Truth and Reconciliation for Group Selection II: The Original Problem
Consider some standard examples of design in nature: the aerodynamic wing of the bird, the concealing coloration of the moth, the dense fur of the polar bear. Darwin's insight was to explain these adaptations as a product of natural selection: individuals vary, some survive and reproduce better than others, and their properties are inherited by their offspring.
All of these adaptations are locally advantageous. Individuals possessing them survive and reproduce better than their immediate neighbors. Now consider some standard examples of social adaptations: the good Samaritan, the soldier who heroically dies in battle, the honest person who cannot tell a lie. We admire these virtues and call them social adaptations because they are good for others and for society as a whole--but they are not locally advantageous. Charitable, heroic, and honest individuals do not necessarily survive and reproduce better than their immediate neighbors who are stingy, cowardly, and deceptive.
How important is this problem? For Darwin, who was formulating the entire theory of evolution, it was one important problem among many. If we restrict our attention to the study of social behavior, however--what E.O. Wilson would later call Sociobiology--it is paramount. Most behaviors that we call prosocial require time, energy, and risk on the part of the prosocial individual. Most behaviors that we call antisocial deliver an immediate benefit to the antisocial individual. If most antisocial behaviors are locally advantageous and most prosocial behaviors are locally disadvantageous, then we have an enormous problem explaining the nature of prosociality, including the nature of human morality, from an evolutionary perspective.
Darwin was aware of this problem and proposed two types of solution. First, he observed that farmers routinely sacrifice some individuals for eating and select their traits by breeding their relatives. Individuals who altruistically sacrifice themselves for their relatives, such as the suicidal sting of the honeybee, might therefore evolve by natural selection. This idea anticipated what later would be called kin selection.
Second, Darwin observed that groups of prosocial individuals will survive and reproduce better than groups of antisocial individuals, even if antisocial individuals have the advantage over prosocial individuals within groups. Here is one of his canonical statements, using human moral virtues as an example (from Chapter 4 of Descent of Man).
It must not be forgotten that although a high standard of morality gives but a slight or no advantage to each individual man and his children over other men of the same tribe, yet that an increase in the number of well-endowed men and advancement in the standard of morality will certainly give an immense advantage to one tribe over another. There can be no doubt that a tribe including many members who, from possessing in a high degree the spirit of patriotism, fidelity, obedience, courage, and sympathy, were always ready to aid one another, and to sacrifice themselves for the common good, would be victorious over most other tribes; and this would be natural selection. At all times throughout the world tribes have supplanted other tribes; and as morality is one important element in their success, the standard of morality and the number of well-endowed men will thus everywhere tend to rise and increase.
In this passage, Darwin clearly demonstrates his awareness that a) moral behaviors are locally disadvantageous or at best deliver no advantage, compared to less moral individuals within the same group, and b) moral behaviors expressed within groups can be decisively advantageous in between-group competition. He didn't comment on the irony that within-group morality might well lead to immoral conduct among groups.
I will return to this issue later, but for the moment let's consolidate our gains. The original problem associated with group selection is foundational for the study of social behavior from an evolutionary perspective. Unlike individual-level adaptations such as the polar bear's thick fur, prosocial behaviors are locally disadvantageous. Fortunately, they are advantageous at the larger scale of whole groups. Prosocial behaviors can evolve by a process of between-group selection, as long as this process is stronger than the opposing process of within-group selection.
Is anyone confused yet? I suspect not. You don't need to be an Einstein to get the basics of group selection. All of us can appreciate that doing the right thing makes us vulnerable to exploitation. We can equally appreciate that united we stand, divided we fall. Why should such simple ideas become the basis of endless controversy?
Yet, as we proceed, I guarantee that you will become confused. One reason that a truth and reconciliation process is needed for group selection is to return to the simplicity of the original problem and Darwin's solution. As Ed Wilson and I put it in our recent review article titled "Rethinking the Theoretical Foundation of Sociobiology": Selfishness beats altruism within groups. Altruistic groups beat selfish groups. Everything else is commentary.
To be continued.
Loading comments…
Want to reply to a comment? Hint: Click "Reply" at the bottom of the comment; after being approved your comment will appear directly underneath the comment you replied to
Here's a reference to a paper that seems to take a different slant on this topic:
Copyright © 2003 Elsevier Ltd. All rights reserved.
Evolution of cooperation: cooperation defeats defection in the cornfield model
Abstract
“Cooperation” defines any behavior that enhances fitness of a group (e.g. a community or species), but which, by its nature, can be exploited by selfish individuals, meaning, firstly, selfish individuals derive an advantage from exploitation which is greater than the average advantage that accrues to unselfish individuals. Secondly, exploitation has no intrinsic fitness value except in the presence of the “cooperative behavior”. - by the simple Prisoner's Dilemma Game (PDG) -. previously shown that koinophilia (avoidance of sexual mates displaying unusual or atypical phenotypic features, such as mutations) stabilizes any inherited strategy in simple or iterated PDG, meaning it cannot be displaced by rare forms of alternative behavior which arise through mutation or occasional migration. In the present model equal numbers of cooperators and defectors (in the simple PDG) were randomly spread in a two-dimensional “cornfield” with uniformly distributed resources. Every individual was koinophilic, and interacted (sexually and in the PDG tournaments) only with individuals from within its immediate neighborhood. This model therefore tested whether cooperation can outcompete defection or selfishness in a straight, initially equally matched, evolutionary battle. - in the absence of koinophilia cooperation was rapidly driven to extinction. With koinophilia there was a very rapid loss of cooperators in the first few generations, but thereafter cooperation slowly spread, ultimately eliminating defection completely.
.Dr Wilson, I’ve recomposed my previous comments with the intention of increasing clarity.
Your statement, “Selfishness beats altruism within groups”, is well accepted as mathematically true for gene based pure altruism where “pure altruism” = reduced fitness for the individual relative to others within the group.
But gene based kin selection, reciprocal, and indirect altruism can beat selfishness within a group since each, on average and in the long term and in the right environment, increases the fitness of the actor’s genes. These behaviors are necessarily encoded in our brains as simple heuristics like “feel rage and an urge to violence if family or friends are attacked by an external agent” which should produce willingness to risk injury and death to defend family or friends.
OCCASIONALLY though, this kind of simple heuristic will result in injury or death that reduces the fitness of the genes responsible for it relative to others in the group. It is only on average and in the long term that these genes increase fitness within the group, not in every case. Occasional ‘bad luck’ in carrying out our simple heuristics for kin selection, reciprocal, and indirect altruism might account for much of the observed levels of gene based pure altruism. And selection on higher group levels could then reinforce, (and perhaps make more altruistic), such imperfect heuristics.
Of course observed levels of pure altruism in humans are also due to culture based influences, but that is another story.
Best regards,
Mark Sloan
“Selfishness beats altruism within groups. Altruistic groups beat selfish groups. Everything else is commentary.”
This is true only in cases where altruism is “pure” altruism, (an individual’s fitness is decreased and fitness of others is increased), with no possibility of returned altruism.
But there are important circumstances where it is not true.
Competitions between completely self interested independent agent computer programs have repeatedly illustrated the evolution of reciprocal and indirect altruism based cooperation when 1) individuals are provokable to retaliation when treated ‘unfairly’ and 2) there are synergistic benefits for cooperation compared to purely selfish behaviors within a group.
Suppose reciprocal and indirect altruism with provokability become evolved heuristics for behaviors based on the selection force of increased benefits of cooperation. These evolved heuristics have, by definition, benefited individual fitness in the past only on average and in the long term. There is no guarantee these heuristics always maximize fitness. We should expect such evolved behaviors to occasionally produce pure altruism, like a soldier jumping on a live grenade to save his friends, just as they have in the past.
So: Evolved cooperative behaviors based on short term altruism that is provokable to retaliation can beat selfishness in a group even though it occasionally produces pure altruism.
I only wish to point out that both group selection and provokable short term altruism can account for the evolution of occasionally purely altruistic behaviors. Both should be included in the discussion and mathematically evaluated for probability.
Mark Sloan
See David Sloan Wilson's Profile
Thanks for this interesting and informed comment. A point that I will elaborate upon in future installments is that when you look at concepts such as reciprocal and indirect altruism in detail, they include the ingredients of multilevel selection within their own structure. Take tit-for-tat, the classic example of reciprocity. It never beats its partner; it only loses or draws. Tit-for-tat succeeds as an evolutionary strategy only because groups of tit-for-tat (=pairs, in 2-person game theory) are more productive than mixed groups or groups of all-defect. More generally, when we realize that n-person game theory is a model that assumes multiple groups of size n within which social interactions occur, it contains all of the elements of a group selection model. It's true that the reciprocity trait can be more fit than the non-reciprocity trait all things considered--that's why it evolves in the total population--but it does not evolve on the strength of a relative fitness advantage within groups.
I considered including in my original comment that reciprocal and indirect altruism based cooperation can be alternately viewed, as you have stated above, as forming a cooperative new subgroup level that can out compete the uncooperative members of the original larger group level.
I don’t remember seeing creation of cooperative subgroups within an original group clearly laid out as an important part of multilevel selection. Discussions and defenses I remember mainly focused on higher group level selection for altruism rather than lower sub-group level selection for whatever the minimum kind of provokable altruism that is required for cooperation. But that could be just my lack of knowledge.
My omission was for the sake of brevity and to emphasize that we should expect evolved heuristics for reciprocal and indirect altruism based cooperation to occasionally produce pure altruism just due to the fallible nature of simple heuristics in complex situations. That is another point I don’t remember seeing clearly stated before.
Thanks for your reply,
Mark Sloan
Sociobiology may need to take the folowing into account:
All organisms have survival strategies and the number and forms of options available will depend on their adaptation to a particular environment. Primarily social animals will arguably have more options that the primarily predatory - in part because there is always an element of within-group competition, depending on resources available to its members. Resources can include the relative abilities of members to gain access to scarce items. And if you are born the weakest physically, you will necessarily opt for strategies that the strongest will necessarily eschew.
The danger comes when you assume that to be born weak is to also be born with the strategies most fitting for survival of the weakest. This may seem to be the case when you are dealing with colonies of insects, but it hasn't been the case when dealing with the branch of life that separates humans from that line.
Insect life seems to have developed an optional set of individuals to carry out diverse strategies rather than, or in addition to, an optional set of strategies to be carried out by diverse individuals.
See David Sloan Wilson's Profile
Dear royniles,
I'm not sure what you are getting at here.
1) Are you contrasting social animals with solitary animals? Many social species are predatory.
2) The idea of condition-dependent strategies could apply to both social and solitary species.
3) I certainly agree that humans are far more behaviorally flexible than social insects. Nevertheless, both can qualify as group-level superorganisms, even if they differ in the proximate mechanisms by which they achieve that status.
I'm basically contrasting insect evolutionary characteristics and traits with those of the animal line that humans are a part of. And I have recognized that social species can be predatory, which is why they may have strategic options that the predominately predatory don't have.
And both insects and humans form group-level superorganisms, but the strategic dynamics of one are not the same or even comparable where individual contributions to the group are concerned. And it was my sense that by equating traits as distinctive to the makeup of selected individuals, you were drawing inferences about human behavior from insect behavior that were just not apt.
Hi David,
You didn't respond to my comment on your first piece. So here it is again. I'd be interested in your comment.
I'm a big fan. I recommend Evolution for Everyone whenever I get the chance. It's a great book about evolution, a great book about you personally, and it makes the case for group selection clearly and convincingly.
What I see as one of its fundamental insight is that everything is both a group and an individual.
But that insight makes it clear why group selection is not as important as you make it out to be. It is not groups that are selected; it is individuals (that look like groups) that are selected. The point is that the concept of individual should be understood more broadly.
As you say, altruistic groups beat selfish groups. Or to put it in the singularly, an altruistic group will beat a selfish group. It's the group as an individual that survives, not the group of people as individual people. In this case the group itself is an individual in much the same way that the cells in our body function as a human being.
What makes group-like individuals different from what we normally think of as individuals is that they are non-contiguous. But so what? They are still individuals. Given this perspective there is no need to continue the fight for group selection. The important question is: what are the individuals in the world, and how do they evolve?
See David Sloan Wilson's Profile
Dear RussAbbott,
Thanks for your kind words about Evolution for Everyone. The concept of major transitions has turned the rejection of group selection on its head by showing that higher-level selection can trump lower-level selection and that the individuals of today are literally the groups of past ages. However, within-unit selection is only suppressed and never entirely eliminated, even when the unit is an individual organism, as many examples of intragenomic conflict attest. I am happy to stress, along with my colleagues Peter Richerson and Robert Boyd, that human groups are at best "crude" superorganisms, with many examples of within-group selection on display. See Peter Turchin's book "War and Peace and War" for a panoramic view of human history as a fossil record of multilevel cultural evolution. To summarize, whenever selection operates at a given level, units of that level become adaptively organized in a way that we associate with the word "individual", as you say, but selection never operates entirely at a given level.
"Charitable, heroic, and honest individuals do not necessarily survive and reproduce better than their immediate neighbors who are stingy, cowardly, and deceptive."
That premise is false and to the extent that your entire thesis is dependent on it, your presentation fails.
Stingy, cowardly and deceptive are not fixed characteristics, but are descriptive terms that originate within these small groups to establish a relative pecking order. Select out and isolate all cowards in a new group, and that group will produce it's own heroes.
Recommended reading as to human nature at its most basic: Don't Sleep, There Are Snakes, by Daniel Everett.
See David Sloan Wilson's Profile
The basic premise is impeccable, based on the basic concept of trade-offs. A group designed to function optimally as a group will require different traits than a group whose members are designed to maximize their fitness relative to each other. It's true that rewards and punishments can alter the benefits and costs associated with prosocial and antisocial behaviors. However, this requires us to explain the evolution of rewards and punishments, in addition to the behaviors that they regulate. More on this in future installments.
You are right that cowardly and deceptive are not fixed characteristics. In game theoretical terms, they are conditional strategies, but that doesn't alter their basic evolutionary dynamics.
But the way the premise was phrased implies that cowardly and deceptive ARE fixed characteristics of those whose immediate "neighbors" are supposedly charitable and honorable by nature. And I don't think there is a problem explaining the evolution of reward and punishment any more than there has been a problem recognizing that trial and error, and consequences thereof, have been among the more important driving forces of evolutionary change from perhaps day one.
And if you are to be the foremost champion of group selection, you have a special responsibility to denounce the wretched output of Kevin MacDonald, whose particular group selectionist paradigm is notorious. There are ugly implications of group selection - genocide, racism, xenophobia, jingoism - and it is incumbent upon promoters of group selection to repudiate those who would attempt to derive "ought" from "is."
See David Sloan Wilson's Profile
I have already acknowledged that group selection doesn't resolve conflict so much as elevate it up the biological hierarchy, from within-group to between-group interactions. Kevin MacDonald was among the first to propose that a religion such as Judaism could be a group-level evolutionary strategy, as he put it. That is my thesis for religions and other social organizations in general, which I develop in Darwin's Cathedral. It has become increasingly mainstream over the years--check out the current issue of the journal Bioeconomics for a review article by Janet Landa and myself, followed by commentaries, on middleman merchant societies around the world as an example of convergent cultural evolution. Between-group conflict happens. So does between-group cooperation. We badly need a theory to explain how they happen and how we can stack the deck in favor of cooperation. A theory that explains between-group conflict does not, or should not, condone between-group conflict. Kevin MacDonald's work, in addition to my own, needs to be evaluated with respect to what "is". To the extent that Kevin MacDonald (or anyone else) attempts to derive moral prescriptions (what we ought to do) from his work, those prescriptions need to be evaluated by conventional moral standards.
See David Sloan Wilson's Profile
The naturalistic fallacy is a complicated subject in its own right--perhaps the subject of a future blog. For an academic treatment, see this article: Wilson, D. S., Dietrich, E., & Clark, A. B. (2003). On the inappropriate use of the naturalistic fallacy in evolutionary psychology. Biology and Philosophy, 18, 669-682.
OK, Dr. Wilson, you were right. You were right in the mid-90s as part of the small group selection contingent within HBES and evolutionary biology in general. As a grad student, I agreed with many of my peers that you were wrong. But I was wrong and you were right. And I don't mind saying so.
That's what this "truth and reconciliation" business is all about, isn't it? Let me guess: From your perspective, your colleagues are failing to give you the credit you deserve and are instead churlishly trying to redefine the issue or revise intellectual history. (But to call it "truth and reconciliation" somewhat trivializes the phrase, considering the massive human tragedies that it is usually applied to.)
I changed my mind on group selection as I further considered the issue of multilevel selection, of which you were not the only advocate. Especially somatic selection. The major transitions work of Maynard Smith and Szathmary was also crucial.
However, just because you were right about selection between groups does not mean that every one of your utterances about group selection - speculations about music and religion, philosophical asides - is correct. I think you have a tendency to be overly adaptationist, but usual critics of adaptationism like Lewontin are inclined to give you a pass on that because these are communal rather than individual adaptations.
See David Sloan Wilson's Profile
Thanks to Bateman for this comment. To briefly respond:
1) Thanks for acknowledging that the rejection of group selection needs to be rethought.
2) Borrowing "truth and reconciliation" is a appropriate for what I am trying to do. It is not trivial to resolve confusion about the nature of social conflict and cooperation from an evolutionary perspective. As for my desire to receive credit, that kind of self-serving motive might--or might not--explain anyone's behavior, no?
3) I never claimed to be the only advocate of group selection. Other people represent me as the lone ranger, but I always credit folks such as Maynard Smith and Szathmary for the concept of major transitions, etc. As I will describe in future installments, there is an entire community of multilevel selectionists out there.
4) Of course every one of my utterances about group selection is not necessarily correct. As with any scientist, every utterance must be scrutinized on a case-by-case basis. Give them your best shot :)
d.
You must be logged in to comment. Log in or connect with
