Settlement and population

The first fact worthy of note, as mentioned in the introduction to this chapter, is that the inhabitants of the Taraco Peninsula seem to have lived in relatively large settled villages from very early in the Early Chiripa phase. As may be see in Table 5.1, the average Early Chiripa sector population index is 77 (phase population index/number of sites=1752/23=77). This probably represents an average of more than 10 households per village. On the face of it, this is not a large number. Compared to other areas of the Titicaca Basin at this time period, however, it is very high.

Table 5.1: Early Formative: Metrics

	EF1 - Early Chiripa	EF2 - Middle Chiripa
Number of sites	9	23
Phase population index	693	1752
Annual population index growth rate	-	0.47
Occupation continuity index	-	63
Site founding index	-	78

As an example we may take sites of the Pasiri Period in the Juli-Pomata area. Pasiri ceramics are somewhat similar to early Chiripa ceramics - though they remain relatively poorly described - and they seem to be diagnostic of Early Formative sites in the southwestern Titicaca Basin ([Stanish et al. 1997]: 40). Pasiri sites average 0.8 hectares in size ([Stanish et al. 1997]: 51). If corrected in the same manner as my own data, this yields an average population index of 32. Early Formative sites on the Taraco Peninsula, therefore, are on average more than twice the size of those in the Juli-Pomata area.

Furthermore, Early Formative occupation is much more dense on the Taraco Peninsula than in Juli-Pomata. Ten Pasiri sites were encountered in the Juli-Pomata survey ([Stanish et al. 1997]: 40). This project intensively surveyed 300 km$^{2}$. Early Formative population density in Juli-Pomata - when the figures are corrected using my method - may therefore be estimated to be approximately 1.0/km$^{2}$. The same calculation on the Taraco Peninsula data (85.22 km$^{2}$ surveyed) produces a figure of 8.1/km$^{2}$.

It is apparent that the Taraco Peninsula was exceptional from the very beginning of its occupational history. No other known area in the Titicaca Basin has Early Formative populations this dense or Early Formative villages this large. This Early Formative exceptionalism clearly set the stage for the continuing precocity of the Taraco Peninsula villages throughout the Early and Middle Formative Periods. It may be explained, as I suggested in the introduction to this chapter, by the 1500 B.C. lake level rise which first created Lake Wiñaymarka. As the lake rose, groups of extensive agro-pastoralists (we may imagine small, mobile communities such as the Late Archaic sites investigated by Aldenderfer and Klink in the Ilave River drainage) occupying the vast grassy plain that then existed would have been increasingly concentrated on the narrow spit of land that was to become the Taraco Peninsula. If we imagine that possibilities for outmigration were limited - by, for example, territorial claims of adjacent groups - these conditions would have created powerful incentives for these people to focus their subsistence activity very strongly on agriculture and lacustrine resources (fish, birds) at the expense of herding or hunting. Such a rapid subsistence shift - this could all have taken place within a generation - could certainly account for the exceptional density of Early Formative population on the Taraco Peninsula relative to areas - such as Juli-Pomata - where the impact of the lake's rise was less pronounced.

Figure 5.4: Early Formative site size distributions

Whatever the case, a stable system of permanent villages was established on the Taraco Peninsula in the Early Chiripa phase. This settlement system was to prove very stable; it changed only incrementally for over 2500 years, until the collapse of the Tiwanaku state.

Figure 5.4 shows that the settlement systems of both the Early and Middle Chiripa phases are characterized by a unimodal site size distribution. The majority of the population in the Early Chiripa phase resided in sites with a population index in the range of 100-200. The Middle Chiripa pattern is basically identical, but the average site size has increased. Most people now lived in sites with a population index between 200 and 300. The unimodal site size distribution in both phases strongly suggests a system of autonomous villages. No single village of group of villages seems to have been even locally dominant during the Early Formative.

Figure 5.5: EF: Sum of phase population index per 0.25 km$^{2}$

• a) Early Formative 1

• b) Early Formative 2

Table 5.1 shows that the Early Formative settlement system was not only very stable (the occupation continuity index is well over 50% for Middle Chiripa), but also that it was growing rapidly. The annual population index growth rate was 0.47%, which is quite close to Hassan's estimate of 0.52% for the maximum growth rate of prehistoric agricultural populations ([Hassan 1981]: 140). This is the highest growth rate in the entire settlement history of the area excepting the Late Horizon and Early Colonial Periods, where there is clear evidence of widespread immigration. Additionally, the site founding index for the Middle Chiripa phase is 78%, a very high value. The general picture that emerges from this information is one of relatively small, autonomous villages growing rapidly and fissioning when their population passes a certain critical threshold. Indeed, two such instances of village fissioning seem to be observable in the Early Formative on the Taraco Peninsula. Interestingly, the two villages in question were the two largest sites during the Early Chiripa phase.

Figure 5.5 is not a settlement pattern map. Rather, it is a population density map. The survey area has been divided into 0.25 km$^{2}$ squares (squares 500 m on a side). The population index values of all of the archaeological sectors falling within each square for a particular time period are summed. This summed population index is represented by the z axis. So higher spikes represent higher local population densities.

A visual inspection of Figure 5.5a reveals two spikes in population density on the peninsula in the Early Chiripa phase. These correspond to the sites of Chiaramaya (T-3) and Cerro Choncaya (T-2). By the Middle Chiripa phase, however, (5.5b) these two sites are no longer the loci of highest population density on the peninsula. Cerro Choncaya has been abandoned, and Chiaramaya has shrunk dramatically in size. Three new sites have come to the fore: Chiripa (T-1), Janko Kala (T-394) and Sonaji (T-271).

Figure 5.6: EF: Change in phase population index per 0.25 km$^{2}$

• Early Formative 2

Figure 5.6 graphically represents the spatial shifts in population that occurred in the Middle Chiripa phase. It was produced by subtracting the sum of Early Chiripa population indices for a given 0.25 km$^{2}$ block from the sum of Middle Chiripa population indices for the same block. Essentially, it was produced by subtracting Figure 5.5a from Figure 5.5b. The vertical axis therefore represents the number of hectares by which the cumulative population index of the block increased or decreased during the Middle Chiripa phase. A raised spike indicates a population increase in a block, which a sunken area represents a decrease. A flat surface indicates no change.

Figure 5.6 shows that only three localities experienced population declines in the Middle Chiripa phase. These were the two sites of Cerro Choncaya and Chiaramaya (mentioned above), together with the site of Sunaj Pata (T-268), located very near to Cerro Choncaya. At the same time, however, a number of new sites were founded and experienced very rapid growth. Among these were Chiripa Pata (T-4), Sonaji (T-271), Kumi Kipa (T-272), Kala Uyuni/Achachi Coa Kkollu (T-232/T-225), Alto Pukara (T-430) and Quiswaran (T-303).

The overall population growth rate for the Middle Chiripa phase was 0.47%. The Early Chiripa villages that were not abandoned in the Middle Chiripa phase experienced growth at a comparable, if somewhat slower rate (0.35% at Chiripa [T-1]; 0.39% at Janko Kala [T-394]). Interestingly, if we combine the population indices of Chiaramaya with those of all the newly-founded Middle Chiripa sites to the east of Taraco (Chiripa Pata, Alto Pukara and Quiswaran), we arrive at a combined growth rate of 0.47%, identical to the gross phase rate. Combining the population indices of Cerro Choncaya and Sunaj Pata with the new Middle Chiripa sites west of Taraco (Sonaji, Kala Uyuni, Achachi Coa Kkollu, Kumi Kipa) produces a growth rate of 0.35%, identical to the growth rate of Chiripa, and similar to the phase average.

To me, these figures strongly suggest a set of actual events; concrete population movements. I would reconstruct them as follows:

• The village of Cerro Choncaya split into two groups. The first of these moved to the east and founded the site of Kala Uyuni and its outlier Achachi Coa Kkollu. The other group moved a short distance to the north and west, combining with the prior population of Sunaj Pata to found Sonaji and its outlier Kumi Kipa.
• The village of Chiaramaya split into two or more groups. The first of these groups moved to the east, leapfrogging over the adjacent site of Chiripa to found Chiripa Pata. Smaller groups may have continued eastward and founded Alto Pukara and Quiswaran. A small percentage of the Early Chiripa population of Chiaramaya remained in the village, which continued to be occupied, though on a much-reduced scale.

The fact that the two villages that fissioned in the Middle Chiripa phase were the two largest in the Early Chiripa phase is almost certainly not a coincidence. This is exactly what one might expect in a relatively lightly-populated and unbounded landscape. The growth of villages led to internal strife and factional conflict (and all of the factors implicated in Johnson's term 'scalar stress' [[Johnson 1982]]), resulting in village fissioning and relocation. The data suggest that the critical threshold - in terms of my population index - in the Middle Chiripa phase was something on the order of 150. When Early Formative settlements surpassed this size, their inhabitants apparently chose to split the community and relocate a portion of it, presumably as a response to increasing internal conflicts. This is a very low threshold indeed, and in subsequent periods we will see it rise considerably as the people of the Taraco Peninsula came to have fewer and fewer possibilities of relocation, and became better equipped to minimize and resolve intracommunity conflicts.