Issues and Problems
A range of issues and problems must be addressed before proceeding. The basic task of this study is to review the archaeological record for the southern Columbia Plateau, and identify continuities and discontinuities in that record that may be relevant to linking the Kennewick skeleton to Indian people living along the Mid-Columbia and Lower Snake Rivers in c. AD 1800. The fundamental problems are 1) what constitutes continuities and discontinuities in the record (indeed, what exactly is a continuity and discontinuity?), and 2) how is their relevance to affiliation to be determined? Perhaps an even more fundamental issue, an elemental issue, is whether the archaeological record is even capable of answering either of those questions. In this section, I will follow what may seem like a rather circuitous route to answering at least the first two questions. It is also perhaps useful at this point to remind the reader that establishing genetic or anatomical connections is not part of this particular study. That is being pursued by others in parallel work. The possible linkages explored here are historical and cultural, and they are linkages between the "early group" and early 19th century groups. These possible linkages then are between groups, or, to be precise, between an archaeological complex dating to the eight millennium BC and human groups of the early 19th century. The relationships between archaeological manifestations and actual ancient human groups or societies that produced them are, at best, indirect.
Recognition of Continuities/Discontinuities and Empirical Gaps in the Record.
In terms of this study, there are at least two major implications of monitoring position. Much of the archaeology discussed below focuses on settlement patterns and land use systems. We can imagine two such systems that are markedly different, but which overlap at only one place or one activity. If, by some chance, we archaeologically monitor both systems primarily from this one point, then we may conclude they are very similar or identical, when, in fact they were not. This could be a place, or an entire region. For example, Swanson (1972) postulated long-term stability in regional cultures in southeastern Idaho based on his excavations in the Birch Creek Valley, a moderately high elevation montane valley in the southern edge of the Northern Rocky Mountains. However, Swanson's archaeological monitoring position was hunting camps, a manifestation that, given the available technology, the terrain, and the animals, was likely to remain quite stable throughout the Holocene, regardless of organizational and cultural changes elsewhere. Thus what looks like long-term stability may only reflect stability in that one place. On the Columbia Plateau, until recently, our monitoring position was from the bottoms of canyons. Thomas (1983) stresses that large regional samples are necessary, though not sufficient by themselves, to control for monitoring position.
Another aspect of monitoring position to be considered here is the degree to which the archaeological data we have informs us about what we want to know. In Prince Rupert Harbor, on the northern coast of British Columbia, the vast majority of artifacts are pointed bone tools of one kind or another. They probably armed spears, arrows, liesters, fishhooks and toggling harpoons, as well as being used as awls, punches, needles and a variety of other ways. They do not change at all over a span of 4000 years or so (Ames 1976), a period marked by a series of major social and economic changes, including ongoing warfare during the past 1000 years or more, which included one period when the harbor was abandoned, and perhaps even one group of people being replaced by another. However, this category of artifact does not monitor those changes. Other categories of material culture do monitor the changes, but this archaeologically most common category does not.
Turning from biology to culture, homologous cultural traits are evidence of the presence of a common cultural ancestry. Analogous traits would indicate independent but similar solutions to similar problems. Their existence also implies that there may be only a limited array of solutions to particular functional problems. Flight is impossible without wings. There is a variety of ways to build wings (insect wings, bird wings, bat wings, airplane wings, helicopter rotors) and ways they can work to power flight, but wings themselves are essential.
While these terms are easy to define, they are much more difficult to operationalize, particularly in dealing with humans, who can borrow traits as other creatures cannot. Whales are not likely to have borrowed their forelimbs from the remote ancestors of humans, for example. The parliamentary systems of Britain and the USA are homologous while they are different in many details, they share a common and now somewhat ancient ancestry. On the other hand, the US constitution has recently been widely borrowed by societies with which the USA has no deep historical connection.
To use an example perhaps more relevant to this study, semi-subterranean houses (pithouses) were a very common form of house throughout much of interior western North America over the past several thousand years. The question is whether these structures are a homologous trait (they were widespread because all of the peoples who built them shared a common, although perhaps remote, cultural ancestor who also built pit dwellings) or an analogous trait (pit dwellings were a common solution to the problem of shelter and were invented many times independently).
The situation is more complex than this short discussion
can cover. I observed above that analogies exist in part because there
may be a limited array of solutions to functional problems there
is no alternative to wings. While not the case for wings, shared common
ancestry and history can be one of the reasons for a limited range of
solutions. Pithouses may be so common in the archaeology of western
North America because they were part of the basic cultural repertoire
of the ancestral culture(s). When the region's ancient occupants were
deciding whether to build permanent shelter or not, and what kind of
shelter to build, they did not have the full array of all human shelters
available to them to choose from, they had the options that their culture
provided, or which they invented. In short, pit houses were common because
1) people had a common functional problem to solve (analogy) and 2)
their shared history (homology) provided them with a common solution
to that problem.
In a broader sense, analogies are examples of equifinality. Processes that are equifinal may begin at very different starting points, but their end products are extremely similar because only a limited number of outcomes are possible. While analogies are found in the living world, including culture, equifinality occurs throughout the material world, living and not.
For the moment, however, it is primarily important to bear the basic distinctions between analogies and homologies in mind: similarities due to common function and to common ancestry respectively. There are also similarities, continuities, due to common environment, to common structures (e.g. Braudel 1972). That is the topic of the next section.
and Matters of Scale: Sources of Stability and Variation
Geography and Resources: Certain features of the region's topography were stressed in the background section on the environment. These features have not changed over the past 9500 years, regardless of the documented changes in climate, stream regimes, and distributions in vegetation and animals. Among these are the distributions of rainfall across the Plateau, environmental patchiness, and possible locations of trails and travel routes.
As noted above, rainfall levels increase across the Plateau from west to east. The low rainfall in the west is a consequence of the rain shadow of the Cascade Mountain range. The increase in rainfall results from increased distance from the mountains and the slow increase in elevation to the east. These circumstances are invariant; therefore, easterly, higher areas have probably always been relatively wetter in dry periods as well as in wet periods. In the same way, the complexity of the topography, and environmental patchiness, increases to the east, (and the southeast) as elevations rise, canyons become relatively deeper, and mountain ranges intervene. From the human standpoint, then, the easterly portions of the Plateau probably have always had the highest ecological productivity in their terrestrial environments. Like terrestrial environments, the region's riverine environments also have some geographic invariances.
Researchers have investigated considerable changes in river regimes during the past 10,000 years, in flow levels, depositional regimes, postulated variations in water temperature and so on (e.g. Chatters and Hoover 1992). There has been particular focus on salmon, and potential causes for variation in salmon productivity, principally declines in productivity. Potential causes of declines include human over-use (Schalk 1987, Campbell 1989, but also see Hewes 1973); increased water temperature and declining water levels during the Holocene thermal maximum (V. L. Butler and Schalk 1986, Chatters et al. 1991) and spawning conditions (V. L. Butler and Schalk 1986). However, some aspects of the riverine environments have not changed much, and cannot change.
One of these is the relationship between the structure of a drainage basin, and the productivity and stability of salmon runs (Schalk 1977). Salmon runs in the main stem or in major tributaries are larger, less subject to variation, and last longer then runs on lesser tributaries. The main stem runs are the aggregate of all the runs in the drainage basin. Runs on tributaries are more subject to failure or decline, in that sense; they are less predictable and less stable. Their fish runs will be of shorter duration. Salmon do not eat once they enter fresh water, so the quality and caloric content of their flesh often declines the farther upstream they are. On the other hand, fish in small tributaries may be much more accessible than those in main stems. This is particularly so in very large rivers, such as the Columbia, where the most productive fisheries were located at a very few places where the river narrowed and went over falls or through rapids (e.g. Hewes 1947, Sneed 1971, Schalk 1977, Romanoff 1992, Kennedy and Bouchard 1992).
V. L. Butler and Schalk (1986) predict possible changes in the salmon runs during the Holocene, based on a broad reconstruction of Holocene climates. They suggest that before 10,000 years ago (Late Pleistocene/Early Holocene) the distribution of salmon was shifted south to an unknown extent. As glacial conditions ameliorated in the basin, salmon behavior may have been similar to that currently exhibited in the most northerly portions of the fish's range: high levels of anadromy (smolts move quickly to salt water) and very short summer runs. Spawning areas may have been limited. V. L. Butler and Schalk (1986) predict that a warmer and drier climate after 10,000 years ago may have had these effects: 1) elimination of some spawning grounds as water levels fell, 2) decline of runs on smaller tributaries 3) anadromy declined (smolts linger in fresh water), 4) runs were spreadout through the year, but 5) overall productivity would have been low because of general water conditions. One effect of lowered overall water levels during this period may have been increased access to fish in areas where currently fish are not particularly accessible. Thus, while the premier-fishing places, such as The Dalles and Kettle Falls, probably continued to be the major fishing localities, they may have been less important in the regional economy, as fish were more generally accessible. For the cool/wet interval between c. 5000 and 2500 BP, they speculate that salmon runs would have become more like those currently to the north, with increased anadromy, and shorter runs concentrated in the summer time. After 2500 BP, in their model, the salmon resource would have achieved its structure during the Early Modern period (AD 1720 1850).
Fresh water mussels are another class of resource that would have been affected by changes in water temperature and sediment load. Mussels were exploited throughout the Holocene. Leonhardy and D. Rice (1970) suggest that their use expanded for a period in the Middle Holocene (see also Lyman 1980). However, mussels were locally important, probably as sources of animal protein.
The rivers and their tributaries are generally deeply entrenched, and have little scope for lateral movement. Unlike the Mississippi, for example, the Columbia, without significant help, cannot dramatically shift its course. "Significant help" here refers to geological processes on the scale of orogeny and the Missoula flood events. In any case, there are a number of consequences of this: nowhere in the region are there the kinds of broad alluvial bottomlands that are often associated with drainages of the scale of the Columbia's. Areas next to, or close, to the rivers have always been close to the rivers, though their suitability for human use will vary according to what the river is doing at that time. Good access to and from the rivers, and travel routes across the Plateau, have also not likely changed in the last 9500 years (at least until the arrival of the horse).
Another consequence of the general restriction of the Columbia's flood plain is that its annual fluctuation in base level is among the most extreme of the world's rivers (Dietrich 1996), even in nonflood years. The strength and timing of this annual fluctuation has probably been affected by regional changes in rainfall amounts and the timing of rainfall, however.
The potential of the river for navigation has also probably varied with changes in effective moisture. It is not known how important navigation and movement of goods on the rivers was to the early peoples of the Columbia Plateau. However, before the horse, people only had themselves and dogs to move freight of any size and weight. In the early 18th century, as amply demonstrated by Lewis and Clark, it was possible to move people and material rapidly downstream along the Clearwater, Snake and Columbia to the upstream end of the gorge. Thus, variation in water levels could potentially affect trade and exchange at least along the river courses.
Changes in terrestrial environments include not only shifting vegetation patterns, outlined previously, but potentially changes in the distribution of crucial animal resources and turn over in fauna (extinctions and additions of new forms) as well. Zooarchaeologists have looked at both issues for the Holocene on the Plateau, but with limited success (e.g. Lyman 1992, Dixon and Lyman 1996). Archaeological sites provide virtually the only source of data about Holocene animal distributions, and therefore questions about sampling and monitoring position affect conclusions that can be drawn as will be illustrated below in the discussion of bison. Given these strictures, the following very general statements can be made. The period covered in this report post-dates the major Late Pleistocene extinction wave in North America that includes mastodon, mammoth, and other megafauna. This event may have occurred as much as 4000 years before the Kennewick individual's life. However, a number of species did disappear locally before the arrival of Euroamerican settlers on the Columbia Plateau. These include wolves (Canis lupus), grizzly bears (Ursus arctos), bison (Bison bison), pronghorn antelope (Antilocapra americana), and mountain sheep (Ovis canadensis) (Schroedl 1973, Lyman 1985c). The reasons for these local extinctions are not currently known. Other than these species, there was probably no other significant turnover in large fauna during the period covered by this report, with one possible major exception bison. This does not mean that the spatial distributions of animals such as deer, elk, and antelope did not alter during the Holocene. They must have. Bison illustrate the issues and problems involved in understanding changing faunal distributions on the Plateau. It is also a problem particularly germane to interpreting the cultural and economic context of the Kennewick individual. Interpretations about bison hunting revolve around questions of climate changes, hunter-gatherer responses to those changes, but also whether the perceived pattern is merely a sampling problem.
Bison (Bison bison) were probably present on the Plateau throughout the Holocene, but in very small numbers during most of that period. Mack and Thompson (1982) demonstrate that the steppe grasslands of the Plateau could never have supported large herds of bison, like those on the Great Plains. However, bison bones do occur in large numbers in archaeological sites in the period between c. 500 BC and AD 500 (Figure 7). Two additional assemblages are slightly younger (at 45AS80), and one is much older. Large numbers of bison bones were recovered at the Lind Coulee site that is located to the east and north of the Kennewick locality4. The Lind Coulee materials are only somewhat older than the Kennewick individual (see below).
The debate here is whether the changing frequencies of bison bones reflects an actual increase in bison abundance between 500 BC and c. AD 500/ 1000 (e.g. Schroedl 1973, Schalk and Olson 1983, Chatters et al. 1995), or sampling (Lyman 1985c). Lyman (1985c) does present evidence that suggests shifting proportions of pronghorn and bison on the Plateau during the Holocene as effective moisture decreased, then increased. Bison declined as the climate became warmer and drier in the Early Holocene and the amount of grass decreased, while antelope increased. As the climate subsequently cooled and became wetter, grasses expanded as did bison, and antelope declined. However, he concludes by arguing that none of the available interpretations is supportable given the present data.
Hunting and Gathering: While Marshall (1999) has recently argued that Early Modern (AD 1720 1859) Plateau economies were, in a manner, horticultural, over the past 9500 years, they were hunter-gatherer economies. However, hunter-gatherers are extremely variable (Kelly 1994) in time and space in their cultures, subsistence economies, and social organization. During the past 10,000 years, worldwide, they were even more variable than is accommodated by Kelly's book. Archaeological practice on the Columbia Plateau over the past 20 years has approached that variability within the framework of Binford's distinctions between "foragers" and "collectors" (Binford 1980, e.g. Ames, 1988b; Chatters 1984a, 1995; Schalk and Cleveland 1983). These distinctions have proved extremely useful for understanding the histories of hunter-gatherer mobility and subsistence patterns on the Plateau. Anthropologists have also drawn other, somewhat similar distinctions among hunter-gatherers. While these distinctions obscure important and interesting variations, they are also heuristically powerful, providing a language with which to characterize and discuss hunter-gatherers, a basis for building models of ancient hunter-gatherer systems and for making predictions about the archaeological record. The first part of this section summarizes some of the distinctions made among hunter-gatherers useful to this study, and the second briefly discusses their relevance here. These distinctions are based on features common to many hunter-gatherer societies across time and space. Such features are thus most probably analogous ones.
Immediate Return/Delayed Return (Woodburn 1980, 1982; Barnard and Woodburn 1988; Burch and Ellanna 1994): Woodburn (1982, 432) defines immediate return hunter-gatherer systems as having:
(T)he following characteristics. People obtain a direct and immediate return from their labour. They go out hunting or gathering and eat the food obtained the same day or casually over the days that follow. Food is neither elaborately processed nor stored. They use relatively simple, portable, utilitarian, easily acquired, replaceable tools and weapons made with real skill but not involving a great deal of labor.
He adds elsewhere (Woodburn 1982, 434):
The social organization of these societies has the following characteristics:
He also stresses that immediate return hunter gatherers are "egalitarian, profoundly egalitarian." Egalitarian societies are those in which there are as many positions of prestige as there are people to fill them, there are no positions that possess power, and everyone has equal access to resources (Fried 1967).
Delayedreturn societies, in contrast, have these
People hold rights over valued assets of some sort, which either represent a yield, a return for labour applied over time or, if not, are held and managed in a way which resembles and has similar social implications to delayed yields on labour. In delayedreturn hunting and gathering systems, these assets are of four main types, which may occur separately but are more commonly found in combination with one another and are mutually reinforcing:
Woodburn also stresses that these are very general characteristics, and that the form that a particular society will take cannot be predicted. They do provide a basis, however, for developing general expectations about past hunter-gatherer societies. Binford's distinctions between foragers and collectors were developed by him to permit developing very explicit expectations.
Foragers and Collectors (Binford 1980): Binford, like Woodburn, classified hunter-gatherers into two ideal types. However, Binford's types, forager and collector, are based primarily on their mobility strategies: on how they moved themselves across the landscape to get access to resources. However, there is a great deal of overlap between the two systems. Foragers have these characteristics:
A number of expectations follow from these points about the archaeological record of foragers. Suffice it to say here that because they may move frequently, their camps may have low archaeological visibility5, and because the activities pursued at each camp reflect only what people were doing at the camp, there may be considerable variability from campsite to campsite. However, there is likely to have been a range of activities pursued at any camp that can be reflected archaeologically. Binford also identifies a second site type localities where low-level bulk extraction might occur. These would also have low-level archaeological visibility (Binford 1980, 9).
Collectors, in contrast, are logistically organized. They position their residential bases centrally, and exploit resources with task groups that make trips to resource localities, where they acquire whatever resource it is, process it, and return to the main residential base. For example, the residential base may be situated to provide access to fishing localities, berry gathering grounds, and root gathering grounds. Task groups leave the residential base, and go, for example, to a berrying locality. The berries are gathered, smoked and dried there, and brought back to the base camp. Fish are caught and dried at the fishing camp, and the dried fish, placed in baskets, are returned to the main camp. Thus, collectors harvest several resources at the same time, and collector strategies are most common in parts of the world where resource availability is limited to one part of the year, and many resources must be harvested at once. Another consequence of strong seasonality is that collectors often practice storage, and therefore, unlike foragers, they harvest and process large amounts for both immediate consumption and to store.
Binford predicts that collectors will generate a very different archaeological record than that of foragers. Their residential bases will be occupied for longer periods, thus they will generate more debris. There are more likely to be facilities such as houses and containers for stores. A wide range of activities will occur at these sites since they are occupied for longer periods than are the residential bases of foragers. Collector localities are archaeologically much more visible, since they were often regularly reoccupied, large volumes of material harvested and processed and so on. Specialized localities are also expected. Places will be consistently revisited by task groups for the same purpose. Binford suggests there may be additional site types as well.
In many respects, Binford's and Woodburn's types are two sides of the same coin: foragers/immediate return and collectors/delayed return. Both authors stress that many groups are not easily accommodated into these ideal types, having some qualities of both.
Discussion: These concepts will be important both in the presentation of the Plateau's archaeological record and in the analysis that follows. Suffice it to say here that the presence through time of either forager strategies or collector strategies on the Columbia Plateau cannot be taken as evidence for homologies. Resource distributions and seasonality on the Plateau are such that collector systems are to be expected, particularly with population increases. However, homologies may exist in the specific arrangements of particular collector systems through time. In contrast, an archaeological search for homologies in forager/immediate return systems may be quite difficult, since one of their major traits is their fluidity. Any forager system is likely to be very fluid, while different collector systems may potentially organize themselves differently on the same landscape. Of course, such organizational differences through time can be the result of the same group reorganizing itself. At least one researcher, however, has seen such differences as indicating cultural differences (Bettinger 1978). This possibility is discussed in more detail in the next section.
Theoretical and Methodological Approaches to Continuities and Discontinuities: Before the 1960s archaeologists often explained what were seen as discontinuities in the archaeological record as evidence for migrations or perhaps diffusion. Such discontinuities could include the replacement of one style of projectile point with another. After the mid-1960s, migration as an explanatory device passed from favor, until quite recently, though some thought was devoted to it (Rouse 1987). In situ development was generally assumed on the Plateau and elsewhere. Discontinuities were generally taken to reflect problems in sampling, or abrupt changes in organization of some kind. The issue here is not demonstrating migration but the identification of continuities. In short, there is not much literature to guide one in identifying and evaluating continuity and discontinuity in the archaeological record. The literature for the Great Basin region of North America is a significant exception. Archaeologists there have been investigating what they term the Numic expansion (Madsen and Rhode 1994a, b). The correlates they have developed for investigating this hypothetical expansion6 provide a set of expectations of varying strength to examine the Plateau's archaeological record. The Thule expansion in the Arctic provides another useful example. The intention here is not to imply that there may have been population movements onto the Plateau, or that all of the correlates are good ones. Hughes (1992, 1994) offers a number of caveats against linking continuities and discontinuities in the archaeological record as indicating continuities and discontinuities in what he terms enthnolinguistic groups.
Hughes (1994a) observes that archaeologists have long assumed that continuity in material culture in an area indicates continuity of people in that area. By people he means ethnolinguistic groups. He argues, however, that much of the continuity that archaeologists see is in the areas of subsistence and settlement systems, aspects of human life that may have little relationship to language and ethnicity. He suggests that archaeologists need to, in his words, "disentangle, or correctly recognize, those aspects of archaeological assemblages that refer most directly to subsistence from those related to ethnolinguistic identity" (Hughes 1994a, 68). He also suggests that claims that attempt to relate material culture to ethnicity or language are very difficult to disprove, or falsify, i.e. they are hard to test. I agree with these cautions. This section is devoted to these issues.
My purpose here is review the kinds of continuities and discontinuities archaeologists have used to provide a set of measures to use against the archaeological record of the Plateau in order not to approach the Plateau record in an ad hoc manner. Some of these measures are stronger than others. Most, if not all, of these are controversial in the discipline. In this study, they act merely as a checklist of the kinds of discontinuities our colleagues look at. They also make explicit the kinds of evidence archaeologists have used, without having to cull that information from the Plateau literature.
The distributional pattern of Numic or Shoshonean languages across the Great Basin and beyond has long suggested a population expansion. The reality, causes, and timing of this expansion have been major research questions on the Great Basin for some time. A recent edited volume on the topic (Madsen and Rhode 1994) provides a compilation and discussion of approaches to the problem. I will here only summarize the kinds of evidence employed. These can be grouped into classes of evidence. Examples from other regions will also be used.
Abrupt changes in Material Culture/entire subsistence economy: Perhaps the most dramatic archaeological example of the replacement of one hunter-gatherer population with another is actually not the Numic spread, but that of the Dorset culture of the Canadian Arctic by Thule culture around AD 1000 1300 (McGhee 1984, Grayson 1994). In this instance, there are major changes in material culture, including subsistence gear and house form, and in subsistence and economic practices, from the animals pursued to overall economic organization. McGhee also argues that the origin point of the migration is known. In most areas, the alteration seems so abrupt as to indicate population replacement. While McGhee describes the change from Dorset to Thule as swift, there are suggestions of periods of interaction between Dorset and Thule peoples (e.g. Fitzhugh 1994), at least in the extreme eastern Arctic. In McGhee's model, the expansion is fuelled by a climate change allowing Thule peoples to expand their range at the expense of Dorset until most of the Canadian Arctic is occupied by Thule groups. This is followed by a period of adjustment and diversification as Thule people adapt to both their new, local environments, and to broader climatic changes that forced them to alter their original economy in many areas. Thus the Thule expansion does appear to meet Rouse's criteria for demonstrating a migration (Rouse 1987).
In the southeast Great Basin, a mixed farming/hunting-gathering economy is replaced by hunting and gathering, suggesting a population change. In other places, one kind of hunting and gathering economy appears to have been replaced by another. It is this latter aspect of the Great Basin research, in fact, that makes it relevant to the Columbia Plateau. The most thorough model building (but still quite controversial [e.g. Thomas 1994]) for this is that of Bettinger (Bettinger 1978, 1994; Bettinger and Baumhoff 1982). Bettinger's approach, which is rooted in evolutionary ecology, is based on yet another binary distinction among hunter-gatherers: travelers and processors (Bettinger and Baumhoff 1982). In their model, and his subsequent exegesis of it (Bettinger 1994), there are significant differences between these two strategies, involving differences in mobility patterns, frequency and distance of movements, decisionmaking and strategies for which resources to exploit and how (sometimes the same resources, but in different ways), technology and the sexual division of labor. These distinctions are much finer grained than those between forager and collector, although travelers are more foragerlike and processors more collector/delayedreturn like. What matters most here is his argument that a shift from traveler to processor is not an adaptive shift by the same people, either by developing it themselves, or borrowing it. Bettinger asserts that the elements of these strategies are so deeply and functionally intertwined that it would be impossible to borrow or invent just one part of the strategy it is all or nothing. In any case, a wholesale change of subsistence economy, even of one hunting-gathering economy by another, may be evidence for discontinuity, if the change is abrupt and if the two economies seem disparate enough that one does not seem as though it could be ancestral to the other.
There is also an implication here that complex economies replace more simple economies, but not the obverse, though Bettinger certainly has never characterized travelers in that way. He argues only that processing will support more people than traveling. In any case, there is a deep-seated notion that things can get more complex, but not more simple. This is not the case. There are well-documented cases of highly mobile foragers swiftly becoming relatively sedentary collectors, and then, after a considerable length of time, some at least becoming foragers again (e.g. Bar-Yosef 1998), with no accompanying evidence of a major cultural discontinuity.
Another implication of Bettinger's work is that economic differences may almost be what could be termed stylistic differences between societies. He made this argument explicitly in 1978 (Bettinger 1978), but has not returned to it. Other archaeologists in other regions of North America are quite explicit about this (e.g. Mitchell 1988). Such differences can include differences in economic emphasis (terrestrial vs. riverine or marine), differences in artifact assemblages (organizational and/or stylistic), and differences in mobility and land use patterns. Generally, one of these may not be seen as sufficient evidence for cultural replacement, though it can be.
Demographic Fluidity: Models of Thule and Shoshonean expansion generally envision replacement of one human population by another through migration, although both populations may be closely related biologically, as in the case of McGhee's Thule model (McGhee 1984). Simms (1994) describes an alternative, which he terms "demographic fluidity." Hunter-gatherers are well known for their mobility, both as individuals and as groups. Nunamuit males in northern Alaska, for example, may cover an area of 300,000 km2 in their lifetimes (Binford 1983). Kelly (1994) discusses factors controlling hunter-gatherer mobility. Ames and Marshall (1980) described levels of group and individual mobility for the southeastern Plateau. Individuals move according to the information they may have about resource productivity, to find mates, to be closer to friends or relatives, and to get away from friends and relatives. They also travel to acquire information about the environment. Such fluidity may be impossible to detect archaeologically, but it might be reflected in some of the patterns archaeologists have sometimes seen as "influences" from different areas, or as diffusion. Such shifts may also be reflected in changes in settlement densities. Reid (1991a, 1991b), for example, argues that in the southeastern Plateau there have been shifts in settlement densities between the canyons and the uplands because of droughts. The canyons are the center of gravity for settlement in wet periods and the uplands in dry periods. These shifts would not be migrations or transhumant movements, but alterations in the movements of individuals, of what Marshall has called "habitation patterns (Ames and Marshall 1980)."
Inferred stylistic differences in material culture, the organization of technology or in "diagnostic" artifacts: "Style" plays a significant role in discussions such as these. It is a vexed issue for archaeologists, and there is a vast theoretical literature that is well beyond the scope of this study. These debates are generally quite "academic" though they now have important implications for the broader society (e.g. Barker and Pinto 1994).
Bettinger et al. (1996) summarize the style literature very crisply. They suggest archaeologists see three causes of variation in artifact form: 1) differences in the utilitarian functions of tools, 2) differences in form due not to function, but to what they call "rote social learning."7 This can be termed "passive style." Among the peoples of the northern Northwest Coast, for example, a wellcrafted wooden object was often given its finishing touch by lightly adzing the surface, producing long, shallow adz grooves. Among European peoples, a good craftsman finishes a piece by sanding it. These differences reflect the tools and conventions of good work: "This is how it is done." Such differences may even have their roots in different motor skills; and 3) what they term arbitrary, iconic and symbolic traits. These latter are widely seen by archaeologists as a form of information or communication, and it is in that way that I will treat it here. This can be termed "active style."
Archaeologists see active style as communicating social information, including group membership. Such style may be "active" in the sense that stylistic differences can arise out of people actively trying to indicate group differences, or it may be "passive" in the sense that stylistic differences may exist simply because of learning different ways of doing something. Thus, some archaeologists (e.g. Croes 1989, Adovasio and Pedlar 1994) argue that basketry is sensitive to group differences. Different groups of people will make the same kind of basket (a storage basket, for example) but these baskets will differ in many details, not only in decoration, but in the techniques used to make them. They also argue that such differences, which at some level may reflect differences in learned motor habits, can persist for long periods. Basketry, then, can differ in both active and passive styles. However, such assertions can be difficult to prove, particularly in the absence of large historical collections of such artifacts, or in the absence of analysis of historical collections.
Such differences can include then not only the shape and decoration of an object, but how objects are made, and are even sometimes extended to entire technological systems. Microblades have long been regarded as evidence for the presence of Athabascan speakers, in western Canada and Alaska, for example. Sometimes, a single artifact type is thought to indicate socio-linguistic differences (e.g. Holmer 1994).
It is essential to be very clear at this point about how "continuities" and "discontinuities" are understood in this work. After considerable thought, I settled on two basic ways. In the first, there are demonstrated (or demonstrable) continuities or discontinuities in a cultural tradition through time. Cultural continuities, at base, means that cultural transmission of a particular suite of cultural traits from one generation to another was continuous. This does not mean that the cultural traits did not change. The first culture might have been quite different from the last. It means that transmission was continuous from one generation to another. A discontinuity means that one cultural system ceased to be transmitted, and was replaced for whatever reason by another, in Hughes' terms, one ethnolinquistic group replaced another. The replacement of Dorset people by Thule people in the central and eastern Arctic seems to be an example of the latter, while the political development of Great Britain since the Norman Conquest is an example of the former. Change by itself then is not evidence for discontinuity; even abrupt change does not necessarily imply discontinuity. Establishing such continuities would seem to require being able to trace changes in active styles (the decorations on baskets), and to be able to distinguish continuity in passive style (homologies) from continuity in function (analogies).
The second kind of continuity/discontinuity is the presence or absence of gaps, breaks in the chain of evidence. The focus in this report is on the latter. Gaps in the record may reflect cultural discontinuities, or they may not. Lengthy gaps between radiocarbon dates are one example of such a discontinuity, abandonment of a site for a time another. Such gaps are scale dependent. In the process of the analysis of the Hatwai site materials, I observed a gap in the radiocarbon record in the houses, particularly House 1 that was not obvious in the depositional record. The question that followed from this was whether the gap was specific to Hatwai the site was abandoned for several hundred years or more general. I found to my satisfaction that it was more general (the temporal break between early houses and later ones), but many such abandonment events occur only at one site. There are very few sites on the Columbia Plateau which contain the full temporal sequence; the sequence is stitched together from bits and pieces from all over. Therefore, some gaps are to be expected. The entire Plateau has a record of continuous occupation through the entire Holocene, but the central Basin appears to have been virtually unused for a few millennia, the Upper Columbia in the area of the Wells and Chief Joseph Reservoirs have no appreciable record for Period I, and there are gaps after that (between the Kartar [4500 1500 BC] and Hudnut [1500 BC AD 1] phases, for example). Assessing cultural continuities/discontinuities is dependent on having a sufficiently continuous empirical record.
Breaks in the record, discontinuities, may be evidence for discontinuities in the historical connection between ancestral and descent cultures; they may be evidence for cultural change; or they may just be breaks in the record. Endzweig (1994) uses discontinuities in site occupations to argue for shifting settlement patterns in the southern uplands. Presumably, if the break represents a cultural discontinuity, cultural traits will differ on either side. Of course, rapid cultural change will also produce such differences. On the other hand, similar traits on either side do not necessarily mean continuity across the gap in the record. An inference of cultural continuity will hinge on the kind of traits on either side of the gap and on our monitoring position from what vantage point are we seeing the gap. Recalling Swanson's argument for cultural stability in the Birch Creek Valley (Swanson 1972), upland hunting camps of different hunter-gatherer cultures may look very much alike: i.e. that particular set of traits may be quite insensitive to culture change, as were bone tools in Prince Rupert Harbor.
A continuous archaeological record can be, but is not necessarily, evidence for cultural continuity. There are many tells in the Middle East that preserve a record of continuous occupation that can include major cultural shifts and even different human populations. Of course, the archaeological record, although continuous, may be markedly different on either side of that change. In the section on long-term structures, I discussed aspects of the Plateau's environment that could produce continuities regardless (within limits) of who occupied it. A good trail down to the river is a good trail down to the river, especially when everyone is on foot, regardless of ethnicity.
Another distinction needed here is between models and evidence. Chatters and I have presented models to explain what we see as real patterns in the occupational history of the Plateau. However, those models are not the evidence itself. Others (Reid 1991a, Schalk et al. 1998) suggest what we see as a major behavioral pattern to be explained may just be a sampling problem, and therefore not requiring explanation beyond that. Such models are not evidence, though they lead us to focus on particular kinds of evidence as opposed to other kinds of evidence. Most models and explanations are the best we can do at a particular time with the data we have and all will be supplanted in time. They are all wrong in some way. They may also come to take on a reality altogether unjustified by the evidence. Debates over models and evidence are part of the normal research process; they are what lead to increased knowledge. Science requires what may be termed a "positive skepticism" about the knowledge claims of one's peers and one's self.
(1993) illustrates three Early Holocene bison localities, but does not
name them. She also illustrates the wide distribution of late bison
sites in the eastern Plateau. I will give the age range at first mention
of a phase in a paragraph, or when its seems necessary. I will not give
them for every mention of a phase.