The Selfish Gene

Dawkins recalls that the “selfish gene” isn’t just one standalone gene, it’s also all of its replicas, distributed throughout the world. A selfish gene is trying to become more numerous in the gene pool. Dawkins says the key point of this chapter is that “a gene might be able to help replicas of itself that are sitting in other bodies.” He thinks this explains why behavior among relatives (such as parents and children) appears altruistic, when it is in fact “brought about by gene selfishness.” 

Consider the gene for being albino. Theoretically, a gene for albinism could do well if it also programmed its survival machines to be altruistic toward other albino people. If the albino gene could make one of its survival machines sacrifice itself to save ten other survival machines with the albino gene in them, the albino gene would be doing well in the gene pool overall. But in reality, albino people don’t go around sacrificing themselves to save other albinos. There is a simple explanation for this. Genes aren’t conscious: they can’t actively choose to make their survival machines act nicely toward others that contain their replicas.  

In order for that to happen, there would need to be a gene that has two functions. First, the gene would need to contain instructions for building survival machines with very pale skin (or, say, green beards). Second, it would also need to contain instructions for building survival machines that are nice to other beings with very pale skin (or green beards, or any other detectable trait). But no such gene exists. The chances of those two functions coming about in one gene by a chance mutation are very low.

Dawkins wonders if some genes can detect copies of themselves in other survival machines, as it’s clear that this ability would very well in the gene pool. He thinks the answer is yes. Close relatives (or “kin”) have a high chance of sharing genes. This means a gene that programs its survival machines to be nice to their kin is more likely to keep replicas of itself alive.

Dawkins wants a more precise way of seeing if the altruism between family members happens in proportion to their shared genes, so he uses Hamilton’s research. Hamilton worked out a system for calculating the odds of two individuals sharing genes. Hamilton figured this out by looking at how many shared ancestors two people have. If one does the math, one’s first cousin is as genetically related to oneself as one’s great-grandchild. A third cousin, on the other hand, is about as genetically close to oneself as any random person.

A gene for altruism toward distant cousins (by sacrificing oneself to save them) would be less successful in the gene pool than a gene for being altruistic to save one’s siblings. Dawkins calculates that a gene for altruism will survive in the gene pool over time if the altruistic person saves two siblings at their own expense.  

A parent taking care of his or her child preserves the same number of genes as when it takes care of his or her orphaned sibling. Both the child and the sibling have 50 percent of the parent’s genes. Dawkins thinks that “genetically speaking,” there’s nothing special about the parent/child relationship when compared to the brother/sister relationship. They are just two examples of gene preservation in survival machines that share 50 percent of their genes. It doesn’t matter that genes aren’t transferred from sibling to sibling the way they are from parent to child, since siblings share replicas of the same genes from their parents. 

Many scientists use the term “kin selection” to talk about altruism between people who are related. Wilson, for example, thinks kin selection is a special kind of group selection. Dawkins disagrees. Hamilton’s research shows that it doesn’t matter so much exactly how two people are related. What matters is how many genes they are likely to share. Hamilton doesn’t have to decide whether or not second cousins count as family to determine their likelihood of being nice to each other. He only has to think about degrees of genetic relatedness and explain altruism in those terms.

Dawkins thinks that of course in real life, organisms don’t go about calculating the percentage of their genetic relatedness to other animals and then making decisions about who to be nice to on that basis. Even if they did, the calculation would be a little messier. They’d have to factor in relative age, life expectancy, environmental considerations and a host of other things. For example, grandparents and grandchildren share 25 percent of their genes. But grandchildren are likely to live longer than grandparents, so it makes sense for grandparents to take risks for their grandchildren, but not the other way around.

In order for altruism to evolve, the genetic payoff has to be higher than the genetic risk of the altruistic act. Although organisms don’t go about calculating these risks and payoffs, behaviors that line up with genetic payoffs will tend to survive the course of evolution. For example, if I see some mushrooms on the ground, the genetic payoff of sharing them with my siblings (but not strangers) might be higher than if I ate them all myself.  Of course, genes can’t make decisions for their survival machines on a day to day basis. They program them with certain hereditary traits—or, “rules for action”—and wait out the ride as the brain interprets those rules each time it acts.

Dawkins wonders what kinds of altruistic “rules” or behavior traits could be effectively pre-programmed into a survival machine. Behavior traits that will be successful over time vary from species to species. In a population with highly mobile organisms, a strategy like “be nice to people that look like you” might work, since it’s more likely for siblings and parents to look alike than strangers. In birds, however, a strategy like “be nice to the others living in your nest” might be a good strategy, since it’s likely the birds sharing a nest are genetically related. 

Chicks that find food, for example, often let out a low twitter that attracts other chicks. On the surface, this seems like altruistic behavior, but Dawkins disagrees. He says that chicks tend to move around by following their mothers, so chicks near each other will likely be siblings. The behavior, “make a low twitter when there’s food” thus benefits the replicated genes living in a chick’s siblings.

However, there are still some behaviors that seem really problematic to explain in terms of their genetic payoff, such as adoption. Even worse, in the animal world, some bereaved mothers steal others’ babies. Genetically speaking, this is doubly bad. The bereaved mother spends her energy on child rearing (at a detriment to the genes inside her own body), and also frees up the biological mother to spend energy further perpetuating her own genes. Dawkins thinks more research is needed on these cases. He speculates that perhaps stealing babies helps mothers learn child-rearing skills for when they mate again.

On the other hand, genetic payoff explains why certain exploitative behaviors between species happen. Cuckoos exploit the behavioral trait of “be nice to birds in your own nest” by laying their eggs in songbird nests. Baby cuckoos stay alive at the expense of the unwitting songbird’s energy, and  mother cuckoos can preserve their own energy for other things. “Cheating” behavior stays in the gene pool because of its genetic payoff to cuckoos.

It turns out that songbirds evolved to recognize markings on eggs of their own species. Since that strategy had a genetic payoff to songbirds, it persisted. But then cuckoos evolved to make their eggs look more like songbird eggs. This strategy had a genetic payoff for them, and remained in the gene pool. Dawkins thinks that mimicking songbird eggs is an evolutionarily stable strategy for cuckoos.

Dawkins returns to the issue of parent/child versus sibling relationships. He thinks that natural selection will favor a degree of altruism that is proportional to the “best estimate of relatedness” between two individuals. It’s important, however, to remember that animals don’t go around calculating their relatedness to each other, they just have pre-programmed traits that either work over time or don’t.

The parental altruism gene is stronger in the gene pool than the sibling altruism gene because a mother is more likely to be correct in assuming a child that she cares for is her own. Siblings could unknowingly be half-siblings (sharing only a quarter of their genes) or adopted siblings (sharing no genes). 

Altruism toward children is also more common than altruism toward parents. The relationship is “asymmetric” because an altruistic child (who spends energy on keeping its parents alive at its own expense) is less likely to survive and reproduce than a selfish child. But, an altruistic parent (who spends energy keeping its child alive) is more likely to keep its altruistic genes in the gene pool than a selfish parent who lets their child die from neglect before it reproduces.