Explanation and prediction in vegetation science

A definition of vegetation science is given, spanning 6 levels of integration and stressing the interrelations among them. The problems of realism are discussed. The selection of levels is related to the adequate correspondence between conceptualization and research aims. Pattern and process are introduced as the central concepts of vegetation science. The perception of reality is dependent on the spatial and temporal scale chosen. The concept of noise is discussed in relation to stochasticity and randomness of events. Traces of essentialism are found both in classification of communities and habitat ecology. Classification is important, particularly the coexistence of alternative classification approaches. Organicism as a basis of vegetation research is rejected because the organismic view is inadequate on higher integration levels. The concept of function is redefined in a non-teleologic way.Present vegetation ecological research is inductivistic. One possible alternative to inductivism is falsificationism. The major domain of this approach is hypothesis testing, which will become more important. Progress can only be reached by a maximum degree of communication among scientific individuals.Predictive ecology is partly based on historic explanation, partly on complementary approaches. Characters of vegetation worthwhile to be predicted are listed and the necessary requirements for vegetation science to become predictive are discussed. A major requirement is the development of succession and life-history theory. A further elaboration of the individualistic concept will be a main task of vegetation science in the near future.