Homogeneity Constraints and Reasoning about Complex Mechanisms (Presented at the Society for Philosophy of Science in Practice, Toronto, June 2013)

Daniel C. Burnston
Department of Philosophy, Center for Chronobiology, and Interdisciplinary Program in Cognitive Science
University of California, San Diego

One vital function of scientific diagrams is to constrain reasoning about the represented domain. Constraints aid in problem solving by limiting search space, and by affording particular hypotheses about the specific system in question. Here, I discuss one particular type of constraint, which I refer to as “homogeneity assumptions,” that plays a role in diagrammatic representation and problem solving in complex scientific domains. Representations of homogeneity assumptions can involve both functional and structural depictions that connote the uniformity of a part type within a mechanism—they can assume that individual components of a particular type are compositionally similar (at a relevant level of abstraction) and/or contribute in similar ways to the operation of the mechanism. Several diagrammatic techniques are used to connote these types of homogeneity, including specific spatial grouping and the repetition of icons intended to characterize the function or structure of a type of part. Homogeneity assumptions can be depicted for different levels of mechanism organization, and help to shape the reasoning of scientists, both within an individual and within a field. I assess three distinct but related roles for representations of homogeneity assumptions in the context of the mammalian suprachiasmatic nucleus (SCN), a structure in the hypothalamus that serves as the central timekeeper for the organism, synchronizing other rhythmic biological processes to time in the external environment.

In mammalian chronobiology, homogeneity assumptions play at least three distinct roles. First, they focus search in a way that directs attention to discovering the capacities and operations of a part of a mechanism. To illustrate this, I will analyze diagrams from a particular period of research in which scientists attempted to uncover the propensities of independent but connected intracellular oscillators within the SCN. Second, homogeneity assumptions can “background” (in the sense articulated by Griesemer) details about a particular part of a complex process, in order to focus attention on the details of other aspects that are “foregrounded.”  I illustrate this via diagrammatic depictions of “peripheral” processes in mammals that are regulated by the SCN. Third, homogeneity assumptions allow for questioning relationships between particular possible functional decompositions of a system. I illustrate with the visual comparison of peptide expression and Period gene expression in different parts of the SCN. Importantly, homogeneity assumptions have different lifespans. The first use can last for numerous years within a field, while the other uses may serve to focus attention through only the course of a single study or reasoning process.

Finally, I will provide some examples of contexts in which questioning standard homogeneity assumptions served to promote conceptual change. Unsurprisingly, diagrammatic practices show noticeable changes as homogeneity assumptions come to be questioned, and I will show examples of this change in diagrammatic practices for representing SCN organization. I conclude that (i) homogeneity assumptions, and their diagrammatic representations, play important roles in cognizing about biological phenomena, both in an individual and a field, and (ii) that analysis of diagrams provides perhaps the most fruitful way to understand developments in these assumptions.