A richer view of complexity
An organisation's position within its environment can matter just as much as its own complexity or simplicity.
To recap previous discussions, it has been established that systems have components, a boundary, and sit within an environment.
We also intuitively understand that systems vary from purely mechanistic, with predictable and observable causes and effects (such as a ball rolling on a pool table after being struck by a cue), through to highly non-deterministic systems with multiple independently acting agents where moment to moment cause and effect is difficult or impossible to establish or observe.
Once a system has polycyclic feedback loops—multiple cycles of interacting components that overlap, share components, or influence each other—the system’s responses become increasingly nonlinear, delayed, and emergent. This is the most common configuration of a complex system. If the system also includes multiple agents that adapt and learn, it falls into the subcategory of a complex adaptive system (CAS).
(Aside: Philosophically, it is my view that quantum indeterminacy makes all complex systems truly non-deterministic, but even if free will is a fallacy we can still stipulate that the limits of human observation make such systems impossible to forecast.)
Critically for our purposes, all organisations are both systems and units that operate within a system. As shorthand, we can term the portion of the environment that exerts detectable system effects on an organisation as its domain.
The boundary of an organisation acts as a transition threshold, where behaviour observed inside can be more non-deterministic than that observed outside, and vice versa. It is therefore meaningful to investigate the characteristics of any organisation, as well as its domain, using the same characteristics set out previously:
All of these characteristics operate on a continuum, and the espoused characteristics of an organisation or its domain may also be meaningfully different from its true characteristics.
While there are impacts of variance within each characteristic that require specific considerations, just looking at a two-dimensional matrix that positions organisations within a complexity landscape is instructive. We do this by (for both the organisation and its containing domain) scoring them on a complexity gradient with designed (traditional) systems at one extreme and natural (CAS) systems at the other.
Consider the examples in the table below:
As illustrated, a university faculty overtly encourages freedom and exploration in research within a highly structured world, whereas a commercial airline aims to deliver predictable service in a domain buffeted by forces outside of its control. Both are designed to harness and respond to complex behaviours but in very different ways.
In a natural environment, we humans exhibit self-organization, innovation, and resilience and adapt, learn, and evolve over time. Things like consumer buying patterns are not dictated by any top-down decision making organisation but can still be observed and interacted with in the aggregate. This is where the language of attractors and fitness landscapes comes into play.
On the other hand, we can also voluntarily participate in highly designed organisations with explicit goals, an important aspect of which is an agreement to act predictably in various ways. Through cooperation, we can also produce valuable outcomes that would never be expected to arise spontaneously.
Both types of system are needed to make sense of and act with purpose in the world. As sentient beings, we naturally and fluidly switch between teleonomous and teleogenic systems every day—sometimes explicitly organisational, sometimes not—and often do not even notice the switch.
Consider a mundane act like driving to the shops to buy some bread and milk. During this time:
I left my house in my car—both the result of teleogenic efforts, but bought through suppliers who started up businesses in a teleonomous free market
I drove on the left-hand side of the road (as in Australian) through a teleogenically orchestrated and enforced effort by the government over many decades, but the specific traffic patterns at a particular day and time are entirely teleonomous
I arrived at my local bakery and queued in a de facto teleogenic system to achieve fair service outcomes, arrived at through teleonomously derived cultural norms
The bakery itself uses highly structured teleogenic systems in response to teleonomous consumer supply and demand patterns
Modern economic theory has established more or less definitively that Smith’s “invisible hand” is the best way to match supply and demand in most contexts, with any command and control attempts simply unable to sufficiently constrain the natural forces in play. As the world has become more entwined and more complex, ignoring the rising tide of complexity inside organisations seems increasingly Canute-like.
In these circumstances, the enhanced adaptivity of CAS systems has become an attractive recommendation to managers, and has been a major reason for the rising popularity of disciplines like systems thinking and complexity science.
However, there has also been a tendency to throw the baby out with the bathwater. There are two opposing questions which need to be asked:
If we encourage a system to organise teleonomously, are we confident that the fitness landscape could lead to naturally arrived-at outcomes that are acceptable?
If we were to design a teleogenic organisation with explicit goals, do we have sufficient power to control and collapse feedback loops to create predictable results?
In many cases, only one of these answers can be answered in the positive, which makes the necessary management approach apparent. If both are positive or both are negative, it will likely be possible to reframe the scope of the system or domain to hone in on where the limits of control lie.
While this dichotomy is essential to understand when planning, it is also not black and white. Teleogenic systems can leave room for emergent outcomes, and teleonomous systems can be significantly affected through changing penalties and rewards for behaviour.
The question then becomes: Which is dominant, and how do we respond? Which we will look at in future posts.