Analysis of the ceramics produced by both general and systematic surface collection of sites was oriented almost completely toward chronological questions. The goal of the fieldwork as a whole was to produce as temporally and spatially fine-grained a settlement dataset as possible, and the goal of the ceramic analysis was the same.
Fortunately, quite a lot of work has been done on the ceramics of the region obviating the need to produce an entirely new chronology. In particular, the Early and Middle Formative period chronology of Steadman ([Steadman 1999]) proved to be quite workable (though with some modifications as outlined below), as did the Pacajes chronology (Late Intermediate Period/Late Horizon/Early Colonial Period) of Albarracín-Jordan and Mathews ([Albarracín-Jordan 1992,Mathews 1992,Albarracín-Jordan and Mathews 1990]). My Early and Middle Formative and Pacajes chronologies as presented here are basically simple adaptations of these previous works.
Other time periods proved to be less tractable. I found that, like other settlement archaeologists working in the Titicaca Basin, I was unable to distinguish between Tiwanaku periods IV and V (see [Ponce Sangines 1981]) on the basis of surface materials. It is becoming increasingly clear that neither the Wila Jawira chronology adapted by Kolata and associates from Ponce's work ([Ponce Sangines 1981]) nor the modified Bennett ([Bennett 1934,Bennett 1936,Bennett 1948]) sequence of Albarracín-Jordan and Mathews ([Albarracín-Jordan 1992,Albarracín-Jordan 1996a]) is a tenable chronology for settlement survey. Goldstein's chronology ([Goldstein 1985,Goldstein 1989a]) from the Moquegua Valley, Peru is likewise unusable in the altiplano, though it does appear to be useful for Moquegua. Burkholder's recent work ([Burkholder 1997]) has promise, but seems - like other chronologies - to rely on relatively intact decorated specimens, always a rarity in surface assemblages. A new highland Tiwanaku chronology designed with mixed and fragmentary surface materials in mind would be most welcome and would be a positive boon to research in the South-Central Andes. Lacking this, however, I was forced to combine all Tiwanaku IV-V materials into a single Tiwanaku phase, as unsatisfactory as this may be.
Also problematic were the Late Formative phases. Ponce originally defined the Tiwanaku I-III phases based on his excavations at Tiwanaku and on the prior definition by Bennett of an `Early Tiwanaku' phase. They have bedeviled archaeologists since. The principal difficulty, as others have pointed out (c.f. [Janusek 2002,Mathews 1992]), is that Ponce's phases were defined in terms of decorated wares: zoned-incised in the case of Tiwanaku I, polychrome painted in the case of Tiwanaku III. This is even more problematic in the case of the Late Formative than in that of Tiwanaku IV-V, as the relative abundance of decorated wares in the later Tiwanaku periods contrasts with their extreme scarcity in the Late Formative. To resolve this difficulty John Janusek and Carlos Lémuz have both produced Late Formative plainware ceramic chronologies for the southern Titicaca Basin based on excavated materials ([Janusek 2002,Lémuz Aguirre 2001]). These chronologies are at present in a tentative state. The rudimentary plainware chronology employed here, and described in the Late Formative chapter (Chapter 7), is my own invention, though produced in close consultation with both Janusek and Lémuz who very generously shared their insights and unpublished data with me. It is consistent with both of their chronologies, though more limited in scope and focused on a restricted range of common paste types.
In short, then, my ceramic analysis was oriented to the identification of assemblages from the various time periods. These assemblages differ radically. Perhaps even more significant, the existing chronologies for the various phases diverge considerably in emphasis and focus. Thus the Early and Middle Formative chronology is focused largely on plainware pastes, while the later period chronologies (Tiwanaku and Pacajes) center on decorated wares. The Late Formative chronology is minimally developed. This situation made it necessary to adopt different strategies for the identification of assemblages of the various phases. Though this matter will be discussed in later chapters as well, here I provide a brief summary of the methods and techniques I developed to distinguish occupations of various phases in mixed surface collection assemblages.