Initial settlement, population growth

No evidence of an Archaic period or preceramic occupation was encountered in the survey. No aceramic sites nor even a single Archaic period projectile point were found. This agrees with the results of many seasons of excavations at Chiripa; though we have found Archaic period points, they seem to have been collected by the Formative period inhabitants; no preceramic levels were encountered, despite the fact that we excavated to sterile in a number of locations. This is probably due to the fact that Lago Wiñaymarka, was almost completely dry prior to 1500 B.C., and the peninsula was no doubt considerably colder and drier than it is today. It seems likely that whatever occupation there may have been of the area would have been located near the meandering streams on the vast pampas. If so, then they are all below the lake at the present time. It should also be mentioned that the entire spine of the Taraco Peninsula was an extensive lithic quarry throughout prehistory, and is thus covered by a low density lithic scatter. It is possible that the remains of short-term, ephemeral aceramic encampments have been lost in this ubiquitous background scatter.

At any rate, the Taraco Peninsula was colonized at the beginning of the Early Chiripa phase (EF1), beginning around 1500 B.C. when the lake rose suddenly to near-modern levels. What is interesting about this colonization is that people were living from the very beginning in nucleated settlements; small villages, with population index values typically between 100 and 200. Four such villages were encountered in the survey, accounting for 82% of the total population in this period. In this, the Early Formative settlement of the Taraco Peninsula contrasts strikingly with that of Juli-Pomata, where people at this time seem to have lived in scattered farmsteads and small hamlets. Most of these early Taraco Peninsula villages (all except Cerro Choncaya) continued to be occupied throughout the Formative and indeed until the collapse of the Tiwanaku state. Already in the Early Chiripa phase, then, we may observe a system of well-established communities that was to endure for two millennia. This is a rather startling finding, not at all expected, and it distinguishes the Taraco Peninsula from other areas of the Titicaca Basin for which we have comparable data.

Of course, the EF1 phase population index represents the population of the peninsula at the end of the EF1. Given the high rate of growth observed in the following LF2 phase, this does not necessarily imply a large starting population. Given the 500 year span of the LF2, and assuming that the EF1 population growth rate was the same as the EF2 rate (0.47% annually), the EF1 phase population index of 693 could have been produced by an initial population of only 66. This would represent a starting population density of only 0.77 persons/km$^{2}$. I would propose that this initial population was concentrated on the peninsula by the 1500 B.C. lake level rise. That is, prior to the lake level rise they were practicing a generalized and extensive agro-pastoralism in the vast grasslands on the Wiñaymarka Basin. The sudden inundation of the pampa forced this population to concentrate on what had become the Taraco Peninsula, and the territorial claims of neighboring groups prevented many of them from relocating. The Taraco Peninsula is not good grazing land, neither today nor in the past. The concentrated population was forced to adopt a sedentary lifeway probably concentrating mainly on lacustrine resources with some intensification of agricultural activity as well.

Sedentary populations normally have much higher population growth rates than do nomadic or transhumant ones (cf. [Hassan 1981]: 221-229, [Lee 1972,Sussman 1972]). The large village sizes and high regional population density of the EF1 Taraco Peninsula relative to other areas of the Titicaca Basin at this time is probably therefore a result of the earlier establishment of a sedentary lifeway. And this in turn could simply have resulted from their concentration on the Taraco Peninsula by the 1500 B.C. lake level rise. This scenario goes a long way toward explaining the cultural and developmental precocity of the Taraco Peninsula within the southern Titicaca Basin. This model would predict that other long, narrow peninsulas surrounded by shallow water - at modern lake levels - would display evidence of a similar precocity. The area that immediately suggests itself for the testing of this model is the Capachica Peninsula, in the northern Titicaca Basin. Virtually no archaeological work has been done there, however.

The Middle Chiripa phase (EF2) begins around 1000 B.C., and comprises the latter portion of the Early Formative period. No significant changes are evident in the settlement pattern. The site size data continue to display a unimodal distribution, though the mode is higher, with population index values between 200 and 300. Population grew rapidly - at 0.47% annually, close to some theoretically derived maxima for sedentary, non-industrial agriculturalists - and most of the Early Chiripa villages continued to be occupied.

If site abandonment rates were low, though, the rate of founding of new sites was quite high. This probably reflects a situation in which villages grew to a certain size and then fissioned into two or more groups, due to scale-related social stresses and conflict. All of these factors are included by Johnson in the term ``scalar stress'' ([Johnson 1982]). Three such fissioning events are documented in the settlement data from the Taraco Peninsula.

1. Near the beginning of the EF2, around 1000 B.C., the village of Chiaramaya (T-3) split into at least two groups and more likely three or four. One of these, probably the largest, remained in place, while another founded the new site of Chiripa Pata (T-4). Other smaller groups may have founded the sites of Alto Pukara (T-430) and Quiswaran (T-303) at the same time. Chiaramaya had a maximum population index value of 186 before this event.
2. Around the same time that Chiaramaya fissioned, the site of Cerro Choncaya (T-2) broke into two groups. One of these moved to the East and founded the site of Kala Uyuni (T-232) which was to become important in the LF1 period. The other segment moved to the North and apparently combined with the population of a smaller village - Sunaj Pata (T-268) - to form the new village of Sonaji (T-271). The site of the Cerro Choncaya village was abandoned. Cerro Choncaya had a maximum population index value of 157 before this event.
3. Near the end of the EF2, around 800 B.C., the site of Sonaji (T-271) - which had itself been the result of the earlier fissioning of Cerro Choncaya in event 2, above - split into at least three groups. One of these reoccupied the site of Cerro Choncaya (T-2), another reoccupied the site of Sunaj Pata (T-268), and another seems to have moved further to the East and combined with the population of the existing village of Yanapata (T-130). Sonaji attained a maximum population index value of 277 probably immediately prior to this event.

It is certainly worth noting that these three events are particularly conspicuous because they took place very near to phase boundaries. They are therefore very visible archaeologically. It is almost certain that other events than these took place during this time. It is probable in fact that all of the new sites founded during the EF1 and EF2 resulted from similar events. In other words, I believe that village fissioning was a fundamental process structuring the development of the settlement system during the Early Formative. Since the sites are all distributed along the paleolake terraces at the base of the Taraco Hills settlement growth was linear. The pattern that resulted from this process (see Figures 5.5a, b and 6.10) is not so different from those observed along river systems ([Flannery 1976b,Reynolds 1976]).

We have ascribed village fissioning to pressures and conflicts which increase as the size of a community increases; that is, to ``scalar stress''. This interpretation is supported by the fact that the three communities which have been shown to have fissioned in the Early Formative were the largest villages in each phase. Chiaramaya and Cerro Choncaya were the two largest EF1 villages, while Sonaji was the single largest EF2 village. As the process of village growth and fissioning progressed, however, the landscape began to fill up. That is, the distance it would be necessary to move in order to found a new village increased as time went on and as the landscape of the Taraco Peninsula became more and more populated. Thus, the costs of community fissioning - which is simply to say the difficulty and inconvenience associated with fissioning - increased through time. This is probably why the maximum community size increased from less than 200 to almost 300 in the course of the EF2.

As the larger communities were increasing to unprecedented sizes, scale-related stresses within those communities were also increasing. It was in such a context, in the EF2, that an array of new material culture associated with public ceremony made its appearance. It was at this time that the earliest known sunken court was constructed at the site of Chiripa. This structure - the Choquehuanca structure - was an open trapezoidal plaza, excavated into the ground, lined with cobbles, and plastered with colored clay. It is associated with a higher than average percentage of serving wares ([Steadman and Hastorf 2001]), suggesting that it was the site of public feasting activity; of what Dietler calls ``commensal politics'' ([Dietler 1996]). At the same time we observe the first appearance of decorated serving ceramics, probably implicated in the same range of activities.

The timing of the appearance of this early public ritual complex is almost certainly not a coincidence. I interpret it as representing experimentation on the part of community leaders and members attempting to discover ways to reduce scale-related stresses and conflicts. The effort seems to have been at least partially successful, since only a few of the largest sites did in fact fission in the EF2, and the ``fission threshold'' increased from around 190 in the EF1 to almost 280 in the EF2. This early period of experimentation was to be followed, in the Middle Formative, by the formalization of a complex of ritual ceramics, architecture, and stone sculpture known as the ``Yaya-Mama Religious Tradition'' ([Chávez and Mohr Chávez 1975,Chávez 1988]). That tradition, however, had its roots in the experimental structures and ceramics which we have observed in the EF2, in the Middle Chiripa phase.