THE FEEDBACK

By Trey Boone
Department of Philosophy
University of Virginia

Abstract
Functional robustness refers to a system’s ability to maintain a function in the face of perturbations to the causal structures that support performance of that function. Modularity, a crucial element of standard methods of causal inference and difference-making accounts of causation, refers to the independent manipulability of causal relationships within a system. Functional robustness appears to be at odds with modularity. If a function is maintained despite manipulation of some causal structure that supports that function, then the relationship between that structure and function fails to be manipulable independent of other causal relationships within the system. Contrary to this line of reasoning, I argue that functional robustness often attends feedback control, rather than failures of modularity. Feedback control poses its own challenges to causal explanation and inference, but those challenges do not undermine modularity—and indeed, modularity is crucial to grappling with them.

Robustness and Modularity was recently published in the British Journal for Philosophy and Science

Commentary from Trey:
Much of my research is focused on understanding causal complexity. One notable feature of complex biological systems, which is particularly salient in neuroscience, is that such systems are remarkably stable over variation in internal and external conditions. For instance, individual neurons with nearly identical electrophysiological profiles can have manyfold variation in the ion channels within their cell membranes. This between-cell variation is surprising because those ion channels quite literally determine those cells’ electrophysiological properties. And yet this sort of stability is not unique to neurons; similar forms of functional stability can be seen across all levels of investigation in neuroscience, from small neural circuits to networks of neural regions. This robustness (or degeneracy) within the brain poses unique challenges to ordinary methods of causal investigation.

The typical way to unearth a causal relationship between some potential causal factor and some outcome of interest is to perform a controlled experimental manipulation (sometimes called an intervention). That is, to establish that some variable A is causally relevant to some variable B, we generally aim to hold everything else fixed and manipulate A to see whether a change occurs in B. Robustness thwarts this sort of causal investigation because experimental manipulation of causally relevant factors may fail to produce any change in the outcome of interest. Note, per the example above, that manipulating the conductances of different ion channels in individual neurons may fail to change the electrophysiological profiles of those cells despite those ion channels being causally relevant to those electrophysiological profiles.

This aspect of robustness has led some to argue that complex biological systems fail to be modular in the sense of having isolable, or independently disruptable, causal relationships. Modularity, in this sense, is a foundational concept for interventionist notions of causation. Thus, those who argue that complex biological systems fail to be modular have argued in kind that interventionist notions of causation are inadequate in biology and neuroscience and that new concepts of causality are needed to understand the complexity of biological systems.

In Robustness and Modularity (BJPS), I argue that the robustness of neural and other biological systems can in fact be reconciled with interventionist accounts of causation. Using a detailed example of individual-neuron robustness in Purkinje cells, I show that disentangling the complex causal structures that enable robustness in fact requires that those structures be treated as sets of modular causal relationships. The appearance of violations of modularity is instead generated from differences in the timescales on which experimental manipulations operate and the timescales on which the compensatory changes that enable robustness take place. As such, robustness should not impel us to adopt radically different concepts of causality, but instead should impel us to incorporate timescale sensitivity into our understanding of both the metaphysics and epistemology of causality.