Archaeologists working in South America are familiar with Steward’s ambitious division of the continent into four parts in the Handbook of South American Indians: Marginal Tribes, Tropical Forest Tribes, Circum-Caribbean Tribes, and Andean civilization (Figure 47.1a; see Chapter 1 in this volume). Later he worked this into: Andean civilization, theocratic and militaristic chiefdoms (circum-Caribbean-like; largely wedged up against the Andes), tropical forest tribes, and marginals. Steward’s quadripartite division was immediately criticized (Figure 47.1b; see Murdock 1951; Steward’s characterization more or less recapitulates Wissler’s [1917] “Amazon culture area” and Cooper’s [1942] “Silval culture”), but the idea of a generalized tropical forest culture (hereafter TFC) has dominated, in spite of evidence for significant diversity. By systematically comparing nine characteristics, Murdock (1951) described 24 areas, with over half in Amazonia (see also Galvao 1967, for a similarly complex rendering of culture areas in Brazil in the twentieth century).
The first attempts to characterize Amazonia culturally were linguistic, for instance Gilij (1780-84), Martius (1867 [1838]), and Steinen (1886), which sought to divide South American languages in phylogenetic terms and locate origin areas for the major groups. The most updated general overview is Dixon and Aihkenvald’s Handbook of Amazonian Languages (1999), but descriptive and historical linguistics are still quite preliminary in Amazonia (Figure 47.2). Several general patterns emerge: (1) most languages are isolates; (2) small language families, largely restricted geographically, have expanded in upland areas; and (3) two large linguistic diaspora—similar to tropical diaspora in the Pacific (Oceanic Austronesian) and Africa (Niger-Congo)—are spread across the entire region (from the Caribbean in the north to the La Plata in the south and the eastern Andean foothills to mouth of the Amazon): Arawakan and Tupi-Guarani (T-G) families (see Chapter 33 in this volume.
Figure 47.1. a) Steward's Cultural Areas (1945).
The early German historicists were the first to draw attention to the Arawak and T-G families as diaspora (although not their word for it) (see Schmidt 1917; Steinen 1886), which has been further developed by several archaeologists, most notably Lathrap (1970). With respect to the Arawak in particular, Lathrap recapitulated Schmidt’s contention that “developed” tropical forest agriculture and navigation propelled Arawak peoples across the
Figure 47.1. (Continued) b) Murdock's Culture Areas (1951).
Lowlands. He was a proponent, however, of a generalized tropical forest culture, which was then enhanced or inhibited in its development by ecological factors, following the expectations of cultural ecology (a view shared by the other major players: Robert Carneiro, William Denevan, Betty Meggers, and Anna Roosevelt). I explicitly adopted aspects of Schmidt’s formulation, but focusing on settlement pattern and landscape, I explored questions of
Figure 47.2. Schematic linguistic map of South American language families and two primary linguistic diaspora: Arawak (A) and TupiGuarani (TG), showing suggested origin areas (hatched).
Social life, namely hierarchy, regionality, and settled life in plaza villages, the anthropogenic landscape, and the hybridizing nature of the diaspora as the foci (Heckenberger 1996, 2001, 2002). One thing is critical: all commentators who have looked at the family from a comparative perspective agree that the Arawak diaspora refers to something real (as noted at a recent conference on Arawakan peoples, see Hill and Santos-Granero 2002).
If the Arawak diaspora was characterized by hierarchy, regional integration, and settled life, even during its “formative” period (when it first dispersed, 500 BC to AD 500), the T-G were, as often as not, an alter-ego to this. Viveiros de Castro (1984, 1992) has outlined basic fundamental issues of a T-G historical identity: other-wordly, mobile, predatory, including the fairly large, heterarchical T-G polities of coastal Brazil, the Paraguai River, and the southern Amazon (e. g., the Pacajas of the lower Tocantins and Para rivers). Viveiros de Castro (1992: 5) compares T-G (and Carib) to Ge, pointing to the sociological locus of cultural energies in these dialectical societies, as opposed to the cosmo-vision of the Amazonians (see also Descola 2001). Not surprisingly, these macro-cultural traditions have material and spatial features that are highly visible in the archaeological record, notably ceramic artifacts and settlement patterns.
Following much debate (summarized in Noelli 1998 and see Chapter 33 in this volume), it seems almost certain that Rondonia, more or less, is the center of dispersal for Tupian languages, and the best evidence to date suggests that T-G separated from somewhere close by to the east or south, and more likely the former (Urban 1997; Viveiros de Castro 1997). T-G languages are mainly distributed in four places: (1) southeastern Amazonia (Xingu-Tocantins); (2) southern Amazonia (Tapajos and Madeira); (2) the Paraguai River (Guarani); and coastal Brazil (coastal Tupi). Migrations are well documented historically. In the southern Amazon, T-G languages are interspersed with Mundurucu and Juruna, among other Tupian families for a near total domination by macro-Tupi languages, except—it is important to mention—along the southern peripheries, which are dominated by Arawakan and related groups (in the west) and macro-Ge (here one might note an interesting correlation between macro-Ge, but not non-Ge family groups, like Karaja, Bororo, Umosina, and Rikbatsa).
The Tupiguarani archaeological tradition, although distinguished from T-G languages by the missing hyphen, is largely co-extensive with the T-G language family, including its dominant presence in much of the southern Amazon, coastal Brazil, and in the Paraguai River, with occasional appearances in central Brazil and eastern Bolivia. This tradition, most clearly hallmarked by its thin line bi-chrome painting (red on white slip), body corrugation, and thin-walled vessel forms, is found in various contexts and relates to the T-G diaspora, ca. 2500 to 1500 BP.
In most cases, hybrid archaeological complexes (if not pluri-ethnic social formations) resulted from cultural interactions between local and immigrant groups. The most notable example of this is the Amazonian Polychrome (see Chapter 20), which takes elements of Bar-rancoid, Tupiguarani, and other (Konduri) complexes and reconstitutes them in the political economy of the Amazonian varzea, the most obvious example of an integrated supra-regional organization, or small “world system” in the lowlands due to its focus on durable prestige goods, especially ceramics. However, during the early diaspora, the time of rapid and extensive population movements, there were clear correlations of language groups and ceramic styles, most notably tied to these two large diaspora (Arawak, characterized by “Barrancoid” ceramics and T-G characterized by Tupiguarani ceramics). Lathrap (1970) recognized this basic relationship, but viewed these largely contemporaneous sloping traditions, which both extend from over 2000 BP to the present in recognizable form, as temporal strata: Barrancoid was early, ca. 200 BC (based on his work on Hupa-iya materials in the Ucayali, and Rouse and Cruxent’s (1961) work at the type sites in the middle-lower Orinoco) (see Chapter 23 in this volume), while T-G was late (Caimito in the Ucayali).
One thing that recent work on large-scale ethno-linguistic entities reiterates is the fundamental difference between riverine and coastal as opposed to upland settings, ecologically, socially, and in terms of community formation, and ethnographic groupings
Figure 47.3. Formative supra-regional systems in South America, including (1) varzea; (2) southern Amazon; (3) Moxos/Chaco; (4) eastern Brazil; (5) Orinocan;(6) northern Andes; (7) Central Andes; and (8) southern Andes.
Favor the uplands (Heckenberger 2006) (Figure 47.3). Two final implications are important to mention, before describing these better known societies in the Brazilian Amazon: one is that “refugia” may relate to anthropogenic processes and second, dynamic language change may be quite different in settled areas than low density areas, and relations between languages, which for instance posit a steady rate of change or dispersal, may overlook important punctuated change, such as cultural contact, or trade-languages (e. g., Moxos [Arawak], Kokama [T-G hybrid], or Nheengatu [“lingua franca,” T-G]).
World History
