According to John Clark (2003: 34), the key concept in trade models is that "Trade has social evolutionary consequences. Without this foundational postulate nothing else matters". While obsidian studies in the Andes can also provide important insights on historical subjects such as the early peopling of the Andean highlands, and the changes in long distance interaction due to camelid caravans, the study of ancient exchange in the south-central Andes has direct bearing on exploring the basis and development of social and political power. Beginning from the assumption that long-distance exchange of products like obsidian in Late Archaic foraging contexts was based primarily on down-the-line exchange mechanisms, it is possible to develop inferences about the contexts of early caravan transport. Drawing on earlier discussions in chapters 2 and 3, the theoretical implications of early long-distance caravan exchange will be considered in terms of three principal variables: (1) herd size and status, (2) the scale of interaction, and (3) competitive behavior and exchange.
The domestication of camelids is not sufficient cause for the development of long-distance caravan traditions. Ethnographic evidence suggests that caravans are organized by herders with sufficiently large herds and the economic stability to depart on a caravan journey for weeks or months. Herders have high labor productivity and scheduling that permits important members of the household to depart on long voyages (Section 3.2.3). Households that conduct long-distance caravans often have a herd that is large enough to contain a viable number of caravan animals, or they will lease animals from others in order to have sufficient camelids to make a journey worthwhile (Browman 1990: 400;Nielsen 2000: 387-388).
Pastoral wealth is primarily manifested in larger herd sizes, where wealthy herders in modern contexts will own hundreds or thousands of animals in the south-central Andes. While an ethos against accumulation may have existed in some contexts, for example among egalitarian foragers, pastoralists worldwide display a similar logic of accumulation with respect to herd size. "In pastoral societies, having more animals is better than having fewer. Because the pastoral economy is based on a continuous demand for successful animal reproduction to replace losses due to predation, disease and use, the accumulation per se of more animals is a socially sanctioned and desirable end in itself." (Aldenderfer 2001: 25). In a human behavior ecology framework, Aldenderfer (2006) suggests that the origins of pastoralism may have its roots in the costly-signaling value of the possession of large herds. Thus, herd size is linked to status and it represents one of the factors contributing to the ability of caravan drivers to organize long distance caravans, and, in turn, to acquire non-local goods as part of caravan activities.
While long-distance interaction through reciprocity networks and high mobility foraging had long existed, it is likely that, along with the development of long-distance caravan traditions commenced a period of unprecedented regularity in interaction over distance. Building on the argument that due to the importance of exotic goods, those who possess such goods, and the ability to transcend regional boundaries as discussed in Section 2.2.2 (Appadurai 1986: 33;Helms 1992: 159), caravan drivers may have been among the early agents able to traverse regional boundaries in a context of increased circumscription. From the perspective of "social distance" using Sahlins' (1972) terms, regular caravan traffic represents a change in the scale of integration extending the "tribal sector" (balanced reciprocity), as well as augmenting the interaction with the "intertribal sector" (balanced or negative reciprocity). Changes are also expected in distance-dependent formulas such as Hayden's (1998) "prestige technology" where goods that were formerly laborious to transport in quantity became increasingly available resulting, over the long term, in a decline in the exotic associations for goods like obsidian.
There is little detailed evidence with which to evaluate the social position of early caravan drivers, but one should consider the socially integrating role of caravans over the larger region (Dillehay and Nuñez 1988), as well as the status and influence that may be accorded such caravan drivers. Itinerant peddlers, roaming herbalists, and other travelers are also common in the south-central Andes, and these persons could have transported goods widely. Furthermore human bearers are well documented in the Andes and in contexts with tight and perhaps exploitative control of labor, human porters can be organized into extremely efficient transport teams. Regular caravan transport is, however, a distinctive development because caravans represent a relatively low-cost mode of transport that is associated with staples and bulk items (Nielsen 2000: 514) and to the extent that caravan transport became linked to key economic cycles like agricultural production, caravans became recurrent and perhaps anticipated along these regular routes.
Changes in the scale of interaction were linked to larger economic transformations that included small-scale food production and the creation of surpluses, widespread technological change, early social inequalities, and greater sedentism; all processes that potentially contributed to the importance of caravan linkages. The integrating role of early caravans, and the position and status accorded to caravan drivers in this process, deserves further consideration.
Evidence of early social differentiation in the south-central Andean highlands take the form of public architecture and ritual features that occur in the latter part of the Late Archaic at Asana during the Qhuna Phase, 3800-3300 BCE (Aldenderfer 1998: 243-261) and exotic grave goods by the Terminal Archaic (circa 2900 BCE) at Jiskairumoko (Craig 2005: 570). While long-distance exchange of singular items, such as a finely knapped projectile tip of non-local material, are well-documented during the Archaic Forager period, these are arguably examples of 'tokens' used to reinforce social relationships across space through buffering and mutualism (Aldenderfer 1998: 301-302;Brown 1985;Spielmann 1986). At the end of the Late Archaic and into the Terminal Archaic there is evidence of a greater emphasis on creating surpluses and perhaps the first evidence of simultaneous or sequential hierarchy (Johnson 1982). The process of increasing complexity was not necessarily linear as, for example, at Asana in the Awati phase (the Terminal Archaic) where there appears to have been simpler social organization at the site focused on herding (Aldenderfer 1998: 275;Kuznar 1990).
What prompted this emergence of competitive behavior discussed above that perhaps lies at the root of the obsidian intensification witnessed at the Chivay source in the Terminal Archaic? These changes could have roots in gradual processes through the Late Archaic that include domestication, low-level food production, and regional packing resulting from population growth. The frictions that develop from circumscription and reduced mobility are argued to have prompted the emergence of more complex social forms. Alternately, highland complexity could have been prompted by the development of complex behaviors on the coast. On the Pacific littoral, albeit 700 to 1000 km north of what is now the Department of Lima, monumental mound building is documented for the period described here as the Terminal Archaic that is argued to be based on surpluses generated from the bounty of the sea, combined with the production of cotton for nets (Moseley 1975). Research in the Norte Chico area shows that during the third millennia BCE monumental construction at coastal sites like Aspero shifted to inland complexes (Haas, et al. 2004;Shady Solis, et al. 2001), perhaps reflecting the growing importance of inland cotton production for the construction of anchovy nets and also for textile trade. Coastal complexity based on cotton surplus may have stimulated highland trade in the production of a surplus in woolen textiles (Aldenderfer 1999: 218-219), as there is evidence for other highland and Amazon basin goods, such as tropical feathers, at preceramic sites on the central coast of Peru (Quilter 1991). While preservation problems in the highlands destroy evidence of preceramic textiles, the consumption of highland and Amazonian products is documented in coastal sites as far south as the Chinchorro III (2500 - 600 BCE) period in northern Chile (Rivera 1991: 15). Cultigens are also shared over a broad region, suggesting that interregional exchange was extensive during the Early and Middle Formative.
Was the intensified production that was observed at the Chivay source, and the regular consumption of obsidian at Qillqatani, a reflection of aggrandizing behaviors? The timing of the increased production and circulation of obsidian would suggest that it was, above all, a surplus productionenterprise that was a reflection of the changing social and economic context of the Terminal Archaic and Early Formative. The evidence from this research shows that the quarrying at the Q02-2 pit was best correlated with consistent acquisition of large obsidian nodules that were probably in excess of 15cm long, yet the finished tools produced during the Terminal Archaic and onward were very small projectile points, often less than 3 cm in length. Thus, the acquisition of large nodules was not out of necessity based on intended tool forms, but probably for some other signaling value which highlights the relevance of the abundance concept. The perception of surplus would have been enhanced through supply lines as large nodules were consumed. The willingness to excavate for obsidian when there was an abundance of smaller nodules available on the surface, underscores the importance of large nodules to ancient obsidian consumers.
The possession of non-local goods are well documented attributes of aggrandizing behaviors in transegalitarian contexts (Clark and Blake 1994;Hayden 1998), yet it is not necessarily an overt expression of status that one should expect in early stages of social differentiation. Instead, the possession and use of irregularly distributed material, which includes obsidian but also might include salt, herbs, and other semi-rare goods, is a visible testament to the regional integration afforded by emerging caravan networks. Thus, while most individual obsidian artifacts may not have been "wealth items" the regular possession of a surplus of scarce "ordinary goods" of cultural value (Smith 1999) had the effect of imparting prestige on the caravan drivers who amassed and circulated these goods. As these exchange networks matured, the availability of non-local goods probably changed dramatically and, as noted above, the scale of what was considered "exotic" probably diminished as well. In that sense, obsidian production and circulation was a reflection of a transegalitarian socio-political context where an individual's successful participation in the regional economy reflected well upon him or her, and that success likely included acquisition of non-local products and manipulation of surpluses.