Language Serves the Group

Steven Pinker has posted an important essay on group selection. You can gather its thesis from the title, “The False Allure of Group Selection.” Since I am on record saying that group selection (really, multilevel selection) was critical to the evolution of language, I read the essay with strong interest. Let me say right off that I was astonished to find that the essay makes no remarks about the evolution of language. Pinker is a famous proponent of language’s evolutionary origins and biological basis, but he says nothing of group selection and language. Instead he criticizes ideas that group selection explains religion, culture, and nations.  I am skeptical of those claims too. Pinker is a fine writer and I got several chuckles out of his examination of various shallow appeals to group selection. Was I laughing at my own doom?

What is in Dispute

My argument is not that Pinker is wrong, but incomplete. We both agree that individuals compete. I go on to say that there are multiple levels of selection. Populations compete and selection occurs at this level as well as the individual level.

Pinker rejects my position by saying, “ I don't think it makes sense to conceive of groups of organisms (in particular, human societies) as sitting at the top of a fractal hierarchy with genes at the bottom, with natural selection applying to each level in parallel ways.”

This statement makes clear to me what Pinker does not get about where I stand. I don’t put genes “at the bottom” of the “fractal hierarchy.” So let’s sort out the basic points about where I stand:

• Genes are the ultimate survivors.
• Genes don’t survive on their own. They are passed down through generations. We can call a member of a generation the carrier.
• Thus, whenever we talk about survival, selection, or fitness we are talking about two things simultaneously—the genes and the carrier. They are as attached as the heads and tails of a coin.

So Pinker has it wrong when he says multilevel selection puts genes at the bottom of the hierarchy. Genes are at every level of the hierarchy, but they are never alone.

Some Peculiarities of Speech

When I began this blog in 2006 I assumed that selection at the individual level would support the whole story. I changed my mind, first because of my concentration on language. Later, a more communal view of Homo led me to notice a series of biological universals that don’t seem explicable in terms of individual selection, kinship selection, or mutual backscratching (a.k.a. reciprocal altruism).

Language functions differently from other signal systems as shown by the speech triangle: two or more individuals pay joint attention to a shared topic.

A common example of the triangle in action is a teaching session. The teacher speaks about a topic; the student listens and learns. Teachers can be kin, but around the world they are not limited to them. Teachers can be expecting a mutual payback from the student, but around the world there are people who teach because they enjoy passing on knowledge. Teaching is essential to maintaining a group’s intellectual capital, but it provides no visible reproductive advantage to the teacher.

The speech triangle is an equalizer. I’m walking along the street and meet a colleague. I volunteer that a store three blocks away has a great deal on watermelons. Thanks, says the colleague who now has some information that was my secret. Pinker would try to explain this in terms of mutual backscratching, but I’m not so sure. Most other animals are not so deeply into sharing information.

We know that captive chimpanzees can learn to use words and phrases but in the wild they never tell one another anything. They communicate to control. This kind of discretion is easy to explain in terms of individual selection. A chimpanzee who knows where there will be some ripe fruit has an advantage over its fellows. A chimpanzee who blabs his news has given up an advantage. The fitness score of the chimpanzee who keeps secrets is almost certainly higher than the blabbermouth’s score. Thus, even though groups might benefit from language, it is not going to evolve among chimpanzees. This kind of reasoning makes it easy to explain why language never evolved with other species, and hard to explain why humans have such a hard time keeping secrets.

.Jean-Louis Dessalles argues that reputation for wisdom, helpfulness, and trustworthiness is the factor that enabled language to evolve. People who never share their knowledge gain poor reputations for either not knowing anything or for not caring about other people. But reputation as a force is peculiar in itself, a social factor that plays a very limited role throughout most of the animal kingdom. Caring what people think about you is a means of social control that in itself is hard to explain outside group-level selection.

Another peculiarity of language is how much more powerful it is than anything else in the animal kingdom. Since captive apes can use words and we can assume that whenever our ancestors began sharing information they were intellectually ready. But how did we get so much smarter than the other apes?

Suppose you have a society of phrase users and a mutation enables a speaker to utter something more sophisticated, perhaps a true sentence. Why would that mutation spread to other individuals? What advantage  does it give? Is there some pressure that pushes individuals to speak in a more complex manner? I’m open to suggestions, but it is much easier to imagine group benefits than individual ones. Groups that can share more complex ideas than their rivals are able to plan and cooperate at a deeper level than groups that converse only at the ape level.

Group competition could push us well beyond the intellectual powers of a chimpanzee. At the individual level there is not much advantage being far superior to the others. The race may go to the swiftest, but there is no point in running twice as fast as the others. But cooperative groups can gain an advantage and replace less cooperative groups. Then some group in that set of winners becomes even more cooperative and outcompetes the more old-fashioned groups.

Other Peculiarities of Human Communities

Evidence of group selection is not limited to language. Let’s notice some other peculiarities of humans that are hard to explain via individual selection but easy to see as group selection.

• Forced sharing. Chimpanzees pay attention to one another’s gaze, but with their black eyes it is hard to tell where they are looking. Human eyes,, however, are largely white and it is easy to see where people are looking. There is no advantage as an individual to constantly reveal where one’s attention is focused, but it is useful to the group. We are the only species I know that has white eyes and it would be very instructive to learn what gene supports the trait and how old it is.
• Trust. Interest in young is very common in primates, but the mother generally keeps others at a distance. It would be too easy for a genetically selfish individual to kill an infant. Such behavior is scandalously common among many species as higher ranking females often kill the young of lower rank mothers. Males kill young too, as a means of bringing the mother back into season more quickly. Mothers have taken this behavior to heart and keep their young away from others. Except for humans where it is common for neighbors to ooh and aah the babies, and to hold them. In some villages it is common for the young to wander freely, subject to the rearing and feeding of any random villager. This kind of behavior is most atypical of other primates and is hard to explain in terms of concern for individual genes.
• Childish dependence: All newborn mammals are dependent on their mothers, but humans are born to a remarkable state of dependence. No other species is so helpless for so long. And we have a surprisingly small kit of innate reflexes. Human newborns must learn what the local conditions are and adapt to them. This situation forces us into a dependence on the group that is quite unknown for other primates, and it forces a teaching obligation on group members that is also unknown in other primates.
• Group Identity: Adolescence (a period of adolescence, a time of sexual maturity but physical immaturity) is another biological difference that is universal in humans and unknown in other species. Why should this be? Adolescence around the world is a time of cultural initiation, often marked by rituals and education. In the United States these are commonly religious rituals—confirmation, bar/bat mitzvah, baptism—and they include an educational component. Once they are initiated, adolescents are typically apprenticed to their new life as members of their group. In the modern world the apprenticeship phase tends to be blurred, as so many opportunities are available. This period of adolescence is a blending of the biological and cultural. The identity, ritual, and education vary from culture to culture, but biology provides the frame. Every culture has to take the frame into account, yet without culture biology cannot survive. I don’t see how this kind of dependence can possibly evolve at the individual level.
• Love: I can only feel pity for anyone who tries to argue that love is a form of mutual backscratching. Obviously they have never known the intensity and power of putting another ahead of oneself.

How Are Such Things Possible?

I am not doubting evolution, gene theory, or survival of the fittest. Nor do I doubt the importance of reproductive success. Reproductive success varies greatly within a population, and evolutionary theory holds that the more successful a carrier or gene is at reproducing itself, the more “fit” it is. Thus, a carrier/gene that produces, on average, 3 offspring is more fit than one that averages 2 offspring. The average fitness (w) is found by taking a population (n) and dividing it into the number of next-generation descendants (n’); that leaves us with this equation:

w = n’/ n

Suppose, for example, that a population has 800 individuals who can run at 30 miles per hour for 10 minutes and 1000 members who can maintain 35 miles per hour for 10 minutes. Then in the next generation we find 850 30 mph runners and 1100 35 mph runners. The calculations are 850/800 and 1100/1000. The results: w₃₀=1.0625; w₃₅=1.1. The faster runners have the higher fitness score and eventually their speed seem likely to predominate throughout the whole population.

In the example above, n is the number of individuals, but it can just as easily be the number of groups. For example, suppose a population has 25 groups with an average of 50 individuals who speak true sentences with a subject, object, and verb. The population also has 123 such groups who speak in phrases. A generation later there are 30 groups speaking sentences and 125 groups speaking phrases. The calculations report: w_sentences=1.2 while w_phrases=1.016. The math tells the story. Groups speaking in sentences should ultimately replace the larger number of groups limited to phrases.

Since, mathematically speaking, the carrier n can be a group or an individual, multilevel-selection deniers are forced to argue that as a matter of practical reality all selection takes place at the group level. I say no, n can be a group as easily as it can be an individual. And remember that whatever the n it represents a carrier and a gene simultaneously.

Pinker replies: “Granted, it is often convenient to speak about selection at the level of individuals, because it’s the fate of individuals (and their kin) in the world of cause and effect which determines the fate of their genes. Nonetheless, it’s the genes themselves that are replicated over generations and are thus the targets of selection and the ultimate beneficiaries of adaptations.”

The first sentence—beginning with Granted—suggests that Pinker sees that the gene and the individual carrier are two sides of the same coin, but he hasn’t digested the fact. That leads him to a vapid conclusion which merely asserts that after all genes are what selection is all about.

His phrase, “targets of selection” strikes me as a truly empty metaphor. Selection is not an agent with purposes (or targets); it is an outcome. Pinker’s phrase is an effort to say the genes are ultimately more important than the carrier, but he has no reasoned justification for favoring one side of the coin over the other.

His second phrase, “ultimate beneficiaries of adaptations” refers to genes alone. I suppose I could say, no, the carriers are the ultimate beneficiaries. After all, the carriers are what actually taste life, but that’s just a matter of viewpoint. Both carriers and genes “benefit” from the adaptations and needless quarrels result from favoring one side or the other.

Comments

[Chris Crawford]

Edward O. Wilson's new book, The Social Conquest of Earth is quite forthright in declaring its support of group selection. Indeed, much of the book is an extended explanation of how eusociality has had such great evolutionary success.

[Raymond Weitzman]

The equation w = n'/n and the examples you present don't really tell us anything about gene selection or at least not unambiguously. The change in population from one generation to the next as to the number of fast runners or the number of people speaking in full sentences could just as easily be accounted for by changes in social learning from one generation to the next. In other words, it can be accounted in terms of changes in training methods, offering more incentives, weight control,and so forth.Only if these social variables were held constant from one generation to the next, would the hypothetical results you got suggest, possibly, that the change was the result of genetic selection.