Dipping a toe into complexity theory and methods.

I’m reading a fascinating, dangerous, and difficult book — Complexity Theory and the Social Sciences: The State of the Art, by Byrne and Callaghan. The book is fascinating because it starts with ontology — what is nature of the world — works through epistemology — how do we come to know the nature of the world — and moves to ‘praxis’ — applying theory or knowledge in the world.

Their position is that the world is complex, that understanding complex systems requires methods that fully account for this complexity, and (here comes the danger) that, because creating understanding necessarily alters the system itself, all (social) science is activist.

The difficult part is unpacking all of this: Understanding the world as complex doesn’t mean that it’s impossible to make sense of or that it is chaotic. It means that the reality of systems is that of nonlinear, recursive cause-effect relations, emergence, autopoiesis (loosely, self-production), and — when humans are involved — agency (people and groups of people want things and act on those wants, although almost never in unanimity).

It gets even more difficult because these have implications. Agency begets structure (people act to organize future situations and resource distribution) and structure influences agency (where someone/something is in the system affects what and how much they can do). By extension, then, not only do agents within the system interact with one another, but the parts interact with the whole and vice-a-versa. What’s more, systems interact with one another. And if that’s not enough, there are systems within systems (people within teams within organizations), system boundaries are less about differences between systems and more about how systems are connected, and systems ‘interpenetrate’ (think Venn diagrams in many dimensions).

What does this mean for our understanding of the world? To begin with, it challenges notions of simple, unidirectional causality, of the existence of general laws (of generalizability itself), of our ability to predict, and the role of experimentation (of deduction based on manipulation). Instead, we are left with contextual, bounded, constructed knowledge (small k). Constructed knowledge? Indeed: Knowledge is created through a ‘diaological’ process in which scientific and system knowledge is shared between those seeking to understand the system and those living within the system — it is through co-production that knowledge is created.

How to gain understanding? Byrne and Callaghan provide a general map. Through qualitative case analysis, cataloguing differences and similarities, and tracing the historical development of the system(s) of interest through narrative, one can arrive at an understanding of some of the forces that have shaped the system as it stands today (a ‘retroductive’ understanding). This can be represented as a phase space through time with associated control parameters (an analogy borrowed from physics).

Essentially, this means that through dialogue and co-production, scientists join with those within the system to come to a shared understanding of what the system is like, how it got to be that way, what a set of possible futures might look like, and how different forces might combine to realize these possible futures.

And here comes the (dangerous) praxis. Acting to realize a particular future is a function of individual and collective agency. And because scientists are part of constructing understanding and therefore expanding and elucidating the set of possible futures and their control parameters, they are inexorably part of system change.

For an applied social scientist like me, this is danger more in the sense of whitewater kayaking than in the sense of bear attack — an invitation to join with the system as it hurtles forward, not lock up my food and stay away.