Tiwanaku dominance, state formation

Around 300 A.D., the beginning of the LF2, the lake rose again to modern levels or even somewhat higher. Lake level was to remain high for the next 800 years. This lake level rise coincided approximately with the collapse of the Pukara polity in the northern Titicaca Basin. Thus, two things happened at approximately the same time: 1) the yungas trade route shifted so that it once more passed through Desaguadero and the Tiwanaku Valley, and 2) any suppression of trade between the southern and western regions of the Titicaca Basin on the part of the Pukara polity abruptly ceased.

These lake level shifts and the collapse of Pukara may or may not have been related, but their combined effect seems to have been dramatic. Together, they would have produced a situation in which the site of Tiwanaku controlled the route of an expanded southern exchange system. If we continue in the explanatory vein established in the preceding sections, we could suppose that this strategic position with regard to an expanded volume of traded goods gave the leaders of Tiwanaku a distinct advantage over those of neighboring polities. That is to say, they became capable of collecting and - to say the same thing - redistributing a much greater quantity of both exotic and staple products than their neighbors. The expanded volume of both staple and exotic goods under their control also would have permitted the leadership of Tiwanaku to greatly expand the scale of the public ceremonialism they sponsored. As a result, Tiwanaku began to attract immigration from adjacent areas. This process is almost certainly responsible at least in part for its greatly increased rate of population growth in the LF2. Tiwanaku grew by as much as 1% annually during this period, an order of magnitude faster than what we have established as a normal rate of increase.

At the same time, the population of the Taraco Peninsula actually declined, the first time this had happened in the Formative period. Other nearby areas such as the Katari Basin and the Khonko Wankane area may have experienced depopulation as well. However, the high population density of the Taraco Peninsula in the LF1 meant that it was the largest population donor, probably accounting for almost 50% of Tiwanaku's LF2 expansion.

During the LF2, Tiwanaku grew to cover about 1 km$^{2}$, probably with a population of more than five thousand. Thus, by the end of the Formative period - ca. 500 A.D. - it was by far the largest site in the Titicaca Basin, and had probably become capable of dominating neighboring populations politically, economically and militarily. The most interesting fact about its early growth, however, is that it probably took place without resort to coercion. When the large-scale population movement into Tiwanaku began, the Tiwanaku Polity probably had a total population well below half that of the Taraco Peninsula Polity, and was probably not its military equal. That is, Tiwanaku regional dominance was an effect rather than a cause of the population movement we have been discussing.

I have just attempted to explain the remarkable LF2 growth of Tiwanaku as a result of two significant changes in the Titicaca Basin environment: 1) a rise in lake level which shifted the route of the southern yungas trade through Tiwanaku and away from the Taraco Peninsula, and 2) an expansion in the activity of this southerly route due to the collapse of the Pukara Polity which hypothetically suppressed external trade in its period of primacy (see my discussion of this possibility in Chapter 7). Can this really explain the unprecedented growth of Tiwanaku? The answer is not clear. After all, the trade route had shifted before on a number of occasions, but never with an effect this dramatic. It is certainly possible that the greater Late Formative population density in the southern basin, the western basin, and presumably in the yungas as well propelled the volume of exchange passing through the southern route beyond some as-yet untheorized critical threshold. However, we should also consider that some new process might have been at work, in combination with those we have just described. I would like to propose that this new process was the incorporation of raised field agriculture into the political economies of Titicaca Basin polities, and of the Tiwanaku Polity specifically.

There are various interpretations as to the manner in which raised field agriculture was integrated into the Tiwanaku political economy. The majority view is that of Alan Kolata, Clark Erickson, and their colleagues, which I have termed the hyperproductivity hypothesis ([Erickson 1985,Erickson and Candler 1989,Kolata 1986,Kolata 1991,Kolata and Ortloff 1989]). These scholars hold that certain properties of raised fields in the Titicaca Basin environment make possible very high yields and short - even zero - fallow periods relative to the dryland techniques employed in the area both ethnohistorically and historically. The most important of these properties are thought to be the use of canal sediments as fertilizer, preventing soil nutrient depletion, and thermal effects of the water in the canals, which provides some protection against frost. My own view ([Bandy 1999c,Bandy 2000]) is what I call the staggered production cycle hypothesis. As I conceive it, the importance of raised field agriculture was not its supposed hyperproductivity but rather the fact that splash irrigation using the water in the canals, perhaps combined with their frost-ameliorating thermal properties, made it possible to cultivate raised fields on an annual agricultural cycle which was offset by as much as two months from the normal dryland agricultural cycle.

This theory is derived from the work of Jürgen Golte, which I regard as absolutely fundamental for understanding ancient Andean political economies ([Golte 1980]). In his slim volume, Golte observes that in areas where the growing season is tightly circumscribed, as in the Andean highlands, a labor scheduling problem arises. In such areas, planting and harvest must take place at certain well-defined points in the annual cycle. There is very little room for variation, since to plant too early would mean that the crops would fail for want of rain, and to plant too late would not allow sufficient time for crops to mature before the onset of killing frosts. Since the most concentrated and intensive labor demands of agricultural production occur at the time of the harvest, a productive unit - a household, for example - may not plant more land than it is capable of harvesting within a short time period; a month, say. In such situations groups often develop mechanisms for spreading their subsistence labor more evenly throughout the year. The cultivation of winter wheat is an example of such a strategy from temperate Europe. Golte calls these `estrategías policíclicas' - polycyclic strategies.

The most fundamental and ancient example of this sort of strategy in the Andes is tuber agropastoralism itself. In this system, the timing of labor investment in the herds is offset from the crunch periods of the agricultural cycle. In this way labor that cannot be invested in agriculture is invested in herding, and vice versa. This results in a more complete and efficient use of labor throughout the year than would be possible in either a purely agricultural economy or in a purely pastoral one.

Another class of related examples are the agricultural systems of the highland valleys, with which Golte is primarily concerned. In these environments, producers practice a form of microverticality, exploiting the agricultural properties of land at varying elevations. In these different elevations, different crops may be planted, with differing growing seasons and labor requirements. In this way a whole series of staggered agricultural cycles are exploited, together with pastoralism in the puna grasslands, to spread productive labor more evenly throughout the year. This is a technique for maximizing labor efficiency on the household level.

In this connection, it is interesting to note that Golte singles out the Titicaca Basin as an area in which the possibilities for polycyclic labor maximization are very limited. This is because the circumlacustrine plain lacks the dramatic vertical relief of the valleys and is therefore unsuitable to microverticality-oriented agricultural strategies. Herding and fishing and more recently wage labor can make up for this to some extent. But, Golte points out, opportunities for maximizing the use of agricultural labor throughout the year remain extremely limited in the Titicaca Basin.

I would like to propose, however, that raised field agriculture represents exactly such a strategy for the staggering of agricultural cycles and the maximization of labor efficiency, and one specific to the Titicaca Basin. The key to this strategy would have been the water in the canals, which, as Erickson has noted, would permit splash irrigation of the fields ([Erickson and Candler 1989]) . This could have allowed planting to take place after the danger of frosts had passed, but still well before the onset of the rainy season. This practice has been observed by Erickson, and he concludes that splash irrigation of planting surfaces with water from the canals can be accomplished with very little effort. Indeed, in one drought year in Huatta, Peru, his raised fields produced good yields while surrounding dryland fields failed completely, due solely to the fact that the raised fields were splash irrigated as necessary ([Erickson and Candler 1989]).

The addition of such a new, offset agricultural cycle would have made it possible for leaders or rulers to exact agricultural labor from the population at large while minimizing the interference of this exaction with normal subsistence activities, and at the same time to put into action a vast labor potential that could not otherwise have been applied to agricultural production. This would have simultaneously 1) reduced conflicts between elite - and eventually state - labor demands and the subsistence interests of the populace and 2) permitted a significantly higher rate of annual agricultural labor extraction and therefore of annual surplus production.

Whatever the exact role raised field agriculture may have played in the Tiwanaku political economy, its centrality is generally accepted. The introduction of raised field agriculture would in any event have greatly increased the quantity of staple goods redistributed by leaders or rulers, and would thereby have permitted a tremendous expansion of existing systems of competitive ceremonialism. Such an expansion is exactly the kind of process which could stimulate immigration and population movement, as I have argued in reference to earlier time periods. Therefore, it is entirely possible that the LF2 growth of Tiwanaku was due not only to its increased importance in the southern Titicaca Basin exchange system, but also - and perhaps more importantly - to the integration of raised field agriculture into its political economy.

I propose that perhaps the crucial factor in Tiwanaku's early success was not its advantageous position viz a viz the yungas trade, nor, at least initially, was it the influence of its religious cult. Rather, the critical factor distinguishing Tiwanaku from other contemporaneous and competing centers was its proximity to the high density populations of the Taraco Peninsula. After the 300 A.D. lake level rise, the Taraco Peninsula communities found themselves in a landscape in which large-scale raised field agriculture was impossible. They were circumscribed by the saline waters of Lake Titicaca, and no appropriate marshy, freshwater landscapes were present within their borders. It is possible that during the second half of the LF1 raised field agriculture was practiced to some extent on the vast pampas that surrounded the Taraco Hills. If such activity did exist, however, it was brought to an end by the 300 A.D. lake level rise. From the beginning of the LF2 the Taraco Peninsula Polity was unable to participate in what seems to have been a general basin-wide increase in the importance of raised field agriculture during the Late Formative (see [Stanish 1994,Stanish 1999]). Its leaders were therefore at a competitive disadvantage relative to those of the Tiwanaku Polity.

Thus, as the importance of raised field agriculture in the Tiwanaku political economy increased, and as the associated social technologies of labor organization and surplus extraction were elaborated and improved, the inhabitants of Taraco Peninsula communities began to exhibit a marked preference for relocation to Tiwanaku ([Bandy 2000]). This, then, is a fact which distinguishes Tiwanaku from other Late Formative polities in the Titicaca Basin. The developing political economy of Tiwanaku had a unique resource: an adjacent area, densely populated whose inhabitants were unable to compete with the emerging Tiwanaku Polity on its own terms. That is, they could not build raised fields on their own land, and could therefore be incorporated into the expanding Tiwanaku political economy on a subordinate basis. No such geographical configuration existed in Juli-Pomata, and perhaps nowhere else in the Titicaca Basin.

The incorporation of the Taraco Peninsula populations into the Tiwanaku political economy seems to have taken place in two distinct stages. The earlier stage, which took place during the LF2 (300-500 A.D.), was characterized by population movement from the Taraco Peninsula into the expanding proto-urban center of Tiwanaku. During this period, the power and importance of Tiwanaku gradually waxed, while that of the Taraco Peninsula Polity waned. However, it does not seem as if the Taraco Peninsula fell under the direct political control of Tiwanaku until late in the LF2.

The later stage began with the political subjugation of the Taraco Peninsula Polity to the Tiwanaku Polity. In this stage the depopulation of the Taraco Peninsula communities ceased, and growth rates returned to normal or near-normal. However, we may assume that the large population of the Taraco Peninsula was still an important source of labor for the cultivation of the vast raised field complexes in the Tiwanaku Valley and the Pampa Koani, and ongoing monumental construction in the civic-ceremonial core of Tiwanaku. The Taraco Peninsula was at this point organized into a heartland province of the Tiwanaku state. By this time Tiwanaku had emerged as the dominant power in the Titicaca Basin and indeed in the entire South-Central Andes .

In this thesis I have identified several potential factors which contributed to Tiwanaku's emergence and eventual dominant position in South-Central Andean prehistory. In the final analysis, though, the principal factor which distinguished Tiwanaku from other Late Formative Titicaca Basin centers may not have been an attribute of Tiwanaku itself; it may simply have been that Tiwanaku was located adjacent to the Taraco Peninsula.

Thus, chance, physical geography and the longue dureé, human agency and especially residential decisions, and human genius and invention can all be seen to have played a role in Tiwanaku state formation. Tiwanaku state formation is here argued to have resulted from the rapid urbanization of the site of Tiwanaku during the LF2. This urbanization itself - in my view - resulted principally from the dramatic success, quite possibly unanticipated, of a political-economic strategy: the strategy of staggered production cycles.