Emergent Causality

For whatever reason, perhaps a pervasive simplicity bias,* we as humans like to think of causality in very basic terms: each event has one and only one cause. Multi-causal explanations seem unsatisfying. We like to know who (or what) to blame or credit. Shared responsibility seems somehow not as real. In cognitive psychology experiments it is well-documented how a crowd of people will stand by watching someone else in distress without anyone offering to help. Yet any one of those same witnesses would invariably take action if nobody else were around. The literature explains this as a sort of “tragedy of the commons” in personal responsibility, i.e. each individual in a crowd of 20 is only 1/20th responsible. Furthermore, everyone assumes that somehow the other 19/20ths of the responsibility will take care of it, if they haven’t already.

We all know on rational level though that most of the time there is more than one cause, sometimes uncountably many. “Shit happens” is a way of acknowledging the inherent complexity in the universe and the impropriety of trying to find someone or something to blame in all cases. Understanding emergent phenomena helps us to comprehend and to put to rest the nagging feeling that someone or something is really responsible, even if we can’t put our finger on it. Conspiracy theorists are those people who refuse to believe that every occurrence doesn’t have have a human agent behind it as a root cause. And because they can’t identify one person in particular, they conclude that there must be a conspiracy of multiple agents working in concert.

But for every erstwhile conspiracy, there is another explanation that doesn’t involve intentional agents. For a political leader to come into power in a democracy, many coordinated efforts have to take place, not the least of which is that millions of individual agents must cast a favorable vote. No one person “caused” the leader to be elected. At times there will be a single agent that is the proximate cause, such as the lone gunman who shoots the president. But in order for the assassin to be able to actually get close enough to carry out the deed, many other things must fall into place. Does this mean that there must be a conspiracy working intentionally towards giving the assassin his shot? No. All it requires is a small amount of apathy (or decreased attention or occasional blind eye) from a number of uncoordinated, unknowing, unintending individuals for the assassin’s path to clear. So who caused the president’s death? The assassin was the single biggest individual factor, but there were also many smaller ones as well, without which the event could not have occurred.

In Hollywood, nobody sets out to make a bad movie, nor is it the intention of anybody in the large web of value-providers and puppet masters required to get a movie completed. Yet bad movies get made all the time, and if you ask any individual working on a stinker they can tell you that it is going to be bad. They can see the train wreck approaching, but they are helpless to do anything about it. Other tragedies of the commons are structurally assured, as illustrated by the Prisoner’s Dilemma. In such cases, the individual incentives are such that an undesirable outcome is inevitable for everyone. But even in the case where a win-win is readily achievable (as it is in the case of movies), small decisions by many different agents, combined with small bits of randomness from the outside, often can lead to a conspiracy-like effect. An intentional story would fit the evidence, at least on the surface, but so would an unintentional one. And it is the unintentional explanations that tend to fit the data better when we look below the surface and ask good questions.

* See future post on simplicity bias