The hidden structure of overimitation

doi:10.1073/pnas.0704452104

The hidden structure of overimitation

*Department of Psychology, Yale University, Box 208205, New Haven, CT 06520; and
^‡Department of Psychology, University of Wisconsin, Brogden Hall, Madison, WI 53706

Edited by Susan E. Carey, Harvard University, Cambridge, MA, and approved October 18, 2007 (received for review May 11, 2007)

Abstract

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.

Footnotes

^†To whom correspondence should be addressed. E-mail: derek.lyons{at}yale.edu
Author contributions: D.E.L., A.G.Y., and F.C.K. designed research; D.E.L. and A.G.Y. performed research; D.E.L. analyzed data; and D.E.L. and F.C.K. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
↵ § The causal opacity of devices like computers is obvious, but the problem extends to much simpler artifacts. As soon as tool use decouples goal-directed actions from immediately observable goal states (such as occurred when early hominids began to use tools recursively), causal learning quickly becomes an intractable inferential problem (1, 2, 26, 27).
↵ ¶ Of course, even simple actions can be encoded in multiple ways, differing in the level of detail that is absorbed from the display (28, 29). We return to this issue in Experiment 1A, comparing the exact “style” of children's actions to that of the adult to determine the granularity with which the hypothesized causal encoding process occurs.
This article contains supporting information online at www.pnas.org/cgi/content/full/0704452104/DC1.
↵ ‖ Neither age nor presentation order ever had a significant effect on overimitation, so subsequent data are all similarly collapsed.
↵ ** Overimitation remained independent of training outcome in all subsequent experiments.
↵ †† Experiments 2A and 2B were run concurrently with the same participants. Because the instructions for Experiment 2A could have biased future responses, the Experiment 2B puzzle object was always presented first.
↵ ‡‡ One might ask why contact principle violations would be expected to block overimitation when violations of the similarly foundational “efficiency principle” (32) (i.e., the adult operating mechanisms in a suboptimal way) did not diminish the effect (Experiment 1A). We return to this matter in the Discussion section of SI Text, arguing that contact principle violations undermine the causal plausibility of the target object in a way that efficiency violations do not.
Freely available online through the PNAS open access option.