Framing, or Relevance Realization, in child and chimp. Possible root for creativity, superstitious behaviour, and effective learning?
Deric Bownd reports on research by Lyons et al. demonstrating that when being taught a new skill, young children are more inclined than chimps to imitate behaviours irrelevant to the task being taught. This suggests that children initially apply relatively inclusive frames in conceiving of novel situations. This could be a generative factor behind creativity, superstition, and learning itself.
In Lyons et al.’s study, adults taught young children and chimps a new skill. Included in the lessons were planned task-irrelevant behaviours. It was found that children would often imitate these inconsequential behaviours, while chimps tended to intelligently ignore them. Children would even do this when they were encouraged to complete the task as quickly as possible. It seems that the children had included the task-irrelevant behaviour in their causal intepretation of the new skill.
A reader of Bownd’s blog, Rick Thomas, made the clever suggestion that this tendency of children to engage in over-imitation may extend to children’s imitation of religious and political behaviours. Bownd considers this to be an interesting idea worthy of consideration. What comes to mind for me is superstitious thinking and behaviour, generally speaking. If people establish a pattern of behaviour early on that includes superfluous behaviours, they may carry the unneeded behaviours with them in the future. This, however, is pure speculation and is hardly a complete explanation of such behaviour. A complete explanation would also require one to account for why many people have trouble extinguishing behaviours that they know to be irrelevant. Moreover, it is well known that individuals beyond their early development can develop supersitious tendencies. A few questions emerge from this. Firstly, would older children, adolescents and adults also imitate more task-irrelevant acts than chimps in learning a new skill? If not, what types of factors account for the development of superstition in more developmentally mature individuals?
The tendency of young children to spread their information net wider than is prudent may also be a source of their lauded creativity. By considering more elements of the situation than is culturally normative, the child may open doors to insights and novel ideas.
At its core, the type of learning studied by Lyons et al. constitutes encoding the identities and relations of factors relevant to some procedure. When the child is first learning something they obviously don’t know what these relevant factors and relations are—this is the starting point of learning. The Frame Problem—that is, its cognitive science application—refers to the problem of navigating in a world in which there is an infinite amount of potentially relevant information. We can’t consider all of it given our finite processing power and time, we can’t consider none of it or we’ll never learn anything, and we cannot just consider arbitrary aggregates of information as statistically we’d be all but guaranteed to fail pretty much all the time. The frame problem is the problem that cognitive agents solve: being able to focus in primarily on what is relevant in a given situation.
The findings of this study seem to gel well with how children have to go about learning. As stated above, they can’t consider everything, nothing, or arbitrary aggregates of information. They also can’t just zoom in on the perfect combination of information, as then there would be no learning—they would already know. It seems that what must happen is that they generally zoom in to a pretty high level of resolution, narrowing down the possible sets of relevant factors and relations substantially but not completely. Through learning they prune out what is irrelevant—though, for some reason they—and indeed people of all ages—can maintain inefficient vestiges from earlier points in the learning process, which we call superstitious behaviours.
How children manage to solve the frame problem—that is, to crop out most of the irrelevant factors and relations—is one of the core problems in cognitive science and artificial intelligence. Clearly, we have evolved in a cognitive niche in which the tendency to attend to certain types of information as opposed to others was selected for, as was the tendency to learn to make certain types of inferences with limited informational input (e.g., that things fall down when not supported, rather than falling sideways or exploding).