It was at the Primate Behavioral Management Conference that I had the opportunity to hear Brenda McCowen speak. Her presentation, “Using systems network analysis for understanding complexity in primate behavior management,” outlined a way of thinking about the world that instantly resonated with me. McCowen studies complex systems through network analysis—a method she contends is crucial to understanding the natural world.
The idea is that, essentially, we can only accurately study intricate systems when all of their components are in place. Many systems consist of not only direct connections or chains of events, but indirect connections that are lost if components are removed from their places in a system. To make an overly simple analogy, take a landline telephone. To understand how a telephone works, you can unplug it and take it apart. You will notice many things about the structure of the telephone, and be able to deduce many of its capabilities, but by removing it from the telephone line, you separate it from perhaps the most important aspect of its function: the network of other telephones it is connected to.
It is easy to see the value of studying biological processes in this way. By leaving the components of systems in situ, you can collect much more data: rather than focusing on the properties of a series of isolated individuals, you can take stock of all of the individuals along with the direct and indirect pathways between. This type of analysis is useful for studying ecosystems, which have many direct and indirect links between organisms, and for social groups for the same reason.
An additional benefit of this kind of analysis is that new information can emerge. Centrality, for instance. Centrality measures the amount that an individual component or node connects others. This kind of information can be critical for primate behavioral management.
Say you need to remove one female rhesus macaque from a social group. Common sense would indicate that you not remove the highest ranking female, because by rank she is most important to the dominance hierarchy, nor the lowest ranking female, because she might be more stressed than other monkeys in the social group, and therefore less resilient to change. So you might choose a middle ranking female: not too important, not too stressed. But shortly thereafter, you find that there has been a huge fight, resulting in a complete restructuring of the previously stable social group. What happened? It could be that you removed a node of high betweenness. She was important not for her rank, but for her role: her centrality connected two subgroups within the larger group. Without her, the larger group lacks cohesion, and splinters.