S01E08 Why feedback governs the world
Control theory, feedback loops in nature, evolution of feedback in man-made systems, circularity of feedback
In last week’s episode, I claimed that policy often misses its intended goals because the policymakers fail to appreciate the complexity of the systems they are trying to influence.
In our quest for organizational performance, we want to gain a better understanding of the interactions that take place in a complex, adaptive system. Ideally, we want to get better at influencing systems such as teams, companies, supply chains, and perhaps even entire markets. Before we jump into the deep end, let’s start by exploring how control works on a more fundamental level.
A good place to start is control theory, a branch of applied mathematics that builds on feedback. Since the geek level is already over 9000, I’ll stay away from the actual mathematics. Fortunately, we don’t need to do the math to appreciate that many systems - both man-made and biological - rely on feedback processes. Some examples from nature:
In biology, feedback loops regulate many organic processes. For example, the hormone insulin plays a key role in regulating blood sugar levels in the body. When blood sugar levels rise, insulin is released, which helps to bring them back down. This is just one example of a self-correcting feedback loop that helps to maintain homeostasis in the body.
Something similar happens on the ecological level. In the population of predators and prey, a decrease in predators will lead to an increase in prey. A boom in the prey population means more food for predators, which will cause their population to rise again.
Fruit ripening is the result of a reinforcing feedback loop. When one apple starts to ripen, it will produce ethylene, a gas that will cause nearby apples to ripen as well, producing more ethylene. As a result, all the apples in a tree will ripen virtually overnight.
Climate change results from interacting feedback processes in the atmospheric carbon cycle. To name just one: melting sea ice exposes more ocean, which in turn absorbs more heat and causes more ice to melt.
Clearly, the concept of feedback is an important building block of biological systems. Why is this concept so foundational? To answer this question, let’s look at the evolution of feedback and control in man-made systems.
Since antiquity, humans have dreamed about automatons or self-operating machines. Truly automated control proved elusive until the 18th century when James Watt adapted an existing device (a centrifugal governor) to control a steam engine. Before Watt’s invention, steam machines needed human operators to manage the load placed on the engine by opening and closing a steam valve.
Watt designed a mechanical control circuit for this valve. It is made up of two leaden flyballs at the end of a pendulum. As the engine turns faster, the hinged flyballs fly outwards. This is a control feedback loop: when the arms of the pendulum separate, a linkage causes the throttle on the steam engine to close. This way, Watt’s governor reliably and automatically controls how much steam can enter an engine’s cylinders based on its desired running speed. This ensures constant revolutions per minute without the need for a human to babysit the machine.
The importance of this invention can hardly be overstated. Watt’s governor was instrumental in taming steam power and kickstarting the Industrial Revolution. The steam engine itself would be relatively short-lived, but the concept of a control circuit to regulate feedback could be extended to speed, pressure, temperature, and other process variables. It proved foundational for all automation (including the combustion engine).
Governors were more than a mechanical innovation; they were the proof of concept for man-made control over matter and energy:
The difference between a car and an exploding can of gasoline is that the car’s information—its design—tames the brute energy of the gas. The same amount of energy and matter are brought together in a car burning in a riot and one speeding laps in the Indy 500. In the latter case, a critical amount of information rules over the system, civilizing the dragon of fire. The full heat of fire is housetrained by small amounts of self-perception.
— Kevin Kelly, Out of Control
In other words, the mechanical governor was a primitive brain of sorts, endowing machines with situational awareness. This innovation redefined our relationship with technology. In lockstep with our worldview, our inventions could now start to shift from a linear and deterministic approach to the fuzzy logic of interdependent, complex systems.
Kevin Kelly’s work has been a major influence on this publication. Two ideas of his were especially foundational:
Technology is increasingly merging natural and artificial characteristics. The world of the born is not separate from the world of the made.
To take advantage of natural principles in technology, we have to let go of our traditional view of control. Supervising complex systems is futile if one disregards the feedback loops that play out between interconnected components in the system.
Feedback is a powerful concept for designing and controlling systems because it takes the system output into account. Before dismissing feedback as a concept you are already familiar with, consider the simple process of pouring a cup of tea. Clockwork thinkers see a linear process: I make tea, and I pour it into a cup.
System thinkers see a circular process, involving a number of variables that contribute to the goals of regulating the temperature, flavor, and hot water level of the beverage. The simple act of pouring hot water is actually a regulating system made up of 5 variables:
the desired level of hot water in the cup
the current hot water level
the gap between the two previous variables
the position of the kettle
the flow of hot water
I won’t bore you with similar diagrams describing the regulation of tea flavor, temperature, sweetness, etc. The point is that we shouldn’t be too quick to dismiss a process as linear. In complex systems, innumerable feedback processes are at play, and each feedback process is a loop of continuously operating cause-and-effect relationships.
This circular relationship - the output of a system being used as input for the same system - has important implications for the behavior that emerges. One of the key characteristics of feedback loops is that they can amplify or dampen the effects of the input on the output, depending on the nature of the feedback. Feedback lies at the root of all shared characteristics of complex systems: emergence, information processing, and nonlinearity.
In the next episode, I’ll take a look at the evolution of feedback through the lens of scientific history. After that, I’ll explore how feedback loops can help us model the systems we create.