Causal opaqueness, imitation and emulation

Posted on May 5, 2009


I am very glad that I went to the St Andrews meeting. In the days since I’ve returned I’ve found myself thinking back to a number of the things discussed there and realising their significance for my own work. As an example of completely non-magical synchronicity, Tom Rees has recently blogged on a couple of papers that cover some of the same ground from a somewhat different angle. The basic issue is that of the mechanisms of the cultural transmission of superstitions.

Among the people I met at EHBE was Claudio Tennie, who is currently finishing his PhD with Michael Tomasello in Leipzig. He is doing comparative research with non-human primates and argued for the Tomasello view that culture is uniquely human – a position which is not actually all that popular among comparative psychology people due to the discovery of numerous examples of what appear to be cultural differences among several non-human animal species. The view is based upon a distinction between imitation and emulation where imitation involves a precise copying of the behaviour of a conspecific while emulation only involves the copying of whatever the final product is. According to Tomasello non-human primates only emulate, with imitation being solely the province of humans. Thus, the seemingly counterintuitive result that human children will ‘parrot’ the way that an adult opens a box, including actions that are unnecessary, while a chimp will generally learn to open the box in the most efficient way.

Both emulation and imitation have their strengths and weaknesses. The important point about imitation is that it makes it possible for children to learn how to interact with objects that are causally opaque – ones that emulation would not be capable of helping us with. So, humans are not only capable of opening a simple box but, after someone has shown them how, they are capable of interacting with a computer running Windows. There is a downside to this, however, that may be obvious from the above example of imitation. Imitation does not necessarily differentiate between relevant and irrelevant actions so that non-effective actions may be imitated along with the significant ones so long as they are part of the behaviour that is being imitated.

This is the issue that is pursued by one of the articles mentioned by Tom Rees. The research, conducted by Lyons, Young and Keil examines the degree to which children imitate irrelevant behaviour. In particular, it tries to distinguish between this thesis and the previously significant thesis that overimitation is due to children’s desire to please the adults they are imitating.

Young children are surprisingly judicious imitators, but there are also times when their reproduction of others’ actions appears strikingly illogical. For example, children who observe an adult inefficiently operating a novel object frequently engage in what we term overimitation, persistently reproducing the adult’s unnecessary actions. Although children readily overimitate irrelevant actions that even chimpanzees ignore, this curious effect has previously attracted little interest; it has been assumed that children overimitate not for theoretically significant reasons, but rather as a purely social exercise. In this paper, however, we challenge this view, presenting evidence that overimitation reflects a more fundamental cognitive process. We show that children who observe an adult intentionally manipulating a novel object have a strong tendency to encode all of the adult’s actions as causally meaningful, implicitly revising their causal understanding of the object accordingly. This automatic causal encoding process allows children to rapidly calibrate their causal beliefs about even the most opaque physical systems, but it also carries a cost. When some of the adult’s purposeful actions are unnecessary—even transparently so—children are highly prone to mis-encoding them as causally significant. The resulting distortions in children’s causal beliefs are the true cause of overimitation, a fact that makes the effect remarkably resistant to extinction. Despite countervailing task demands, time pressure, and even direct warnings, children are frequently unable to avoid reproducing the adult’s irrelevant actions because they have already incorporated them into their representation of the target object’s causal structure.

The methodology seems particularly careful to me and I find the results very convincing. The overimitation appears to be very robust even in situations where the social pressure appears to be unequivocally against imitating non-effectual actions. The relevance for superstitions is fairly obvious as this seems to provide a possible mechanism by which superstitious practices are perpetuated. Even so, a number of questions have to be asked. The first is to what degree what it true of children is also the case with adults? Clearly, adults are more capable of distinguishing effective and non-effective actions. The dimension of causal opaqueness of whatever is being interacted with is going to be key here, it seems. The experiments were actually carried out with set-ups that were unfamiliar but not causally opaque to children of the age that was tested. This is known as controls were able to open the boxes without adult instruction and with fewer non-effective actions. Not surprisingly, superstitions appear to involve situations that are causally opaque – a notion that appears to be potentially connected to loss of control that is also known to be correlated with superstitious responses.

There is a further issue raised by the article, however. The researchers found that the overimitation effect was affected by the non-effective actions being performed on objects that were not physically connected to the box that was to be opened. So, it seems that some general ideas from folk psychology place a limit upon what is seen as relevant. The question for me is how this interacts with the postulation of supernatural causes that is part and parcel of superstitions proper. The relevant experiment seems to be relatively easy to do. One could simply check if priming with stories of the magical affects this as well as checking if being presented with such a situation changes responses to tests of supernatural beliefs such as the Tobacyk scale.