Emergence - when the sum exhibits properties not observed in its parts - is a double-edged sword.

When the emergent properties are beneficial to us, it’s awesome. It feels like ordering a pizza and getting a Lamborghini, with the pizza on top. In fact, it’s so awesome that certain classically insoluble problems have trivial (to implement, not to understand) emergent solutions. Centrally planning large economies is tough, and the results typically suck. But by setting up some basic conditions, suddenly global resource allocations become near-perfectly efficient, even though no one is specifically aiming to do that.

But when the emergent properties are harmful to us, the problems they cause can be absolutely awful. We all know, for example, that when left alone, the free market could easily result in people doing terrible things that completely mess up the planet, to everyone’s detriment. These problems are usually considered to be failures in coordination, and have been widely discussed. But something I’ve been thinking about only recently is how they often represent the emergent versions of some corresponding “classical”, non-emergent problem.

If I don’t maintain my own driveway, and eventually it decays to the point where I can’t get my car in or out of the garage, I’m just being stupid. If we as a society don’t maintain the public roads we share, and eventually they decay to the point where nobody can go anywhere, we are just being stupid collectively. Tragedy of the commons is actually just emergent stupidity. Of course, nobody is being an idiot individually, so this problem is somewhat different. Yet on some level, it’s the same as the classic problem of individual stupidity.

Standard Oil threatening to cut off supplies to the distributors that did business with its competitors is the classic problem of monopoly. When websites try to conserve bandwidth and voluntarily limit web crawling to just Google’s bot, we get an emergent version of monopoly. There is no explicit coercion from Google, no deliberate collusion, yet the outcome is recognizably similar. In both cases, the result is a massive reduction in the fairness of market competition.

The government limiting what I can or cannot say under the threat of violence is the classic problem of freedom of speech violations. Some angry Twitter mobs influencing what I can or cannot say, through the threat of torrents of online abuse and sometimes physical danger, is an emergent version of the same freedom of speech restrictions. Again, there is no single person or group doing the “threatening of abuse” (okay, sometimes there is explicit coordination, but usually not), but the consequence - the narrowing of the range of allowed discourse - is eerily similar to the classic case.

I suspect that often when people disagree on certain issues, they are actually disagreeing on whether an emergent version of a problem belongs in the same category as the classic version.

The two types of problems do look very different. Unlike their classic counterparts, in the emergent versions of these problems: no single person is being outright stupid; Google is not threatening website owners to only allow their crawler; no malicious orchestrator is directing the digital mob at certain people to “cancel” them. If you look at the local level, everyone involved is just doing innocent, harmless (good, even), and perfectly reasonable things. But when you zoom out and look at the globally emergent properties, the two types of problems undoubtedly share the same signature. Their effects are often extremely similar, and if you squint your eyes until the group becomes a single hypothetical person, they downright become identical problems.

The series of tech “de-platformings” from January this year, and people’s varied reactions to it, serve as a good illustration of this (here I’m specifically referring to the question of whether the tech companies were in the right to kick people off of their platforms). The ones seeing it more through the “classic problem lens” do not see anything wrong. There is no issue to speak of really, just the free market working as intended. But those who are more sensitive to the “smell” of problems recognize something there. The emergent consequences do feel severely troubling. These seemingly irreconcilable, polar opposite views of “no problem here at all” and “some serious things may have been violated” are actually both correct. Translated properly, we get the more complete diagnoses: “there is no classic problem here” (no evil tech CEO scheming and acting with malice) but “there are some emergent problems” (for which no one in particular can be blamed directly for).

The frustrating thing with emergence properties is that we have almost no way of directly and reliably controlling them. When it works in our favor, like in the case of the self-organizing free market economy, this isn’t a big deal. We mostly just go “yay” and reap the benefits. We do try to understand and tweak it, but even with a lot of effort by some of the smartest people, it’s only sometimes that we get what we want without making things much worse in ways we didn’t anticipate.

When emergent properties work against us, we can’t just ignore them and move on, since they will continuously bring trouble. So we have to try our best to solve them, even though we can’t really do a good job at it given their difficulty. This is why individual stupidity (of the “failing to maintain my own driveway” kind) is not at all a problem, but tragedy of the commons, though partially solved, still continues to plague us. A key limitation is our lack of suitable mental tools and frameworks for dealing with emergent phenomena.

Our legal system, and more fundamentally our system of ethics, is based on the assumption that events are causally linked to sovereign agents with will. These agents can be individual people, or groups of people sharing a common set of goals who act together (e.g. corporations, governments). But regardless, when something goes wrong, we need to be able to point the finger at some definitive entity as being responsible, and compel them to fix it. The oil market is becoming like a cartel? Seems like Standard Oil is largely responsible, as they are threatening distributors to push out competitors, let’s break them up. The government is trying to mandate some dubious restrictions on free speech? Let’s sue them, win, and stop them.

This works well for classic problems, because when an agent can be clearly determined to be at fault, we can easily fix the problem by modifying the agent’s behavior, either through force or incentives. But emergent problems completely defy this. By definition, they are the result of the collective behaviors of the whole, with no single agent being responsible. We don’t really know what to do in these cases. People are very well aware of this, and often exploit it to avoid being held accountable for problems. Straight up blacklisting websites as a search engine would be too obvious, people would immediately recognize the classic problem of monopoly. But encouraging everyone to implement a standard you created, which gets preferential treatment on your search engine? That turns the problem a little more into an emergent version, where you can’t be blamed for it easily (though antitrust litigators are also familiar with the technique, and in this case have caught on).

I don’t really have good answers for dealing with emergent problems in general. My sense is that there are no silver bullets here, and that each problem probably requires its own tailored solution. Sometimes the effective techniques for solving the classic version of the problem can be used. Maybe we can just break up or heavily penalize monopolies, even the ones that are emergent. But other times I’m not so sure. I don’t know if emergent free speech problems are best tackled with the same top-down solutions as used for the classic version. Either way, building mental frameworks that help better understand this class of problems will be required.