In a commentary published earlier this month in Nature, Harvard professor Sarah S. Richardson and six co-authors caution scientists, journalists and the public against drawing hasty conclusions from findings concerning epigenetic effects on human development.
Epigenetic effects are those that involve heritable changes to gene expression that don't arise from changes to the genotype itself. For example, a fetus's intrauterine environment can influence gene expression in later life, with some effects that extend beyond one generation.
To illustrate how epigenetic effects are often reported, Richardson and colleagues quote the following headlines:
"Mother's diet during pregnancy alters baby's DNA" (BBC)
"Grandma's Experiences Leave a Mark on Your Genes" (Discover Magazine)
"Pregnant 9/11 survivors transmitted trauma to their children" (The Guardian)
Notice a trend? It's all about moms. And where mom is portrayed as the cause of some negative outcome, it's a simple step to blame. That's what worries the authors, whose major lesson is summed up in the title of their piece: "Don't blame the mothers."
The thing is, epigenetic effects can also arise from "paternal effects" (they contribute to the fetus, too!) or other aspects of a child's environment. And even where effects are "maternal," the lesson is rarely that we should blame or regulate pregnant women and mothers — often what's called for is better maternal and social support. The authors sum up:
"[Research on the developmental origins of health and disease] provides a rationale for policies to improve the quality of life for women and men. It must not be used to lecture individual women, as in a 2014 news report from the US media organization National Public Radio on an epigenetics study in mice: 'Pregnancy should be a time to double-down on healthful eating if you want to avoid setting up your unborn child for a lifetime of wrestling with obesity.' How are women who lack time or access to healthy foods to act on such advice?"
The NPR story they reference is here.
The authors of the Nature commentary go on to recommend four ways in which epigenetic effects on human development can be more responsibly reported: (1) by presenting extrapolations to humans from other animals with appropriate qualifications, (2) by emphasizing both maternal and paternal effects, (3) by conveying the complexity of the causal processes involved, and (4) by recognizing the need for societal — not (only) individual — changes.
Oversimplified science and a critical focus on mothers are nothing new — we've been blaming mothers for decades (and probably longer). In fact, Richardson and colleagues note plenty of historical precedents. "Refrigerator mothers" were once blamed for autism, and women have been judged, regulated and even criminally prosecuted for alcohol and drug use during pregnancy based on relatively flimsy evidence concerning its harmful effects.
So why is it that the complex science of human development, in particular, is so readily distilled into this single, unhelpful message: "It's all about mom"?
Of course, science is influenced by values in all sorts of ways: in the questions we address, the conclusions we prioritize, and the applications we pursue. But when dealing with complex causal processes and the assignment of causal responsibility ("it's the mother!"), values can affect the conclusions we draw from science in an especially pernicious way. That's because we think of causal claims as simple descriptive facts about the world — as value-free. But a growing body of empirical work shows they're not. In fact, the way we make causal claims depends a lot on how things normally happen and on how we think they should happen.
Before considering the complex case of epigenetics and motherhood, take a simpler and less contentious example. A husband and wife both work, but the husband works at home and the wife works out of the house. Like many of us, they are chronic users and losers of pens — they can never seem to find the one they had an hour ago. So they keep the kitchen junk drawer stocked with pens. They also have the following arrangement: The husband is allowed to swipe these pens for work (after all, they won't leave the house), but the wife is absolutely not (she can get pens at the office). Nonetheless, they are both chronic swipers — in fact, the wife takes pens just as often as the husband.
One morning, the husband and wife each swipe a pen for work. Later that day, the phone rings in the kitchen. Someone answers and needs to take an important message but there aren't any pens left. Who caused the problem?
In a 2008 paper, Joshua Knobe and Ben Fraser used vignettes like these to get at people's judgments of causal responsibility. Overall, they found that people thought the wife caused the problem and that the husband did not, even though both causally contributed to the pen shortage at the crucial moment. (Their vignettes actually involved professors and administrative assistants swiping pens, but the logic was the same.)
This finding, among others, suggests that outcomes are more likely to be attributed to causal factors that deviated from how things "ought" to have occurred. In this case, the "ought" comes from an explicit rule about who can take pens, but in many cases it has more subtle origins in our views about what's moral, natural or desirable. For example, swap the pens with bottles of beer, add that the wife is pregnant, and make the problem a crucial beer shortage when an important guest arrives for dinner. You might be more inclined to think the wife caused the problem with that beer she had with lunch, even if the husband had one too.
Having laid out these more familiar examples, we can return to where we began — with the tendency to blame mothers for negative epigenetic effects. We rarely have explicit rules about a mother's role in gestating and raising her children, of course, but we have plenty of strong expectations about how things typically do — and should — occur. In fact, you need look no further than Mother's and Father's Day cards from Hallmark to find evidence of our widespread assumptions about the different roles of male and female parents: Mothers are thanked for being nurturing (e.g., being a friend, making the child feel loved), fathers for supporting achievement (e.g., teaching useful skills, being the breadwinner). Despite changing norms and different realities for many families, the current expectations are clear: Mothers nurture (intensively); fathers provide.
These assumptions about each parent's roles and responsibilities transfer all too readily to research on the developmental origins of childhood health and disease. If we implicitly assume that mothers ought to have sacrificed, nurtured and known best, we'll tend to see them as "the cause" when a suboptimal outcome occurs — even when other causal factors were also at work and even when an individual mother may not be the most appropriate locus for intervention.
Values infuse science in all sorts of ways, for better and for worse. The influence of values can be dangerous, however, when it slips in under the radar. In the kinds of cases Richardson and colleagues cite, it's all too easy for people to take themselves to be making value-free descriptive claims about what causes what. "That's just what the science tells us!" is the tempting but naive response.
When we talk about whether there's a causal influence of one variable on another, we're typically engaged in descriptive science. And when we talk about moral responsibility or public policy, we know we've entered the domain of values. But when we go beyond claims of causal influence to claims about "the cause" of something, or to claims that "X causes Y," we've entered the danger zone — a place where values play a role that often goes unrecognized. It doesn't mean that we should never make such claims, but it does mean that we need to check which values we're importing and to critically examine whether or not they belong there.
You can keep up with more of what Tania Lombrozo is thinking on Twitter: @TaniaLombrozo