Systems: An Ostensive Definition

Tuesday, May 20, 2025

#SystemsTheory

I'm currently preparing a presentation about Systems Theory, and I figured it would help me structure my thoughts if I write down some of the things I intend to present - starting with the basics: What is a System?

Instead of finding a formal definition, we're going to approach this by investigating an example.

Let's assume you're doing your groceries. You approach the checkstand, but you're not the only one who wants to pay, so you are queueing up. Congratulations, you have just entered a social system.

What exactly belongs to the system - does the system include the cashier or only the customers? does it include non-human parts such as the register itself or not? - doesn't need to be clearly defined, it ultimately depends on what you want to analyse or investigate.

Observing this simple system already surfaces a lot of interesting aspects. For starters, there are unwritten rules at play that everyone follows. Whenever people join the system, they queue in at the back, for example. This requires no alignment or discussion. Then, there are loops at play, where the system observes itself and adapts based on what it senses. For example, if the queue gets too long, either the cashier will notice it or someone else will point it out, and then another cashier will be called. There are also dormant loops which are not visible most of the time, but which become active when needed. For example, if someone enters the queue at the wrong position, the system will make them aware of their mistake. Or, if someone takes a very long time to pay, the system usually goes through several escalation mechanisms such as poisonous glares, loud sighs, muttered "I can't believe it" up to a yelled "HURRY UP ALREADY!", which sometimes (not always) leads to a speed-up of the payment process.

There are more fascinating aspects of the system. If we would be able to magically replace every single person - the cashier and every one in the queue - by a completely different person (still doing their groceries), the system would behave pretty much the same way. The individuals who make up the system don't matter here. ^[1] It gets better: If you replace all of the existing people with random people who also occasionally enjoy watching soccer matches live in a stadium, all of them would still stick to the unwritten rules of the system "grocery store", they would not suddenly start cheering on the cashier, or doing The Wave on every successful payment.

This hints at an aspect I want to spell out more clearly: You can't understand systems by reducing them to their individual parts. Systems express certain properties only as a system, not as a collection of people. A soccer fan alone at home doesn't cheer the same way as they would as part of the system Soccer Audience. Yet, there is no one in the stadium steering, not a single mind saying "Now we do the wave", or "now we yell at this-and-this volume". Everyone knows this from work contexts as well: There are teams of medium-skilled people who are able to perform on a level that is above and beyond, and there are teams of highly-skilled people who just fail to perform together.

This has interesting implications for leadership. You can't reason about your team as a collection of FTEs. You can't split up your team and expect the performance of all parts to be the exact same as the performance before the split. You can't double your FTEs and expect double the performance. All these changes change the system, and any change in the system changes their emergent properties, such as cheering in the example above.