cial agency was already well organized by the end of the 1940s. Through the strong presence of this agency, wildlife management was centralized early on, even in northern Ontario, which is occupied predominantly by aboriginal groups. By contrast, in Quebec, the government management agency was only weakly present in the north, even as late as the 1970s. Perhaps as a result, local institutions for wildlife management were strongly present in the Cree areas of northern Quebec as late as the 1980s and effectively managed wildlife (Drolet et al., 1987). By contrast, in the Cree areas of northern Ontario, such institutions were almost nonexistent, presumably because they had been swept away by centralization (Berkes et al., 1991).
The replacement of local institutions by centralized ones often involves a change in the way knowledge is used for management. Local institutions tend to use their own folk knowledge, often referred to as local knowledge, indigenous knowledge, or traditional ecological knowledge, whereas centralized management agencies tend to use internationally accepted scientific practice and often assume away local knowledge and practice (Berkes, 1999; Williams and Baines, 1993). The shift of knowledge systems is one of the major impacts of government-level institutions on local institutions because it is often accompanied by a change in control over a resource. The differences between the two systems of knowledge can be substantial in the way resources are viewed.
A case in point is caribou management in the Canadian North (Berkes, 1999). A number of studies indicate that aboriginal hunters from the Arctic and the Subarctic monitor caribou distributions, migration patterns and their change, predator presence, individual behavior, sex and age composition of the herd, and fat deposits in animals. The Western science of caribou management also monitors much the same things, but there is a fundamental difference: decision making in scientific management is based primarily on population models. The aboriginal system, by contrast, is based on local observations and ethics, assumes that caribou are not predictable or controllable, and does not try to use harvest or population size estimates. Rather, it pays relatively high attention to fat content (an excellent integrative indicator of caribou health) and uses a qualitative mental model that provides hunters with an indication of trends over time.
This qualitative model reveals the direction (increasing or decreasing) in which the population is headed, without requiring the estimation of the population size itself (Berkes, 1999). This locally developed, aboriginal approach to management has potential to result in good resource management, but it is different from scientific management. Centralization of management leads to a shift in the knowledge system used. Government management of resources, based on universal science rather than on locally developed knowledge, undermines the knowledge systems, as well as the institutions, of northern aboriginal groups. Hence, the centralization of resource management and the assertion of “government’s science” becomes a political tool for the control of the local indigenous populations (Freeman, 1989).