Intro

At a high level, we are going to talk about systems design and sociology. The idea here is that average behavior (of people or animals or technology) is the result of the way the overall system is set up, so how do you set up a system with a targeted behavior in mind?

​

This isn't really an area with a lot of work done on it in academia. But we are going to start with what has already been done in the overlap between studying systems design (control systems) and studying human behavior (sociology).

​

I am writing this in an attempt to make this overlapping area more accessible, and add some original commentary in the form of hot takes, social commentary, and more math. The spicy shit. 🌶️. If math makes you uncomfortable, do not worry. It is not my focus and I will do my best to explain necessary concepts in simple ways (or at least keep the crazy stuff brief)

​

This involves some math concepts that most math programs don't teach but are surprisingly intuitive. First let's get a bit of understanding then apply them to social behavior.

Turn on Cruise Control

Bear with me here

In cars, cruise control is a system for controlling the speed of the car automatically. It is a simple system which lets the driver set a reference speed and then the car tries to go that speed. More advanced cars also use sensors to break before they hit anything, but let's just take a look at the simple case.

Here is how cruise control works

1. The driver sets a desired speed
2. The car looks at the desired speed and the current speed and calculates the amount of additional power needed
3. The car knows how much fuel (etc.) to give the engine to make that much power
4. The car supplies the engine
5. The engine supplies the power
6. The car speeds up or slows down

​

By making a really good calculation in step 2, you can handle more complicated situations like

• Going up or down hill
• You are going a bit too slow but speeding up so fast that you are about to go too fast (this uses acceleration and is called a 2nd order system because acceleration is the derivative of speed)

​

"Control Theory" (AKA 'negative feedback loops') is a type of system design that allows a given system to behave like some 'reference' using only:

• the reference
• information about the outside world from measurements ('signals' and how they are changing)
• the initial setup, which is one with knowledge of how the system itself functions

​

In this case, the system was:

Doing this makes a closed loop, where the last element 'speed' is used to determine something upstream. We can do this largely because we know how the car works and it still likely took a lot of trial and error.

​

Other control systems include:

• Thermostats (desired temp → AC/Heating → actual temp)
• Rocket Gimbals (impressive because all you can control is thrust and gimbal angle, and you have to land a rocket on a barge in the ocean)
• Drones (can give a drone an xyz location and a 'facing) and they can get there without crashing while only controlling the amount of voltage supplied to their rotors

​

You can imagine how difficult some of these things would be able to do manually with 'stability' (ie. not crashing the rocket), which is why this is a design science.

​

Now in Sociology, 'negative feedback' is used prevalently to describe things like interpersonal behavior, relationships, crowd behavior, and organizational behavior.

​

This paper is an excellent shotgun blast of applications of this field of mathematics to sociology at a deeper theoretical level.