Phase definition

Late Formative ceramics in the southern Titicaca Basin were first identified by Wendell Bennett in his excavations at Tiwanaku ([Bennett 1934]). In deep midden buried by later deposits, Bennett identified a group of ceramics he called ``Early Tiahuanaco''. These consisted mostly of plainwares, together with ``clay buttons'' (Figure 7.1v-w), bowls with horizontal handles attached to the rim, and two sets of decorated wares which we now call ``Kalasasaya'' and ``Qeya.'' Early Tiahuanaco levels were subsequently excavated by Rydén and Kidder at Tiwanaku and elsewhere ([Kidder 1956,Rydén 1947]), though they contributed little to the definition of the phase. The Bennett sequence was later refined by Dwight Wallace's formal seriation ([Wallace 1957]).

Ponce Sanginés's later chronology took the work of Bennett as a starting point, but considerably modified and extended Bennett's original phase definitions ([Ponce Sangines 1981]). As far as the Late Formative is concerned, Ponce divided the ``Early Tiahuanaco'' phase into three separate phases, Tiwanaku I, II, and III. While Ponce's sequence was explicitly designed around a conception of Tiwanaku political evolution, and therefore was not strictly a ceramic chronology, he did note correlations between ceramic style and his broader cultural and political epochs. He correlated Tiwanaku I with the first of Bennett's Early Tiahuanaco stylistic groups, the Kalasasaya style ([Ponce Sanginés 1976]). This group was characterized by zoned-incised pieces, clearly related to the northern basin Pukara tradition. His Tiwanaku III was associated with Bennett's second group, Wallace's Qeya style, distinguished by a distinctive vessel shape assemblage and color palette. Ponce's Tiwanaku II apparently contained only undecorated sherds, and was poorly defined ([Ponce Sangines 1981]).

Recent investigators have tended to dismiss the Tiwanaku II phase as a fiction, while debating the chronological relationship between the two stylistic groups of Early Tiahuanaco, Tiwanaku I and III. The problem has been particularly acute for settlement research, since decorated ceramics of both the Kalasasaya and Qeya styles are quite rare and are very infrequently encountered on the surface of sites. Further, in contrast to the earlier Chiripa and later Tiwanaku decorated ceramics, the distribution of Late Formative decorated materials - Kalasasaya and Qeya wares - seems to have been highly restricted. That is, they occur only infrequently at major sites and political centers, and almost never at smaller habitation sites. Reliance on this sequence for chronological control has been the source of major problems for some recent Tiwanaku-area settlement surveys ([Albarracín-Jordan 1992,Albarracín-Jordan and Mathews 1990,Albarracín-Jordan 1996b,Mathews 1992]).

Application of the Ponce Tiwanaku I-III sequence to regional settlement survey was so problematic, in fact, that Mathews advocated a return to Bennett's inclusive Early Tiahuanaco phase ([Mathews 1992]: Chapter 4). Based on his survey and limited excavations, he suggested that Tiwanaku I and III were two components of a single ``cultural stage,'' and that the categories had no real chronological significance. This interpretation was shared by Albarracín-Jordan ([Albarracín-Jordan 1992,Albarracín-Jordan and Mathews 1990]).

More recent research, and more careful consideration of a range of datasets, has shown Mathews's conclusion to be incorrect. It is indeed possible to distinguish between earlier and later ceramic complexes within the Late Formative. What is more, thanks to the work of John Janusek ([Janusek 2002]) and Carlos Lémuz ([Lémuz Aguirre 2001]), it is now possible to speak of differences in plainware pastes and vessel shape, in addition to changes in the decorated wares. Following Janusek, I have divided the Late Formative into two phases, the Late Formative 1 (LF1), which can be equated with Tiwanaku I at Tiwanaku, and the Late Formative 2 (LF2), including Tiwanaku III and Qeya ceramics. While my own work draws heavily on that of Janusek and Lémuz, I have adapted their sequences to the ends of settlement research. What follows, then, will emphasize a few common plainware pastes at the expense of a detailed seriation of attributes such as vessel form or firing.

In comparison to the Chiripa phases, the Late Formative phases are poorly-defined. This is obviously a priority for future research in the region. It should be possible to subdivide the LF1 at least, since Bermann's work at Lukurmata has demonstrated substantial changes in assemblage composition during this period ([Bermann 1990,Bermann 1994], see also [Janusek 2002]).