Two kinds of perspectives governing the provision and preservation of knowledge, a universal and an ecological perspective, are discussed in this paper. In the first case, scientific observations are represented through a semantic interpretation of facts. This is illustrated with a series of experiments on semantic feature perception in the recall of pictures described by descriptor terms. In the second case, scientific observations are conceived as part of a constitutive context. Consequently, the researcher's development toward structurization and his use of precise concepts and well-defined conceptual relations requires an orientation toward the cooperative dimension of the context. This is illustrated by an ecologically oriented study of an information system that measures intentionality and orientation by means of an Agent-action-Objective (AaO) formula. Within the AaO formula a dependency between the cooperatively operating and interacting components creates the absolute terms (i.e., the invariants); these, in turn, serve as a point of departure for information synthesis. It is demonstrated that scientific reporting has the intermediary function between context and knowledge representation. Intentionality and orientation are fundamental in the development of concepts and conceptual relations. (21 references) (Author/DB)