Infants will give away food to a stranger even when they’re hungry, a new psychology study finds.

One of the most famous — if contentious — studies in social science is known as the marshmallow test. In the 1970s, that experiment probed kids’ ability to delay gratification when faced with a yummy snack. It turned out some kids can resist the temptation to eat one marshmallow immediately if you tell them that waiting will earn them two marshmallows later.

But will a kid pass up a yummy snack altogether — not just delaying gratification but forgoing it entirely — if there’s a stranger nearby who seems to want the snack?

It turns out even a baby will engage in this sort of altruism, according to a new study conducted by the University of Washington’s Institute for Learning and Brain Sciences and published last week in Scientific Reports. Altruism, it seems, may well begin in infancy.

Researchers studied how nearly 100 babies, all 19 months old, behaved when presented with sweet fruits like blueberries and grapes. When a researcher pretended to drop a fruit onto a tray and reach for it unsuccessfully, signaling a desire for the snack, 58 percent of the babies picked up the fruit and gave it to the researcher. (When the researcher didn’t bother reaching for the fruit, only 4 percent of the babies tried to help out.)

The researchers were also curious about whether the babies would be so generous if they were hungry. After all, part of altruism is helping others even when it means incurring a personal cost. So the researchers brought in a different sample of infants just before their scheduled snack or mealtime, when they were likely to be hungry, and repeated the experiment.

Even under these conditions, an impressive 37 percent of hungry babies picked up the fruit and offered it to the researchers who pretended to have trouble reaching it.

“The infants in this second study looked longingly at the fruit, and then they gave it away!” said co-author Andrew Meltzoff in a press release. “We think this captures a kind of baby-sized version of altruistic helping.”

It’s worth noting that the infants had a clear path for escaping with the food (the researcher’s path to the baby was blocked by a table) so they easily could have retreated and eaten the snack instead of giving it up. Plus, the babies received zero verbal prompting (no “Can you give me that food?” or “Thanks!”) and zero facial feedback (the researcher maintained a neutral expression no matter what the baby did). And the researcher never reciprocated by giving anything to the baby. Still, the babies kept on giving.

Previous psychology studies have highlighted two of the factors that seem to influence kids’ social behaviors: having siblings and being raised by parents from a cultural background that emphasizes our connections to one another (some Asianand Hispanic cultures tend to be particularly big on interdependence).

So the researchers in this study asked the parents about their infant’s siblings and cultural background, and examined whether these factors could account for some of the variance in the infants’ tendency to help strangers. It turned out they could, which suggests that children’s altruistic behavior is malleable.

Although the findings of this study are consistent with previous studies on how culture and siblings influence infant helping behavior, we shouldn’t draw any sweeping conclusions about our kids just yet. It will be interesting to see if future studies manage to replicate its results — or if, as in the case of the marshmallow test, some aspects fail to replicate.

The authors note that their study has limitations, including “the small sample size, the skewed sample of parents who come into research laboratories, and the desirability in the future of video recordings of infants’ sociocultural experiences at home.”

It’s also important to note that one inherent limitation of studying babies is that we can’t say for sure what motivation is driving them; we can determine what they do, but not why. And the question of why human beings at any age ever act altruistically is a puzzle that has long perplexed philosophers and scientists.

Why do human beings ever behave altruistically?

When we step back and look at behavior from an evolutionary perspective, the fact that we sometimes help perfect strangers — even when it means giving up our own resources or risking our lives — seems to make zero sense. Aren’t we supposed to act in a way that maximizes our genes’ chance of survival?

Human altruism seems even stranger when we consider it against the backdrop of the animal kingdom. Chimpanzees, our primate cousins with whom we share a recent common ancestor and 98 percent of DNA, do not voluntarily hand over food. Some mammals and birds do share food with their kin, but they don’t generally engage in that behavior unprompted with non-kin or strangers.

So why do we humans?

Philosophers, psychologists, biologists, primatologists, and others have debated this question for centuries. It remains contentious today, with a few dominant hypotheses emerging.

Some evolutionary psychologists argue that we act altruistically to boost our reputation or prove we’ve got resources to spare, which may help us win an attractive partner with whom we can then reproduce. Others say we do kind things for others in the hopes that they (or the broader culture whose norms we’re helping to shape) will reciprocate one day when we’re in need.

Another hypothesis suggests we engage in altruism because it causes us emotional turmoil to see someone in need and not help them; in other words, we’re trying to alleviate our own pain, evoked by empathetic impulses. Alternatively, we may be trying to alleviate our fear, acting generously because we believe other people or perhaps a watchful deity will punish us if we’re selfish.

What all these explanations have in common is the underlying notion that when we act altruistically, it’s not really “pure” altruism — we’re getting some long-term benefit out of it, even if we’re not consciously aware of it.

But others hypothesize that altruism may not serve an evolutionary purpose now that we live among non-kin in huge groups like nation-states; rather, it’s a leftover trait from when we lived in small groups where everyone was genetically close to us. The moral conditioning that we receive — from our parents in particular and our culture in general — may reinforce the belief that we must act altruistically.

Why babies might act altruistically

The authors of the infant study lean toward that last explanation. “We speculate that certain childrearing practices and values … convey the expectation to infants that people tend to help others and may engender in children a generalized feeling of interpersonal obligation towards other humans in need,” they write.

In tandem with this cultural influence, the authors add, there may be an evolutionary mechanism at work: By giving away food to strangers, an individual might promote affiliation with another individual as well as broader group cohesion, ultimately promoting the success of their species.

Both biological and cultural influences play a role in altruism, but to find out exactly what those influences are and how they interrelate, we need a lot more experimental research.

“Continued research would be desirable to more fully understand what maximizes the expression of altruism in young children,” said the study’s lead author, Rodolfo Cortes Barragan. “If we can discover how to promote altruism in our kids, this could move us toward a more caring society.”

Vox