6. Dynamics, Feedback, and Homeostasis
Source: ebook ch. 10–11 (“Systems Dynamics”, “Homeostasis”); slides 18–19 (“Nonlinear Paradigm”, “Nonlinear Thinking”).
Every chapter so far has treated systems relatively statically. This chapter introduces how
systems change over time — the Linear → Nonlinear paradigm shift previewed in
01-worldview-and-paradigm.md.
System dynamics
System dynamics is the branch of systems theory that models and understands the dynamic behavior of complex systems — dealing with internal feedback loops and time delays that affect the behavior of the entire system. First developed by Professor Jay Forrester at MIT as a management method, it has since been applied to systems from earth science to the economy to political regimes.
Analytical thinking sees the world in terms of linear cause and effect. Systems thinking instead looks for the interplay between elements — the feedback loops through which elements are interconnected in producing a joint outcome.
“The world is made of circles and we think in straight lines.” — Peter M. Senge
Feedback loops and causal loop diagrams
A feedback loop can be defined as a channel or pathway formed by an effect returning to its cause, generating either more or less of the same effect. Example: a dialogue between two people — what one says now affects what the other says, which in turn feeds back as input to what the first will say in the future.
Not every effect loops back, though: an effect that lands on another agent or system without ever returning to its own source is an externality, not feedback. Feedback is self-regulating by construction — the system that caused the effect is also the one that feels it come back. An externality is not self-correcting in this way and typically requires some outside governance (regulation, a shared norm) to be accounted for at all — pollution costs borne by a community downstream of a factory are the classic example.
System dynamics uses causal loop diagrams to represent this: a simple map of a system with all its constituent components and their interactions. By capturing interactions — and therefore the feedback loops between them — a causal loop diagram reveals the system’s structure. By understanding not just the structure of these relations but also their nature, it becomes possible to model and simulate a system’s behavior over time.
Feedback loops come in two kinds:
- Positive feedback loop: values associated with the two linked nodes change in the same
direction — if one decreases, so does the other; if one increases, so does the other.
- Example: economies of scale between a business and its customers. More products sold → more revenue → more investment in scaling production → lower costs → more customers purchase → (repeat). This is a virtuous cycle where one party’s gain is also the other’s.
- Positive feedback loops cannot go on forever — they are typically associated with unstable processes likely to crash at some point (e.g. a financial bubble that grows exponentially, then crashes). The named mechanism behind this crash is overshoot: because stocks and flows respond with a delay (see below), a growing quantity can keep drawing down a slow-replenishing stock past the point the stock can actually sustain — a boom that outruns its own resource base, only for feedback to catch up once the deficit becomes unavoidable (a town’s population booms after a gold discovery, then collapses once the gold runs out).
- Negative feedback loop: the two linked nodes change in opposite directions — if one
increases, the other decreases, and vice versa.
- Example: predator–prey dynamics. More predators → fewer prey → (via reduced food supply) fewer predators → more prey → (repeat).
- Negative feedback loops are typically associated with an overall stable and sustainable pattern of development — a wave-like graph, bounded within upper and lower limits, with relatively smooth fluctuations over a prolonged period.
Stocks and flows
For a more detailed, quantitative analysis, a causal loop diagram is transformed into a stock and flow diagram, useful for studying a system quantitatively, typically via computer simulation.
- A stock is any entity that accumulates or depletes over time — a simple variable measured as a quantity. Example: a water reservoir, measured by the volume it contains.
- A flow is the rate of change in a stock — measured over an interval of time (like electrical current, telling us how fast something is flowing). Example: a tap on the side of the reservoir, pouring water out.
A quick test for telling the two apart: if time stopped, would it still exist? A stock (an accumulation) would — the reservoir’s water is still there. A flow (a rate) would not — “the water is currently draining” is meaningless without time passing.
“Systems thinkers see the world as a collection of stocks along with the mechanisms for regulating the levels in the stocks by manipulating flows.” — Donella H. Meadows
Nonlinear thinking, in short
| Linear thinking | Nonlinear thinking |
|---|---|
| Events are the result of simple linear interactions; for every effect we search for a mediate cause | Events are the product of a complex of interacting parts, where relations are often cyclical, with feedback loops |
| Analytical thinking’s default mode | Systems thinking’s default mode — think about the feedback loops in the system and cyclical causation |
Homeostasis
Many systems require both a continuous input of resources from their environment and the
capacity to export entropy back to it, in order to maintain a specific level of functionality
(see also 07-energy-entropy-efficiency.md). A tractor needs
periodic fuel input and must export heat and gases; a business needs continuous revenue and must
externalize waste material.
Homeostasis (Greek homos, “similar” + stasis, “standing still”) is the state in which a system’s internal variables are regulated so that they remain stable and relatively constant, despite changes in the system’s external environment — its “normal” or equilibrium state. (It’s worth flagging early that equilibrium is not the only game in town — see the note on attractors at the end of this chapter.)
To maintain homeostasis, a system needs a regulatory mechanism, also called a control system, which regulates both the system’s internal and external environment to ensure conditions stay within the parameters needed for the system’s internal processes to function. Any control system can be broken into three parts: a sensor (measures the relevant variable), a controller (compares the measurement to the target and decides what to do), and an actuator (carries out the action). Cybernetics (from a Greek word meaning “to steer or guide”) is the area of systems theory that studies these regulatory mechanisms — designed to guide the system toward the environmental parameters best suited to maintaining homeostasis.
A control system can only regulate what it can distinguish. Requisite variety (Ashby’s Law) states that a regulator must have at least as many internal states as the states it needs to control in its target system or environment — a thermostat with only “on” and “off” cannot regulate a target that needs five distinct temperature bands. Undersupplied variety is a common, nameable failure mode of control systems generally, not just thermostats.
The homeostatic control loop
- If the system is within its homeostatic parameters, it simply continues its previous course of action.
- If one or more monitored parameters fall outside those parameters, the system performs some operation to affect the state of its environment.
- The control system then waits for feedback — information from the environment about how the previous action affected the desired parameters.
- Depending on whether this information signals the system moving away from or back toward homeostasis, it reacts accordingly — and the loop repeats.
Worked examples:
- A thermostat: switches heaters or air conditioners on/off in response to a temperature sensor, regulating the environment to maintain conditions suited to the human body.
- Driving a car: while cruising, we simply continue what we were doing, while continuously monitoring feedback loops. As soon as information signals we’re approaching a homeostatic limit (e.g. drifting toward the side of the road), we react by adjusting the steering wheel, then wait a fraction of a second to monitor the effect of that action, and react again — all in the service of returning to (or staying within) the desired homeostatic condition.
This concept of homeostasis is a powerful model for capturing the development of any adaptive system — its course of development is the product of continuously acting on, and reacting to, feedback loops. Two or more adaptive systems reacting to each other’s behavior over time produces increasingly complex, evolutionary-like dynamics — underpinning phenomena like international politics, free-market economies, and social relations generally.
One nuance worth flagging before moving on: homeostasis/equilibrium is not always what a healthy
system is aiming for. Real complex adaptive systems often deliberately operate away from strict
equilibrium — homeostasis is better understood as one attractor among potentially several a
system could settle into, and under enough strain a system can bifurcate into a qualitatively
new attractor altogether rather than snapping back to its old one. This deeper dynamics-of-change
territory — attractors, the “edge of chaos,” evolution, resilience — is developed in
10-complex-adaptive-systems.md.