Multilevel selection

The theory of multilevel selection aims to explain that there are different levels on which natural selection can attack. The common neo-Darwinian theory that only the individual is the object of selection (or - with Richard Dawkins - only the gene ) is replaced by an overarching theory: According to this theory, selection forces act both below the individual level (organs, cells , genes) as well as above (group, population ) - possibly even simultaneously.

The theory was developed in the 1990s by the evolutionary biologist David Sloan Wilson and the philosopher Elliott Sober .

Prehistory and fundamental criticism of group selection

Charles Darwin already knew the idea of group selection and used the idea. But he could not solve the problem workers - insects without an own fertility. At the time he was not aware that worker insects are more closely related to their siblings than to the queen.

It was only the Briton Vero Wynne-Edwards who worked this out in the 1960s on the theory of group selection . According to her, traits that are good for the group may evolve even if they are not fitness-promoting for the altruistic individual himself . Altruism , analogous to William Donald Hamilton's theory of kin selection, is an important element, but Wynne-Edwards did not achieve success with his concept. He couldn't show how the groups can stay stable. Critics such as the American George C. Williams accused him that what is explained at the group level can always be explained at the individual level: "One should not use the adaptionist idea above the level of the individual." Williams did allow group selection as a theoretical possibility, but rated it rarity in empirical environments.

According to the critics, a group always consists of individuals and it is the individuals whose genetic makeup or behavior inevitably maximizes the fitness of a group. After all, Hamilton had shown a way with kinship selection. The group selection, however, wants to consistently show beyond family relationships that the group can be seen as an evolutionary unit for which the selection can in principle play as well as on the individual level. If necessary, it should be demonstrated that natural selection acts simultaneously on the different levels.

Wilson & Sober published an essay in 1998 in which they again advocated group selection. They called the concept MultiLevel Selection Theory .

Examples of group behavior in people

DS Wilson makes the problem situation clear: What happens when you bring a good and a bad person together on an island? Second, what happens when you put a group of good people on one island and a group of bad people on the island next door? And finally, thirdly, what happens if you allow a bad person to swim to the island with the good guys?

"Good" is equated with: "altruistic, friendly, loving, loyal, forgiving" etc. "Bad" is equated with "egoistic, insidious, hateful, avaricious, cowardly, treasonable, angry". Darwinian natural selection is based on fitness differences in a group, which express the good and the less good. Natural selection is based on differences in fitness between groups, which, it is argued, leads to the development of good traits:

In the first case: the bad kills the good, packs his provisions and flees the island. In the second case: The good team look for a solution to get away from the island or they set it up there and live in peace while the bad kill themselves on their island. In the third case, there are many good ones and one bad. Is the bad guy taking advantage of their behavior? What is happening to him? Can one or a few bad people put a group of otherwise good people at a disadvantage compared to another group of only good people? The evolutionist asks: Can fitness differences arise when a group of sincere people is undermined by one or a few troublemakers who ruthlessly exploit their good intentions? Exactly here lay the problems of instability in the group, which were previously insoluble in theory of group selection, with which Wynne-Edwards got stuck, and which is why the idea of group selection always met with rejection.

Optimization of laying hen batteries

Wilson uses another example to describe how to get an enclosure of 20 hens that in total lay the most eggs (Wilson 2007,33f). In the past, breeders looked for the most productive hens from a larger group, repeatedly selecting the one or two dozen best laying hens for a few generations until, after a number of generations, the best were determined. However, this had the unpleasant effect known among breeders that the remaining best laying hens in the group did not tolerate any competition and killed each other. The American William Muir discovered the connections: If everyone gives their best in the state, it is not necessarily best for everyone.

The search for patterns for group-specific fitness behavior

It has to be demonstrated that a group of individuals produces a higher level of fitness through their behavior than through the fact that each individual in the group maximizes their fitness. That means that the reproductive maximum of the whole is only reached through the reproductive restriction of its members. But not among close relatives as with Hamilton. This thinking goes beyond Darwin's teaching, according to which selection at the individual level is fitness maximization at the individual level. However, this is not the case here.

If every hen turns back her laying frenzy a little for the good of the enclosure, this becomes more evident in the end in a few eggs overall, apart from the fact that all the chickens stay alive.

So far we have been dealing with a changed behavior of individuals. Wilson further asks: Can the changed behavior of the group members lead to a new type of behavior of the group as a whole, to a behavior that is not shown in that of the individuals? He wants to know: Can you imagine group behavior as a new quality that is different from the sum of the behavioral characteristics of the members of the group?

Neural analogy: Consciousness does not exist in a neuron

A similar problem arises when asked about the location of consciousness: consciousness is not found in a neuron. You can measure your electrical voltage exactly, analyze the transmission of the biochemical messenger substances, but you can find neither consciousness, memory nor feelings. The organism creates such for itself only at the higher level. And then you have to do with something other than the sum of the information stored in the neurons. Is there such a thing in groups of individuals? And can it help improve fitness, i.e. play a role in evolution?

Wilson distinguishes within-group selection , i.e. natural selection within a group, from between-group selection , the selection between groups of a species.

Swarm intelligence in bees

Wilson explains the example of swarm intelligence ( collective intelligence ) in bees. A beehive must come to decisions that will ensure its daily survival. If the swarm cannot collectively, quickly, and correctly decide when searching for food, when food shortages or when looking for a new home for a part that has split off from it, the survival of the swarm is at risk. Democratic voting methods are inefficient, too lengthy, too costly. Nevertheless, alternative offers must be “obtained”, “compared” and “weighed up”; more urgent or better alternatives must be given preference over less good ones. The “preference” means: It takes an unmistakable, unambiguous decision for everyone. Individual preferences have to be transformed into coherent instructions for action. Evolution must have found ways here for an insect strain of bees to come up with effective, reliable instructions.

Wilson draws on Thomas Seeley's research on swarm intelligence. The German zoologist Karl von Frisch discovered the dance language of bees in 1920 . Bees can communicate in this way. A bee returning from a successful foraging for food performs a certain dance. Their conspecifics recognize from the figuration of the dance, the length, the speed, the wagging of their rear end and other patterns, in which direction and distance a new food source is, how productive it is, etc. The corresponding signals are given by the other bees recognized and interpreted and they know the way to the better feeding place.

However, swarm intelligence is more than the intelligence of the bee. Seeley suspected and searched for behavioral patterns of individuals that are transmitted to the entire swarm of tens of thousands of bees in such a way that the swarm carries out actions that cannot be derived from the signals of the dancer (s). It is not that some people observe the dancers, compare and evaluate signals in order to pass on signals until in the end everyone knows what is going on. It is essentially different. It is about the behavior of the entire swarm, which through specific behavior can then also become an object of selection as a collective, as a superorganism , because this behavior increases the swarm's ability to survive or its fitness. Less fitness without group behavior, higher fitness with group behavior.

The bees' dance forms do contain specific signals to their conspecifics about the quality of food sources, etc. If other bees implement these specific signals "correctly" and follow up on the information, we are evolutionary at the level of individual selection. According to Wilson, she is always involved. The additional question is now: Can it be overwritten by selection on the next level? How does the swarm work as a closed, integrated unit?

The dance contains information that Karl von Frisch correctly described. The length of the dance is proportional to the sugar content of the food source found. The bees have to watch the bees dancing in the trunk, compare their information. But you also come to the nutritional information on an additional, not individual, route. According to Seeley, the length of the dance also leads in another way, namely through the fact that more bees perceive the longer dance, purely statistically motivated bees to fly out that would otherwise not fly out at all. According to his theory, none of these bees compare the different information content of dance lengths, although this information is actually available. A bee simply picks a dancer by chance. The length of a dance alone generates a statistical direction that more bees seek a good source of food.

No animal in the swarm understands the whole, and yet: Each one contributes its part to success. There is no commander in chief, no central authority, no administrative apparatus. The queen is not involved. Instead, there are evolved, highly efficient, situation-dependent coordination processes. The crush responds to challenges and finds the solutions that a single member cannot find.

Foundation of a new bee colony

Seeley also writes about how bees are creating a new state. From a certain swarm size, the queen moves out with around half of her colony and sits down on a branch of a nearby tree. Scouts go in all directions and look for the new nesting place. Criteria such as the cavity are examined. It must be big enough and at the correct height above the ground. There is a clear conception of the size of the entrance hole, and the presence of dead conspecifics that may have already lived in the searched place is another criterion. Ultimately, the new nesting site must not be too close to that of the old swarm in order to prevent conflicts. Once a scout has found a place he thinks is good, he stays there for a while. Other scouts also find the place. If a certain number of scouts, a certain threshold value, is reached at this very place, they all fly back to the swarm. No more than a hundred bees determine where tens of thousands will move. The returnees use specific signals to tell the swarm what they have scouted and where it is. In less than a minute, the entire swarm has detached itself from its branch and flies in a closed formation to the new location that the scouts have chosen. According to Seely, nobody knows exactly how the swarm will find their way there, which is up to two kilometers long.

A fraction of the crush made the right decision for everyone. It is implemented without hesitation. Few of them decide on the behavior that is essential for survival for their entire swarm. No individual knows the decision made.

Implications for the theory of evolution

No altruism required

The multilevel selection theory does not need a self-sanctifying altruism as with Hamilton. Family relationships become unimportant and take a back seat. Behavioral variations between groups can also be large, although the genetic variation between them is small, for example when members of a group imitate others or adopt certain social norms. Individuals do not have to primarily worry about the welfare of the group, as was originally seen by Wynne-Edwards. The independence from altruism is given because the specific forms of group behavior (see examples) can themselves promote fitness. Accordingly, there is no discussion about the fitness stability of the group as with group selection .

The importance of swarm intelligence

From the swarm behavior it can be seen that the selection can attack at this level. The requirements for group selection are met here. The idea that swarms develop different degrees of fitness over many millions of years because they adapt to different degrees when searching for food, when finding nests, i.e. with tasks that they can only solve as a swarm and not as individual individuals, is only very good It is difficult to imagine that such characteristics do not represent a form of group selection, i.e. characteristics that arise from the social behavior of all its members and can also be described as such. The fact that selection exists on different levels, from genes to cells, organs and organism to small or large groups, which today is a must to reject, takes a lot of effort. The hunger of a bee colony cannot be traced back to the hunger of any individual bee. The intelligence of a higher-level unit cannot be found in any of the parts, but rather emerges from the interaction of the parts. This expresses the multilevel selection.

Matryoshka dolls

With the multilevel selection theory, according to the understanding of its authors, a unified theory of natural selection has emerged, which is based on the idea of a nested hierarchy. Think of it like Russian matryoshka dolls. That is how the authors want it to be understood. Group selection is seldom the only force acting on a characteristic expression; individual selection is almost always present. That is why hierarchical theory explains both. Adaptation at every level of biological hierarchy requires a process of natural selection at that level. That is the core thesis of Wilson & Sober. Williams' claim to apply the adaptionist idea exclusively at the individual level is fundamentally wrong.