Archaeological research conducted in the Lower Amazon demonstrates that this region was populated by one of the largest Amazonian polities during late precolonial times. The Tapajo people, described in contact chronicles, occupied a large settlement at the mouth of the Tapajos River called the Santarem site by archaeologists. Three models linked to specific theoretical frameworks and ideological aspirations have been formulated by archaeologists in order to explain the socio-political organization of the Tapajo. These models rely on descriptions provided by contact chronicles, archaeological data coming from sites located at the mouth of the Tapajos River, especially ceremonial contexts and ceramics. However, recent surveys and excavations in the upland Belterra plateau provide new data on ceremonial and domestic contexts and call for a re-evaluation of these models of social complexity. This dissertation compares riverine and upland settlements in order to challenge previous models of social complexity and formulate a new model that looks at the Tapajo occupation at a broader regional scale. I use seven variables in order to investigate the relationship between upland settlements and well-known riverine settlements: 1) environment and subsistence, 2) chronology, 3) settlement size and demography, 4) domestic and ceremonial contexts (i.e. domestic house floors, fire pits, buried vessels, and refuse pits), 5) domestic ceramics, 6) ceremonial ceramics (i.e. vessels and figurines), and 7) lithic industries. Furthermore, this work provides new data on daily-life practices at the Cedro site, located on the Belterra plateau 30 km away from the mouth of the Tapajos River. Based on a technofunctional analysis of 1,424 sherds, my research uses a chaine operatoire approach in order to look at all the stages of pottery making, use, and discard. This anthropological approach is used for assessing bodily and daily-life practices as well as social contexts. It contrasts with previous ceramic analyses conducted in the Brazilian Amazon which were based on classifications inherited from culture history, equating decoration with pottery style as well as variation in style with ethnic groups. Based on results of the Cedro site ceramic analysis and the comparison between riverine and upland sites, I argue that the Santarem and Cedro sites were organized as primary and secondary centers and maintained a heterarchical relationship during the late precolonial period.