When I was about to conclude the chapter on selection of my book I came across a problem of definition I’ve never thought about before. Almost without planning, leaving the ideas flow freely to the paper (actually, to the hard disc), I had begun to write about a defensive behaviour found in certain African elephants (which I’ll describe in detail below), when I found myself between a rock and a hard place: how to correlate the concept of cultural transmission to the concept of selection currently used by the vast majority of evolutionary biologists? The rock would be deleting all paragraphs regarding cultural transmission; the hard place would be changing the definition until they fit my needs, something a science writer should refrain from (freely) doing.
Before I begin to explain my problem, we should clearly define two things:
- What culture (or cultural transmission) is;
- What selection (or natural selection) is.
Selection can be defined in various ways, and different texts use distinct definitions of selection. But be warned: contrary to the second law of thermodynamics, which can be defined in dozens of different ways, the definition of selection is much more restricted, and many of the definitions that we read and hear out there are completely misleading. In a very simple yet technical definition, selection is “a non-random and consistent difference in the reproductive rates of genetically distinct biological entities”. In other words, selection occurs when variant A has a higher reproductive rate than variant B and, consequently, frequency of alleles from variant A rises while frequency of alleles from variant B decreases.
These two points are fundamental: differences in reproductive rates must be consistent and non-random. Why? To discard the phenomenon of genetic drift (for those who like statistics, when the genetic composition of a population varies over time genetic drift is the explanation associated with the null hypothesis, while selection is the explanation associated to alternative hypothesis). Furthermore, differences in reproductive rates must be due to genetic differences among the variants considered, because only this way we’ll have a change in population’s allele frequencies over time and, thus, the cumulative effect of selection. Let me illustrate this point: If you and your monozygotic twin brother (genetically identical) decide, for particular reasons, having different number of sons — for example, you have just one child, while he have four children — we cannot assert that there was selection, since this difference in reproductive rates between you and your brother is not due to genetic differences between you two (and also it is not consistent from a statistical point of view).
Once we understand the concept of selection, let’s move on to the concept of cultural transmission of a given behaviour. A lot of people think that culture is a uniquely human phenomenon. This is a mistake. All mammals I had the chance to study until today have cultural behaviours, and there are records of culture in many other animal groups. In a nutshell, we can say that a behaviour or knowledge is cultural when it was learned by a member of the population and taught by another Member (usually older, but not necessarily its progenitor). For example, a monkey who learns how to wash a food for removing dirt by viewing other monkeys washing their foods, or a crow who learns that a unknown approaching animal is an enemy after being warned by another crow, or the famous example of chimps who learn how to crack nuts by observing the complex nutcracker skills of older chimps.
The behaviour of African elephants that I mentioned earlier in this post is described in James Gould’s book “Ethology” (1982). This is a fantastic example of cultural transmission: in 1914, in an area of South Africa where is now the Addo Elephant National Park, a group of 140 elephants was intensively hunted (mostly for bringing damages to crops), and in only a year less than 20 elephants remained. When the Park was created in 1931, there were only 12 elephants from the original group, two males and ten females. These elephants mated, and the number of elephants in the park began to grow. Although they have not been hunted since park creation and despite the original survivors have died a long time ago, the fourth generation of elephants in the Park remains shy, cautious and aggressive against the humans that approach. How can they know the danger that a human represents? How can they know what happened decades ago, before they were born? What happens is that young elephants learn from older ones, which sound their species-specific alarm when they see an approaching tourist. Thus, the older elephants tell to the young, by cultural transmission, “that particular animal (humans) is dangerous, avoid him”. Much work has been done, and, from what I’ve researched, nowadays (almost 30 years after publication of Gould’s book) elephants in the Addo Park are less worried and aggressive toward humans. In fact, it is relatively simple to interrupt the advance of a culturally transmitted information, or meme (attention: here I am using the correct definition of meme, as proposed by Richard Dawkins. Meme is a unit of cultural information, not a funny video that went viral on internet). Just stop the younger from receiving information from the older animals. In the case of Addo Park, for example, if we were forbade the entry of humans for several decades, until the last elephant that had received an alarm against humans die, the population of remaining elephants would be (relatively) docile towards humans, because they would have never received a warning from the elders telling them that humans are a dangerous being.
The importance of this type of learning for those poor elephants is obvious, and doesn’t need any further explanation: it allows to elephants know that human beings are as dangerous species, without having to have a previous experience with a rifle or a shotgun, which is almost always fatal.
And it was here that I came to my problem. In most of commonly used definitions of selection, like the definition I just provided above, it is an essential, sine qua non condition, that the morphological, anatomical, biochemical or behavioural differences between variants are due to genetic differences between them, because only in this situation there would be changes in the allelic constitution of the following generations. And, without this change, there would be no evolution at all.
And what cultural transmission of information has to do with all this? Using the definition of selection previously given, we are not allowed to say that there is selection in cases of cultural transmission, because the variants do not differ genetically. Undoubtedly, however, cultural transmission changes frequencies in the following generations, but not allele frequencies! Imagine two populations of elephants, one that teaches the young that humans are dangerous (there is cultural information) and other that does not alert their young (there is no cultural information). Surely, over time, there will be a higher reproductive rate in the population that transmits the alert compared with the population that does not transmit the alert. Cultural transmission of information changes the frequency of variants in a population, but the variants here are “who transmits the information” and “who does not transmit the information”, and the difference between them is not genetic.
So, we conclude that there is change in the frequency of “individuals with a given behaviour” versus “individuals without a given behaviour.” Clearly, there is descent with modification. But, at the same time, there is no change in allele frequencies over time. What to do then?
Changing the definition of selection and removing the genotypic component from it doesn’t seem as a proper option to me. As I said before, we cannot assert that any difference in reproductive rates among variants of a given population is a selection process: to be considered selection, these differences in reproductive rates must have a cause, whatever it may be, whose frequency (the occurrence of this “cause”) changes in the next generation.
The solution that I propose here is quite simple. What is the similarity between a cultural behaviour and a genetically determined trait? Both are hereditary, transmitted from one generation to the next. So, if we want to have a definition of selection that takes account of cultural transmission, we could use something like the following:
Selection is a non-random and consistent difference in the reproductive rates of biological entities that are anatomical, biochemical, physiological or behaviourally distinct, and whose distinctions are hereditary.
This may be not the most common definition of selection, but is adequate. In fact, it is quite similar to the definition used by Stearns in his “Evolution”, a textbook that I really enjoy: “selection is the differential reproduction and survival correlated with heritable traits”. How else could handle selection of cultural behaviours? It’s worth remembering that hereditary and genetic are not synonyms. This, however, will be the theme of a future post.