In 1969, when Niklas Luhmann was admitted to the newly-founded Faculty of Sociology at Bielefeld University, he stated as his research project: “Theory of Society; duration: 30 years; costs: none”. 28 years after this application, in 1997, he published his work Die Gesellschaft der Gesellschaft (the society of the society), which can be regarded as a comprehensive theory of society; he died shortly afterwards.
Luhmann was a systems theorist. He differentiated between four kinds of systems – machines, organisms, social systems and mental systems – and three kinds of social systems: interactions, organisations and societies. So what kind of systems does digital technology belong to?
Hardware clearly is a machine, but without software it would be dead. Thus, a digital machine consists of hardware and software. We can abstract away one or the other and analyse software or hardware in isolation. But what about the user? Their interaction with the machine constitutes a social system, not a machine.
The same is true for the organisation around both the interaction and the machine: organisations are social systems as well. For digital technology and digital products, this is where the beef is. Digital products aren’t machines; they are social systems. And that’s why they are services. The opposite is also true: because they are services, digital products are social systems.
Treating digital products as machines is the shortest way to failure. That’s what the classical waterfall product development approach aimed for. If you need to design a machine, the waterfall approach not only makes sense, it is necessary. By contrast, a social system can be neither fully specified and designed in advance nor implemented as designed and specified.
The proof is in the pudding
Some have tried – usually authoritarian systems, like dictatorships. It is, again, China that puts our modern Western assumptions about social systems and their capabilities to the test. Meanwhile, Donald Trump conducts another authoritarian experiment under Western conditions.
This aside, digital products as social systems are complex. To fully specify a complex system would require us to absorb its complexity into the specification, adding tremendous overhead. Even the user stories of modern software development don’t try to overspecify user behaviour. User stories are limited to a specific context, and not every user story ever written gets implemented.
Viewed as social systems, digital products are their own specification. The proof is in the pudding. Systems theorists call these self-referential systems.
Luhmann views systems as regulated by the difference between system and environment.
From this perspective, an ecosystem isn’t a system, but rather an eco-complex (a term coined by Luhmann that apparently didn’t take off). If we would see the Earth’s climate as an eco-complex, rather than an ecosystem, it wouldn’t be easy to ignore its complexity.
Climate change has enormous impact on our social systems. But this relationship is itself complex. For social systems, the climate is part of the environment. It seems as if the overheating of the planet’s climate in turn causes our social systems to overheat. To put it in Luhmann’s terms: complexity means mandatory selection, mandatory selection means contingency, and contingency means risk.
A systemic challenge for modernity
The difference between system and environment is crucial. Since the environment is always more complex than the system, a system needs to reduce complexity. The overheating of social systems due to awareness of climate change signals a failure of complexity reduction. Remember, the climate is part of the environment, not of the system, and must be treated accordingly.
When a system observes its environment overheating, it’s no solution to overheat itself. Instead, systemic changes are necessary. The system must adapt itself. But while the difference of system and environment can change, it cannot disappear. There is no system without environment, and vice versa.
Climate change, as our societies currently treat it, reintroduces the complexity of the environment into the system. This reintroduction is only possible if the system either builds up new complexity or reduces other parts of its own complexity.
The first option roughly translates into a scientific, engineering, innovation approach, while the second option is the sacrifice, degrowth, retrograde path. In other words: climate change poses a systemic challenge for modernity itself.