The Dangers of Building Historical Databases

Wednesday and Thursday I ran a workshop in my lab. I call these gettings-together ‘micro-workshops,’ because there may be as few as 4-5 people, or as many as 10 (but not more). Many years ago, when I was hired by my university, I asked for a laboratory that was set up not as a usual biological lab, but as a space where I could run such micro-workshops. It’s basically a room with a big table in the middle,  chairs, whiteboards, and a counter with computers and peripherals. This is the room where we have been ‘brainstorming’ for two days.

In several of my recent posts I wrote about how we can answer such questions as what role (if any) climate played in the rise and collapse of complex societies. I have criticized some previous studies and promised to offer a better way of answering such questions. Well, what we have been doing during the last two days bears directly on this issue, because we have been designing a historical database of social and cultural evolution. This database will eventually enable us to test empirically all kinds of possible explanations about how large-scale complex societies, including the one we live in, evolved. One kind of hypothesis that we will be able to test are those that invoke climate as the main engine of change, although there are also many more interesting theories that cry out to be tested.

What makes this approach possible is that different regions of the world vary greatly in when (and whether) they evolved complex societies. Some parts of the world were very precocious in getting complex societies – states and empires. Other regions never evolved complex societies, until they were annexed by one or another European Power in the last century or two. In yet other regions social complexity achieved a middling level, and was stuck there (actually, the typical pattern is that of back-and-forth cycling between, say, simple and complex chiefdoms). There are many instances of complex societies first rising, and then collapsing. When we look at the whole world over the last several thousand years, we have a rich database for answering all kinds of questions about social evolution.

For example, as long-term readers of this blog know, my pet theory is cultural multilevel selection CMLS), in which the primary driver of evolutionary transitions to ever more complex societies is competition between cultural groups. Put simply, warfare. Because until recently (and some would say even today) the chief form of intersocietal competition has been war, the theory predicts that intensification of warfare should precede in time an increase in social complexity. The time dimension is critical, because it allows us to distinguish CMLS from an alternative theory that something else causes the transition to more complex societies, and those, in turn, begin waging more intense warfare. In the second case, intensification of warfare should follow, rather than precede, an increase in social complexity.

The time dimension is why we need a historical database for testing theories about social evolution. Currently the most used databases, such as the Standard Cross-Cultural Sample, or the Ethnographic Atlas, lack such a time dimension – they code cultural variables for a particular society at a particular point in time. So we have to build our own.

To give you a general idea of how we will analyze the data once we have them, suppose that we have measures of both social complexity and warfare intensity in many world regions (hundreds of them) for each century going back as far into the past as we can push it (in some exceptional cases going back five or six thousand years, at worst just a century or two). How we actually measure ‘social complexity’ and ‘warfare intensity’ are huge questions, which I will defer to future blogs. But suppose we solve this problem. Then, we use statistical methods to determine whether warfare intensity can predict how social complexity changes in the following centuries. And this will tell us whether the CMLS theory is correct.

If the authors of the recent Science paper about Mayan collapse, which I criticized in my previous blog, had the data on social complexity that we will eventually develop, they could have done a similar kind of analysis by looking at how changes in climate can predict the increase or decrease of social complexity. Instead, they had to rely on just one indicator, the number of dated inscriptions (and they did not do the statistical analysis). What I am trying to say is that although our research group has several hypotheses that we want to test (the effects of warfare, ritual, and religion are at the forefront right now), we want to build the database in such a way that it will be useful to others in testing their theories.

Balancing the demands of designing this database in such a way that we can start entering data of primary interest to us, and build in flexibility so that other kinds of questions can be addressed with it in the future, has proved a much more difficult task than we at first anticipated. In fact, in our first meeting in Vancouver in August we couldn’t solve this problem.

So I was really worried that we would end up without a consensus again. But, fortunately, it did not happen. Not that it was easy – at some point one member of our group tried to strangle me, while another was looking meditatively at the Storrs Cemetery across the road, selecting a nice spot to lay my mortal remains to rest:

Despite such regrettable excesses by my excitable colleagues, by the end of the workshop, we really hammered out a fairly simple, but flexible scheme that pleased all of us. Next week I am running another microworkshop where we will be meeting with two historians who will be pouring their specialist knowledge into the data structure that we designed.