The methodology described in the preceding chapter is an excellent
procedure for precisely defining the distribution in space of a determinate
set of ceramic attributes. Concentrations of sherds with these attributes
which satisfy an explicitly-stated test (in this case, the 0.1 sherds/m
rule) are recorded with respect to their location and a range of ancillary
variables (topography, architecture, etc.). The register of these
concentrations, or `sites,' together with a much smaller list of other
culturally significant spatial loci (agricultural features, rock art,
the like) constitutes a settlement dataset, of which Appendix A is
an example.
The dataset at this point has a strictly descriptive function. As yet it has acquired no interpretive significance. Between settlement data and settlement analysis yawns the chasm of ``middle-range theory'' or ``bridging arguments.'' For the purposes of settlement analysis, crossing this chasm requires that the objective ceramic attribute distributions recorded in the dataset be correlated with relevant analytical categories. Standard practices exist for bridging the gap in settlement analysis, and these have become sufficiently transparent to archaeological practitioners that they are frequently allowed to remain implicit. The bridging arguments I intend to use here, however, differ somewhat from received theoretical practice. For this reason, I wish to describe them as explicitly as possible.
As I emphasized in the introductory chapter, I am interested primarily in the analytical categories of time and of population. Connecting ceramic attributes with the dimension of time is a relatively straightforward business, well-studied in archaeology (for a review see [Marquardt 1978]; for recent south-central Andean examples see [Alconini Mujica 1995,Goldstein 1985,Steadman 1995]), and my chronology will be a basically conventional one. Ceramic chronology will be discussed in the various chapters to follow. Connecting ceramic attribute distributions to measures of human population, however, and derived measures of demographic process, is a more difficult and disputatious matter. It is to this question that the present chapter is devoted.