Spaghetti, cumulative cultural evolution and ritual traditions

Posted on December 9, 2011


Back in 2009 I went to a European Human Behaviour and Evolution Conference in St Andrews. One of the highlights of that conference and definitely the bit of it that has stuck in my mind was a paper presented by Christine Caldwell. Caldwell talked about a methodology that she’s been using to test claims about cumulative cultural evolution. In the paper, which was published in Evolution and Human Behavior, she discusses using what Alex Mesoudi calls the replacement method. It is best to describe the method using one of the examples from her paper. A small group of people was told to build paper planes, the aim being to build planes that fly the furthest. Each person got to build their own paper plane but they had the opportunity to see what the other members of the group did to build their planes. The replacement part of the method comes into the picture when after a short time one of the people in the group is replaced by a new member, such replacement being repeated until all the members of the group have been replaced a couple of times. The idea is to test for the appearance of a ‘tradition’ that continues across ‘generations’ of group members. Indeed, this was what Caldwell found, with later planes flying further and with planes made by one group being more alike when compared with planes built by other groups.

Caldwell ran the experiment on two different tasks, the other being to build the tallest tower out of spaghetti and modelling clay. What struck me about the paper was that here was a very simple methodology that nonetheless made it possible to make far-reaching conclusions, i.e. to show cumulative cultural evolution at work. As I have mentioned previously, it is precisely the papers that are able to do so much with so little that I find particularly fascinating from the philosophy of science point of view.

An even simpler method, also discussed by Alex Mesoudi in his paper in JSECP, is the transmission chain method. In that case, there is no group, all that happens is that the output of one generation of subject is given as the input for the very next generation – very much like the game of chinese whispers. So, a story might be told to the first subject, who is then asked to tell the story to the second subject and so on. Afterwards, the researchers can check to see how the story was changed by the generations of subjects. Not surprisingly, this very method was used by Justin Barrett to check that it is the case that minimally counterintuitive concepts are passed on more reliably than other concepts. The conclusion of that paper – pretty much, yes.

Apart from liking these methodologies for their ‘complexity to power ratios’, I find both of them potentially useful for testing claims about ritual traditions. Of course, this assumes that it should be relatively easy to get people to engage in ritualised behaviour in the lab but I happen to think that this is very much the case.