Linearity enigmas in ecology

Abstract Two issues concerning linearity or nonlinearity of natural systems are taken up. Each is related to one of the two alternative defining properties of linear systems, superposition and decomposition. The superposition property exists when a linear combination of inputs to a system results in the same linear combination of outputs that individually correspond to the original inputs. To demonstrate this property it is necessary that all initial states and inputs of the system which impinge on the output in question be included in the linear combination manipulation. As this isfficult or impossible to do with real systems of any complexity, nature appears nonlinear even though it may be linear. A linear system that displays nonlinear behavior for this reason is termed pseudononlinear. The decomposition property exists when the dynamic response of a system can be partitioned into an input-free portion due to state plus a state-free portion due to input. This is a characteristic of all linear systems, but not of nonlinear systems. Without the decomposition property, it is not possible to distinguish which portions of a system's behavior are due to innate characteristics (self) vs. outside conditions (environment), which is an important class of questions in biology and ecology. Some philosophical aspects of these findings are then considered. It is suggested that those ecologists who hold to the view that organisms and their environments are separate entities are in effect embracing a linear view of nature, even though their belief systems and mathematical models tend to be nonlinear. On the other hand, those who consider that the organism-environment complex forms a single inseparable unit are implictly involved in non-linear thought, which may be in conflict with the linear modes and models that some of them use. The need to rectify these ambivalences on the part of both groups is indicated.