There’s an old wives’ tale that having a child ages a woman. And why wouldn’t it? When a woman becomes pregnant, her body undergoes a massive transformation. She gains weight and her metabolic rate spikes. Her pulse quickens. Her uterus expands and presses on surrounding organs and blood vessels. Her estrogen and progesterone levels surge, reaching astronomical levels. The volume of gray matter in her brain shrinks.

While all these changes are necessary or advantageous for childbirth, many of them can cause problems later in life. Sex hormones improve reproductive odds but can cause an increased risk of breast cancer. A greater amount of mitosis, or cell division, means greater cellular damage. The high metabolism required for pregnancy and lactation increases oxidative stress, believed to accelerate aging. Building up the baby’s bones depletes the mother’s calcium. Hosting an “alien” organism forces her body to suppress its immune system, changing her response to infection.

It should come as no surprise, then, that some studies have linked childbirth to reduced lifespan. A 2007 study of population data from late 19th-century Utah, for example, found a correlation between the number of children a woman had and her mortality risk. A 2006 historical study of women in rural Poland also linked having more children with a shorter lifespan.

What may come as a surprise is all the evidence pointing in the other direction. In 2015, a team of researchers led by Thomas Perls, founding director of the New England Centenarian Study, found that subjects in the United States and Denmark who had a child after age 40 had four times the chance of living to age 100. Historical studies of Sami villagers, Amish families, and other populations also observed a correlation between having children at a later age and living longer. A 2009 examination of Dutch records from 1850 to 2000 found that increased fertility was associated with increased survival past the reproductive years, even after controlling for selection effects by factoring in conditions in early life.

And a study published last year of a rural population of Mayan women in Guatemala found that women who had a larger number of surviving children had longer telomeres (controlling for their initial length). These are caps at the end of chromosomes that get whittled down with age, bringing on cell death once they wear away entirely. The paper’s lead author, Pablo Nepomnaschy, is a biological anthropologist and epidemiologist at Simon Fraser University in Canada. He explains that the telomere effect may be a result of estrogen production, which has been shown to protect against oxidative stress and telomere shortening.

Or, as a group of researchers at Hebrew University in Israel suggested in a 2015 review paper, it could be that the state of pregnancy itself has a rejuvenating effect, at least in women who have children at a later age. A 2010 study found that the livers of aged pregnant mice regenerated twice as fast as those of their non-pregnant counterparts—and by an entirely different process. And several other studies, published in 1994, 2004, and 2006, showed that pregnant women diagnosed with multiple sclerosis had a lower relapse rate than they’d had the previous year, exhibited fewer and smaller active white matter lesions, and were at lower risk for the disease if they had many children than if they’d had none at all. Similarly, pregnant mice in which lesions had been chemically induced had greater nerve regeneration and improved paralysis controls.

Pregnancy clearly seems to have some kind of effect on longevity, even if the literature disagrees on the direction of that effect. What, though, should we expect the direction to be? How would evolution have programmed us to respond to pregnancy?

It wasn’t until nearly a century after Charles Darwin proposed his theory of evolution that scientists began to formulate evolutionary theories of aging in earnest. Until then, they struggled to make sense of why natural selection would continue to put up with aging at all. Why didn’t evolution find a way to maintain the health and vigor enjoyed during youth? What purpose could senescence serve?

In the 1950s, evolutionary biologist Peter Medawar formulated a working principle of aging, known as mutation accumulation theory. Selection, he argued, does not concern itself with an individual’s fitness in late life, beyond reproductive age, because that stage of life has a reduced impact on the survival of the organism. It therefore ignores mutations that go into effect in old age, whether they are advantageous or harmful.

In 1976, Richard Dawkins published The Selfish Gene, in which he argued that the struggle to survive takes place not at the level of the individual, but at the level of the gene. The selfish gene framework further emphasized that genes that helped an organism produce offspring would not be designed to either help or harm her later in life. Passing our DNA on to subsequent generations, he said, took precedence over our health and longevity.

Both Medawar and Dawkins decoupled any genetic adaptations relevant to pregnancy from any that affect aging. Given that it is unlikely that an adaptation that helps with childbirth will also coincidentally help with longevity, their framework was inconsistent with childbirth increasing a mother’s longevity. But neither did it support a negative impact on longevity.

There is a set of theories, though, that do. One was developed by the biologist George Williams, who wrote in a seminal 1957 paper that “Natural selection may be said to be biased in favor of youth over old age whenever a conflict of interest arises.” Building on Medawar’s work, he argued that selection would be bound to favor genetic variants that have opposite effects at different ages. According to this theory, called antagonistic pleiotropy, mutations that are beneficial during reproduction could have negative consequences after an organism ceased reproducing.

A recently discovered potential example of antagonistic pleiotropy is the p53 gene, which helps to protect against cancer early in life, but accelerates aging later in life. The hypothesis could also explain why higher fecundity is associated with Huntington’s disease, or why the overproduction of sex hormones that improve reproductive success also increases the risk of developing certain cancers.

Two decades after this theory was introduced, the biologist Thomas Kirkwood suggested yet another hypothesis consistent with an inverse relationship between childbirth and longevity, called the disposable soma theory. It stated that when the body devoted some of its limited amount energy reserves for reproduction, it scaled back on cellular repair and maintenance functions, leading to somatic disintegration and aging. Giving birth, in other words, was expensive, and the mother paid the cost with a shortened life.

To listen to these theories, then, there would be little reason to expect childbirth to make a mother live longer.

But theoretical biologist Josh Mitteldorf thinks there’s something deeper going on. Many of the prevalent theories of the evolution of aging are focused on evolution at the individual or gene level. Now, he says, it’s time to examine the issue from a new perspective: according to group selection. Group selection favors the fitness of communities and populations over the fitness of individuals.

A long-living mother has certain advantages that could apply a positive selective pressure. For one thing, she has the chance of becoming a grandmother. In 1997, anthropologist Kristen Hawkes proposed that human females survive as long past their reproductive age as they do so that they can take on the role of a grandmother, supplementing caregiving, ensuring the survival of their grandchildren, and freeing up mothers (their daughters) to have more children of their own. Hawkes’ studies of the Hadza women in Tanzania supported the grandmother hypothesis with the observation that the older women collected more food than women of childbearing age did.

Lorena Madrigal, another anthropologist, promotes an alternative thesis, called the mother hypothesis, that also supports a positive selective pressure for longevity. “A women stops reproducing,” Madrigal says, “because she’s going to put all her effort into making sure her offspring survive, period.” Since it becomes far riskier to have children later in life, evolution eventually put a stop to a woman’s ability to reproduce past a certain point (the age of menopause), allowing mothers to direct their energy to caring for their children. Madrigal examined historical data from families who lived in Costa Rica from the 16th century to the 20th and found a positive correlation between a woman’s fertility and her longevity, but an inverse relationship between a grandmother’s age and the number of grandchildren she had.

Whether it’s the grandmother hypothesis or the mother hypothesis, however, the implication for longevity is the same: An evolutionary pressure exists for women to have an extended lifespan after their reproductive years have come to an end.

Women with many children may also live longer for reasons that lie outside of evolution altogether. Mothers often receive greater social support, which goes a long way to increasing life expectancy. A team of researchers at Deakin University in Australia, for instance, surveyed nearly 800 childless women in the country and found them to report perceived social exclusion. And a paper published last year argued that social support from adult children later in life played a significant role in extending mothers’ lifespans.

“If women receive more social support,” Nepomnaschy hypothesizes, “they can spend less energy on reproduction. They can invest that energy in their soma, in preserving their bodies from the aging process.” He cites other studies that show a correlation between increased social support and a “buffering effect” against telomere shortening. Social factors can also translate in psychological factors, Mitteldorf says, “which can in turn translate into metabolism and make a huge difference in how long someone lives.

The bottom line is—this all remains “very complicated,” as Madrigal puts it. “We should be very careful about finding a simple explanation for something as complex as the evolution of post-menopausal longevity.”

The effect of having a baby on an individual woman will be some balance among some or all of these effects—and, as it turns out, on what the mother believes to be the case. Studies have shown that our beliefs and attitudes can, in themselves, have an effect on how quickly we age, and pessimism has been linked directly to shortened telomere length. While the science doesn’t prove that optimism is correct, it may, in a sense, be warranted.