The Foundation of Moral Reasoning in Infants

Figure 1. Maybe babies know more than we think they do!

The work of Lawrence Kohlberg was an important start to modern research on moral development and reasoning. However, Kohlberg relied on a specific method: he presented moral dilemmas and asked children and adults to explain what they would do and—more importantly—why they would act in that particular way. Kohlberg found that children tended to make choices based on avoiding punishment and gaining praise. But children are at a disadvantage compared to adults when they must rely on language to convey their inner thoughts and emotional reactions, so what they say may not adequately capture the complexity of their thinking.

Starting in the 1980s, developmental psychologists created new methods for studying the thought processes of children and infants long before they acquire language. One particularly effective method is to present children with puppet shows to grab their attention and then record nonverbal behaviors, such as looking and choosing, to identify children’s preferences or interests.

A research group at Yale University has been using the puppet show technique to study moral thinking of children for much of the past decade. What they have discovered has given us a glimpse of surprisingly complex thought processes that may serve as the foundation of moral reasoning.

EXPERIMENT 1: Do children prefer givers or takers?

In 2011, J. Kiley Hamlin and Karen Wynn put on puppet shows for very young children: 5-month-old infants. The infants watch a puppet bouncing a ball. We’ll call this puppet the “bouncer puppet.” Two other puppets stand at the back of the stage, one to left and the other to the right. After a few bounces, the ball gets away from the bouncer puppet and rolls to the side of the stage toward one of the other puppets. This puppet grabs the ball. The bouncer puppet turns toward the ball and opens its arms as if asking for the ball back.

This is where the puppet show gets interesting (for a young infant, anyway!).  Sometimes, the puppet with the ball rolls it back to the bouncer puppet. This is the “giver puppet” condition. Other times, the infant sees a different ending. As the bouncer puppet opens its arms to ask for the ball, the puppet with the ball turns and runs away with it. This is the “taker puppet” condition. Although the giver and taker puppets are two copies of the same animal doll, they are easily distinguished because they are wearing different colored shirts, and color is a quality that infants easily distinguish and remember. It looks like this:

Each infant sees both conditions: the giver condition and the taker condition. Just after the end of the second puppet show (i.e., the second condition), a new researcher, who doesn’t know which puppet was the giver and which was the taker, sits in front of the infant with the giver puppet in one hand and the taker puppet in the other. The 5-month-old infants are allowed to reach for a puppet. The one the child reaches out to touch is considered the preferred puppet.

Remember that Lawrence Kohlberg thought that children at this age—and, in fact, through 9 years of age—are primarily motivated to avoid punishment and seek rewards. Neither Kohlberg nor Carol Gilligan nor Jean Piaget was likely to predict that infants would develop preferences based on the type of behavior shown by other individuals.

Work It Out

The puppet show is over and the experimenter is holding the two dolls—the giver puppet and the taker puppet—in front of the infant. The reaching behavior of the infant is being videotaped for later analysis.

What do you think? Make a prediction about the results of this study—which should reflect your own theory of an infant’s ability to judge and care about the types of behavior others display. Do you think infants will choose the taker or the giver puppet? Do you expect the results to be significant?

INSTRUCTIONS: Adjust the pink bar on the left to show the percentage of infants who reached for the giver puppet. The yellow bar on the right will automatically adjust to make the total (sum of both bars) equal 100%.

Show Answer

Here are the results from Experiment 1:

Figure 2. Results from Experiment 1.

Experiment 1 suggests that 5-month-old infants are not just passive observers. They notice what others do and, if we are interpreting the results of experiments like this one correctly, they distinguish helpful behaviors (“prosocial behaviors”) from behaviors that hurt others (“antisocial behaviors”). But they do more that that. They are attracted to those who are acting in a prosocial way, and they reject those who act in an antisocial way.

These researchers also tested infants who were only 3-months old. These infants were so immature that they did not yet have good control of their arms, so the experimenter could not use “reaching for one of the puppets” as the dependent variable, as they did with the 5-month-olds. Three-month-old infants can control where they look quite well, and previous research has indicated that very young infants will look longer at objects they want. The researchers showed these very young infants the same puppet shows that were described above and then, during the choice phase, they recorded which puppet (giver or taker) the 3-month-olds looked at longer. The results were very similar to those found with the 5-month-olds. A strong majority of younger infants (92%) looked longer at the giver puppet than the taker puppet.

But this isn’t the end of the story…

EXPERIMENT 2: Do infants judge others based on their behavior?

In the research world, the early attempts to study something, when the researchers work to develop a solid and reliable research procedure, is often the most challenging time. Once the researcher works through initial problems and issues and begins to get consistent results, they can gain a deeper understanding by adding new variables or testing different groups of subjects (e.g., older children or children with some interesting psychological characteristics).

The study you just read about is an example of a simple, basic study. The researchers found that infants preferred puppets that help another puppet (the puppet in the giver condition) over puppets that are not nice to another puppet (the puppet in the taker condition). A common sense interpretation of this simple result is that infants like nice behavior and they dislike hurtful behavior. And perhaps that is as complicated as an 8-month-old infant’s thoughts can be. But maybe not.

Dr. Hamlin and her colleagues wondered if infants might consider more factors when judging an event. Adults generally prefer situations where good things happen to someone rather than something harmful. However, when adults see someone do something bad, they may find satisfaction in seeing that person punished by having something bad happen to them. In a nutshell: good things should happen to good people and bad things should happen to bad people. This is what is called “just world” thinking, where people get what they deserve.

In the study we will call Experiment 2, Hamlin’s team tested 8-month-old infants and repeated the procedures from Experiment 1 with a major addition. In Experiment 1 (described above), the puppet bouncing the ball was a neutral character, neither good nor bad. In Experiment 2, the infants saw 2 different shows. First, they saw the bouncer puppet either helping or hindering another puppet. Then, they watched the same ball-bouncing puppet show. Here is what happened:

• Puppet Show #1:  A puppet is trying to open a box, but cannot quite succeed. Two puppets stand in the background. For some infants, as the first puppet struggles to open the box, one of the puppets in the back comes forward and helps to open the box. This is the helper puppet. For other children, as the first puppet struggles, a puppet comes from the back and jumps on the box, slamming it shut. This is the hinderer puppet. Each infant sees only a helper or a hinderer—not both. Here is a video showing the helper puppet situation:

• Puppet Show #2: Just after the infants have watched the first show, the second puppet show begins. This is the show that you read about in Experiment 1. The only thing that is new is that the bouncer puppet, the one that loses the ball, is either the helper puppet from Puppet Show #1 or the hinderer puppet from Puppet Show #1. Each infant sees this puppet lose the ball to a giver, who returns the ball, and to a taker, who runs off with the ball.

This video demonstrates show #2.  The elephant in the yellow shirt from the first show is now bouncing a ball. After dropping the ball, the moose in the green shirt gives it back to him, while the moose in the red shirt takes it away.

So far we have concluded that even young babies prefer the “nice” puppet and show a preference for a puppet who helps another puppet. But this only happened when the bouncer puppet was the helper from the first puppet show. What if, instead of the nice elephant in the yellow shirt bouncing the ball, the elephant in the red shirt (the one who jumped on the duck’s box, remember?) was the one bouncing the ball? Imagine the same scenario: the mean elephant in the red shirt is bouncing the ball, he drops it, and the moose in the green shirt gives it to him or the moose in the red shirt takes it away.

Figure 3. This bar graph shows the results of Experiment 2 for 8-month-old infants. The blue bars show the preferences for the infants who saw the helper from the first show as the bouncer in the second. Bar A is taller than Bar B, showing the greater choice of the giver than the taker puppet. The red bars show the reverse effect. The babies strongly preferred the taker (Bar C) to the giver (Bar D) when the puppet bouncing the ball had been the hinderer, who jumped on the box in the first show.

So now things are getting interesting, right? Do 8-month old infants understand the concepts of revenge or justice? We must always be careful when labeling behaviors of children (or animals) with characteristics we use for human adults. In the description above, we have talked of “nice puppets” and “mean puppets” and used other loaded terms. It is tempting to interpret the choices of the 8-month-olds as a kind of revenge motive: the bad guy gets its just desserts (the hinderer puppet has its ball stolen) and the good guy gets its just reward (the helper puppet is itself helped by the giver). Maybe that is what is going on, but we encourage you to consider these very sophisticated types of thinking as merely one hypothesis. Remember the facts—what did the puppets do and what choices did the infants make?—without committing yourself to the adult-level interpretation.

The researchers believe that this type of thinking, which is remarkably sophisticated, takes some cognitive development. They tested 5-month-olds using the same procedures, and the results with these younger infants were different. The 5-month-olds showed an overwhelming preference for the giver puppets, regardless of who was bouncing the ball. Maybe it is too complex for them to understand that the bouncer puppet in the second show was the same puppet from the first show. Or perhaps their memory processes are too fragile to hold onto information for that length of time. Maybe the revenge motive is too advanced. Or maybe something else is going on. What is clear is that 5-month-olds and 8-month-olds respond differently to the situations tested in the second experiment.

EXPERIMENT 3: Do infants judge others based on their preferences?

Across the first two experiments, infants appear to prefer puppets (and, by extension, maybe people, as well) that are helpful over those that are not helpful. Experiment 2 complicated our story a bit, but it still appears that prosocial behavior is attractive to infants and antisocial behavior is unattractive. But another experiment, again using the bouncing ball show, suggests that infants as young as 8-months of age may have some other motives that are less altruistic than the first two experiments indicate.

In a study by Hamlin, Mahanjan, Liberman, and Wynn from 2013, 9-month-old infants watched the bouncing ball show, but with a new twist.

At the beginning of the experiment—Phase 1—the infants were given a choice between graham crackers and green beans. The experimenters determined which food the infant preferred.

You can view the transcript for “Graham Cracker Choice” here (opens in new window).

Then, in Phase 2, the infants watched a puppet make the same choice. For half of the infants, the puppet chose the same food that they preferred, saying “Mmmm, yum! I like ___(graham crackers or green beans)!” and saying “Eww, yuck! I don’t like _____ (graham crackers or green beans!”  This was called the SIMILAR condition because the puppet was similar to the child in its food preference. For the other half of the infants, the puppet chose the other food, choosing graham crackers if the infant preferred green beans and preferring green beans if the infant liked graham crackers. This was the DISSIMILAR condition.

You can view the transcript for “Similar / Dissimilar Puppet Preference Example” here (opens in new window).

Why did the experimenters do this? They wanted to know if young children form in-groups and out-groups by perceiving some people as being like them and other people as being unlike them. The experimenters noted in their research introduction that we (adults) are influenced by our perception that others are similar to us or not like us. We tend to project positive qualities—being trustworthy, intelligent, kind—on people we perceive as similar to ourselves, and people we see as unlike us are seen as having negative qualities—being relatively untrustworthy, unintelligent, and unkind.^[1]

Of course, there is a big difference between claiming that adults use similarity to make judgments about others and saying that infants less than a year of age do the same thing. However, the researchers note that some recent research has suggested that infants less than a year old are more likely to develop peer friendships with other infants who “share their own food, clothing, or toy preferences” compared to those who don’t.

So, back to the experiment. In Phase 3, the infants either saw a similar puppet (one that chose the food the baby preferred) or a dissimilar puppet (one that chose the food the baby did not prefer) bouncing the ball. As in the other experiments, the ball got away from the bouncer and rolled to the back of the stage. In one instance, the giver puppet returned the ball and, in the other instance, the taker puppet ran away with the ball. Finally, in Phase 4, the 9-month-old baby was shown the giver and taker puppet and the experimenters recorded which of the two puppets the baby preferred (reached out to touch). This video shows the dog in the light blue shirt giving the ball back to the red bunny that preferred graham crackers.

Here is a summary of the four phases in Experiment 3:

• Phase 1: The infant chooses graham crackers or green beans.
• Phase 2: The bouncer puppets choose graham crackers or green beans.
  • Similar condition: The bouncer chooses the same food that the infant chose.
  • Dissimilar condition: The bouncer chooses the food that the infant did not choose.
• Phase 3: This is the same bouncing ball experiment that you have been reading about.
  • Remember that each child sees both the Giver and Taker shows.
• Phase 4: This is the same choice—Giver or Taker—that was the final phase in the other two experiments

Work It Out

Now make predictions for the results. Here is a matrix picture of the design of the experiment:

INSTRUCTIONS: Adjust bars A and C to make your predictions. Bar A represents the “nice” puppet who gave the ball to the bouncer puppet that liked the same food as the child, while bar B represents the “mean” puppet who took the ball away from the bouncer puppet who liked the same food as the child. Bar C represents the “nice” puppet who gave the ball back to the puppet who did not like the same food as the child, and bar D represents the puppet who took the ball away from the puppet who did not like the same food.

As before, move the bars on the left to indicate the percentage of infants preferring the giver puppet in the similar condition (purple bars) and in the dissimilar condition (green bars). The bars on the right will adjust to make the total in each of the similarity conditions equal 100%.

After you have recorded your predictions, click the “Show Answer” link to see the results from the experiment.

Show Answer

Here are the results from Experiment 3:

Figure 4. Results from Experiment 3.

These results are similar to those for the 8-month-olds in the previous experiment. But remember that, in this experiment, the variable that distinguishes the two bouncer puppets was a food choice, not the prosocial or antisocial behavior in Experiment 2. If we take the results from Experiments 2 and 3 together, the results here suggest that the similar puppet is being treated as if it is nice or good. Puppets that treat this similar puppet in a nice way are preferred. Conversely, the dissimilar puppets are treated as if they have done something negative and puppets that treat these dissimilar puppets badly are preferred.

The experimenters also tested an older group of babies that were 14-months-old. The results for these older babies were consistent with the 9-month-old and, if anything, the effects were stronger. Their results showed that all infants preferred when the giver puppet was nice to the puppet that was similar to them and all infants preferred when puppets were mean to the puppet that was dissimilar to them.

Figure 5. These bar graphs show the results of the experiment when 14-month-olds were tested. One hundred percent of the children chose the puppet that gave the ball back to the puppet that was similar to them, and 100% of children chose the puppet that took the ball away from the puppet that had a different preference than they did.

CONCLUSIONS

This exercise started with a reminder that Lawrence Kohlberg found that children went through a long developmental process in their moral reasoning. Based on children’s reasoning aloud about moral dilemmas, Kohlberg concluded that children younger than about 8 or 9 years of age make moral decisions based on avoiding punishment and receiving praise. Neither his research nor that of most others in the 1970s and 1980s suggested that young children would use multiple sources of information and judgments about the meaning of behaviors in their thinking about what sorts of behaviors are better or worse.

If Dr. Hamlin and her colleagues are right, then infants are much more sophisticated and complex in their thinking about the world than these earlier researchers thought. In Dr. Hamlin’s view, infants like good things to happen to good puppets and people, and bad things to happen to bad puppets and people. Experiment 3 suggests that they make judgments about more than helping and harming behavior. They prefer others who are like them (green beans vs. graham crackers) and they don’t mind if others who are not like them have unpleasant experiences.

The research we have been reviewing is just part of an impressive set of research on infant thinking. The ideas that the researchers have developed are intriguing and they are consistent with the modern view of the infant as an active, creative thinker. At the same time, remember that science doesn’t rest on an early set of explanations based on a small set of complicated experiments. Science pushes beyond what we currently know and believe. This starts with curiosity on your part. Are the experimenters correct in interpreting reaching behavior as showing a preference or is something else going on? Do infants really prefer prosocial behaviors to antisocial behaviors, or is there some other explanation for their preferences? How else could we test the moral judgments of infants without using puppet shows? The next generation of creative scientists will push beyond what we know now, with new research methods and new ideas about the mind.

We’ll give Dr. Hamlin the last word. Here is part of her conclusion section from an article that summarizes some of the research we have been studying: “In sum, recent developmental research supports the claim that at least some aspects of human morality are innate…Indeed, these early tendencies are far from shallow, mechanical predispositions to behave well or knee-jerk reactions to particular states of the world. Infant moral inclinations are sophisticated, flexible, and surprisingly consistent with adults’ moral inclinations, incorporating aspects of moral goodness, evaluation, and retaliation.“ (Hamlin, 2013, p. 191)