Algorithmic Politics

I am fascinated by the ideas behind the “Tragedy of the Commons”. It describes the story of a few farmers acting in their self-interest, and failing to maximize their common good by depleting their shared resources.

This is as much a reflection on computer systems, long before they were invented, as it is on politics. (Its original formulation, in 1833, was intended as an economic study.)

The “Tragedy of the Anticommons”, a story of starvation caused by an excessive number of property owners, is similarly striking. Resource starvation, right?

It ties into politics even more deeply. In algorithmics, there is a concept of “greedy algorithms”, which are made to be simple to understand. They are often efficient. They work by seeking local optimality. Whatever gives immediate best results for us is the path we take.

They are notable for failing to find the global optimum in fitness landscapes which require foresight — in particular, NP-complete problems. Of course, maybe there is a greedy solution to all NP-complete problems — maybe there even is one to all NP-hard problems. After all, a concept of local optimality in one universe can be different from one in another. Transforming the problem may yield a greedy algorithm that finds the optimal solution, which it didn’t reach before.

Politics seem mostly divided on that particular approach. The right-wing seems convinced that the capitalistic self-interest of every individual must be pursued greedily, that global optimality will be reached through this approach. What “self-interest” means is left open for interpretation; a self-interest in one universe can be transformed in another universe. What is torture according to one can be interrogation according to another.

Conservatism stems from the idea that up until now, things have been looking up; maybe we really did reach global optimality by following this approach. There is no reason, then, to part ways with the status quo. Tradition must therefore be preserved. Often, this is also tied to religion: by acting greedily, things have worked out all right; surely someone must be taking care of the order of things. This line of thinking protects us from fearing the future: if all of this is part of a deity’s plan, despite the issues we face, we will eventually prevail, because we as animals have more value than animals that have gone extinct in the past.

On the other hand, the left-wing is generally afraid of the greedy approach, preferring complex rules and regulations that guarantee global optimality in the end. Unfortunately, those algorithms are usually not efficient, imperfect, and costly. Sometimes, there actually hasn’t been any good solutions found on a particular problem. Sometimes, the solution assumes the good will and obedience of all citizens (cough cough, communism). Sometimes, we must compromise so much that we forget what we fight for.

Of course, in politics and in computer science alike, there is no silver bullet. The werewolves always win.