How children see desire and belief in others.

by Jonathan Freedlander
Science and Consciousness Review, May 2004

"Agency" is the idea that an individual’s actions are motivated by their own beliefs and desires. Children learn to understand other people as agents pursuing their own goals and driven by their thoughts. This is fundamental to social interaction; without it, one would be unable to predict and anticipate the actions of others, and would therefore have great difficulty functioning socially. As such, much research has examined the ways in which children learn to understand human agency.

The understanding of human agency in the first several years of life typically comes in three stages. In the first year or so of life, children are thought to acquire the understanding that people do not act blindly, but rather follow a principle of rational action; that is, people behave in rational, meaningful ways. By age two or three, children begin to attribute behavior to simple motivations; action is interpreted as being the result of desires. Lastly, around age four or five, children incorporate rational action with representational attributions, that is, they come to realize that agents’ desires are acted upon in congruence with their beliefs. Exactly when the transition to this last, and most important, stage occurs remains controversial. Some research has observed evidence for representational reasoning in three year olds, though most data suggests that this skill isn’t stable until age five.

The majority of the research in this area has examined this process in European-American and East Asian children; few studies of children of other cultures have been done. Additionally, most of the available research has focused on children’s acquisition of agency concepts as they pertain to other humans; however, we do not only see people as agents. Animals, particularly our pets, are often attributed at least some degree of agency, as are a variety of disincarnate entities, such as deities, ghosts, angels, etc. It has been thought that children first develop human agency concepts, and then apply them as templates for other beings.

However, some research suggests otherwise, that the development of non-human agency is a distinct process from that of human agency. The present study indicates that this non-similarity paradigm, as Knight calls it, is evident cross-culturally. Forty eight Maya children from small villages in the Quintana Roo state of southeastern Mexico were divided into four age groups and tested on two false-belief tasks. A false-belief task is one of the more common ways of attaining insight into children’s understanding of beliefs. In one false belief task, known as the Sally-Ann test, a child is typically shown a scene in which two dolls, representing humans, are animated by experimenters. Previous research has shown that using dolls as proxies for real humans does not affect the outcome of the experiment. The two dolls enter the stage, and one of them, Sally, places an object in one of two containers. Sally then exits the stage, and while she is away, the other doll, Ann, moves the object to the other box. Finally, Sally reenters the stage, and the experimenters ask the children where they think the she would look for the object.

The present study examines Maya children’s responses to a similar false-belief task, known as the “surprising contents” task. In the prototypical surprising contents task, children are shown a cracker box, with a picture of crackers on it, and are asked what they think is inside. An experimenter then opens the box, revealing not crackers, but some completely unrelated object, such as a rock. The experimenter closes the lid of the box and confirms that the children understand what’s really in the box. A doll is then brought on stage who has not seen the inside of the box, and the children are asked what the doll thinks is in it. To make this exercise culturally relevant to the children in this study, a ho’ma, a container made of a dried squash used to keep to tortillas warm, was used in place of a cracker box.

The children were presented the ho’ma with the opening covered by a piece of cardboard, so they couldn’t see the contents. They were asked what they would normally expect to find in it, and naturally answered “tortillas”. When the cover of the ho’ma was removed, however, the true contents were revealed to be in fact, a pair of shorts. The children were then presented with two dolls; one, named Soledad, was to represent an ordinary person, and the other was to represent the Catholic god (which the Maya incorporated into their pantheon a few centuries ago). Finally, the children were asked what they thought each doll would probably expect to find in the ho’ma.

Answers for the human doll showed a statistically significant positive correlation with age. As age increased, Maya children were more likely to ascribe false beliefs to humans; that is, they realized that humans would likely incorrectly guess that there were tortillas in the ho’ma. However, as predicted by Knight et al, a significant correlation was not apparent between age and responses for the God doll. There were significant differences between responses for the God and human dolls in 5 year olds and in 7 year olds, though not for the 4 and 6 year old groups.

The results of this study lend support to Knight et al's assertion that children may not need to use an understanding of human behavior to think about God. Children, in multiple cultures, seem capable of understanding that certain non-human entities such as God as having special attributes, and the development of this understanding appears to follow a different pattern than the development of human false-belief understanding. This does not necessarily imply that children truly understand God as a different sort of agent than humans, rather than as just a human with certain special powers. But it does offer further evidence against Piaget’s assertion that children cannot think about non-human agents in notably different ways than humans. Indeed, some research has found that though 4 year olds understand that humans make machines, not God, when asked who made rocks or mountains, they answer God’. Furthermore, children have been found to view magicians as a special type of agent, able to perform acts that defy rational explanation. However, the exact nature of the development of these non-human agency concepts remains elusive, and warrants further study.

Copyright © Jonathan Freedlander, 2004

Jonathan Freedlander

University of Maryland, Baltimore County

More from Jonathan Freedlander : How we see ourselves.

Intro. slides to psychological social development [pdf]

Knowing about Knowing [pdf]

Does an autistic child have a theory of mind?

Psychological and social characteristics of autism

Theory of Mind and Autism

Knight, N., Sousa, P., Barrett, J. L., Atran, S. (2004). Children’s attributions of beliefs to humans and God: cross-cultural evidence. Cognitive Science, Vol. 28, 117 126.