nox. This resulted in an orientation 8 N ofeast, or as it is often given, 8 W of north

(Flannery and Marcus 1994:3133; Marcus

and Flannery 1996:87). Complex A at La

Venta had a similar orientation (Drucker

et al. 1959), but since the Oaxaca Mens

Houses antedate Complex A by 500 years,

one can hardly credit the Olmec withMexicos first solar or astral alignments.

Trait 3. Although admitting that we

lack precise data on the size of Olmec

polities, the authors of OA argue that the

territories controlled by Olmec centers

may have been significantly larger than

those of their contemporaries. The truth

is that we also lack precise data on the size

of their contemporaries polities, makingthe whole topic speculative.

Trait 4. The Olmec, OA asserts, had a

highly sophisticated symbol system ex-

pressed in a coherent art style. We defer

our discussion of this trait to a later sec-

tion, where we show that San Lorenzo had

only a subset of the repertoire of symbols

used throughout early Mexico.

Trait 5. The Olmec invented monumen-

tal stone carving, which was a defining

characteristic of every Mesoamerican civ-

ilization. We agree that monumental

sculpture was a defining characteristic of

the Olmec; the question is, how accurate

an indicator of sociopolitical complexity is

it? We have already shown that Easter

Island, a modest chiefdom by Polynesianstandards, produced 100 times as many

colossal heads as are known from San

Lorenzo.

Trait 6. Predictably, the authors of OA

use the colossal heads for a second trait.

Both the heads, and the wooden busts

found in the spring at El Manat, are

thought by them to be portraits of rul-

ers. Again, this is pure speculation. Likethe statues of Easter Island, the Olmec

colossal heads might represent chiefly an-

cestors. As for the busts o f El Manat, they

might be (1) ancestors, like some Maori

woo dcarvings, or (2) surrogate sacrificial

victims tossed into a spring.



8/37

Trait 7. This trait, special ritual locales,has already been discussed under Trait 1.

Trait 8. The ballgame, OA claims, findsits oldest known e vidence in Olmec de -posits; San Lorenzos Palangana m ound

complex (600400 b.c.) is the first known,purposefully constructed ballcourt. Thisassertion is contradicted by Hill et al.(1998), who claim to have found a 1400 b.c.ballcourt at Paso de la Amada, Chiapas.The game itself is surely older than theOlmec; we even have one preceramiccamp site with a boulder-lined area thatcould be for ballgames (Marcus and Flan-

nery 1996:5859).The most convincing evidence for an

early Mexican ballgame comes from rub-ber balls preserved by waterlogging in thespring at El Manat (Ortz and Rodrgue z1989, 1999). The discovery of these balls,however, is an accident of good preserva-tion. We cannot assume that similar ball-games were unknown in the highlands;

after all, there are very early figurines ofballp laye rs at El Ope no, Michoacan (Oliv-eros 1974).

Trait 9. The authors o f OA use El Manatfor a second trait: the first ritual use ofrubber. It makes sense that the first ritualuse of rubber might occur on the GulfCoast, where rubber tree s are nativejustas it makes sense that the first ritual use of

obsidian and magne tite might occur in thehighlands, where those raw materials arenative. The point is, every region hassomething it did first.

Trait 10. But wait; El Manat g ets use dfor a thirdtrait. It provides the Olme c withthe oldest eviden ce for infant sacrifice inwater-related rituals.

The truth is that by the time El Manat

was occupied, infant sacrifice had existedin Mexico for thousands of years. Severalinfants were sacrificed (perhaps even can-nibalize d) in Level XIV o f Coxcatlan Cavein the Tehuacan Valley, an occupationdating to 5000 b.c. (MacNeish et al. 1972:266 270). The fact that the occupants of

the arid Teh uacan Valley used a dry cave

for such sacrifices, while the occupants of

the humid Gulf Coast used a spring,

hardly seems earth-shaking.

Trait 11. The Olmec had extensive

trade networks. While they admit thatmost Formative cultures had extensive

networks, the authors of OA insist that the

Olmec moved a greater quantity and

more different kinds of goods than their

contemporaries (they then pad the list

with probable exports for which we

have no physical evidence.) The fact is

that we currently have no objective, quan-

tified measure of goods moved by any

Formative society, especially in the case of

perishables.

We cannot resist pointing out the irony

of the OA authors position on Trait 11: All

Formative cultures had trade, but the

Olmec had the most trade. Doesnt that

make the Olmec primus inter pares?

TRAITS CONSPICUOUS BY THEIR

ABSENCE

As interesting as the 11 traits given in

OA are the firsts the authors do not list

for the Olmec. These include the first use

of lime plaster, adobe brick, and stone

masonry, three materials emblematic ofClassic Mesoamerican civilization. OA

cannot list these as Olme c innovations be -

cause their first use occurred in the Mex-

ican highlands. In the Valley of Oaxaca,

for example, lime plaster was used in

Mens Hou se s as early as 1350 b.c.; adobe s

were used in public buildings by 1000 b.c.;

and stone masonry platforms up to 2.5 min height were in use by 1000 b.c. (Marcus

and Flannery 1996:87, 109110). By the time

such construction techniques reached Com-

plex A, La Venta (Drucker et al. 1959), they

had been used in the highlands for cen-

turies.



9/37

A D ETAILED LOOK AT TRAIT 4

We now look at Trait 4, the highly so-phisticated Olmec symbol system/artstyle. OA asserts that this style spreadover all of Mesoamerica between 1150 and850 b.c., and its principal componentswere monumental, three-dimensional

stone sculpture; hollow whiteware figu-rines depicting babies; and CalzadasCarved pottery (Diehl and Coe 1996:23).The OA authors insist that these elem entsare indigenous in San Lorenzos InitialOlme c period culture and appear as intru-sive elem ents at San Jose Mogote in the

Valley of Oaxaca; Tlatilco, Tlapacoya, and

Las Bocas in central Mexico; several sites

in Guerrero; and Abaj Takalik, La Blanca,

and the Mazatan region in the Pacific

coastal region of Chiapas and Guatemala

(ibid.). As we shall see, the available data

do not support the notion that carved pot-

tery and hollow baby dolls are intrusive

in the highlands of Mexico.It is now clear that wide spread regional

styles existed in Mexico even before the

Olmec rose to prominence. Between 1400

and 1150 b.c., as pointed out by Clark

(1991: Fig. 8), Mexico was divided into at

least two ceramic style provinces (Fig. 3).

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 eme rged at

14001150 b.c. (Clark 1991: Fig. 8) and remained intact through 1150850 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 Co xcatlan Cave ; 9, Teopante cuanitlan; 10, Oxtotitlan Cave ; 11, No chixtlan and

Etlatong o; 12, Cuicatlan; 13, San Jose Mogo te and Tierras Largas; 14, La Ven ta; 15, San Lorenzo; 16,

El Manat; 17, Las Limas; 18, Mirador-Plumajillo; 19, Chiapa de Corzo; 20, Paso de la Amada.



10/37

The Basin of Mexico, Morelos, Puebla, theTehuacan Valley, the Valleys of Oaxacaand N ochixtlan, and the Cuicatlan Can-ada all shared red-on-buff bowls, bottles,

and jars (Fig. 4). East of Tehuacan andOaxaca, this red-on-buff complex gradu-ally gave way to one linking southern Ve-racruz, Tabasco, and Chiapas. This low-land complex featured tecomates orneckless jars with bichrome slips, fluting,or crosshatching (Fig. 5).

Despite these regional differences, afew pottery types were present on bothsides of the style boundary. One of thesewas a pure white product called kaolin

ware, believed on the basis of petro-graphic analysis to have been made in twoto three different regions (Fig. 6). Alsofound on both sides of the boundary weretecomates decorated with rocker stampingin zones (Fig. 7). Such vessels make thepoint that plastic decoration was already

FIG. 4. A comp lex of red-on-buff ves sels characterized the valleys o f the h ighland style province

at 14001150 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) He mispherical bowls from Tierras Largas. (f, g) He misphe rical

bowls from Tlapacoya. ( h, k) Hemispherical bowls from Ajalpan. (Redrawn from Grove 1974;

Niederberger 1976; Flannery and Marcus 1994; MacNeish et al. 1970.)



11/37

popular at 14001150 b.c., in what we as-sume the Olmec-centrists would have to

consider Grandmother Cultures.

The So-Called Early Horizon

Sometime around 12001150 b.c., in thewords of Tolstoy (1989:275), conditionsover much of Mesoamerica evidently fa-

vored demographic growth, craft special-

ization, increased interregional exchange,

greater disparities in social rank, andmore elaborate ceremonialism. Certaincommunities (often the largest in each re-

gion) seem to display these characteristics

more than others. The increased interre-gional exchange mentioned by Tolstoy in-volved obsidian, marine shell, iron ores

FIG. 5. Southern Veracruz, Tabasco, and Chiapas were part of a lowland style province at

14001150 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) Cotan Red flutedtecomate, 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.)



12/37

and pigme nts, jade, m ica, stingray spine s,turtle shell drums, and pottery. Oftenflamboyant, the pottery came in white,black, gray, red, red-and-white, andblack-and-white. Its plastic decoration,while still including rocker stamping, nowfeatured delicate fine-line incising, deepexcising or carving, and combinations of

these. Many of the carved and incised m o-tifs of 1150 b.c. were so stereotyped andpan-Meso american that some scholars as-sign them to an Early Horizon (seeGrove 1989 for discussion).

Olmec-centrists want us to believe thatthis style was created by the Olmec andimposed on the rest of Mexico. There areseveral reasons why that is unconvincing.

One reason is that Mexico did not, in fact,become one uniform style province be-tween 1150 and 850 b.c. Ceramic assem-blages from the Basin of Mexico, Puebla,Morelos, and Oaxaca compone nts of theold red-on-buff provincestill resem bledeach other more than they did the assem-

blages of the lowlands. Assemblages fromsouthern Veracruz, Tabasco, and Chia-pascomponents of the old lowlandprovincestill found their strongest tieswith each other. Another reason theOlmec-centrists model will not work isthat many of the ceramic features theyattribute to the Olmec appear earlier, are

more abundant, and/or are better made atTlapacoya, Tlatilco, Las Bocas, and SanJose Mogote than at San Lorenzo or LaVenta (Grove 1989).

Almost 30 years ago, Joralemon (1971)assembled an inventory of 176 allegedlyOlmec motifs. While widely cited byOlmec-centrists (e.g. Coe and Diehl 1980a,b), this study has two flawed assumptions:

(1) a belief that every motif was Olmec nomatter what region it came from, and (2)the notion that every motif was a deity.Joralemon created a pantheon of al-leged Olmec gods, but he did so relyingheavily on decorated wares from Tlatilco,Tlapacoya, Las Bocas, and other sites in

FIG. 6. Despite Mexicos division into style provinces at 14001150 b.c., some luxury pottery

types s how ed up everywhe re. Above , collared te comates in kaolin ware from San Jose Mogote (a,b)

and San Lorenzo (c). Belo w, kaolin bottle s from San Jos e Mo gote (d) and San Lorenzo (e). (Redrawn

from Flannery and Marcus 1994; Coe and Diehl 1980a.)



13/37

the Mexican highlands, rather than onGulf Coast pottery. As a result, a funny

thing happened to the pantheon on theway to San Lorenzo: most of its godsdropped out. Disappointed Olmec-cen-trists failed to realize that this was be causethe bulk of Joralemons motifs were not Olmec

at all, but highland Mexican.

A few years later, Pyne (1976) studied595 examples of decorated pottery from1150 650 b.c. in the Valley of Oaxaca.

Pyne identified 18 free-standing motifs,the full inventory of which can be foundin Figs. 12.512.6 of Flannery and Marcus(1994). Rather than referring to these asgods, Pyne simply called them Motifs1-18. She did point out that Motifs 16resembled a being Coe had called the

fire-se rpent or sky-dragon, whileMotifs 8 14 resembled another being,the were-jaguar (Pyne 1976:273). Be-cause the ceramics Pyne studied couldbe linked to house floors, burials, or

features (an advantage Joralemon didnot have), Pyne was able to point outthat Motifs 1 6 and Motifs 8 14 weremutually exclusive, that is, associatedwith different households or residentialwards (Pyne 1976:278).

Eve ntually, by com bining O to-manguean ethnohistory with an evenlarger sample of ceramics, Marcus (1989)

concluded that most of the motifs werenot gods at all, but references to thegreat world-divisions Earth and Sky(Flannery and Marcus 1994:136 149).Motifs 16 depicted Sky in its angryform, Lightning, a serpent of fire inthe sky (Figs. 8ac). Motifs 814 de-picted Earthsometimes as an Earthmask (Fig. 9) but often in its angry

form, Earthquake, complete with a clefthead representing a fissure in the earth(Fig. 8d).

The reason such motifs were wide-spread in early Mexico is because Earthand Sky we re parts of an ancient cosmo -logical dichotomy, not because of any-thing the Olmec did. Grove (1989) sug-gests that much of the symbolic content

existed before 1150 b.c. and is morelikely to reflect the common ancestry ofFormative cultures than the ingenuity ofone culture. By the time the motifs firstappeared on ceramics, they were al-ready stylized an d had reg ional variants.Fo r e xamp le , wh ile Earth was o fte nshown as Earthquake in the tremor-prone highlands, other artisans referred

to Earth by rendering the foot of theGreat Crocodile on whose back they be-lieved they resided (Fig. 10; see Marcus1989).

In s um, de spite references to the pe riod1150 850 b.c. as an Early Horizon, Me xicowas still divided into roughly the same

FIG. 7. Some decorative techniques were shared

by the h ighland and low land style provinces at 1400

1150 b.c. Here we see tecomates with zoned rocker

stamping, in Matadamas Orange from Tierras Lar-gas (a), Tatagapa Red from San Lorenzo (b), and an

unspecified ware from Tlatilco (c). (Redrawn from

Flannery and Marcus 1994; Coe and Diehl 1980a;

Porter 1953.)



14/37

stylistic provinces seen at 14001150 b.c.(Fig. 3). Ties between the Basin of Mexico,Morelos, Puebla, and Oaxaca remainedstrong, with San Jose Mo gote and Tlatilco/Tlapacoya using similar distinctive arti-facts (Figs. 11, 12) and displaying similarmotifs on similar vessels (Fig. 13). Tiesbetwee n Veracruz/ Tabasco and Chiapasalso remained close; for example, a Brain-

erd-Robinson matrix calculated byAgrinier (1989) shows strong similarity ince ram ic as se mblage s be tw ee n SanLorenzo (Veracruz) and Mirador-Pluma-jillo (Chiapas). Furthe r artifact similaritiesbetween those two sites include thou-sands of iron ore lug nuts or multi-

drilled cubes (Fig. 14). These unusual ar-tifactspresent also at Las Limas, Vera-cruz (Agrinier 1989:21)are virtually ab-sent to the west of the style boundary.

EVALUATIN G THE CLAIM OF

INTRUSIVENESS

Having shown that the major stylistic

provinces of early Mexico were un-changed by the rise of the Olmec, let uslook at the claim in OA that monumentalthree- dimensional stone sculpture; hol-low whiteware figurines depicting babies;and Calzadas Carved pottery were in-trusive elements at highland centers

FIG. 8. Representations of Sky/Lightning (ac) and Earth/Earthquake (df) on the pottery of1150850 b.c. (a) Lightning as a serpent of fire, Tlatilco. (b) Pynes Motif 1 (a stylized version of

Lightning in which the eyebrow flames are sine curves and the serpents gums are inverted Us),

San Jose Mog ote . (c) Motif 1 set at a 45 angle, as it often was in the highlands. (d) Angry ve rsions

of Earth with its head cleft by a seismic fissure, Tlapacoya. ( e) Stylized Earth m ask w ith cleft head

framed by music brackets, Tierras Largas. (f) Pynes Motif 13, Earths cleft head, as it often

appeared on white ware in the highlands (see Fig. 19).



15/37

(Diehl and Coe 1996:23). How does one

evaluate such a claim? To argue that a

specific area was the center of origin for

an artifact category, we believe youshould to be able to show that it occurred

first in that area; that it was more abun-

dant in that area; that it displayed greater

variety in that area; and/or that it was

more skillfully made in that area. Let us

see if these criteria are met.

Monumental Three-Dimensional Stone

Sculpture

The Gulf Coast does indeed have mon-ume ntal stone sculpture in greatest abun-

FIG. 10. Alternative ways of depicting Earth on

pottery, 1150 850 b.c. (a) Angry Earth (with cleft

head and anthropomorphized world directions) in-

cised on a Pilli White vessel from Tlapacoya. No

vessel with a motif this complex has been found in

the San Lorenzo phase, which lacked an incisedwhite ware comparable to Pilli White. (b) The hide of

a crocodile, as depicted on a human figure from

Atlihuayan. The foot of the crocodile (often mistak-

enly called a paw-wing motif) was widely used as

a symbol for Earth (see text). (Redrawn from Nied-

erberger 1987: Fig. 439; Benson and de la Fuente

1996:187.)

FIG. 9. Fine-line incised versions of Earth on pot-

tery, highland style province (1150 850 b.c.). (a)

Earth mask on Pilli White vessel, Tlapacoya; the

crossed bands in the mouth are Pynes Motif 7. (b)

Stylized Earth mask on Leandro Gray vessel, Tierras

Largas. (Redrawn from N iede rberger 1976; Flanne ry

and Marcus 1994.)



16/37

dance and variety. San Lorenzo alone hasproduced more than 70 stone mo nume nts,including 10 colossal heads (Cyphers1997). To be sure, since many heads werefound reused, rededicated, defaced, re-

worked, or out of context, we cannot besure how many actually date to the EarlyHorizon. Many similar monuments fromLa Venta are thought to be Middle Forma-tive (850500 b.c.) in date (Drucker et al.1959; Hammond 1988; Graham 1989;Grove 1997).

The real question is, how often doessuch sculpture appear as an intrusive el-

ement in the Mexican highlands? Teo-pantecuanitlan (Guerrero) has somethree -dim ensional monumen ts (MartnezDonjuan 1985, 1994), but mo st of these areMiddle Formative and might have beeninfluenced by the much nearer highlandsite of Chalcatzingo (Grove 1987). Oax-acas Early Horizon sculptures, such asMonume nts 1 and 2 o f San Jose Mogote,

are not Olmec in style (Marcus 1989:165;Flannery and Marcus 1994: Fig. 18.9). Inthe Basin of Mexico neither Tlatilco, Co-apexco, nor Tlapacoya has produced stonemonuments imitating those of the Olmec.Thus, w hile conceding a Gulf Coast originfor colossal heads, we find little evidence

for their intrusion into the Mexicanhighlands.

Hollow W hiteware Figurines Depicting

Babies

Hollow white-slipped baby dolls ap-pear to have been present at every majorMexican site of 1150500 b.c. Tlatilco, Tla-pacoya, Gualupita, Las Bocas, Teopante-cuanitlan, San Jose Mogote, Etlatongo,San Lorenzo, La Venta, and Paso de laAmada have all produced fragments orcomplete specimens.

For hollow w hite dolls, w e lack detailedstatistics comparable to those for thecarved pottery discussed below. It is in-structive, however, to examine examplesfor which proveniences are known or al-leged. Consider the catalogues for two re-cent exhibits of supposedly Olmec art:(1) one held by the National Gallery of Artin Washington, D .C. (Benson and de la

Fuente 1996) and (2) one held by The ArtMuseum of Princeton University (1996).

The National Gallery catalog illustratesseven hollow white dolls of young indi-viduals. All are masterpieces; none arefrom the Gulf Coast. Two are from Tla-tilco, two are from Tlapacoya, one is from

FIG. 11. Diagnostic of the highland style province at 1150850 b.c. were white-slipped spouted

trays, used to mix and pou r pigm ents. From L to R, thes e examp les com e from Tlatilco (Porter 1953),

Gualupita (Vaillant and Vaillant 1934), and Tierras Largas (Flannery and Marcus 1994). Diameter of

a, 15.5 cm. Such trays w ere not a significant part of the Gulf Coast inven tory; Coe and D iehl (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). Su ch indiscriminate application of the term Olmec waters d own whatever

regional and cultural significance it might have had, reducing it to a synonym for pretty.



17/37

Las Bocas, one is from Atlihuayan (More-los), and the last is said to be from Xo-chipala (Guerrero). Pages 130139 of the

Princeton catalogue illustrate nine morewhich look relatively authentic.5 The al-

leged proveniences are: three from Las

Bocas; two from Guerrero; one each from

Tlapacoya, Morelos, and the central

highlands of Mexico; and one listed sim-

ply as Mexico, although it was once at-

tributed to Las Bocas (Coe 1965: Fig. 184).The dolls illustrated in both catalogues

exceed in craftsmanship any hollow figu-

rine found by Coe and Diehl (1980a:261

279) at San Lorenzo or by Drucker et al.

(1959) at La Venta. This reinforces what

we learned 35 years ago with Coes (1965)

publication of The Jaguars Children: if you

want masterpieces in the hollow white

baby doll genre, turn to the Mexican high-lands. Such baby dolls were neither de-

monstrably earlier, nor more abundant,

nor more varied, nor more skillfully made

on the Gulf Coast;indee d, one could make

a case that their epicenter was Mexicos

central highlands.

Coe is aware of this fact, and has tried to

dismiss it by arguing that while fragments

of hollow white dolls occur in householddebris at San Lorenzo, they were treated

as prized burial furniture at highland

sites like Las Bocas or Gualupita (Coe

1989:77). The archaeological data do not

support this contrast. The restored baby

doll shown in Fig. 15 was pieced together

from fragmen ts in occupational refuse

at Tlapacoya (Tolstoy and Paradis 1970:

347). Pieces of hollow white dolls occur

regularly in ho uses and midde ns at Valley

of Oaxaca sites, even hamlets as small as

Tierras Largas (Marcus 1998b: Figs. 10.25,

11.14, 11.44, 12.7, 12.15, 12.22, 14.15, 14.34,

15.2). An d at Etlatongo in the N ochixtlan

Valley, a broken hollow doll was swept

5 It is disturbing to see how many of the objects in

the Princeton e xhibit resulted from looting. Page af-

ter page of the catalogue attributes pieces to private

collections. Do nt look for the names of any Me xican

archaeologists in the table of contents; they wo uldnt

have been caught dead participating in this display

of their stolen patrimony. It is perhaps forgivable

when a peasant farmer plows up an important piece

in his field and sells it to feed his family. It is unfor-

givable wh en a professional archaeologist or art his-

torian, who knows better, validates looting by au-

thenticating and glamorizing such pieces.

FIG. 12. Also diagnostic of the highland province

at 1150 850 b.c. were enigmatic ground- stone yugui-

tos or tiny yokes. Some scholars believe they werepart of the paraphernalia for a ballgame. (a) From

San Jose Mogote (Flannery and Marcus 1994:Fig.

13.9). (b) From Tlatilco (Porter 1953:Pl. 13H). Width of

(b), 15.5 cm.



18/37

into a trash-filled pit with the remains of adead dog (Blomster 1998).

Pottery with Pan-Mesoamerican Carved

Motifs

Finally, let us examine the claim in OAthat Calzadas Carved pottery was an in-trusive element at s ites like Tlapacoyaand San Jose Mogote. Since key strati-graphic units from Tlapacoya, San JoseMogote, and San Lorenzo have been pub-lished in detail, we can compare all threesites to see if the evidence supports thisclaim. We will look at several aspects ofthe pottery with pan-Mesoamerican mo-tifsits abundance in terms of she rds percubic meter;the percentage of the ceramicassemblage it makes up; its diversity insurface color and vessel shape; and thevariety of pan-Mesoamerican motifspresent in each region.

The Basin of Mexico

We begin at Tlapacoya in the Basin ofMexico (Nie de rberger 1976, 1987). N iede r-

FIG. 13. Between 1150 and 850 b.c. pottery as-

semblages of the highland style province shared

similar combinations of vessel shape and motif.

Here we see dark bottles with crosshatched sun-

burst motifs from Tlatilco (a) and San Jos e Mogo te

(b). (Redrawn from Porter 1953:Pl. 6I; Flannery and

Marcus 1994:99). Height of (a) 16.2 cm.

FIG. 14. Just as highland Mexican sites of 1150

850 b.c. shared ground- stone yuguitos, many lowland

sites shared multidrilled iron ore cubes or lug

nuts. These examples, averaging 3.1 cm thick, come

from Mirador-Plumajillo (ac) and San Lorenzo (df).

(Drawn from photographs in Agrinier 1989:25; Coe

and Dieh l 1980a:242.)



19/37

bergers unit Zohapilco-Tlapacoya IVwas a stratigraphic trench o ne me ter wide ,dug by natural stratigraphy. The relevantphases are Nevada (Levels 13, 12), 1350

1250 b.c.; Ayotla (Levels 11, 10, 9), 12501000 b.c.; and Manantial (Levels 8, 7, 6),1000 800 b.c. Profile drawings suggestthat 39 m of the trench were opened to thedep th of Level 9, while no m ore than 30 mwere opened to Level 13 or below (Nied-erberger 1976: Pl. 3, 4).

The trench itself was quite long, andfew of the levels ran for its full length. For

example, Level 9 ran for 20 m and was 60cm thick, so approximately 12 cubicmeters were removed (1 20 0.6 m).Roughly 12 m 3 of Level 8 were excavated;the volume removed from Level 7 wasonly 4.4 m 3 (1 11 0.4 m). We havechosen to highlight Levels 97 at Tlapa-

coya because they have the highest fre-quency of pan-Mesoamerican motifs. Thevolumes of earth removed can be com-pared with those at San Jose Mogote andSan Lorenzo (see below).

Now let us look at Fig. 16, the graph ofsherd frequencies in Niederbergerstren ch, and Table 1, the she rd counts of allTlapacoya pottery types bearing pan-Mesoamerican motifs. Note, first of all,that 6 pottery types at Tlapacoya bearpan-Mesoamerican motifs. Such motifsoccur on local dark gray wares (TortugaPolished and Volcan Polished ); gray ware

possibly imported from Oaxaca (AtoyacFine Gray); white-rimmed black ware(Valle White-rim Black); white-slippedware (Pilli White); and resist white ware(Paloma N egative).

Next, note how common many of thesetypes were. Tortuga Polished was the sec-ond most abundant ware of the Ayotlaphase, outnumbered only by the sherds of

utilitarian jars (Chalco Smoothed ). Tor-tuga Polished was 20% of the sherds inLeve ls 13-6; there we re 7728 she rds o f it inLeve l 8 alon e. Volcan Polished, a relatedware constituting le ss than 5% of the pot-tery, reached a peak of 787 sherds in Level9. These two gray wares bore many differ-ent motifs, from carved versions of PynesMotifs 1, 2, and 7 to fine-line incised or

hachured versions of her Motifs 12 and 15(Niederberger 1976: Pl. 35, 37).

Third, note that Tortuga Polished andVolcan Polished we re local types; they didnot appear suddenly at 1150 b.c., as if in-troduced from elsewhere. Both werepresent throughout the Nevada phase(1350 1250 b.c.), with Tortuga Polishedrepresenting m ore than 20% of the classi-

fied sherds at that time. Valle White-rimBlackanother type already prese nt atTlapacoya by 1350 b.c.was also used forpan-Mesoamerican motifs (Nied erberger1976: Pl. 45).

What adds to the variety of Tlapacoyapottery is the fact that white-slipped

FIG. 15. Pilli White h ollow doll from Tlapacoya.

This doll was found in pieces in household refuse; ithas since been restored as shown. Height, 41.5 cm.

(Drawn from a photo in Benson and de la Fuente

1996:185.)



20/37

FIG.16.Frequ

enciesofclassifiedsherdsin

Levels136atZohapilco-TlapacoyaIV,BasinofMexico.

Potterytypesbearingpan-M

esoamerican

motifsareprinte

dincapitalletters.(Basedon

Niederberger1976:Pl.32.)



21/37

wares were also used for pan-Mesoameri-can motifs. Pilli White appeared in nu-merous bowl forms, some carved withPynes Motif 1 and others showing herMotifs 10 and 15 in fine-line hachure(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 SanLorenzo (Fig. 17).

Finally we come to Atoyac Fine Gray, animported ware decorated with Pynes 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 andWayne Lambert, who consider them tohave been made in Oaxaca (Weaver 1967:30; Lambert 1972; Niederberger 1987:564;Flannery and Marcus 1994:259 262). Wesuspect that many of these vessels belongto a Oaxaca type called Delfina Fine Gray.However, other Atoyac Fine Gray vesselsillustrated by Niederberger (1976: Pl. 46)

were probably made locally.To summ arize: carved and incised pan-

Mesoamerican motifs were neither rarenor intrusive at Tlapacoya. The darkgray wares on which they occurred hadbeen among the most common local typesat 1350 b.c., and the motifs themselves

were common by 1250 b.c. In Level 8,whose volume was 12 m 3, there were 7728sherds of Tortuga Polished; in Level 7,amounting to o nly 4.4 m 3, there were 2569sherds of that type. Moreover, pan-

TABLE 1

Zohapilco-Tlapacoya IV. Total Sherd Counts of the 6 Pottery Types Bearing Pan-Mesoamerican Motifs,

Level s 13-6, Phases N evada Through Manantial. (Source: Ni ederberger 1976: 164.)

Type

N evada Ayotla Manantial

13 12 1011 9 8 7 6

Tortuga Polished 442 176 2954 4318 7728 2569 286

Volcan Polished 44 13 531 787 394 39 5

Pilli White 7 5 217 461 424 39 2

Paloma N egative 1 38 58 16 12 3

Valle White-rim Black 31 19 220 600 522 28 1

Atoyac Fine Gray 1 4 73 340 142 14 1

All classified sherds 1954 971 15637 20418 30914 11828 2278

FIG. 17. Four angry versions of Earth/Earth-

quakeone for each of the four great Mesoamerican

world d irections circle this bowl from Tlapacoya.

The type, Paloma Negative, combines (1) white slip

and (2) resist white over pale brown. Locally made atTlapacoya, Paloma Ne gative was traded as far as

Oaxaca. Highland vessels like this should not be

called Olmec. Coe and Diehl (1980a) report no

sherds of this ware from San Lorenzo and illustrate

no vessel approaching it in sophistication. (Drawn

from a photograph in Benson and de la Fuente 1996:

202.)



22/37

Mesoamerican motifs (sometimes bril-liantly executed) also occurred on Tlapa-coyas white-rimmed black, white-slipped, and resist white wares.

San Jose Mogote

The Valley of Oaxaca lies 330 km fromTlapacoya, but only 210 km from SanLorenzo. If the Olmec were truly thesource of inspiration alleged in OA, Oa-xacas Early Formative ceramics shouldresemble San Lorenzos more than Tlapa-coyas. In fact, the reverse is true (Flan-

nery and Marcus 1994).The relevant periods in Oaxaca are the

Tierras Largas phase (14001150 b.c.) andSan Jose phase (1150 850 b.c.). Fig. 18show s the changing freque ncies of potterytypes during the course of these periods,including the crucial Tierras Largas/ SanJose transition. All provenie nces use d inFig. 18 come from San Jose Mogote and

Tierras Largas, two sites excavated by nat-ural stratigraphic units. Complete sherdcounts can be found in Flannery and Mar-cus (1994).

Four pottery type s of the San Jose phasew ere use d as the m edium for pan-Mesoamerican motifs. One, LeandroGray, rese mble s Tlapacoyas Tortuga Pol-ishe d and Volcan Polished. An other, San

Jose Black-and- White, resem bles Tlapa-coyas Valle White-rim Black. Still an-other, Atoyac Yellow-white, resemblesTlapacoyas Pilli White. Finally we cometo Delfina Fine Gray, an export warewhichas we saw abovewas traded to(and imitated by) Tlapacoya.

Leandro Gray was one of the most com-mon pottery type s of the San Jose phase ,

usually exceeded in frequency only byutilitarian cooking jars (Fidencio Coarse).Leandro Gray grew out of Tierras LargasBurnished Plain, the most common utili-tarian ware of the Tierras Largas phase.The changes producing Leandro Gray,which e me rge d during the Tierras Largas/

San Jose phase transition, s imp ly requiredburnishing the ware twice instead of once,then firing it in a reducing atmosphere(Flanne ry and Marcus 1994:157165). Le-andro Gray went on to constitute 23% of

all sherds in m iddle San Jose times , a per-centage comparable to that of TortugaPolished in Tlapacoyas early Manantialphase.

Table 2 gives the actual counts of Lean-dro Gray, Delfina Fine Gray, San JoseBlack-and-White, and Atoyac Yellowwhite she rds from an excavation in Area Aof San Jose Mogote (Flann ery and Marcus

1994: Figs. 14.1, 14.4). We have chosen tofeature this excavation because it covered12 m 2, virtually the same area as an im-portant excavation at San Lorenzo whichwe will discuss below. The stratigraphiclevels consist of a midden (Zone D) andthe remains of four superimposed house-hold units (Units C4C1). The details canbe found in Flannery and Marcus (1994:

Table 14.1).The Zone D midden was roughly 40 cm

thick. The volume excavated was 4.85.0m 3, s lightly g reater than that of Leve l 7 atTlapacoya. The number of Leandro Grayshe rds from Zone D (2332) is similar to thenumber of Tortuga Polished sherds fromTlapacoyas Level 7 (2569). On the otherhand, the number of Delfina Fine Gray

sherds from Zone D (106) is greater thanthe number of Atoyac Fine Gray sherdsfrom Tlapacoyas Level 7 (14). This is rea-sonable, since petrographic evidence sug-gests that such gray ware is native to Oa-xaca.

Household Units C4-C1 each producedfewer sherds than Zone D, since the vol-ume of earth removed from each was on

the order of 2.4 m3

. Nevertheless, eachhouse hold produced 674 to 1667 sherds ofLeandro Gray, and 16 to 43 sherds ofDelfina Fine Gray. Such quantities ofsherds are consistent with w hat might beexpected from volumes of earth half thatof Tlapacoyas Level 7. Like Tortuga Pol-



23/37

FIG.18.Frequ

enciesofclassifiedsherdsfrom

eightproveniencesatSan

JoseMogote(SJM)andTierrasLargas(TL),ValleyofOa

xaca.Pottery

typesbearingpa

n-Mesoamericanmotifsarep

rintedincapitalletters.H.16,

House16;H.C3,HouseholdU

nitC3;C/D2,AreaC,LevelD2;C/E,Area

C,LevelE;LTL-3,HouseLTL-3;C/F,AreaC,

LevelF;C/G,AreaC,LevelG;C/G2,AreaC,LevelG2.(R

awdatafromFlanneryandMarcus1994.)



24/37

ished, Leandro Gray was produced in a

wide variety of vessel shapes: cylinders,

outleaned-wall bowls, tecomates, bol-stered-rim bowls, spouted trays, vertical-

necked jars, and many others. The variety

of pan-Mesoamerican motifs was also

great, including both carved examples

(Pynes Motifs 17) and fine-line incised

examples (Pynes Motifs 811 and 1518).

Atoyac Yellowwhite, a ware almost as

popular as Leandro Gray, was also use d as

a medium for pan-Mesoamerican motifs

(Marcus 1989). Like Leandro Gray, it first

appeared during the Tierras Largas/ San

Jose phase transition and g rew o ut 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 Graywhich

was m ost often use d for depictions o f Sky/LightningAtoyac Yellowwhite was most

often used for depictions of Earth/Earth-

quake (Pynes Motifs 810, 12, and 14).

The Valley of Oaxaca w as one of the

earliest regions to feature the double-

line-break, an incised motif in which par-

allel lines turn up or down at intervals(Flannery and Marcus 1994: Figs. 12.19

12.22). Yellow-w hite sherds of the San Josephase suggest that the double-line breakoriginated as a sim plified ve rsion of Earth,with its cleft head and associated musicbrackets (Fig. 19). This is significant for

TABLE 2

San Jose Mogo te, Area A. Total Sherd Counts of the 4 Pottery Types Bearing Pan-Mesoamerican Motifs.

The Stratigraphic Units (All Belonging to the San Jose 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

All Leandro Gray sherds 2332 1160 1667 949 674

Decorated Leandro sherds 298 149 151 74 46

Excised Leandro sherds 282 141 132 57 38

All Delfina Fine Gray sherds 106 43 43 27 16

Decorated Delfina sherds 16 5 4 5 5

Excised Delfina sherds 9 5 4 5 5

All San Jose Black-and-White sherds 150 36 51 19 14

Decorated B/ W sherds 2 1 2

Excised B/ W sherds 2 1 2

All Atoyac Yellow-white sherds 1904 762 1073 670 676

Decorated Atoyac sherds 117 51 81 51 46Excised Atoyac sherds 38 14 2 1 1

Double-line-break rims 9 3 23 24 27

All classified sherds 11356 4546 6876 3893 3361

FIG. 19. As early as 1150 b.c., abstract versions of

Earth/Earthquake were incised on white-slipped

pottery in the highland style province. This sherd ofAtoyac Yellowwhite from the Valley of Oaxaca

show s the cranial fissure (Pynes Mo tif 13) and mu-

sic brackets often associated with depictions of

Earth (see Fig. 8). No comparable white ware with

incised Earth/Earthquake motifs has been found in

1150850 b.c. levels at San Lorenzo. (Drawn from a

photograph in Flannery and Marcus 1994:147.)



25/37

three reasons. First, it reinforces the asso-ciation of white ware with Earth/Earth-quake motifs. Second, it emphasizes theties between Oaxaca and the Basin ofMexico, where similar motifs occur on

Pilli White (compare N iede rberger 1987:Figs. 475476 with Flannery and Marcus1994: Fig. 19.1). Third, it suggests that thedouble-line break variant of the Earth mo-tif originated in the Mexican highlandsaround 1150 b.c. Not until three hundred years later, in the Nacaste phase, did a com-

parable incised white ware show up at San

Lorenzo (Coe and Diehl 1980a:194).

Finally we come to San Jose Black-and-White, Oaxacas version of TlapacoyasValle White-rim Black. Such ware was notpresent in Oaxaca until 1150 b.c.; oncepresent, however, it was carved withPynes Motifs 7 and 11.

San Lorenzo

We turn now to San Lorenzo, the al-leged wellspring of pan-Mesoamericanmotifs. In their report on the Yale project,Coe and Diehl (1980a: Tables 4-1 to 4-4)publish the sherd counts from four strati-graphic excavations at San Lorenzo. Weassume that these were their best strati-graphic units, since they chose to publishthem in detail.

Our first surprise is that Coe and Diehldefine only one pottery typeCalzadasCarvedwhich bears pan-Mesoame ricanmotifs. Their white-rimmed black waredoes not bear such motifs, and even moresignificantly, the San Lorenzo phase has noincised white ware analogous to Atoyac Yel-

lowwhite or Pilli White. This fact has be enconfirmed by Ann Cyphers (personal

communication, 1996) following her re-cent excavations at San Lorenzo. Owing tothis lack of incised white wares, the SanLorenzo phase has surprisingly few pan-Mesoamerican motifs featuring Earth/Earthquake.

Calzadas Carved appears abruptly at

the start of the San Lorenzo phase (Coeand Diehl 1980a: Fig. 97), rather than hav-ing a long previous history like Tlapa-coyas Tortuga Polished. Equally surpris-ing is the fact that Calzadas Carved seems

to be relatively rare, not exceeding 4% ofthe classified sherds. Having been shownCyp he rs n ew co lle ctio ns fro m SanLorenzo, we have no doubt that she willone day be able to divide Calzadas Carvedinto (1) a softer and darker gray ware likeLeandro Gray/Tortuga Polished, and (2) aharder and lighter gray ware like DelfinaFine Gray/Atoyac Fine Gray. At this writ-

ing, however, we are limited to Coe andDie hls type s. Let us therefore look at theirfour published stratigraphic units.

SL-PNW-St. II, a major stratigraphicunit for which Coe and Diehl prese nt botha frequency graph and a sherd count, be-gan as a 12 m 2 excavation (Coe and Diehl1980a: Fig. 51, Fig. 97, Table 4-1). In itslower levels the excavated area was twice

reduced, but its upper levels are compa-rable in volume to the 12 m 2 excavation inArea A at San Jose Mogote. Leve ls O-K1are attributed to the pre-Olme c Bajoand Chicharras phases (1300 1150 b.c.);K2 is mixed; and Levels JF are assignedto the San Lorenzo phase, 1150850 b.c.(Fig. 20, Table 3).

Calzadas Carved, regarded by Coe and

Die hl (1980a:159) as 100 pe rcent Olme c,occurred in Levels K2-F. What stands outis the small number of sherdsonly 29 inall of SL-PNW-St. II. Level K2, whose vol-ume was somewhere between 3 and 6 m 3

(3 2 0.51.0 m ), produced 1617 classi-fiable sherds, of which only 19 were Cal-zadas Carved. Level F, whose volume wasroughly 69 m 3 (4 3 0.50.75 m), pro-

duced 133 classifiable sherds, of whichonly 5 were Calzadas Carved. Nor do thesurprises end there: the total number ofCalzadas Carved sherds produced by theYale projects four published stratigraphiccuts was only 38 (Coe and Diehl 1980a:Tables 4-1 to 4-4).



26/37

FIG.20.Frequ

enciesofclassifiedsherdsin

LevelsODofStratigraphicUnitSL-PNW-St.IIatSan

Lorenzo,Veracruz.Thelone

potterytype

bearingpan-Mesoamericanmotifsisprintedincapitalletters.(BasedonC

oeandDiehl1980a:Fig.97.)



27/37

We know, of course, how Coe and Diehlwill rationalize these low counts; they will

argue that poor preservation of sherd sur-faces at San Lorenzo made it impossible toidentify those sherds of Calzadas Carvedvessels that did not bear the actual excis-ing (Coe and Diehl 1980a:131). We re-spond to this apologia by giving the countsof excise d Leandro Gray she rds from AreaA of San Jose Mogote in Table 2. The ZoneD midden alone, with a volume of no

more than 5 m3

, produced 282 excisedsherds of Leandro Gray and 9 more ofDelfina Fine Gray. Household Unit C4,with a volume of only 2.4 m 3, produced141 excised sherds of Leandro Gray and 5more of Delfina Fine Gray. Even if wecount only those gray sherds bearing ac-tual excising, Area A produced 678.

To be sure, since Cyphers has opened

up larger areas of San Lorenzo, her sam-ple of Calzadas Carved is now larger thanCoe and Diehls. No amount of earthmoved, however, will make up for theaforementioned lack of white-slippedware comparable to Pilli White andAtoyac Yellowwhite. At Tlapacoya andSan Jose Mogote, such white wares bearfully half the pan-Mesoamerican motifs;

take away the white ware and one losesmost of the depictions of Earth/Earth-quake. Area A of San Jose Mogote hadmore than 300 white-slipped sherds withvariants of pan-Mesoamerican motifs;SanLorenzo phase levels in SL-PNW-St. IIhad none.

White-rimmed black ware began at SanLorenzo in the Chicharras phase (1200

b.c.). But Perdida Black-and-White, therelevant San Lorenzo type, was neithercarved nor incised (Coe and Diehl 1980a:Fig. 156). It most closely resembles Coate-pec White-rimmed Black from the Te-huacan and Oaxaca Valleys (MacNe ish,Peterson, and Flannery 1970:108; Flanne ryand Marcus 1994:274), and may in fact bethe same ware. Coatepec White-rimmed

Black, while extremely well-made, wasnot carved. This lack of carved white-rimmed black ware contributes to thesmaller repertoire of pan-Meso americanmotifs at San Lorenzo.

Thats right: San Lorenzo displays fewer pan-Mesoamerican motifs than either Tlapa-

coya or San Jose Mogote. We should not besurprised, since Tlapacoya had 6 pottery

types bearing such motifs, San Jose Mo-gote had 4, and San Lorenzo had only one.We have already stressed San Lorenzoslack of Earth/ Earthquake motifs on white-slipped ware; even Calzadas Carved,however, shows an impoverishment ofmotifs. Aside from a sunburst motif (con-fidently called God III, an eagle by Coeand Die hl 1980a:166), m ost illustrated mo-

tifs on Calzadas Carved are versions ofPynes Motif 1 (e.g. Coe and Diehl 1980a:Figs. 138, 143).

When one focuse s in detail on the use ofPynes Motif 1 by the makers of CalzadasCarved, one sees another difference be-tween the highland and lowland style

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 Bajo Through N acaste. (Source: Coe and D iehl 1980a:

Table 4-1.)

Type

Bajo Chicharras Mixed San Lorenzo N acaste

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



28/37

provinces. At Tlapacoya and San JoseMogote, Motif 1 was usually placed onbowls at a 45 angle (Figs. 21ab). At SanLorenzo and various Chiapas sites, on theother hand, the same motif was usuallyplaced horizontally (Figs. 21cd).

The occasional exceptions to this pat-

tern are interesting. In 1972 Pyne, afterstudying hundreds of carved sherds fromOaxaca, e xamined the Yale collection ofCalzadas Carved from San Lorenzo.While most of the carved motifs were sethorizontally on bolstered-rim bowls, Pynenoticed eight which we re set at a 45 angle

on cylindrical bowls. Pyne was allowed totake small pieces off these sherds so thatWilliam O. Payne, the Oaxaca project ce-ramicist, could examine them under themicroscope. Four of the eight fragmentsappear to be Leandro Gray, one resem -bled Delfina Fine Gray, and two others

contained decomposed gneiss or alteredpegmatite like that present in FormativeOaxaca clays (Flannery and Marcus 1994:262263). This contrasts with locally madeCalzadas Carved, which is tempered withfine, quartzite sand (Coe and Diehl1980a:162). Thus at least 7 examples of this

FIG. 21. In the highland province, Pynes Motif 1 was usually carved onto vessels at a 45 angle,

as shown in (a) and (b). In the lowland p rovince, Motif 1 was usually carved horizontally, as show n

in (c) and (d). (a) Volcan Polished bolstered- rim bow l from Tlapacoya (Nie de rberger 1976:170). (b)

Leandro Gray cylindrical bo wl 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 ne ck from Chiapas with Mo tif 1 at a 45 angle, do ne in the

way typical of jar necks at San Jose Mogo te.)



29/37

allegedly 100 percent Olmec ware maybe from Oaxaca.

It is significant that Pyne was successfulat picking intrusive Oaxaca sherds out ofthe Yale collections based solely on vessel

shape and the 45 placement of m otifs; itconfirms the relationship between stylisticpreference and region of origin. (Paren-thetically, we did not notice any additionalOaxaca-like sherds in Cyphers collec-tions of Calzadas Carved, which camefrom different proveniences at SanLorenzo.)

THE N EED FOR A MORE

RESTRICTED D EFINITION OF

OLMEC STYLE

In sum, only for colossal sculpture can acase be made that it is indigenous to theGulf Coast. Even Olmec-centrists turn tothe central highlands of Mexico when

their art e xhibit requires lots o f comp lete,well-made examples of hollow white babydolls. We should stop calling these dollsOlmec, since to d o s o results in the par-adox pointed out by Serra Puche et al.

(1996:39): it leaves us with more objectsof [allege d] Olmec style in the highlandsof Mesoamerica than on the coasts of Ta-basco or Veracruz.

In the case of pottery carved with Earthand Sky motifs, the notion that it is in-trusive in the Mexican highlands is non-sense. At Tlapacoya, pan-Mesoamericanmotifs occur on six different pottery typesranging from dark gray to fine gray, white,white-rim black, and resist white. Thosetypes represe nt more than a quarter of thepottery assemblage, occurring at densities

of up to 769 sherds per cubic meter. Vesselshapes are diverse, and at least 6 of Pynesmotifs (nos. 1, 2, 7, 10, 12, 15) were com-mon. Earth/Earthquake mo tifs from Tla-pacoya include some of the most bril-liantly executed masterpieces of EarlyHorizon art, and many of them are on

white (or resist white) wares unknown in

the San Lorenzo phase.

At San Jose Mogote, pan-Mesoame rican

motifs occur on four different pottery

types, ranging from dark gray to fine gray,

white, and white-rim black. Becausewhite-slipped ware is so common in Oa-

xaca, these four types represe nt more than

a third of the pottery assemblage, occur-

ring at densities of up to 1180 sherds per

cubic meter. Sky/Lightning motifs (nos.

1 6) w ere more common on gray ware,

while Earth/Earthquake motifs (nos. 8 10,

12, 1415) were more common on white

ware.At San Lorenzo, pan-Mesoamerican

motifs occur only on Calzadas Carved.

Carved sherds represent less than 4% of

the pottery assem blage, occurring at den -

sities of 36/m 3 in Coe and Diehls four

best stratigraphic proveniences. The de-

sign repertoire is essentially limited to

Pynes Motifs 1, 7, 11, and a sunburst. In

part because the San Lorenzo phase lacks

a white ware equivalent to Pilli White or

Atoyac Yellow-white, it also lacks most of

the Earth/Earthquake motifs so wide-

spread in the highlands. Motif 1, when

present, is usually set horizontally; a few

sherds with motifs set at a 45 angle

turned out to be made of clays like those

used in Oaxaca.What would an impartial obse rver con-

clude from this? That the Basin of Mexico

has so far produced the most abundant,

varied, and skillfully produced assem-

blage of vessels with pan-Mesoamerican

motifs, and that the farther away you

travel, the more impoverished the assem-

blages are in surface color, vessel shape,

range of motifs, and quality of execution.Many regions contributed to the richness

and diversity of Early Horizon ceramics,

and San Lorenzo never had more than a

subset of the shapes and motifs. We

should stop calling such p ottery Olme c,

and restrict that term to the chiefdoms of



30/37

southern Veracruz/Tabasco and the ob-jects most diagnostic of that region.

D OWNSIZIN G THE OLMEC

PANTHEON

Our data on carved pottery do morethan refute the notion that it was intrusivein the highlands. They spell Gotterdam-merung for the pantheon of Olme c gods.

Recall that in his study of Formativeiconography, Joralemon (1971) drewheavily on the Basin of Mexico, Morelos,Puebla, and Guerrero in assembling his

inventory of motifs. Each motif was thenassumed to be a god. Expecting to findall these gods at San Lorenzo, Coe andDieh l (1980a:166) expressed disappoint-ment w hen they found only two. The rea-son for their disappointment should nowbe clear: it was highland Mexico that hadthe greatest repertoire of pan-Mesoameri-can motifs. Far from being the so urce of all

Early Horizon iconography, the Olmecwe re out of the loop relative to the Basinof Mexico, Morelos, Puebla, and Oaxaca.Had Joralemon bee n restricted to motifs actually found on pottery at San Lorenzo, hewould have had to conclude that theOlmec were nearly monotheistic.

SEWALL WRIGHTS MOD EL:

AN ALTERN ATIVE TO THEOLMEC-CEN TRIC VIEW

It should now be clear why our positioncannot be described as primus inter pares.The Olmec may have been first amongequals in sculpture; some Olmec chief-doms may even have been first in pop-ulation size. But they were not the first to

use adobes, stone m asonry, and lime p las-ter, nor to lay out buildings 8 N of east.Nor were they first among equals in theproduction of white-slipped baby dolls orcarved pottery with pan-Mesoamericanmotifs. Olmec chiefly centers were big,but not as big as Cahokia; they built

mound groups, but no more impressivethan Tongas; they set up colossal heads,but not as many as Easter Island; theycarved jade and wood, but no more skill-fully than the Maori. The Olme c we re im-

pressive enough not to need the hyper-bolic claims of Olme c-centrists. Yet like allchiefdoms, they were a product of theirtime, their place, and their interactionswith their neighbors.

It is no accident, we believe, that 1150850 b.c. was a period of rapid social evo-lution in Mesoamerican prehistory. It wasa period during which many competing

chiefly centers were concentrating man-power, intensifying agriculture, exchang-ing sumptuary goods, and borrowingideas from each other. We believe it wasthe intensity of this competitive interac-tion, rather than the supremacy of any oneculture, that made social evolution sorapid. The social landscape of Mexico wasone in which dozens of emerging chiefly

centers were (1) sufficiently isolated tofind the best adaptations for their respec-tive regions, yet (2) sufficiently in contactto borrow relevant innovations from otherregions as they arose . This is analogous toa situation the biologist Sewall Wright(1939) once identified as favoring rapidevolution.

Wright modeled a hypothetical adap-

tive landscape of very rugged charac-ter on which a genetically flexible popu-lation might live (Fig. 22). The rises in thislandscape we re not hills, but peaks of po s-itive selective valuesthat is, gene com-binations that would be selected for.Those peaks were in turn separated bysaddles of low selective values.

Let us imagine, Wright (1939:42) said,

an indefinitely large but freely inter-breeding species living under conditionswhich have not changed . . . for a longtime. As the result of natural selection,that species has come to occupy a certainfield of variation around one of the adap-tive peaksin the case of Fig. 22a, the



31/37

peak in the upper left corner of the map.There is, however, an even higher peak ofselective values available in the lowerright corner. The evolutionary problem

then becomes: how may the species con-tinually find its way from lower to higherpeaks in such a system (ibid.)?

Wright considered six possible scenar-

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).



32/37

ios for change in this system. If mutation

rates were to increase, or selection pres-

sure were to be reduced (Fig. 22a), the

population might spread downhill from

its peak and, by further spreading across a

saddle, fortuitously reach a higher peak.The cost of such a sp read, how ever, would

be an average lowering of the popula-

tions adaptive level. Wright then consid-

ered the opposite scenario, one in which

the mutation rate decreased or selection

pressure increased (Fig. 22b). In this case,

the population might shrink to cover only

the top of the peak, strengthening its ad-

aptation but de creasing its chance of cap-turing a neighboring pinnacle.

Were the environment to change (Fig.

22c), so w ould the adaptive landscape; the

species might now find itself in a saddle.

Unde r conditions o f severe se lection pres-

sure, the species would merely be kept

continually on the move (S. Wright 1939:

44). It wou ld, over time , be shuffled out of

low peaks more easily than high ones,

and thus should gradually work its way

to the higher general regions of the field

as a whole (ibid.). Wright called this an

evolutionary process of major impor-

tance, but one which requires a long pe-

riod of time and continual environmental

change.

Two of Wrights scenarios, shown inFigs. 22d and e, relate to inbreeding. Un-

der conditions of close inbree ding (d) cer-

tain alleles gradually become fixed, re-

g ard le ss o f whe th er o r n ot th ey are

selectively advantageous. The result is a

population low in variability, one which

soon moves erratically downward from its

peak, and finally becomes so homoge-

neous that its change is slow and largelynon-adaptive. Under conditions of only

slight inbreeding (e), the species tends to

wander continuously around its peak

without leaving it entirely. It may find a

higher adaptive peak during this wander-

ing, but its rate of progress is extremely

slow and depends largely on trial and

error (S. Wright 1939:45).

Finally we come to the scenario of most

relevance to this article, Fig. 22f. This is

the case of a large species that is subdi-

vided into smaller local races, each breed-ing largely within itself but occasionally

cross-breeding (S. Wright 1939:46). With

many local races spread out over the rug-

ged landscape, there is a good chance that

at least one will come unde r the influence

of another peak, acquiring a preadapta-

tion useful to the species as a whole. Bet-

ter still, several of the local races may ac-

quire preadaptations. Those races willexpand in number and, by cross breeding

with the others, make useful preadapta-

tions available to all. This in turn pulls the

whole species to a higher position. Fine

division of a sp ecies into partially iso lated

local populations, said Wright (1939:46),

provides a most effective mechanism for

trial and error in the field of gene combi-

nations and thus for evolutionary advance

by intergroup sele ction.

To be sure, Wrights conclusions

emerged from research on genetics, and

are most relevant to biological evolution.

We believe, h owe ver, that Wright had dis-

covered a deeper underlying principle,

one relevant to sociocultural evolution as

well. That principle can be stated as fol-lows: One of the most favorable sce narios

for rapid evolutionary change is the divi-

sion of a large population into numerous

smaller units, all adaptively autonomous

but still periodically in contact. Their au-

tonomy increases the likelihood that each

will adapt successfully to its own social

and environmental setting; their periodic

contact increases the likelihood that anybeneficial innovation will eventually be

picked up by the entire population. We

believe that Wrights principle underlies

several successful middle-range theories,

including Renfrew and Cherrys (1986)

pee r-po lity interaction.



33/37

CONCLUSIONS

Competitive interaction is a bettername for our position than primus interpares. In a scenario analogous to Wrights

Model f, chiefdoms in the Basin of Mexicoadapt to the humid plains of their lakesystem, export obsidian, and excel atcarved pottery and hollow dolls. Thechiefdom at Chalcatzingo builds damsand irrigation canals, e xports w hite kaolinclay, and excels at carving bas-reliefs onliving rock. Chiefdoms in the Valley ofOaxaca deve lop canal and we ll irrigation,

export magnetite mirrors, and becomeprecocious in the use of adobe, stucco, andstone masonry. The Olmec chiefdomsfarm river levees, excel at three-dimen-sional sculpture, and build colored claymounds. It is the adaptive autonomy andfrequent competitive interaction of suchchiefdoms that speed up evolution andeventually make useful technologies and

sociopolitical strategies available to all re-gions.

The anachronistic notion of Olmec im-perialism spreading outward from theGulf Coast, colonizing highland culturesand converting them to foreign ways, issimply not supported by the archaeologi-cal record. Set aside the fact that theOlmec had no expertise at arid highland

agriculture. Set aside the fact that theOlmec lacked most of the Earth/Earth-quake symbols so important in highlandcosmology. Conside r only that the Olme c-centric modellike Wrights Modeldleads to homogeneity, which wouldactually slow down evolution. It would doso (1) by eliminating autonomy, whichhad allowed each region to produce the

innovations best fi tting its natural and so-ciocultural setting, and (2) by reducingcompe tition among regions, one o f the e n-gines that drove social evolution.

We recommend removing the MotherCulture from life-support and laying herto rest beside the Kulturkreise and 19th-

century diffusionism. The Mother Cul-tures dwindling band of defenders thinkthat if they can just inflate their estimatesof Olmec site sizes, find more basalt mon-uments, and convince us that San Lorenzo

was shaped like a cosmic bird, wellchange our minds. But we have all seenTeotihuacanthat quintesse ntial exam-ple of Classic Mesoamerican civilizationan d it is n ot a co lle ctio n o f e arth enmounds, colored clay platforms, and co-lossal heads. It is made of adobes, ston

El Mito de La Cultura Madre FLANNERY

Documents

Transcript of El Mito de La Cultura Madre FLANNERY