The questions addressed in the research involve the diachronic and regional political organizational
changes occurred as a consequence of the relationship between native local societies and the Tiwanaku polity.
These questions aim to identify the complex processes involving the objectives, demands, and political strategies
of two interacting polities. A set of more specific questions related directly to the material correlates is presented
The Tiwanaku polity developed in the Titicaca Lake altiplano in the South-Central Andes (ca. A.D.
300-1100; Figs. 1 and 2). As a consequence of unidentified processes Tiwanaku ceramic sherds and other
artifacts are widely distributed in the South-Central Andean region. This distribution is interpreted as a process
of territorial expansion (ca. A.D. 350) towards mesothermal Andean valleys, but such process has never been
The research will explore what the distribution of Tiwanaku materials represent by identifying changes
in the political organization of settlements of the Valleys of Cochabamba. Associated with the appearance of
Tiwanaku-style materials, these changes will be documented for the transitional period prior and subsequent to
the influence by the Tiwanaku polity. Changes will be perceived through the identification and comparison of
the distributional patterns of Tiwanaku artifacts (i.e. settlement patterns, architecture, and ceramics) in two
systematically surveyed areas of the region located in different ecozones.
Theoretical Considerations. Analysis of expansion of Andean imperial polities has concentrated on
the Wari (Schreiber 1992) and the Inca polities (D'Altroy 1992; Morris 1988). These analysis agree that the
interaction strategies of Andean polities depend on the local political and economic conditions of the provinces.
The interaction features correspond to Complementarity mechanisms (Salomon 1985). Complementarity is the
articulation of productive zones located in different ecozones and at varied distances. Its variants correspond to
cases with different degrees of centralized control and number of political units involved.
The five alternatives considered as modes of interaction with the Tiwanaku polity are: (1) Hegemonic
strategy; (2) Territorial strategy (Luttwak 1976 in D'Altroy 1987); (3) Archipelago control strategy; (4) Prestige-
good economy; and (5) Simple exchange. These alternatives enlist different means of organization, and therefore
the possibility exists to find different shifts and changes associated with them (e.g. land production areas). Their
differences are expected to be documented in the empirical record. The hegemonic strategy is
implemented where pre-existing political hierarchies are sufficiently centralized to fill the administrative role for
the conquering polity. Native resistance or acceptance of imperial rule is conditioned by provincial political
conditions; if competing polities exist, alliances may bring benefits to a particular polity in detriment of others.
It is an indirect, low-control, and low-extraction resource system. It implies a superficial presence of the
dominant polity and rather a domination through client-elites. Native elites control the exploitation of resources
through tribute burden for the benefit of the dominant polity. This strategy creates a prestige-good economy of
heavily dependent native local elites.
In contrast, the territorial strategy entails a direct occupation and rule of the provincial region. It will
be implemented where native local political fragmentation impedes control, or where native groups are resistant
to the conqueror's domination; it is then able to generate warfare. It is a high-control, high-extraction strategy
where the dominant polity is in charge of the exploitation of resources. This strategy implies the presence of the
dominating polity elites concentrated in one or several major administrative settlements, and a continuous control
of the region.
In third place, the Archipelago control strategy (Murra 1972) is characterized by direct control of niches
in different ecozones in a discontinuous territory, hence dispersed regional control, and high extraction of
resources in those niches. This altiplano controlled centralized type of Complementarity,
documented ethnohistorically for pre-contact societies, corresponds to an organization by large states with
complex administrative implications (Dillehay 1979).
The fourth alternative corresponds to the creation of a prestige-good economy in the hands of native
elites. The access to multiple resources occurs through multiple alliances in a dynamic of central administered
trade (Salomon 1985). The trade is accomplished with no direct control of enclaves but with alliances established
with independent craft specialized polities, localized at the margins of a large polity. Moved also by mechanisms
of alliances, this alternative is extremely similar in provincial contexts to a hegemonic strategy of control. The
differences of both prestige-good economies reside in the independent relationship the latter one has. Additionally,
Tiwanaku materials will be differently channeled into the hierarchical levels of the provincial economy reflecting
the heterogeneity in the political hierarchy and differential access to goods. The amount and types of exotic goods
the specialized polities could have controlled would have depended on the amount and importance of the products
they traded. The establishment of new trade networks generated by the Tiwanaku demands would have occurred
in this alternative.
The fifth alternative is the existence of simple trade between the dominant polity and the native local
polities. Simple transactional relations might not have created regional prestige economies, given the non-
existence of alliances for resource procurement. Trade might have corresponded to transactions made at the
household level. In this alternative, Tiwanaku might not have controlled the networks of distribution, using pre-
existing channels for movement of goods.
A set of more specific questions to direct the analysis of the material correlates of the presented
alternative strategies are: (1) What are the changes in the settlement patterns, and the specific differences the Tiwanaku polity implied in
terms of hierarchy, types and location of sites, and the land-use of the region?;
(2) Are there material differences to distinguish elites from common groups? and Tiwanaku people from native
(3) Is there craft-specialization in the region?, and what is the evidence for an increase in the specialization of
the native local sites to supply the Tiwanaku polity?;
(4) How do the Tiwanaku goods in the Cochabamba region compare to assemblages described for other regions
with Tiwanaku artifacts?
Archaeological correlates of the theoretical alternatives. In the hegemonic strategy alternative, the
indirect rule, the "conqueror" facilities are either absent or integrated in pre-existing administrative native local
centers. The control lies in the hands native officials acting in name of the Tiwanaku polity. There should not
be exclusive Tiwanaku habitational areas in the already existing sites; Tiwanaku materials will have a small
proportion and an exclusive range of ceramics types, and will be combined with native ceramics. The endorsed
control to the native elite should create a prestige economy correlated to the productive or strategic aspects of
one or various local polities.
The second alternative, the direct control territorial strategy, would be manifested in the archaeological
record by the relocation of sites and population shifts, and new administrative Tiwanaku sites. These sites may
be identified by the architectural layout (Goldstein 1990), and construction techniques. The ceramics would
exhibit specific surface treatments, designs, and forms typical of the Tiwanaku nuclear area; and native local
ceramics would have influence of Tiwanaku forms and decorative elements. Tiwanaku ceramics would be found
in the context of state administrative facilities, and in a limited number and contexts in non-Tiwanaku sites.
The Archipelago control alternative will be reflected in Tiwanaku additions to native local villages,
forming multiethnic settlements, or less complex and exclusive habitational structures
in new settlements. It implies the permanent presence of altiplano populations in restricted colony sites of the
region. These sites would be correlated with a higher proportion of pure Tiwanaku ceramics, or pots
with local clays with a consistent Tiwanaku iconography, reflecting the origin of the potters. A large
proportion and types of pure Tiwanaku artifacts would reflect a continuous relationship maintained with the core
area. Differences in the range of ceramic types and of contexts would indicate the presence of different Tiwanaku
hierarchical levels in the colonies. Exclusive storage areas would reflect the management of resources for
distribution in other areas of the polity.
The fourth alternative of a trade model between the Tiwanaku polity and specialized centers at its
margins would be reflected by the presence of Tiwanaku artifacts in restricted areas of the region, where the
accumulation by a native controlling elite was made. This elite establishes the existence of a prestige-good
economy. This alternative would principally have to find the existence of such specialized centers and document
their origins, and their evidence for transactions with Tiwanaku. This prestige economy would vary within the
region depending on the amount and type of goods that were exchanged by each independent polity. This would
be reflected in polities controlling and accumulating altiplano goods in different proportions. Mujica et al. (1983)
have suggested a multi-ethnic port-of-trade settlement for the exchange.
The fifth and last alternative is of a long distance simple trade economy would leave dispersed evidence
of Tiwanaku goods but in lesser proportions and less circumscribed contexts, with no elite in control. In fact
there would be an exponential decrease in the frequencies with the distance from the center of Tiwanaku.
Similarly, this and the fourth alternative would be supported based on the degree of presence of artifacts from
other regions that would indicate a non-exclusivity to exchange with the Tiwanaku polity.
All these alternatives might have been related to the transaction of the same goods, but under different
kinds of administration. Hence they would produce different proportions and distribution in the archaeological
record. The altiplano products that may have been brought to the Cochabamba area are wool, Tiwanaku fine
pottery, and ritual paraphernalia. The mesothermal products that may have been supplied to the altiplano are
maize, coca leaves, fruits, and perhaps minerals.
Preliminary research and location of areas to investigate. In the summer of 1992, a reconnaissance
field season was undertaken in order to select the two areas (of 35 kmý each) to conduct the proposed research.
The first area is an narrow alluvial valley located at an elevation of 3200 m. with a transitional ecology
combining mesothermal and altiplano cultigens (e.g. maize and grains) and in a crossroads location to the
highlands; the second area is located at 2500 m. in a valley arm that leads to the lowlands.
Field strategy. The systematic survey of each of the two areas, the mapping and surface pottery
collection of the sites, and limited number of test pits in selected sites will be conducted in order to obtain
evidence indicative of Tiwanaku-native local relations, and test the above alternatives. These tasks will allow
us to establish the spatial and temporal distribution of Tiwanaku artifacts in each of the two areas.
Preliminary site location and maps of visible sites based on air photos will be prepared. Information will be
obtained to map the agricultural soil properties, the clay sources, and the ecological diversity of the areas.
Observations of specific architectural features or subsurface features visible will be noted for the determination
of types of sites.
The record of location and size of sites in each area is critical to perceive the magnitude of Tiwanaku
influence through the following evidence: demographic shifts in the region; the change in site locations in the two
periods concerned; the clustering of sites at selected altitudes; their catchment areas and control of resources;
the distribution of types of sites; the spatial relation of Tiwanaku and native local sites; and the settlement
hierarchical organization in the two areas.
The surface sherd will be conducted for pre-determined site sectors using a sampling strategy
which ensures uniform coverage. Ceramics will be analyzed to determine the relative distribution of Tiwanaku
and local wares, and the probable function of sites of similar and different cultural affiliation. In
a third stage a limited number of test pits (16 pits) will be conducted in midden areas of large Tiwanaku and
native local sites. The 1x1 m. test pits will permit (1) the analysis of surface-subsurface artifact relations in order
to evaluate the reliance on surface remains to interpret site sizes and population densities; (2) the establishment
of the chronological sequence and the Tiwanaku-local polities relations; (3) the gathering of soil samples in order
to establish the first corpus on heavy and light botanical remains of the area and initially address the issue of
exploitation of resources in the two areas; and (4) the intersite and interegional comparisons of the artifact
Schedule. The research project is programmed from June 1993 to May 1994:
1) 15 June-7 July: Cochabamba, Administrative details; study of aerial photography and preparation of the
2) 8 July-11 November: Survey, mapping of sites and surface collection in the area of Independencia; laboratory
analysis of collected data in order to define the location of the test pits in midden areas;
3) 15 November-21 March: the same tasks in Aiquile as in previous area;
4) 25 March-15 June: Final laboratory analysis in Cochabamba: artifact processing, flotation of soil samples.
The project was discussed with both Cochabamba and Bolivian authorities. Bureaucratic and logistic
arrangements, including permission request for fieldwork in the 1993 season, were initiated.
3. Significance of the research
Previous research on the Tiwanaku polity. The Tiwanaku interaction and relationships with the
Valleys of Cochabamba is identified based on the presence of burial contexts and Tiwanaku sherds. Stratigraphic
data from excavations in the region have revealed different temporal relations between Tiwanaku evidence and
local traditions; their significance, however, has not been analyzed.
This research questions the explanations proposed for Tiwanaku artifacts dispersion. The few provincial
regional descriptions of Tiwanaku evidence impedes to define the polity as an empire, or as an expansionist state.
Similarly, it is speculative to fix Tiwanaku territorial limits and consider it as a"controlled" territory. The
Tiwanaku territory is described to be integrated by a trading network with camelid caravans (Browman 1984).
Sites in different territories were defined as colonies representing direct control by Tiwanaku (Goldstein 1990).
The Tiwanaku State has been defined based exclusively on features of its nuclear area: the large "capital" with
monumental buildings, the extensive agricultural fields (Kolata 1986), and the settlement hierarchy in the
Contributions of this research. The proposed research in the Valleys of Cochabamba will make the
following contributions to Central Andean archaeology: (1) Establish a ceramic chronology the Cochabamba
cultural development; (2) Produce a complete inventory of Cochabamba's archaeological remains; (3) Present
a valuable case study of provincial interaction of the Tiwanaku polity at a regional level; (4) Set the first
problem-oriented systematic archaeological investigation in the region; and (5) Analyze process of Tiwanaku
political interaction through testing of five alternatives of interaction, each with its respective archaeological
This research, by providing a diachronic perspective on the interaction of a large-scale polity and small-
scale polities for resource exploitation purposes, will be of interest to a broad range of scholars in the Andes and
elsewhere. It will provide new insights in the processes of adaptation and cultural evolution in areas marginal
to the nuclear development of large-scale polities.
I consider that a major contribution of the research is the analytical strategy. Rather than assuming a
specific type of interaction in Cochabamba, this research has proposed five different political organizational
alternatives of interaction and will tests how that interaction might have been structured. These alternatives will
be examined through a particular set of archaeological correlates. The archaeological data will be used to support
one --or none-- of these alternatives.