Before proceeding, I remind the reader of the discussions in Section 3, particularly of the issues relating to monitoring position; the distinctions between analogies and homologies; of Hughes' strictures of making inferences from data about subsistence and settlement patterns to ethnolinquistic groups, and the two levels of continuity and discontinuity identified at the end of that section. I distinguished between historical continuities and discontinuities (presence or absence of cultural links between ancestral and descendant groups) and continuities and gaps in the archaeological record. This report has focused on gaps and continuities in the record itself.
Based on my review of the available evidence, the empirical gaps in the record preclude establishing cultural continuities or discontinuities, particularly before about 5000 BC. I will provide one example of the difficulties. The Lind Coulee site, the closest major site both geographically and temporally to the Kennewick individual contains a somewhat distinctive variety of stemmed Windust points (they are not the style of the point embedded in the Kennewick individual's hip). These are also about the last such points on the Plateau. Recently, Davis recovered very similar points26 at the Copper's Ferry site in Idaho that are 2000 years older than the Lind Coulee site. Does this mean continuity in this point style on the Plateau during this long period? Alternative explanations are that Davis' dates or the Lind Coulee dates are wrong. I noted above problems in firmly dating the Lind Coulee site. If Davis' dates prove to be reliable, then it will be necessary to reevaluate the age of Lind Coulee. If there was a big sample of Lind Coulee points from multiple, well-dated sites, this issue might not even arise. However, there is no such sample.
The sample of radiocarbon dates for this early period is small and marked by gaps. These gaps may reflect the sample's size, or, like the gap in house floor dates at Hatwai, point to something else. The sample is also sensitive to archaeological sampling practices. It is also sensitive to different quantification and graphing techniques that produce differing apparent patterns in the data. The sample is therefore not very robust and conclusions drawn from it must be viewed cautiously. In spite of these problems, all available data do suggest very low populations for the period before the Mazama ash fall.
The period's material culture has a number of features which do not carry forward, including bola stones, edge ground cobbles, core and blade technology, the fine bone needles, and perhaps the Ft. Rock-style sandals among others. Their subsequent disappearance is probably (to my mind) due to culture change, not to cultural replacement. However, I cannot preclude that possibility. On the other hand, there are continuities in the forms of end scrappers, and the presence of utilized flakes and cobble tools, among others. However, these may be analogous or functional similarities, and therefore not evidence for cultural continuity. Methodological issues, such as whether classifications of artifacts are consistent, also cloud some possible long-term continuities (e.g. edge-ground cobbles present in a very few late assemblages). A key issue here is the general absence of occupation of the central Basin, except immediately along the river. The riverside occupation in the immediate area is probably somewhat later, however, than the Kennewick individual's lifetime. The occupation record is continuous to the east, along the Lower Snake River and its tributaries.
The major changes that occurred after 4000 B.C. also make it exceedingly difficult to trace connections forward in time. Many of these changes, such as the initiation of pit house construction, are probably related to broad-scale changes in western North America, and so do not indicate population replacement. However, they do point to a period of significant culture change and reorganization. Other changes include the reorientation of the region's interaction patterns from apparently tending towards the south to the west and the coast. There are also puzzles during this period, such as microblades, and what they represent.
Chatters' proposed explanatory model for the changes between Pithouse I and Pithouse 2 approximates the expectations of (and shares some common ideas with) Bettinger's model for what the Shoshonean expansion should look like archaeologically (Bettinger 1994). However, these remain models, not evidence. Nor is there any suggestion in the evidence of a migration. However, it does seem evident that there was a major reorganization of subsistence and settlement systems during this period.
The evidence is overwhelming that many aspects of the "Plateau Pattern" were present between 1000 BC and AD 1, and that the Pattern itself, as described by Nelson (1973) and Walker (1998), was probably fully in place by AD 1, although change continued to occur (e.g. the appearance and spread of long-lodges). However, evidence from the southern uplands (Endzweig 1994, Schalk et al. 1995, Hess 1997) indicates considerable movement and shifting around of peoples in the region. These changes may include aggregating down along the Columbia in the last 1000 to 1500 years in large villages, with an accompanying reduced use of the uplands.
These conclusions emphatically do not mean that to my mind there was not cultural continuity between the people of the Columbia Plateau in1800 and earlier peoples on the Plateau. At the beginning of this section, I wrote that the empirical record precludes establishing cultural continuities or discontinuities across increasingly remote periods. However, if the available evidence cannot be used to show continuity, it is equally refractory for demonstrating discontinuity. Such evidence, either way, might be developed by an extensive review of both the published literature and of museum collections, and through continued fieldwork.
26I have not seen the points. Davis and Jerry Galm have independently informed me of the similarities between Lind Coulee and the Copper's Ferry points.