In journalism you look for one thing and find another that confounds your expectations. It's what make makes this gig so frustrating and fun. I went looking for reassurance in Mothers and Others: The Evolutionary Origins of Mutual Understanding (Harvard University Press, 2009) by the anthropologist Sarah Blaffer Hrdy and found something scary.

Hrdy is one of my favorite evolutionists. She's unpredictable, iconoclastic, passionate about her work—unafraid of mining her own life for insights. She earned a PhD at Harvard in 1975, the glory days of sociobiology, and she remains committed to that discipline's goal of understanding primate behavior in evolutionary terms. She rejects knee-jerk liberal/feminist hostility toward sociobiology (or evolutionary psychology, the preferred term these days) but calls out male colleagues when they're being sexist dolts.

No one can accuse Hrdy of trying to feminize or pacify our evolutionary past. She helped win acceptance for selfish-gene interpretations of infanticide among primates, including humans. Male langurs, for example, may kill an infant they suspect is fathered by another male. To prevent male infanticide, a female langur may mate with many males, who then refrain from killing infants because they think they may be the father. But females may also abandon or kill infants that they lack the resources to raise.

In Mothers and Others, Hrdy decried theories of human nature that emphasize "demonic" male aggression. (I slammed the "demonic-males" thesis in a recent post.) The key to our humanity, Hrdy contended, was the emergence of group child-rearing—also called cooperative breeding, or "allocare"—some two million years ago. Among all the ape species—chimpanzees, bonobos, gorillas, orangutans—mothers rear offspring without seeking or receiving help. Indeed, ape moms keep infants away from other females and males—with good reason, because others may hurt or kill the infant.

Child-rearing is radically different in hunter–gatherer societies like the African !Kung, Hadza and Aka, who are thought to live more or less as our ancestors did for 99 percent of our evolutionary history. Mothers in these societies get lots of help from other females, including grandmas, sisters and friends, who may even breast-feed an unrelated child. Dads and other males often hold, feed and play with children, too, which ape males never do.

There is a dark underside to all this group nurturing. A human mother's care for her infant is more contingent on circumstance than the care of ape moms. If a hunter–gatherer mother feels she's not getting enough support from others, she may abandon or kill her newborn. Natural selection thus favored babies who excel at "mind-reading"; they can intuit and manipulate the emotions of their mothers and other potential caregivers, to ensure that they get the care they need to survive. Empathetic kids become empathetic adults. In this way cooperative breeding promoted the emergence of our extraordinary "hypersocial" intelligence.

I find this theory of human nature more plausible—and, yes, palatable—than those emphasizing violent competition. But Hrdy's book ends on a disturbing note: She pointed out that many modern children—far from growing up surrounded by doting kin—don't even see much of their busy, working parents. Kids receive much of their care from non-kin, whether babysitters or preschool teachers. Then there are all the children who grow up with only one or no parent or abusive parents. As a result, many kids suffer from "disorganized attachment," which means they have a hard time understanding and trusting others.

If children receive poor care, their innate capacity for caring may not be fully expressed; they may become uncaring parents, colleagues, citizens. Even more alarmingly, the genes underpinning our prosocial impulses may dwindle because they are no longer favored by natural selection. We may lose our innate empathy and compassion, Hrdy speculated, just as cave-dwelling fish lose their eyesight. To an evolutionary theorist "surveying humans 20,000 years hence," Hrdy wrote, "our powerful impulses to empathize with others, to give, to share and seek reciprocation, might seem like nothing more than transient phases in the ongoing evolution of the species."

This coda initially struck me as the kind of hyperbole that science authors (like me) rely on to provoke attention. But then I read in The New York Times that researchers at the University of Michigan at Ann Arbor had found a 40 percent drop in the empathy of college students over the past 30 years. The decline has been especially rapid in the past decade.

This is just one self-report study (pdf). I still have faith that we're headed toward a more peaceful, caring future. My two teenage kids aren't cold narcissists, nor are the college students I teach. But I'm haunted by the image of a future in which our descendants have become as morally blind to each other as cave fish.