#3: The perils of basin bias
On attractor basins, stability landscapes, and the tools we have to map them.
I’m Luke Craven; this is another of my weekly explorations of how systems thinking and complexity can be used to drive real, transformative change in the public sector and beyond. The first issue explains what the newsletter is about; you can see all the issues here.
Hello there, reader,
Complex systems tend to oscillate within steady states, which are called "attractor basins". The various basins that a system may occupy, and the boundaries that separate them, are known as a “stability landscape”. Most attempts to map a complex system do so from within particular basins (what I call "basin bias") rather than across the landscape. We need to challenge basin bias in our work to help achieve meaningful systems change.
Complex systems are non-linear, dynamic, unpredictable, but they also tend to settle down into one of several possible steady states. These steady states are called “attractor basins”. Think of a bowl containing a ball bearing. The ball bearing will move around the bowl until eventually it comes to rest at the lowest point. We can say that it is “attracted” to that point, so each part of the bowl can be regarded as leading to that stationary point, and the whole bowl is what we call the basin of attraction of that system.
In contrast to the ball bearing, social and economic systems are continuously buffeted by the decisions of actors, which tend to move the system away from that stationary point. These systems often shift and oscillate within a given basin of attraction (what we call “equilibrium”), rather than tending directly toward an attractor. There is often more than one basin of attraction for any given system. The various basins that a system may occupy, and the boundaries that separate them, are known as a “stability landscape”.
In this graphical representation of a “stability landscape” [source], green depressions represent regions of stability (the basins). Two system states are represented. The one on the left is disturbed and returns smoothly to its original position. The one on the right amplifies the displacement, either returning eventually to its original position, or possibly transitioning to another basin, or alternate state.
“Basin bias” is where attempts to understand a system do so from within particular basins rather than across the landscape. Take different forms of systems mapping, for example. Causal loop maps tend to ask questions like “what drives obesity?” or “why don’t people have access to decent housing?” which focus on single attractors (obesity, homelessness, etc.) within a system. Asking questions like these is a powerful way to uncover the forces that drive a system toward a single attractor, but they teach us very little about the other steady states that are possible within a stability landscape; the boundaries between those states; or how the landscape itself could be destabilised to create new possibilities. Even when causal loop maps do justice to the basin of interest, the policy response is commonly to nudge the trajectory of the ball, rather than to attempt to reshape the basin. The system absorbs the disturbance of that new trajectory and reorganises while retaining essentially the same constraints. The landscape remains unchanged and the system oscillates within an established equilibrium.
Basin bias is particularly nefarious because it often occurs alongside the perception that systems mapping is the road to enduring, transformational change. Making change off the back of a detailed systems map makes us feel like might actually achieve something more than a “band-aid solution” when really, we are only temporarily disturbing an established equilibrium. Don’t me wrong, there is value in producing detailed systems maps that attempt to understand why something (obesity, homelessness, etc.) happens in a system, but this is quite different to mapping a stability landscape. And, without a more detailed landscape view, it is hard to decide whether efforts to change a system should occur within the stability landscape of a system or to the landscape itself. Of course, it is impossible to take the entire landscape into account, but that shouldn’t stop us from trying to lift our heads above the parapet of individual basins. Seeing and imagining different landscapes can only enhance our ability to make possible different trajectories for particular systems.
Basin bias, to some degree, is an unavoidable product of the way we experience the world. But keeping ourselves conscious to it and its impacts is an important part of being an effective systems change practitioner. We can start by embracing some simple habits:
Think critically about the boundary of analysis. Does it match the question being asked? Do you want to understand the forces driving a system toward a single attractor; the different states possible within a stability landscape; or what could shift a landscape to help new possibilities emerge? Each of these questions requires a very different boundary to be drawn to capture the relevant dynamics and constraints.
Search for patterns of equilibrium. There’s truth in the old saying that the more things change, the more they stay the same. Identifying these circular patterns in a system is one way to make visible the different basins in the landscape and the thresholds between them. Sharing what we discover about these patterns helps to build collective knowledge about what is possible and how we can nudge the system in different directions.
Don’t settle for nudges. Small change can have a large impact, but more commonly, it doesn’t. The more we focus our attention on shifting the system within established patterns, the less effort we can dedicate to looking across the entire landscape and dreaming big about what else is possible. This doesn’t mean change that occurs within a single basin can’t play an important part in reshaping a system, but it is very unlikely to be the route to enduring, transformational systems change.
Not unrelated miscellany…
The initial parts of this issue were drawn from an earlier piece I wrote for the Centre for Public Impact, which reflected on how COVID-19 highlights limitations in language used to express change, especially in complexity and systems practice.
A “strange attractor” is a driver of change in a complex system that can cause it to shift on a radically divergent trajectory. Referenced heavily in Tyson Yunkaporta’s work Sand Talk, they have popular resonance through the well-known metaphor of the butterfly effect (Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas?).
The human mind struggles to think in greater than three dimensions, which can make it challenging to conceptualise a multidimensional stability landscapes (like the one in the gif above). In evolutionary biology, where “fitness landscapes” are used to are used to visualise the relationship between genotypes and reproductive success, there has been a push for the seascape metaphor to displace the more static landscape concept.
By the way: This newsletter is hard to categorise and probably not for everyone—but if you know unconventional thinkers who might enjoy it, please share it with them.