HUMAN IMPACTS ON THE ENVIRONMENT: UNPREDICTABILITY AND THE PRIMACY OF PLACE

Abstract

Landscapes, environments, and ecosystems (i.e., places) are historically and spatially contingent and therefore unique. Generalizations and laws are based on “all other things being equal.” Owing to historical and spatial contingency, all other things are never completely equal and often do not approach that ideal closely enough to allow for reliable prediction. The problem is illustrated by examining effects of grazing and fire suppression on transitions between grass and woody vegetation communities. In three different environments in the southern United States (central Texas, southwestern Virginia, and eastern North Carolina), three different relationships exist between grazing and fire regime, and vegetation. Other examples are common in the geographic and environmental literature. This suggests that the search for a single applicable generalization or predictive model of a given human-induced environmental change that can be applied without accounting for the particular characteristics of places may be futile. Several general ways to address this issue are suggested. First is the recognition that place matters. Conditions specific to a location or region will influence what, how, or even if generalizations apply. Second is the adoption of a “demographic” approach. This is based on the idea that for places, like people, it is possible to make aggregate or probabilistic generalizations or predictions, but these cannot be applied to deterministic prediction of individuals. Third is a synoptic approach based on forecasting not via simple cause-effect relationships, but based on a typology of situations that recognizes and incorporates the unique characteristics of site and situation. Several converging threads of inquiry in geography and geosciences explicitly or implicitly address the issue of historical and spatial contingency. [Key words: landscape, environment, place, prediction.]