Ceramic Classification

The ceramic classification created for this project serves the main purpose of allowing the chronological placement of the settlements identified in the regional survey. It was designed to permit the chronological identification of sherds collected without a stratigraphic context and without association with absolute dates, and is based on the excavation of stratigraphic tests. The vast majority of materials collected in the regional survey are small non-diagnostic sherds without decoration. In the eyes of some these are worthless, because classification should focus on the more reliable diagnostic and decorated sherds (yet these are still considered a poor substitute for complete ceramic pieces). Although it is certainly the case that form and decoration are very important chronological markers, using them alone in this case would have implied ignoring the bulk of the ceramics collected, and therefore, most of the sites in the regional survey could have not been assigned to a chronological period. This would have been a very unfortunate outcome, equivalent to treating only large sites with architecture as informative and worthy of the efforts of archaeologists. This lack of attention to less flashy material findings may explain why none of the publications that review the work of Porras mentions the presence of Papallacta Ordinario and Borja Ordinario in his excavations. Despite extensive attention to the ceramics of the region (in reality only to Cosanga pottery), no one has questioned how Porras arrived at the conclusion that there had been only one occupation in the region, or why he did not make much of any of the different types that accompanied Cosanga. Thus, undecorated sherds different from Cosanga have simply not been dealt with, either because they are not informative in the eyes of some archaeologists or because their quantities are comparatively so low that they are not perceived to be worth the effort. In this project they received equal treatment.

The process of classifying ceramics is always guided by the goals of a project. From the infinite factors that can enter a matrix of classification criteria, one generally chooses those that are most likely to provide the necessary information to answer specific research questions. In that sense, a classification that accounts for all of the possible dimensions of ceramic variability is never possible, and therefore no classification is ever “complete” (Sinopoli 1991:44). The dimensions of variation I used in this project are surface, paste, temper, form, and decoration. When the attributes of more than one ceramic type were present on a single sherd, the tendency was to favor attributes related to general appearance to classify it. Drennan (1993) has suggested that because people are ultimately concerned with appearance when making and using pottery, attributes that relate to it are more useful than others such as temper—which may reflect just variations in the distribution of minerals in a region.

The types proposed below, as types that predominated during different time spans, were defined through the analysis of materials from 46 stratigraphic tests spread throughout much of the survey region. Of these 46 tests, 15 2x1 m tests were excavated with the specific aim of establishing a ceramic chronology that could be used to analyze the materials of the entire region. The remaining 31 1x1 m tests were excavated as part of a program focused on the recovery of botanical remains, yet many of them were informative in terms of chronology, as they strengthened observations derived from the former set of tests. The methods and rationale for selecting sites for excavation of 2x1 m test pits are explained in the "Stratigraphic Test" section of this electronic document (details about the rationale for the placement of 1x1 m test pits excavated with the main purpose of recovering botanical samples is provided in Chapter 6 of the published volume from which this electronic document is derived). In the process of analyzing ceramics and sorting them out into types with chronological significance, survey materials were used as referents against which to compare what appeared in the excavations and test the utility of the typology in terms of accounting for the vast majority of ceramics collected in the survey. This process of defining types based on materials collected with a stratigraphic context, and then using the typology to classify survey materials was repeated, back and forth, until the process of classifying survey materials ran more or less smoothly. The types presented here were initially subdivided into more types, as explained below, and finally grouped together again in the way that seemed most appropriate in terms of accounting for temporal variation. Future research may lead to the conclusion that some of those sub-types are chronologically distinctive, but for the moment the evidence does not seem strong enough to make that assertion. The schema utilized here seemed the most unproblematic and straightforward in terms of making possible the classification of survey materials, and has strong support in the analysis of excavated ceramics.

The 46 stratigraphic tests yielded 9,506 sherds. 2,121 of the 2,256 lots of the survey (133 lots contained only lithic material) were classified using this typology (23,585 sherds). Some sherds did not seem to fit any of the type definitions and therefore were not classified. These comprise less than 1% of the survey materials, and could only have a negligible impact on the interpretation of regional settlement patterns. A research assistant and I classified all of the sherds.

Return to "Ceramic Chronology and Stratigraphic Testing" Main Page