I should have loved sociology / Sociology 101 at KIT

Published: October 2, 2026
URL: https://toranm.me/post/2026-10-02-i-should-have-loved-sociology/

The course Introduction to Sociology (dt. Einführung in die Soziologie) at KIT was amazing. If you are into prisoners' dilemmas (https://en.wikipedia.org/wiki/Prisoner%27s_dilemma), decision-making, or group coordination, sociology might be for you. I actually think many people are prejudiced against sociology; they think it is a soft science for weed smokers who don't want to do math. But that is not true. Sociology has a lot in common with rationalist thinking (https://www.lesswrong.com/).

If you are a student at KIT, I highly encourage you to take this course. Even if you don't do it as part of your degree, it will provide you with insights and methodologies that translate into your field, the business world, and everyday life.

On a personal note, I regularly visit the University of Frankfurt (https://www.uni-frankfurt.de/) Campus Westend and stood on Theodor-W.-Adorno-Platz, walked on Max-Horkheimer-Str., and was at the building for the excellence cluster Normative Orders (https://www.normativeorders.net/).

# Why Sociology Matters

Sociology is the study of society, social relationships, and social institutions. This field is distinct from other sciences, including psychology (https://en.wikipedia.org/wiki/Psychology), economics (https://en.wikipedia.org/wiki/Economics), and political science (https://en.wikipedia.org/wiki/Political_science). Sociology provides tools to analyze social behavior. Think of any interaction where two or more people interact. It includes social norms and discrimination. A lone individual can be studied, but sociology is more about the collective. Individual behavior influences society, but society also influences individual behavior. Sociology is about understanding the complex interplay between individuals and society.

Sociology starts at fictitious prisoners' dilemmas (https://en.wikipedia.org/wiki/Prisoner%27s_dilemma) and ends at the large-scale, pluralistic, and complex societies that we live in.

Computer science and sociology are deeply connected.

Uber developed apps to innovate in facilitating a social interaction. Technology enables new forms of social interaction. Sociology provides a framework to understand these interactions at a level that goes beyond tech stacks and algorithms. Computer science is about how to build systems; sociology is about understanding the social implications of these systems and shaping the technical requirements.

Social Choice theory (https://en.wikipedia.org/wiki/Social_choice_theory), think Arrow's impossibility theorem (https://en.wikipedia.org/wiki/Arrow%27s_impossibility_theorem) or the Balinski–Young impossibility theorem (https://en.wikipedia.org/wiki/Balinski%27s_and_Young%27s_apportionment_method) are other examples of computer science and sociology colliding.

# How to Get into Sociology as a Mathy Person

Sociology can be very mathematical and quantitative. Of course, statistics (https://en.wikipedia.org/wiki/Statistics) is used a lot, but also game theory (https://en.wikipedia.org/wiki/Game_theory), algebra (https://en.wikipedia.org/wiki/Algebra), and graph theory (https://en.wikipedia.org/wiki/Graph_theory).

# Concepts to Highlight

## Collective-Action Problem

Prisoners' dilemma (https://en.wikipedia.org/wiki/Prisoner%27s_dilemma) is a classic example. Games of this type are known as collective-action problems (https://en.wikipedia.org/wiki/Collective_action_problem) and can be brought into a normal form. This framework resembles Karnaugh–Veitch diagrams (https://en.wikipedia.org/wiki/Karnaugh_map). The number of players is the number of dimensions. Each cell in the diagram represents a possible combination of strategies. Each cell contains a payoff value for each player. Prisoners' dilemma is a special case of the collective-action problem with two players.

These are the four payoffs in a two-player game:

- $T$ (Temptation): The payoff for defecting while the other cooperates.
- $R$ (Reward): The payoff when both players cooperate.
- $P$ (Punishment): The payoff when both players defect.
- $S$ (Sucker's payoff): The payoff for cooperating while the other defects.

[image: Game Representation](../assets/post/2026-09-24-I-should-have-loved-sociology/game.svg)

Various games are realized by different orderings of these payoffs. The following table summarizes some common games:

| Payoff Inequality | Game |
| --- | --- |
| $T > R > P > S$ | Prisoner's Dilemma |
| $T > R > S > P$ | Chicken |
| $R > T >= P > S$ | Stag Hunt |
| $R > T > S > P$ | Harmony Game |
| $T > P > R > S$ | Deadlock |

## Changing Behavior

Three main ways to change behavior: incentives, nudging, and institutions.

Incentives are pretty straightforward. You can change behavior by changing the payoffs. Good ol' carrot and stick.

Nudging (https://en.wikipedia.org/wiki/Nudge_theory) is about changing how the decision is perceived.

> "Do you want your burger vegan or with meat?"
> vs.
> "Do you want your burger normal or with meat?"

The second question nudges the person to choose the vegan option, but the payoffs are the same. Nudging is the exploitation of cognitive biases in the decision architecture to change behavior.

Institutions are the most sociological. Social norms (https://en.wikipedia.org/wiki/Social_norm) are an example of an institutional mechanism. Imagine all people around you do not eat meat and never have. When you go out with them, will you still order meat when it's on a restaurant's menu? What about when eating out alone or getting takeout? The payoffs are the same, even the nudges are the same, but the social norm has changed. You probably will not eat the meat.

Imagine a world where there are so few people eating meat that restaurants or supermarkets do not offer it. The social norm influences the economics.

## Axelrod

You might remember "The Evolution of Trust (https://ncase.me/trust/)" by Nicky Case (https://ncase.me/) (2017). It is a fun experience, if you don't know it, you are one of todays lucky 10,000. It provides a great introduction to evolutionary game theory and the iterated prisoners' dilemma.

It is underpinned by Robert Axelrod (https://en.wikipedia.org/wiki/Robert_Axelrod)'s seminal work on the iterated prisoners' dilemma (https://en.wikipedia.org/wiki/Prisoner%27s_dilemma#Iterated_prisoner's_dilemma).

## NetLogo

NetLogo (https://ccl.northwestern.edu/netlogo/) is a tool for agent-based modeling (https://en.wikipedia.org/wiki/Agent-based_model). You create a world with agents that follow your rules. It has predefined models for many social problems, including the prisoners' dilemma. You can run simulations and observe emergent behavior (https://en.wikipedia.org/wiki/Emergence).

One really interesting insight is the emergence of discrimination. In a simulation, there are blues and greens that play repeated prisoners' dilemma against each other. No one knows what their own color is, but they see their partner's color. The payoffs are asymmetric to the detriment of the green. The following will happen: Blues cooperate with each other, Blues will defect against greens, but most interestingly, greens will also defect against other greens. Essentially, greens will be discriminated against.

## Externalities

Externalities (https://en.wikipedia.org/wiki/Externality) are the effects of someone's actions on someone unrelated. Climate pollution (https://en.wikipedia.org/wiki/Climate_change) is an example. You get a payoff, say cheap energy, while someone else has to pay for the cleanup.

## Liquid Democracy

Liquid democracy (https://en.wikipedia.org/wiki/Liquid_democracy) is a mixture of direct and representative democracy. It allows individuals to delegate their voting power to trusted representatives, while still retaining the ability to vote directly on issues they care about. Your representative can also delegate their voting power to someone else, creating a dynamic and flexible system of representation. For example, you might align with a political party on most issues, but you might trust a friend to make decisions on environmental policy, but vote directly on economic policy. LiquidFeedback (https://liquidfeedback.org/) is one software implementing such a system.

LiquidFeedback is a software platform that implements such a scheme.

## Multi-Armed Bandits

Imagine you are in a casino with "one-armed bandits," aka slots. You don't know which machine has the best expected payoff. The payoff of each machine is on a distribution that you don't know. The act of using a machine to sample its distribution is called exploration. The act of using the best one you currently know is called exploiting. Multi-armed bandit (https://en.wikipedia.org/wiki/Multi-armed_bandit) analysis provides a strategy to optimally balance exploring and exploiting: the Gittins index (https://en.wikipedia.org/wiki/Gittins_index).

Lots of scenarios can be analyzed in these terms or variations thereof.

## Honorable Mentions

Here are a few honorable mentions; I won't go into detail. You can ask your chatbot of choice about them and learn something.

- Durkheim's Social Facts (https://en.wikipedia.org/wiki/Social_fact)
- Tragedy of the commons (https://en.wikipedia.org/wiki/Tragedy_of_the_commons)
- Wisdom of Crowds (https://en.wikipedia.org/wiki/The_Wisdom_of_Crowds)
- Micro and Macro (https://en.wikipedia.org/wiki/Macro%E2%80%93micro_link)
- Emergence (https://en.wikipedia.org/wiki/Emergence)
- Of course, we also derived Bayes' theorem.

# Conclusion

Sociology is fascinating and, in a way, not well enough recognized. A lot of concepts are mirrored in the rationality community, but without looking at the heritage of the field. Thank you to my professor, Mr. Mäs, for this lecture.

Also see I should have loved biology (https://jsomers.net/i-should-have-loved-biology/) by James Somers.