The theme of Smaldino's target article is the importance of developing a theoretical understanding of how group-level traits evolve. Group-level traits are not social organizations, but rather are phenotypic effects that emerge from social organizations. He focuses on a class of social organizations from which group-level traits emerge but which are not reducible to aggregates of individuals. Smaldino does this by analyzing social organizations using Wimsatt's (Reference Wimsatt1997) four properties of aggregate systems. If a system is not reducible to an aggregate system, then it is an organization capable of producing emergent properties and behaviors. Smaldino argues that group-level traits emerge from social organizations that are strictly non-aggregative.
I believe that strict non-aggregativity is too strong a condition for social organizations capable of producing group-level traits. Strict aggregativity excludes paradigmatic examples of social systems, such as flocks of birds, that have emergent group-level traits. Although it is largely correct that flocks of birds have Wimsatt's (Reference Wimsatt1997) aggregative properties (1) through (3), the failure to have property (4) – “there are no cooperative or inhibitory interactions among the parts of the system that affect this property” – is crucial. I do believe, however, that properties (1) to (3) do demarcate an important class of social organizations and that it is fruitful to distinguish between weak and strong group-level traits that can emerge from these two classes of social organizations. The strong group-level traits emerge from strictly non-aggregative social organizations whereas weak group-level traits emerge from systems that may fail to have properties (1) to (3) but still have (4). Smaldino has done us a favor by pointing out that strong group-level traits should not be lumped in with weak group-level traits and require new models.
Indeed, strong group-level traits are especially problematic for the development of models of cultural evolution. Smaldino is well aware that strong group-level traits pose problems for developing models of selection and transmission of the social organizations that produce these traits. To reinforce and extend this point, I will devote the rest of this commentary to elaborating how difficult this problem is by connecting Smaldino's discussion of selection and transmission of group-level traits to Donald Campbell's (Reference Campbell1956a; Reference Campbell1956b; Reference Campbell1960; Reference Campbell and Schilpp1974b; Reference Campbell, Ayala and Dobzhansky1974c; Reference Campbell1975) notion of vicarious selection mechanisms.
Campbell (Reference Campbell1975) recognized that in early human cultural evolution (e.g., the early cultural evolution of tools and weapons), the physical environment played an important role, and what tools and weapons evolved were directly affected by biological fitness. However, for the evolution of group-level traits (e.g., social organization, inhibitory moral norms, and beliefs in transcendent gods), Campbell's view was that biological evolution plays a minimal role at best. For many cultural traits, and especially group-level traits, vicarious selection mechanisms are required. For example, flintknapping practices for making stone tools and weapons can be selected and transmitted by imitating or learning from those who are successful flintknappers. Individual learning is a biologically evolved mechanism, which also serves as a vicarious selection mechanism for rapidly evolving cultural practices such as flintknapping.
Vicarious selection mechanisms require variation to operate on, but Campbell's (Reference Campbell1960) requirement for blind variation-generating mechanisms in culture is too strong and not essential for models of cultural evolution (Sternberg Reference Sternberg1998). Unlike genetic heritability mechanisms in biological evolution, there are many heritability mechanisms in cultural evolution. Campbell, like Smaldino, identified many retention and transmission mechanisms in culture such as imitation, individual learning, indoctrination systems, writing, and printing. There are, of course, others especially with the introduction of the Internet, and most importantly, they are still culturally evolving.
Mechanisms for selecting, generating variation, and transmitting cultural traits are often intertwined. For example, imitating the practices of one group also involves the selection of which practices are partially imitated. In addition, the selection and imitation of practices can introduce new variation resulting from imperfect imitation as with the cultural evolution of cargo cults. Unlike natural selection in which mechanisms of heritability, mutation, and selection are independent or nearly so, mechanisms of selection, variation, and transmission in cultural evolution can be properties of a single organization.
To this entanglement of mechanisms of selection, variation, and transmission, it must be added that as cultures evolve and become increasingly complex, new vicarious selection, variation, and transmission mechanisms (VSVTMs) evolve and appear at new levels in culture. The cultural evolution of military strategies, tactics, and organization by war games is one example. War games allow new tactics and strategies for warfare – group-level traits – to evolve without conducting actual war. War games are social organizations, but they are also VSVTMs for evolving new social organizations with tactics and strategies for actual wars. The evolution of VSVTMs does not end there; in modern militaries, computer-simulated war games introduce a new layer of VSVTMs (see Wagenhals & Alexander Reference Wagenhals, Alexander, Sisti and Trevisani2001, for a discussion of simulated war games).
As cultures become increasingly complex, so do the number and kinds of VSVTMs, which evolve to facilitate rapid evolution. As they evolve, they can become deeply entwined with other social organization and VSVTMs. Thus, I would speculate that as cultures evolve with social organizations that have strong group-level traits, they also evolve VSVTMs with entanglements that produce selection-transmission thickets. For example, computer-simulated war games may be a group-level trait of war-game organizations that incorporate new individuals and technologies that facilitate the development of war-game simulations. These simulated games are group-level traits that function as VSVTMs for strategies and tactics used in war games and wars.
Selection-transmission thickets pose obstacles to modeling the cultural evolution of social organizations with strong group-level traits. They are often group-level traits that require cultural evolutionary explanation. Their component mechanisms (selection, variation, and transmission) are not aggregative (e.g., for the evolution of cargo cults, there is no clear distinction between selection, variation, and transmission). In population and quantitative genetic models of biological evolution, selection, variation, and transmission can be represented with additive or multiplicative models. For cultural evolution, the problem is how to cut through these thickets to develop tractable selection-transmission models at cultural levels.
You have
Access
The theme of Smaldino's target article is the importance of developing a theoretical understanding of how group-level traits evolve. Group-level traits are not social organizations, but rather are phenotypic effects that emerge from social organizations. He focuses on a class of social organizations from which group-level traits emerge but which are not reducible to aggregates of individuals. Smaldino does this by analyzing social organizations using Wimsatt's (Reference Wimsatt1997) four properties of aggregate systems. If a system is not reducible to an aggregate system, then it is an organization capable of producing emergent properties and behaviors. Smaldino argues that group-level traits emerge from social organizations that are strictly non-aggregative.
I believe that strict non-aggregativity is too strong a condition for social organizations capable of producing group-level traits. Strict aggregativity excludes paradigmatic examples of social systems, such as flocks of birds, that have emergent group-level traits. Although it is largely correct that flocks of birds have Wimsatt's (Reference Wimsatt1997) aggregative properties (1) through (3), the failure to have property (4) – “there are no cooperative or inhibitory interactions among the parts of the system that affect this property” – is crucial. I do believe, however, that properties (1) to (3) do demarcate an important class of social organizations and that it is fruitful to distinguish between weak and strong group-level traits that can emerge from these two classes of social organizations. The strong group-level traits emerge from strictly non-aggregative social organizations whereas weak group-level traits emerge from systems that may fail to have properties (1) to (3) but still have (4). Smaldino has done us a favor by pointing out that strong group-level traits should not be lumped in with weak group-level traits and require new models.
Indeed, strong group-level traits are especially problematic for the development of models of cultural evolution. Smaldino is well aware that strong group-level traits pose problems for developing models of selection and transmission of the social organizations that produce these traits. To reinforce and extend this point, I will devote the rest of this commentary to elaborating how difficult this problem is by connecting Smaldino's discussion of selection and transmission of group-level traits to Donald Campbell's (Reference Campbell1956a; Reference Campbell1956b; Reference Campbell1960; Reference Campbell and Schilpp1974b; Reference Campbell, Ayala and Dobzhansky1974c; Reference Campbell1975) notion of vicarious selection mechanisms.
Campbell (Reference Campbell1975) recognized that in early human cultural evolution (e.g., the early cultural evolution of tools and weapons), the physical environment played an important role, and what tools and weapons evolved were directly affected by biological fitness. However, for the evolution of group-level traits (e.g., social organization, inhibitory moral norms, and beliefs in transcendent gods), Campbell's view was that biological evolution plays a minimal role at best. For many cultural traits, and especially group-level traits, vicarious selection mechanisms are required. For example, flintknapping practices for making stone tools and weapons can be selected and transmitted by imitating or learning from those who are successful flintknappers. Individual learning is a biologically evolved mechanism, which also serves as a vicarious selection mechanism for rapidly evolving cultural practices such as flintknapping.
Vicarious selection mechanisms require variation to operate on, but Campbell's (Reference Campbell1960) requirement for blind variation-generating mechanisms in culture is too strong and not essential for models of cultural evolution (Sternberg Reference Sternberg1998). Unlike genetic heritability mechanisms in biological evolution, there are many heritability mechanisms in cultural evolution. Campbell, like Smaldino, identified many retention and transmission mechanisms in culture such as imitation, individual learning, indoctrination systems, writing, and printing. There are, of course, others especially with the introduction of the Internet, and most importantly, they are still culturally evolving.
Mechanisms for selecting, generating variation, and transmitting cultural traits are often intertwined. For example, imitating the practices of one group also involves the selection of which practices are partially imitated. In addition, the selection and imitation of practices can introduce new variation resulting from imperfect imitation as with the cultural evolution of cargo cults. Unlike natural selection in which mechanisms of heritability, mutation, and selection are independent or nearly so, mechanisms of selection, variation, and transmission in cultural evolution can be properties of a single organization.
To this entanglement of mechanisms of selection, variation, and transmission, it must be added that as cultures evolve and become increasingly complex, new vicarious selection, variation, and transmission mechanisms (VSVTMs) evolve and appear at new levels in culture. The cultural evolution of military strategies, tactics, and organization by war games is one example. War games allow new tactics and strategies for warfare – group-level traits – to evolve without conducting actual war. War games are social organizations, but they are also VSVTMs for evolving new social organizations with tactics and strategies for actual wars. The evolution of VSVTMs does not end there; in modern militaries, computer-simulated war games introduce a new layer of VSVTMs (see Wagenhals & Alexander Reference Wagenhals, Alexander, Sisti and Trevisani2001, for a discussion of simulated war games).
As cultures become increasingly complex, so do the number and kinds of VSVTMs, which evolve to facilitate rapid evolution. As they evolve, they can become deeply entwined with other social organization and VSVTMs. Thus, I would speculate that as cultures evolve with social organizations that have strong group-level traits, they also evolve VSVTMs with entanglements that produce selection-transmission thickets. For example, computer-simulated war games may be a group-level trait of war-game organizations that incorporate new individuals and technologies that facilitate the development of war-game simulations. These simulated games are group-level traits that function as VSVTMs for strategies and tactics used in war games and wars.
Selection-transmission thickets pose obstacles to modeling the cultural evolution of social organizations with strong group-level traits. They are often group-level traits that require cultural evolutionary explanation. Their component mechanisms (selection, variation, and transmission) are not aggregative (e.g., for the evolution of cargo cults, there is no clear distinction between selection, variation, and transmission). In population and quantitative genetic models of biological evolution, selection, variation, and transmission can be represented with additive or multiplicative models. For cultural evolution, the problem is how to cut through these thickets to develop tractable selection-transmission models at cultural levels.