Human groups are not aggregates of interchangeable parts. The “who” of groups matters. Diversity across individuals produces collective group phenotypes that are both variable across groups and impossible for any single individual to express. Smaldino insightfully argues that such emergent group-level phenomena (i.e., collaborative interdependence) drive human cultural evolution. He further proposes these units of selection are inadequately represented in current models of cultural multilevel selection (cMLS) that follow an inclusive fitness approach. We submit, however, that an MLS model for group-emergent traits well ensconced in inclusive fitness reasoning already exists. This is social heterosis – where group diversity, itself, can be the mutually advantageous trait (Nonacs & Kapheim Reference Nonacs and Kapheim2007). Under certain conditions, individuals in heterogeneous groups achieve higher fitness than they would in homogeneous groups, and diverse groups experience more beneficial collective properties than uniform ones do. Social heterosis models examine emergent group-level fitness through nonadditive benefits that individuals gain from being in specific differentiated and organized units. Although originally applied to the maintenance of genetic diversity, social heterosis is also relevant to cultural evolution. The possibility that groups can have collaborative interdependence is an intrinsic part of social heterosis.
Collaborative interdependence can arise within social heterosis theory through social genomes. A social genome forms across group members that possess complementary alleles at key genes (i.e., fitness epistasis across individuals). For example, Nonacs and Kapheim (Reference Nonacs and Kapheim2012) accurately modeled the evolutionary progression of HIV to AIDS within human hosts. As Smaldino describes for collaborative interdependence, replicative fitness of HIV increased nonadditively when genetically different clones evolved complementary capabilities and could co-infect the same cell. In other words, once an HIV social genome evolved, immune system collapse followed. Just as the social genome is an example of intergenomic epistasis, collaborative interdependence is an example of cultural epistasis. Cultural epistasis may indeed provide the emergent fitness properties that help maintain cultural diversity.
Culturally based social genomes could arise in multiple contexts in humans, and as Smaldino persuasively argues, perhaps too much weight has been given to overt cooperation, with its implicit assumptions of altruistic or self-sacrificial behavior. Social heterosis can arise without explicitly cooperative interactions or any direct interactions at all. For example, consider Smaldino's collaboratively interdependent “seal-hunters” and “kayak-builders.” Social heterosis is evident as cultural epistasis: For example, successful groups require both skills to be present and payoffs are nonadditive. If hunting is always more prestigious with higher payoff, then the ratio of hunters to builders will increase in all villages, as hunters can support more children (if hunting is genetically heritable) or more children will decide to hunt (as a cultural preference from observing outcomes). Villages with too many hunters, however, will flounder as a result of few and poor boats, leading to fewer offspring for repopulating and export. In contrast, villages with healthier mixes of occupations will export more new hunters and builders. In general models of such scenarios (Nonacs & Kapheim Reference Nonacs and Kapheim2007; Reference Nonacs and Kapheim2008), population trait diversity can be maintained if the mean fitness of traits disadvantaged within groups exceeds the fitness of advantaged traits when they are in less diverse groups (i.e., builders in diverse villages do better than hunters in less diverse villages, although within any given village, hunters always do better). Therefore, between-group variance across village cultural genomes would maintain skill set diversity in populations without requiring either kin-biased nepotism or reciprocated cooperation.
Smaldino proposes novel mechanisms are needed to explain human group-level traits. However, the three proximate mechanisms Smaldino proposes for the emergence of group-level traits are encompassed in the emergence and maintenance of a social genome through social heterosis. For the HIV example, leadership arises because epistasis across loci appears to evolve in a predictable order, division of labor arises when complementary alleles give rise to new function, and repeated assembly occurs via cell death and infection. Nonadditive benefits of intergenomic epistasis enhance the probability that group-level properties will be preserved across repeated assemblies. Analogously, collaborative interdependence leads to and maintains emergent properties of cultural groups.
Social heterosis and social genomes are MLS models and fit comfortably in an inclusive fitness framework such as Hamilton's Rule (i.e., traits are selectively favored when benefits provided to relatives exceed costs to the actors, or rb – c > 0). The key to social heterosis is that relatedness, benefit, and cost are not independent variables; instead, benefits provided (b) or costs incurred (c) correlate with relatedness (r) at the group level. For example, Smaldino highlights the role high genetic relatedness plays in social insect evolution. Although the initial evolution of cooperative breeding and specialized worker castes correlate with high relatedness through strict monogamy (Boomsma Reference Boomsma2013), subsequent evolution has often led to low relatedness, as polygamy has independently appeared in more than 20 different taxonomic lineages (Hughes et al. Reference Hughes, Oldroyd, Beekman and Ratnieks2008). Genetic diversity and “who” resides in colonies significantly affects colony-level fitness (Wray et al. Reference Wray, Mattila and Seeley2011), such that higher b and lower c associate with lower r. Close relatedness is disadvantageous when group diversity benefits are positive and nonadditive. Indeed, diversity-producing social genomes of low relatedness may be the hallmark of behavioral complexity and ecological success throughout the social Hymenoptera (Nonacs Reference Nonacs2011a; Reference Nonacs2011b).
In conclusion, social heterosis and social genomes models can predict the evolution of genetic traits, and we propose they similarly apply to cultural ones. It does not matter if “kayak-building” is cultural rather than genetic because the trait is still transmitted with fidelity. Unlike genes, culture transmits both vertically and horizontally across individuals and can be malleable within individuals over their lifetimes. Hence, cultural traits increase in frequency through differential net rates of phenotypic conversion within populations. Just as the logic of natural selection applies to how gene frequencies change across generations, the same logic can apply to the spread and evolution of human culture, behavior, and practices. Our rejoinder to Smaldino's view that, “models are needed that capture the difference between the social spreading of a particular individual-level trait and the emergence of group-level behaviors that rely on differentiation and organization” (sect. 8, para. 2), is that they already exist – we just need to use them.
Human groups are not aggregates of interchangeable parts. The “who” of groups matters. Diversity across individuals produces collective group phenotypes that are both variable across groups and impossible for any single individual to express. Smaldino insightfully argues that such emergent group-level phenomena (i.e., collaborative interdependence) drive human cultural evolution. He further proposes these units of selection are inadequately represented in current models of cultural multilevel selection (cMLS) that follow an inclusive fitness approach. We submit, however, that an MLS model for group-emergent traits well ensconced in inclusive fitness reasoning already exists. This is social heterosis – where group diversity, itself, can be the mutually advantageous trait (Nonacs & Kapheim Reference Nonacs and Kapheim2007). Under certain conditions, individuals in heterogeneous groups achieve higher fitness than they would in homogeneous groups, and diverse groups experience more beneficial collective properties than uniform ones do. Social heterosis models examine emergent group-level fitness through nonadditive benefits that individuals gain from being in specific differentiated and organized units. Although originally applied to the maintenance of genetic diversity, social heterosis is also relevant to cultural evolution. The possibility that groups can have collaborative interdependence is an intrinsic part of social heterosis.
Collaborative interdependence can arise within social heterosis theory through social genomes. A social genome forms across group members that possess complementary alleles at key genes (i.e., fitness epistasis across individuals). For example, Nonacs and Kapheim (Reference Nonacs and Kapheim2012) accurately modeled the evolutionary progression of HIV to AIDS within human hosts. As Smaldino describes for collaborative interdependence, replicative fitness of HIV increased nonadditively when genetically different clones evolved complementary capabilities and could co-infect the same cell. In other words, once an HIV social genome evolved, immune system collapse followed. Just as the social genome is an example of intergenomic epistasis, collaborative interdependence is an example of cultural epistasis. Cultural epistasis may indeed provide the emergent fitness properties that help maintain cultural diversity.
Culturally based social genomes could arise in multiple contexts in humans, and as Smaldino persuasively argues, perhaps too much weight has been given to overt cooperation, with its implicit assumptions of altruistic or self-sacrificial behavior. Social heterosis can arise without explicitly cooperative interactions or any direct interactions at all. For example, consider Smaldino's collaboratively interdependent “seal-hunters” and “kayak-builders.” Social heterosis is evident as cultural epistasis: For example, successful groups require both skills to be present and payoffs are nonadditive. If hunting is always more prestigious with higher payoff, then the ratio of hunters to builders will increase in all villages, as hunters can support more children (if hunting is genetically heritable) or more children will decide to hunt (as a cultural preference from observing outcomes). Villages with too many hunters, however, will flounder as a result of few and poor boats, leading to fewer offspring for repopulating and export. In contrast, villages with healthier mixes of occupations will export more new hunters and builders. In general models of such scenarios (Nonacs & Kapheim Reference Nonacs and Kapheim2007; Reference Nonacs and Kapheim2008), population trait diversity can be maintained if the mean fitness of traits disadvantaged within groups exceeds the fitness of advantaged traits when they are in less diverse groups (i.e., builders in diverse villages do better than hunters in less diverse villages, although within any given village, hunters always do better). Therefore, between-group variance across village cultural genomes would maintain skill set diversity in populations without requiring either kin-biased nepotism or reciprocated cooperation.
Smaldino proposes novel mechanisms are needed to explain human group-level traits. However, the three proximate mechanisms Smaldino proposes for the emergence of group-level traits are encompassed in the emergence and maintenance of a social genome through social heterosis. For the HIV example, leadership arises because epistasis across loci appears to evolve in a predictable order, division of labor arises when complementary alleles give rise to new function, and repeated assembly occurs via cell death and infection. Nonadditive benefits of intergenomic epistasis enhance the probability that group-level properties will be preserved across repeated assemblies. Analogously, collaborative interdependence leads to and maintains emergent properties of cultural groups.
Social heterosis and social genomes are MLS models and fit comfortably in an inclusive fitness framework such as Hamilton's Rule (i.e., traits are selectively favored when benefits provided to relatives exceed costs to the actors, or rb – c > 0). The key to social heterosis is that relatedness, benefit, and cost are not independent variables; instead, benefits provided (b) or costs incurred (c) correlate with relatedness (r) at the group level. For example, Smaldino highlights the role high genetic relatedness plays in social insect evolution. Although the initial evolution of cooperative breeding and specialized worker castes correlate with high relatedness through strict monogamy (Boomsma Reference Boomsma2013), subsequent evolution has often led to low relatedness, as polygamy has independently appeared in more than 20 different taxonomic lineages (Hughes et al. Reference Hughes, Oldroyd, Beekman and Ratnieks2008). Genetic diversity and “who” resides in colonies significantly affects colony-level fitness (Wray et al. Reference Wray, Mattila and Seeley2011), such that higher b and lower c associate with lower r. Close relatedness is disadvantageous when group diversity benefits are positive and nonadditive. Indeed, diversity-producing social genomes of low relatedness may be the hallmark of behavioral complexity and ecological success throughout the social Hymenoptera (Nonacs Reference Nonacs2011a; Reference Nonacs2011b).
In conclusion, social heterosis and social genomes models can predict the evolution of genetic traits, and we propose they similarly apply to cultural ones. It does not matter if “kayak-building” is cultural rather than genetic because the trait is still transmitted with fidelity. Unlike genes, culture transmits both vertically and horizontally across individuals and can be malleable within individuals over their lifetimes. Hence, cultural traits increase in frequency through differential net rates of phenotypic conversion within populations. Just as the logic of natural selection applies to how gene frequencies change across generations, the same logic can apply to the spread and evolution of human culture, behavior, and practices. Our rejoinder to Smaldino's view that, “models are needed that capture the difference between the social spreading of a particular individual-level trait and the emergence of group-level behaviors that rely on differentiation and organization” (sect. 8, para. 2), is that they already exist – we just need to use them.