emergence and levels of analysis

This post is 99.44% nerd, I’m afraid, with not so much sex in it. But it has become urgent to post it, since lack of this idea has generated a number of comments and emails whose answer is… this post.

Sadly, this is one of my favorite things to talk about, but I’ll try not to be too pedantic. Really I will.

When I talk about levels of analysis and things that are “only available at the population level” I’m talking about emergent phenomena usually. These are characteristics of a group or a population or a system that are not descriptive of any individual or unit WITHIN that group, population, or system.

Flocks of birds are the classic example. How does flocking emerge? Well, each bird has a mechanism in its brain saying, “stay about yay far from your neighbor and make your way south.” (Roughly – for details see Craig’s beautiful work on “boids”.) The actual mass of movement we observe is a whole other thing. There is no leader, no followers, no DECISION on the part of the birds to create a flock. It just HAPPENS. It emerges.

You can say many, many things – indeed you can say EVERYTHING – about an individual bird without saying anything about a flock. And vice versa.

Let’s take a human example. Have you been to a big theater to see a performance and afterward we all clap for a long time and gradually all of us in the audience end up clapping together, like a chant, and then we gradually desynchronize, like we planned it that way, and then if we continue to applaud long enough, we clap together again, and then phase out again? That phasing in and out of synchrony is an emergent phenomenon. No one is leading it. It just happens. Take any massive group of humans anywhere or anywhen and put them in a giant room and make them clap for 15 minutes. It’ll happen.

Patterns emerge at the system level that can only be described by describing the SYSTEM PER SE, rather than by describing the individuals in the system or their behavior. Flocks of birds, herds of bison, schools of fish, and of course groups of humans.

Not a lot is understood about complexity in human social systems, and even less is understood about complexity in human sexual systems. But it’s true that patterns emerge at the population level that don’t describe (or only incidentally describe) the behavior of any individual in the group.

We know know, for example, that even just the barest minimum of a preference for people who are like oneself can, at the population level, result in widespread and absolute segregation of two or three different populations. All the other emotional, political, and sociological drama around discrimination may be sufficient, but it is not necessary.

What’s important about this is that it tells us that discrimination doesn’t require design or purpose; it doesn’t require a leader or a conscious decision of any kind. All it takes is a lot of people, all of whom, given 100 choices about who to sit next to, choose to sit next to someone like themselves 51 times and someone not like themselves 49 times. From that we get absolute segregation. That’s all it takes. The segregation is the RESULT of the problem, not the problem itself.

In human sexual systems, one of the population level characteristics that I think is an emergent property of the dynamics of the system is the existence of sexual “hubs.” My own extremely nerdy and boring dissertation found that hubs – that is, individuals with an exponentially higher number of partners than their consepecifics – result when there is a normal distribution of sexual motivation in the population. If the distribution is skewed – like, if pretty much everyone has high sexual motivation (high SES, low SIS) OR if pretty much everyone has low sexual motivation (low SES, high SIS) these hubs don’t emerge. I made it happen in a computer model, but the neat part is that it parallels real life. There is a normal distribution of sexual motivation, and we have sexual hubs in the population. Cool, eh?

Humans being the extraordinary creatures that we are, we can see a result of these dynamics and declare them good or bad and then make a law or policy to prevent or enforce the dynamic, as with discrimination (this, btdubs, is what’s fundamentally wrong with don’t ask, don’t tell (IMO); it reinforces rather than prevents the dynamic of discrimination). We can make intellectual moral declarations about what the result of the dynamic means. But maybe the most important thing we can do is understand what dynamic gave rise to the emergent property.

It’s important to understand complexity in human social systems because that’s how we’ll make policies and laws to create positive change. I can talk about this more at some other point, but surely you’ve all stopped reading by now anyway.