Can a realistically-parameterized model tell us why our brains are so big?

Nature “Sizing up human brain evolution”

Nature “Inference of ecological and social drivers of human brain-size evolution”

This is an interesting study that I simultaneously think is really cool and has some major flaws.

What’s cool about this study is that it trys to get at this question with a model that’s (reasonably) constrained by observed parameter values. It’s a great approach to try to model real human interactions and look at the evolutionary consequences of those interactions. Obviously the first concern is that these interactions and consequences will be “made up” and therefore have no real-world meaning. And in this study, these interactions and their evolutionary consequences are not entirely “made up”: the actions that humans are simulated to take are reasonable, and the parameters that govern the consequences are at least somewhat constrained by real observations.

So does that mean I would “trust” the findings of this study that ecological drivers are the best explanation of human brain and body size? Not really. My big worry here is the arbitrary nature of the model simulation structure. What humans actually experience is so much more complex that these simple simulations, and that begs the question: is this result robust to changes in the complexity or other structural assumptions? It doesn’t appear that this study did much more than vary one parameter, so it is very possible that its results are idiosyncratic to assumptions that are not varied in other parameters or just the structure of interactions in this model.