The social sciences have recently become a common battleground for debates about group selection. From Jonathan Haidt’s use of group selection to explain “groupish” traits in humans, to some of the recent rumblings about cultural group selection (as expressed in a debate triggered by Steven Pinker), and even at last years Consilience Conference, group selection has undergone a revival in the social sciences.
One social science in which group selection has re-emerged is economics, with the most recent occurrence in the Journal of Economic Behavior & Organization special issue, Evolution as a General Theoretical Framework for Economics and Public Policy.
In that special issue, group selection is one of the central threads of an article by John Gowdy and colleagues, who take apart three Western “cosmologies” that the authors consider have influenced economic thought. These cosmologies are that “natural man” is a rational, self-sufficient and egotistical individual, that competition between individuals can lead to a well-functioning society, and that there exists an optimal state of nature (equilibrium). I will deal with each of these three cosmologies in separate posts, with this post on the nature of “natural man”.
Gowdy and colleagues hint at some of the previous debate over whether man is purely self-interested, pointing to the origins of the cosmology in work by Pareto, Jevons and Walras, along with critiques of this view by Edgeworth, Veblen and some modern writers. However, the bulk of the argument developed in the paper relates to the nature of groups.
Their opening argument is that group level phenomena can affect individual behaviour, which can then affect the economic system as a whole. This is a fair point, and was one of the central themes of An Economic Theory of Greed, Love, Groups and Networks that I recently reviewed. Humans are highly social and highly responsive to group-level incentives.
This point, however, leads Gowdy and colleagues to group or multilevel selection. As they frame the question:
How can natural selection favor traits that are “for the good of the group” when they are selectively disadvantageous within groups? The answer is, by a process of between-group selection. Groups of solid citizens outcompete other groups, even if solid citizens are not selectively advantageous within their own groups.
The authors note the arguments of George Williams and friends of the 1960s, who generally considered that while group level selection can theoretically occur, between group selection was invariably weak as there was too much mixing between groups. Selfish individuals within a group undermine any potential for group selection. At this point, however, Gowdy and colleagues rewrite history when they suggest that the second of these points - the weak potential for group selection - has been overturned. They go as far as stating that:
The consensus among evolutionists that humans are a highly group-selected species (as conceptualized within multi-level selection theory) challenges the individualistic assumption of economics at its core.
This is where I need to be careful with language, particularly given the mention of multilevel selection theory.
The phrase “group selection” is often used in reference to selection between populations, and this was the context for the debates between Wynne-Edwards, Williams, Maynard Smith, Dawkins and others. In the mid to late 1970s, David Sloan Wilson (one of the authors of the paper the subject of this post) developed a different conception of group selection, generally termed multilevel selection, which looks at the development of individual traits within group structured populations. The groups are not population size groups, but are rather “trait groups” within populations that occur through non-random assortment of altruistic genes. These are not groups in the sense that people typically use the word. For example, when I engage in a trade with someone, we could be considered a group under this multilevel selection framework. (I have written extended posts explaining the concept of multilevel selection here and here.)
In multilevel selection theory, selection occurs at all levels and the multilevel selection framework allows you to partition the selection effects between these levels. Importantly, multilevel selection theory is generally accepted as a different way to conceptualising the same evolutionary processes as are captured by the concept of inclusive fitness (kin selection). Gowdy and colleagues hat tip to this concept by noting that altruistic actions can be made to appear selfish by altering the frame of comparison.
But when Gowdy and colleagues flick back and forth between the group and multilevel selection terms, and particularly in their reference to Wynne-Edwards and Williams, I become confused about which group selection concept they are talking about. Are they talking about the old group selection concept as debated in the 1960s? If so, then their statement that evolutionary biologists accept that humans are a highly group selected species would be wrong. That idea appears as contested today as it ever was.
But what of the newer multilevel selection theory? Yes, biologists would agree that multilevel selection occurs in humans, although they differ in their opinion as to its usefulness as a frame of analysis. But since it is just a frame of analysis, what does it mean to say that humans are a group selected species? By using the other frame of reference, are we also a highly kin selected species?
The ability to change the frame of comparison can be seen in the examples of traits they use to illustrate that humans are a highly group selected species. We are other-regarding as to kin. There are clear benefits to self through having a social brain and the theory of mind to put ourselves in others' shoes. And there can be huge personal benefits to being cooperative (particularly with kin).
Now, this is not to say that I do not agree with the authors' overarching point. Despite many occasions where other-regarding considerations have been included in economics analysis (economics is a huge field with many practitioners), it does not happen as often as it possibly should. Pressures within groups, our social natures and our “groupish” traits have major effects on our actions.
But I am not certain that the multilevel selection approach is the optimal way to sell the concept that these groupish, social traits need to be considered in economic analysis. If evolutionary biologists tend to disagree on that point, are you going to be able to sell it to the economists? A starting point would be to at least get the nature of those traits on the table, present the design process from both frames of reference, and make it clear that it is not the old group selection concept that you are talking about.
As an end note, the papers in the special issue are easy to read and are not full of the usual mathematical signalling contained in economics journals. I recommend reading them yourself.
My series of posts on the Journal of Economic Behavior & Organization special issue, Evolution as a General Theoretical Framework for Economics and Public Policy, are as follows:
Four reasons why evolutionary theory might not add value to economics - a post on David Sloan Wilson and John Gowdy’s article Evolution as a general theoretical framework for economics and public policy
Economic cosmology - The rational egotistical individual (this post) - a post on John Gowdy and colleagues' article _Economic cosmology and the evolutionary challenge _
Economic cosmology - The invisible hand - a second post on _Economic cosmology and the evolutionary challenge _
Economic cosmology - Equilibrium - a third post on Economic cosmology and the evolutionary challenge
Design principles for the efficacy of groups - a post on David Sloan Wilson, Elinor Ostrom and Michael E. Cox’s article Generalizing the core design principles for the efficacy of groups