Innate Morality? Human Babies Weigh In.

by Kenneth W. Krause.

Kenneth W. Krause is a contributing editor and “Science Watch” columnist for the Skeptical Inquirer.  Formerly a contributing editor and books columnist for the Humanist, Kenneth contributes regularly to Skeptic as well.  He may be contacted at

In 1762, Rousseau characterized the human baby as “a perfect idiot.”  In 1890, William James judged the infant’s mental life to be “one great blooming, buzzing confusion.”  We’ve learned much about early human cognition since the nineteenth century, of course, and the current trend is to assign well-expanded mental capacities to young children.  But intellectual battles continue to rage, for example, over the possibility of an innate and perhaps nuanced moral sensibility.

Indeed, psychologists split last summer over the question of whether preverbal infants are capable of evaluating the social value of others.  Back in 2007, Yale University researchers led by J. Kiley Hamlin claimed to have demonstrated that infants can morally assess individuals based on their behavior toward third parties (Hamlin, et. al. 2007).  Those findings were challenged last August, however, by postdoctoral research fellow, Damian Scarf, and his colleagues from the University of Otago in New Zealand (Scarf, et. al. 2012).

Hamlin’s pioneering study deployed three experiments on six- and ten-month-old babies to test her team’s hypothesis that social evaluation is a universal and unlearned biological adaptation.  In all trials infants observed characters shaped like circles and either squares or triangles moving two-dimensionally in a scene involving an artificial hill.  Parents held their children during the program, but were instructed not to interfere.

In experiment one, the characters were endowed with large “googly eyes” and made to either climb the hill, hinder the climber from above, or help the climber from below.  With looking times carefully measured, the infants observed alternating helping and hindering trials.  The question here was whether witnessing one character’s actions toward another would affect infants’ attitudes toward that character.

When encouraged to reach for either the helper or the hinderer, twelve of twelve six-month-olds and fourteen of sixteen ten-month-olds chose the helper.  But might the babies have responded to superficial perceptual, rather than social, aspects of the experiment?  For example, perhaps the infants merely preferred upward or downward movements.  In an attempt to rule out that possibility, Hamlin modified a single test condition and gathered a second group of children.

In experiment two, the object pushed was represented as inanimate.  Its googly eyes were detached and it was never made to appear self-propelled.  If the infants had chosen based on mere perceptual events in the first experiment, Hamlin proposed, they should express an analogous preference for the upward-pushing character in the second.  But that didn’t happen.  Only four of twelve six-month-olds and six of twelve ten-month-olds picked the upward-pushing shape.

So the team decided that three possibilities remained.  The infants might positively evaluate helpers, negatively evaluate hinderers, or both.  To determine which, Hamlin assembled a third group of children, reattached the googly eyes, and altered the experimental design to include a neutral character that would never interact with the climber.

In the final experiment, then, children first observed either a helper or a hinderer interact with a climber as in experiment one.  Thereafter, they witnessed a neutral, non-interactive character that moved uphill or downhill in the same way.

When prompted to choose, infants reacted differently toward the neutral shape depending on the character with which it was paired.  Seven of eight babies in each age group preferred the helper to the neutral character and the neutral character to the hinderer.  Hamlin thus inferred that her subjects were fond of those who facilitate others’ goals and disapproving of those who inhibit them.

“Humans engage in social evaluation,” the Yale researchers concluded, “far earlier in development than previously thought.”  The critical human ability to distinguish cooperators and reciprocators from free riders, they agreed, “can be seen as a biological adaptation.”

Having viewed recorded portions of these experiments, I felt compelled to question some of the program’s most basic assumptions and methods.  Can infants fathom, for instance, what artificial landscapes represent, or what “hills” look like?  Can they grasp the symbolic significance of squares, triangles, and circles adorned with “googly eyes”?

Although groundbreaking in its own right, Hamlin assured me that her 2007 study was built on a solid foundation of previous experiments employing both a hill and a helping/hindering paradigm.  Numerous analyses, she insisted, have shown that infants will interpret even two-dimensional animations as real, and often attribute goals and intentions to basic shapes engaging in apparently self-propelled movement—with or without artificial eyes.

I also wondered how the infants were “encouraged” to choose.  In the video, characters were shaken by the person holding them.  Could that have affected the outcome, perhaps combined with verbal inflection?  Was one character ever held closer to an infant than the other, or at least closer to the infant’s dominant hand?

Her presenting colleagues, Hamlin answered, were always blind to the condition—i.e., ignorant of which character was helper or hinderer for that particular baby.  So, if differences in proximity or emphasis existed, their effects would have been divided randomly across subjects.  Also, she noted, parents were instructed to close their eyes during choice phases.

Scarf responded quite differently.  He sees no reason to believe, for example, that six- and ten-month-olds would interpret Hamlin-esque displays as landscapes, or that they would be familiar with the concept of a hill.  Nor could they distinguish between helping and hindering, he argued.  And while infants may attribute intentions and goals to animate objects, he added, no convincing data suggests they might assign relevant feelings to them as well.

Five years passed before Scarf’s team would offer a conflicting explanation—the “simple association hypothesis”—for the infants’ remarkable behavior.  While inspecting Hamlin’s videos, Scarf distinguished “two conspicuous perceptual events” during the helper/hinderer trials: first, an “aversive collision” between the climber and either the helper or the hinderer, and second, a “positive bouncing” when the climber and helper reached the hill’s summit.

Rather than rendering complex social evaluations, Scarf proposed, Hamlin’s babies may have simply been reacting to a visual commotion.  The hinderer was perceived negatively, he hypothesized, because it was associated only with an aversive collision.  The helper, by contrast, was viewed more positively because it was linked with an optimistic bouncing in addition to a collision.

To test their suspicions, the New Zealanders devised two experiments.  In the first, eight ten-month-olds would be presented with googly-eyed characters on a Hamlin-esque stage.  Scarf would eliminate the climber’s bounce on help trials and then pair the helper with a neutral character.  If infants choose based on social evaluation, he reasoned, they should select the helper.  But if infants find the helper/climber collision aversive and react instead via simple association, they should pick the neutral character.

In the second experiment—this time involving forty-eight ten-month-olds—the team would manipulate whether the climber bounced during help trials (at the top), hinder trials (at the bottom), or both.  They would then present the children with a choice between hinderers and helpers.  Again, Scarf proposed, if infants choose based on social evaluation, they should select the helper universally.  But if driven by simple association instead, they should select whatever character was present in the trials when bouncing occurred, and show no preference in the bounce-at-both-top-and-bottom condition.

The results were striking.  In the first experiment, seven of eight children chose the neutral character over the colliding and non-bouncing helper.  In the second experiment, twelve of sixteen picked the helper in the bounce-at-the-top condition, another twelve of sixteen opted for the hinderer in the bounce-at-the-bottom condition, and an equal number (eight of sixteen) chose the helper and hinderer in the bounce-at-both condition.

Thus, Scarf resolved, simple association can explain Hamlin’s 2007 results without resorting to the comparatively complicated notion of an innate moral compass.  In fact, he continued, his results were entirely inconsistent with Hamlin’s core conclusions.  Infants do not perceive collisions between hinderers and climbers as qualitatively different from those between helpers and climbers, and they do not prefer helpers regardless of bounce condition.

Invoking Darwin, Scarf claimed to add momentum to a movement in developmental psychology toward more parsimonious interpretations of infant behavior.  There is much “grandeur in the view,” he philosophized, that complex adult cognitive abilities can be discovered through a more sober comprehension of “these simple beginnings.”

On August 9, 2012, Hamlin—now at the University of British Columbia—posted her team’s unyielding response.  Generally, they found Scarf’s account “unpromising,” and were “bemused” by the New Zealanders’ attempt to recruit Darwin—who “wrote extensively about the powers (and the limits) of our inborn moral sense”—to their cause.  Hamlin criticized Scarf’s experimental design as well, and his failure to adequately address results she had obtained and published after 2007.

By Hamlin’s lights, Scarf’s stimuli had differed from her own in ways that left the climber’s goal—and thus the insinuation of being helped or hindered—unclear.  First, the googly eyes attached to Scarf’s climber were not fixed in an upward gaze.  Second, Scarf’s climber moved less organically, as if able to climb easily without the helper’s assistance.

Hamlin emphasized too that she had replicated the 2007 results in studies involving no climbing, colliding, or bouncing whatsoever.  In 2011, for instance, she found that infants prefer characters who return balls to others who drop them over characters who take them and run away (Hamlin, et. al. 2011).

More recently, Hamlin’s new team claimed to demonstrate that, like adults, babies interpret others’ actions as positive or negative depending on context (Hamlin, et. al. in press).  In this particularly chilling report, infants were found to prefer both individuals who helped others with attitudes (tastes in food) similar to their own, and individuals who hindered others with different attitudes.

Scarf stands firm, however.  He finds it implausible, for example, that ten-month-olds would consider such small differences between stimuli significant.  Regardless, his team had also replicated Hamlin’s 2007 results when the climber was made to bounce at the top of the hill—an unlikely outcome, Scarf chides, if their stimuli were—as Hamlin claims—somehow deficient.

He argues as well that the Canadian’s more recent experiments, though admittedly altered in design, suffer from the same general confound as the originals.  In one case, the protagonist was made to dive toward a rattle in helping trials—an “interesting” event, according to Scarf, while the hinderer was made to slam a box closed in hindering trials—an “aversive” event.

Though already extensive, Hamlin’s explorations into infant prosociality will continue.  For his part, Scarf intends to author a review of the existing literature.  Defending parsimony is an honorable cause, of course, and the New Zealanders have succeeded in raising important questions for further research.  Are we innately moral, or is prosociality primarily learned?  Are we naturally discriminatory and intolerant, or must those behaviors be taught and learned as well?


Hamlin, J.K., Mahagan, N., Liberman, Z., and Wynn, K. (in press). Not like me = bad: Infants prefer those who harm dissimilar others. Psychol Sci.

Hamlin, J.K., and Wynn, K. 2011. Young infants prefer prosocial to antisocial others. Cog Dev 26(1): 30-39.

Hamlin, J.K., Wynn, K., and Bloom, P. 2007. Social evaluation by preverbal infants. Nature 450: 557-560.

Scarf, D., Imuta, K., Colombo, M., and Hayne, H. 2012. Social evaluation or simple association? Simple associations may explain moral reasoning in infants. PLoS One 7(8): e42698.


2 thoughts on “Innate Morality? Human Babies Weigh In.

  1. F. Binkham, PhD.

    Interesting. Sometimes groups of scientists find exactly what they want to find. Not saying that’s what happened here. But we always need to be aware of such possibilities. Nice job, Krause.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s