Monday, December 10, 2012

Infants and imitation


There is a pendulum that swings back and forth in the study of child development.  The work of Piaget generally assumed that infants have very little knowledge and react to the sights and sounds of the world as their movement abilities improve.  Gradually, children learn more complex concepts, and in the end adults have an impressive capacity to reason.
Starting in the 1980s, there was a challenge to this view.  Studies of young infants suggested that they have a much greater understanding of the world than Piaget proposed.  For example, studies by Renee Baillargeon and her colleagues suggested that infants as young as 5-months-old understood that objects that were hidden by a screen were still there.  These studies led to a flurry of research exploring infants’ abilities.
More recently, the pendulum has started to swing back.  The difficulty with doing studies with infants is that you can’t ask them specific questions, because their language abilities are just developing.  Instead, you have to find indirect ways to explore what they are thinking. Then, you have to interpret what infants’ must have been thinking in order to have acted that way.  It is easy to fall into the trap of rich interpretation, which you (as an adult) think about what would have made you act that way.  That can lead you to overestimate what infants know.
This pendulum can be seen in research on imitation in infants. 
Anyone who has been around an infant for any length of time can see their imitation skills in action.  I have a great picture of my brother holding one of my kids who was about 5-months old.  Both of them are sticking their tongue out at the camera.  This picture was the result of an hour in which my brother stuck his tongue out, and then my son stuck his tongue out in response. 
How does this imitation happen?
One theory, which falls in the “infants don’t know that much” camp, assumes that imitation is based mostly on the way that infants represent movements.  This theory suggests that the way that people understand the movements of other people is based in part on using the parts of the brain that are associated with making movements.  In essence, when we watch someone else performing an action, our own movement system is engaged in simulating what the other person is doing.  Infants imitate, because this activity of the movement system prepares them to make the same movements they just observed.
A second theory, which is more in the “infants know a lot” camp, assumes that infants are trying to figure out what the person they are imitating was trying to do.  This view suggests that infants are constantly reasoning about why people do what they do, and then they imitate what someone intended rather than what they actually did. 
The view that infants are imitating the goals of someone else got support from a 2002 study in the journal Nature by Gyorgy Gergely, Harold Bekkering, and Ildiko Kiraly.  They had an adult sit in front of a box with a light on the top.  The light could be turned on by pressing on it.  The adult sat across from a 14-month-old infant (who was sitting on a parent’s lap).  In one condition, the adult had her hands on the table, and then bent over and pressed the light with her forehead to turn it on.  In the other condition, the adult was wrapped in a blanket so that her hands were not available to use, and she bent over and pressed the light with her forehead.  After watching the experimenter, the infant was allowed to play with the lamp.

The idea here was that if infants just imitate actions, then they should press the light with their forehead in both conditions of the study.  However, if infants are reasoning about what the adult is trying to do, then when the adult’s hands were under the blanket, infants should realize that the adult just wanted to turn the light on and used her head because her hands were not free to use.  Consistent with the idea that children were reasoning about goals, the infants were much more likely to touch their head to the light when the adult had her hands on the table (so that she clearly chose to use her head) than when the adult had her hands under the blanket.

A study in the July 2011 issue of the journal Child Development by Markus Paulus, Sabine Hunnius, Marlies Vissers, and Harold Bekkering suggests that children might not have been reasoning about the goal of the adult after all.

These researchers note that for infants to use their head to turn on the lamp, they have to put their hands on the table to steady themselves, because their stomach muscles are not strong enough to hold them up when leaning over without support.  That action is quite similar to the one the experimenter makes when putting her hands visibly on the table in the original study. 

To see whether children are focusing on actions they are able to perform, these researchers added three
more conditions to the study I just described (in addition to the conditions in which the adult had her hands on the table and the one where her hands were trapped under a blanket). 

In one condition, the adult had her hands under a blanket, but the blanket had an obvious button that was holding it closed so that the adult could easily use her hands if she wanted to.  If the infants are just reasoning about the adults’ goal, then they should use their heads to press the button, because the adult clearly chose to use her head.  However, if it was important that the adult had her hands on the table, then the infants would not use their head in this case.  In fact, infants rarely used their head to turn on the light when the adult had her hands under the blanket with the button holding it closed.

In another condition, the adult had her hands in the air and then bent over and pressed the button with her head.  Again, children reasoning about the adult’s goals should use their heads, because the adult clearly chose to use her head rather than her hands.  However, children can’t lean over with their hands in the air, so if they just wanted to imitate an action, they wouldn’t use their head in this case (because they would fall over).  Once again, infants rarely used their heads.

In the third condition, the adult started the task by playing with two foam balls.  Then, she held one ball in each hand and put her hands on the table while holding the balls and bent over and touched the light with her head.  Children reasoning about goals should not use their heads in this case, because they should see that the adult’s hands were occupied holding the balls and so she used her forehead out of convenience.  Children who are imitating an action they can perform should use their heads, though, because the adult had her hands on the table, and if the infants put their hand on the table, they could support themselves as they bent over.  In this condition, the infants often pressed he light with their head.

Putting all of this together, then, it looks like the pendulum has swung back toward a focus on the influence of movement ability on imitation.  Early on, infants are trying to imitate actions they can perform themselves.  These findings are quite consistent with a lot of research about the way that adults think about actions.  Generally speaking, even adults understand actions by thinking about how we could perform them ourselves.  It is just a short step from that understanding to doing the same thing.    In that way, infants may not be that different from adults.  Perhaps we’re guilty of rich interpretation of adult performance as well.