Who do newborn babies resemble?

Imagine that you are a genome, and you are about to write instructions for how to create a brand new baby. Half of your genes come from the mother, and the other half come from the father. You are equally related to both parents. Now, if you have a choice between making the baby you are creating look like the mother or look like the father, what would you do? Would you make the baby resemble the mother or the father or both equally?

As I note in a previous post, or more extensively in Chapter 2 of our book Why Beautiful People Have More Daughters (“Why Are Men and Women So Different?”), because of the sexual asymmetries in reproductive biology, the possibility of cuckoldry exists only for men. Men can be cuckolded and unwittingly invest their limited resources in someone else’s genetic children, whereas women could never be cuckolded. In other words, paternity can never be certain, while maternity is always certain. This is well expressed in the common saying “Mommy’s Baby, Daddy’s Maybe.”

Men who are cuckolded do not manage to transmit their genes to the next generation and so they achieve no reproductive success. Men are therefore selected to be very sensitive to cues to possible cuckoldry and to attempt to guard against the possibility. A man would therefore only invest in his mate’s children if he was reasonably certain that they were genetically his. In the absence of DNA tests (which did not exist in the ancestral environment), how could men ever be certain that their children were genetically theirs?

The child’s physical resemblance would be one clue available to men in the ancestral environment. If the baby looks like the father, it is more likely that it is genetically his, whereas if the baby looks nothing like him, or, worse yet, looks a lot like his neighbor, then it is doubtful that he is its genetic father. This reasoning leads evolutionary psychologists to predict that, holding constant the probability of cuckoldry, babies who resemble their father are more likely to survive than babies who do not resemble him (or resemble the mother), because the father of babies who resemble him is more likely to be convinced of his paternity and to invest in them, thereby increasing their chances of survival. In contrast, the father of babies who do not resemble him (or resemble the mother) is less likely to be convinced of his paternity and to invest in them, thereby decreasing their chances of survival. Over many generations throughout evolutionary history, genes that make babies resemble the father therefore survive, whereas genes that make them resemble the mother do not, and so more and more babies come to resemble the father, until most babies are born resembling the father, not the mother.

This is precisely what two psychologists at the University of California, San Diego, Nicholas J. S. Christenfeld and Emily A. Hill, discover in their ingenious study, published in Nature in 1995. Christenfeld and Hill show the subjects in their experiment a picture of a child at ages 1, 10, and 20, and a set of three pictures of adults, one of whom is the real parent (mother or father) of the child. They then ask the subjects to match the child with the correct parent. Christenfeld and Hill’s subjects therefore have .33 probability of selecting the right parent by chance. If the child truly resembles the parent, then the subjects should be able to match the two pictures at a much higher probability.

A major finding in Christenfeld and Hill’s experiment is that children in general do not physically resemble their parents. The subjects are not able to match the picture of the child at any age to the picture of either the mother or the father better than expected by chance. The only exception, however, is the matching of one-year-old babies to their father. The subjects are able to match both baby boys (.505) and baby girls (.480) to their father (though not to their mother) at a statistically significantly greater rates than by chance. That means that one-year-old babies resemble their fathers, as might be expected from the evolutionary psychological logic presented above.

Christenfeld and Hill’s finding was widely reported in the media, but it has also become one of the most controversial contentions in evolutionary psychology, not least because, although their explanation had impeccable logic, their finding could not be replicated. To date, attempts at replication have shown that newborn babies objectively resemble mothers more than fathers, and infants and children resemble both parents equally. Thus, the question of whether newborn babies objectively resemble the father more than the mother must be treated as an open one until more experiments are conducted.

While the question of whether babies objectively resemble fathers more than mothers remains an open question, there is a related question about baby’s resemblance which is more satisfactorily answered and more clearly established empirically. I’ll talk about it in my next post.