This data us used to suggest and test theories about how social units -- their scale, complexity, integration and boundedness change over time. This works in roughly the following way.
The assumption underlying serration is that , over time, in the study area as a whole, any specific ceramic type will initially appear in relatively low frequency, grow in frequency for a time, then decline in frequency for a time and ultimately disappear.
Thus one may attempt to order sites (characterized by relative frequency of ceramic types) in such a way that the relative frequency of each ceramic type exhibits this characteristic pattern with respect to the order. Given enough sites and ceramic types, there will usually be only one way to do this.
Ceramic types themselves may be ordered by the order this imposes on the sites at which they have the highest relative frequencies.
Additional information must be employed to impose a temporal direction of the order obtained by this serration process. Typically, this is stratification obtained from excavation. Ceramic type recovered from lower strata are assumed to be earlier than ceramic types recovered from higher strata. The result of this is a ceramic sequence.
This ceramic sequence is used to identify and temporally order occupations of sites. Roughly, each ceramic type present at a site corresponds to an occupation of the site and the occupations are temporally ordered according to the temporal order of the types.
For the mathematically inclined, a somewhat more precise discussion of ceramic serration is provided.
The scale ordering on contemporaneous occupations is used to identify social units and the part of relation on them. The idea is simple. A set of spatially contiguous, contemporaneous occupations is taken to be a social unit with it’s (set theoretic) members as parts just when there is one member of the set that is strictly larger than the rest on the scale ordering. For example, a set of spatially proximate contemporaneous occupations will be identified as a village-unit just when there is one occupation (the village) that is of significantly larger scale than the others (hamlets) . The village will be identified with the village unit and the hamlets will be taken as parts of the village.
This is somewhat oversimplified. Some kind of social interaction among the parts of social units and/or the wholes and their parts is also required. In a fully precise treatment, the part of relation on social units and the interaction relation among them as well would together be required to satisfy some requirements. Just how to make this explicit is not easy to derive from the discussion in Blanton, et. al. ‘93,
There is an inferential step required to move from artifactual-contextual data to attribution of activities to occupations. This step requires some general principles or “laws” relating activities to configuration of debris they produce.
Assignment of activities to contemporaneous occupations provides one of the basis for inferring a complexity ordering on the social units identified from analysis of the settlement pattern data. Together with the part of relation on social units provided by the settlement pattern analysis, one may draw conclusions about both horzontal and vertical complexity.
While it is most natural conceptually to view social integration as inferred from assignment of social interaction to occupations, it is also possible to infer social integration directly from the distribution of ceramic types across contemporaneous occupations and the cost ordering on ceramic types. This is discussed in Blanton, et. al. ‘93,
