Culture-Oriented History

I’ve been traveling a lot for six weeks; one more week to go, then I’ll stay home for a while. In that time, I’ve given many talks, including two on cultural drift (to be posted soon), a topic I’ve continued to read and think lots about. I feel a bit out of practice writing now, and I’m struggling with how to tell you about all the changes in my thinking.

I guess the biggest thing I want to say now is that the cultural drift thesis seems rich with implications for history and current affairs. But these implications may be a lot of work to tease out, and I fear I’m likely not very well qualified to this work. However, let me at least outline the task here.

Good historians generally try to explain what has happened to humans in time, using all the promising social models and abstractions they can find. Instead of focusing on explaining the details of individual lives, they naturally prefer to focus on explaining large scale patterns of behaviors, only considering individual details in a few especially important cases. All of which makes sense.

Looking at history through the lens of cultural evolution, we note that we want to categorize all this human behavior in terms of the biological concept of “adaptation”. That is, we want to say which of these behavior patterns are more (vs less) prone to win out against rival patterns, and to enable correlated behaviors and DNA to also win out against rivals. The more adaptive patterns should tend to win out in history, and be expected to win out in the future.

However, tracing out which patterns we should expect to win in the long run seems really hard. We should of course try anyway. But it might help to start with a somewhat easier task: assigning relative weights to different causes of pattern changes. After all, we should expect the changes caused more by selection to be more adaptive, and others to be at more risk of being maladaptive. 

For this purpose, the most important type of change is, of course: direct selection. Some behaviors cause associated hosts to live or die, reproduce or not, grow or decay in resources, be eagerly or reluctantly emulated, etc. Usually via war, famine, disease, or fertility differences. The patterns we see as growing more due to selection are also those we expect to be more adaptive, and to become more common in the future. 

When our environments change, that often induces a second key cause of change: conditional adaptations. Plants and animals long ago evolved habits of changing their behavior with time of day, day of year, or cycle in life. Brains allowed for far more of that conditionality, and human brains even more so. For example, human and other animal behaviors plausibly vary in standard ways with density, disease provenance, mortality rates, associate relatedness, sex ratio, and resource wealth. 

In addition to directly executing simple inherited conditional adaptations, our brains also let us calculate more complex context dependent adaptations. To do this, brains have some goals that they seek to achieve, and they try to calculate what behaviors are more likely to achieve those goals in the current environment. 

Changes caused by conditional adaptations should be expected to be adaptive if those tendencies are actually still adaptive. But contexts may have changed to make them less adaptive. Calculated behaviors can change when environments change, when goals change, or when calculation strategies change. As it is packages of goals and calculation strategies that are selected together, changing calculation strategies while holding goals constant is often maladaptive. 

An important category of conditional adaptation is: cultural learning. Cultural evolution was made feasible by a human capacity to copy the behaviors of others, together with heuristics on who is worth copying on what when. The first heuristic was probably to just copy the high status, but with time humanity collected many more specific and stronger heuristics. When such heuristics are adaptive, their cultural learning applications will also on average be adaptive. But maladaptive status markers can lead to maladaptive copying of high status behaviors.

Finally, many rivals in conflicts are tied to packages of cultural elements, packages which rise or fall when those rivals win or lose their conflicts. For example, World War II was the most culturally influential event of the twentieth century. Yet the rise of Hitler to run Germany and Germany losing the war were pretty random events. So the big cultural changes caused by that war are also pretty random changes, and thus mostly not adaptive. Cultural activists also often seek glory via leading culture change crusades, but which of the rivalrous movements win out can be pretty random. 

To summarize, selection should pretty consistently make culture more adaptive, as should cultural learning using strong heuristics. Cultural learning with crude heuristics, and conditional adaptations, should also tend to be adaptive, so but less consistently so. Finally conflicts among rivals will mostly be random, but may weakly tend in selective directions. 

Note that if culture starts out at a pretty adaptive location in culture space, and then moves due to substantially random forces, it should on average move to less adaptive places, even if those change forces have a weak selective tendency. This is because culture space can be high dimensional, with highly adaptive locations taking up only a tiny fraction of the space.

And that’s the hard, but maybe not maximally hard, problem of describing history in cultural evolution terms. If we can identify the relative weights of different kinds of forces changing different aspects of culture, we can guess which ones are getting more vs. less adaptive. Which might help direct our efforts at cultural reform.