Abstract

Reductionism in science is a viable program, indeed a vital perspective. A realistic and functionalist, as opposed to a reconstructionist, approach to reduction in practice reveals much about scientific reasoning and theory development that is obscured in the traditional reductionist/anti-reductionist debates. Instead, I analyze reductionism as a term used to describe scientific reasoning, and as a concept functioning to guide scientific research, an approach which allows the influence of reductionism to be apparent in the formulation of questions, with presuppositions of parts and mechanisms, and in the explanations evolved from theories developed to answer those questions. I further describe how the theory develops as a series of models, each articulated by means of reduction-promoting heuristics of theory development. This laminar theory structure offers both a method for generating particular microstructural accounts to be related to macroscopic phenomena and a means of generalizing the resulting explanation to apply to various specific systems. I suggest that the development of a theory proceeds from the first, most general, conceptual model, which I call a "C-model", since it demarcates a class of possible processes underlying the phenomenon. Then the C-model is further specified to describe a process model. The process model is developed through explorations of one or several particular models . The models of the series are explanatorily developed by specifying their elements for the biological system and articulating the relations. I show that reductionism is also influential in the heuristics that guide the specification in a reductionist direction by suggesting simplifications in the physical interpretation and representation of the system under study. The reductionist heuristics can be seen even in the development of ostensibly non-reductionist theories, such as Wolpert's positional information. As extended examples, I analyze Wolpert's positional information concept in terms of reductionist heuristics, and the Bryant, Bryant and French polar coordinate model in terms of the articulation of a series of models. My development of functionalist reductionism allows for partial and point reductions, rather than demanding global reductions. When a series of models has been articulated and process described that accounts for the phenomenon, so that a request for explanation can be answered, then the theory treating the phenomenon is reduced at the point of inquiry. The explanation, supported by the evidence for the successive steps of specification of the articulated layers of models, is a point-reduction of the theory