False belief for chimps

One of the classic experiments in developmental psychology revealed a fundamental shift in social cognition that occurs around 4 years old. The Sally-Anne task was used to test whether children understood that another person could have a belief about the world that was untrue - a so-called false belief. The task was simple. The experimenter presents the child with two dolls, Sally and Anne. Sally and Anne decide to hide a piece of chocolate in a drawer (for example). Next Sally leaves the room and, while she is gone, Anne moves the chocolate hiding it under a pillow. The question for the child is: when Sally returns where will she look for the chocolate. The correct answer is the drawer, because Sally did not witness the change in location - in other words, she has a false belief about where the chocolate is. Young children, however, respond that she will look under pillow. They do not grasp that she can have a false belief.

Now, much work has been done on this topic modifying tasks to get younger children to give a correct answer. But the limit was pushed recently by a non-verbal task used with infants by Rene Baillergeon. The results are controversial and hence in need of replication. So of course, the researchers here tried a version on the chimps. Chimpanzees certainly understand certain kinds of mental content. They will look where you are looking, following your gaze behind barriers - a sign that they know you can see things they cannot. But full false belief is difficult to demonstrate non-verbally.

The study required training chimps to associate bananas with one color container (blue) and grapes with another (white). [This simple association took about 500 trials per chimp. Actions not part of their normal ecology are very difficult to learn.] The apes were shown the food first and then watched as the experimenter placed it in a yellow box. An occluder was raised while the experimenter placed the yellow box under the appropriate (overturned) container. The subject then gets to guess where the food is. So if the chimp is shown banana, he should point to the blue container to get the food.

Once the subjects have this down, a further step is added. After the experimenter hides the yellow box with the food (lets say banana again), she leaves the room. A different experimenter enters, exposes the yellow box and changes the contents to a grape, then places it under the banana container again. The first experimenter enters again and lets the chimp choose. Now, this is tough to follow but the ape needs to understand that although the food is now grape, the experimenter thinks it is banana and expects it to be where bananas usually go - the blue container. But the apes point to the grape container instead (white). I'm surprised they get this far given the task complexity.