Systemic Analysis and Functional Explanation: Structure and Limitations

Abstract

The systemic approach to function proposes to ground every use of functional language in biology in a systemic analysis. A systemic analysis is a kind of explanation or procedure that parts from a capacity of a system and shows how this capacity is exercised by the system by decomposing it into other sub-capacities that are implemented in the (parts of the) system. Our goal in this work is to examine the adequacy of this proposal. We provide a formal reconstruction of systemic analysis based on the structuralist metatheory. This will help us clarify its structure and explicate its key concepts, which will be useful for a fine-grained conceptual discussion in the two sections following it. After this, we examine the question of whether systemic analysis can account for every use of functional language in biology. We argue that, in order to be biologically significant (to avoid the charges of promiscuity frequently raised against it), systemic analysis as a whole requires that there exist some criteria for distinguishing biological functions that are independent from it. In order to better discuss the relationship between systemic analysis and functional explanation, we focus on the issue of the explanandum of functional systemic explanations. We argue that in systemic analyses that have biological functions as their top-level capacities, these functions explain the structure of the traits that exercise them, as is usual in functional explanations. We show that this fact has not been appreciated for two reasons. Firstly, because the order of determination is reversed from the usual cases, in the paradigmatic examples that systemic proponents had in mind (i.e. in the neurosciences, one typically first has access to the function and then seeks the traits that implement it) and, secondly, because the proponents refer to the top-level capacity as the explanandum capacity (since the components and sub-capacities also explain how it is exercised).