BLIND AND INCREMENTAL OR DIRECTED AND DISRUPTIVE? ON THE NATURE OF NOVEL VARIATION IN HUMAN CULTURAL EVOLUTION

Many scholars have rejected cultural evolutionary theory on the grounds that cultural variation is directed and intentionally created, rather than incremental and blind with respect to function, as is the case for novel genetic variation in genetic evolution. Meanwhile, some cultural evolution researchers insist that cultural variation is blind and undirected, and the only directional force is selection of randomlygenerated variants. Here I argue that neither of these positions are tenable. Cultural variation is directed in various ways. While this does not invalidate cultural evolution, more attention should be paid to the different sources of nonrandomness in culturally evolving systems. Introduction Does human culture evolve? Can we draw useful parallels between genetic evolution and cultural change? Can we use similar tools, methods, concepts and theories to understand cultural change as biologists use to understand genetic evolution? The rapid increase in research adopting a cultural evolution framework in the last few decades (Youngblood and Lahti 2018) shows that increasing numbers of scholars are answering “yes” to these questions. Yet in the grand scheme of academic research, cultural evolution remains a fringe pursuit. The majority of the social sciences and humanities reject any kind of evolutionary theorizing for understanding cultural phenomena, including cultural evolution (Slingerland and Collard 2011; Perry and Mace 2010). Even within the evolutionary human behavioral sciences, the theory of cultural evolution is often treated with skepticism (Daly 1982; Atran 2001; Pinker 1997). While there are many reasons for this rejection and skepticism, one common point of contention surrounds the issue of randomness and directionality in the generation of novel cultural variation. This is also a common source of disagreement amongst cultural evolution researchers (Lewens 2015; Claidière et al. 2014). In this paper I argue that there are genuine differences between genetic and cultural evolution in this sense. While some novel cultural variation seems to be generated randomly with respect to selection, akin to the generation of novel genetic variation, there are also several forms of directed cultural variation. I argue that while the latter do not invalidate the theory, approach or methods of cultural evolution, This content downloaded from 92.72.139.75 on Tue, 26 Jan 2021 16:19:44 UTC All use subject to https://about.jstor.org/terms 8 / AMERICAN PHILOSOPHICAL QUARTERLY more attention should be paid to the different sources of directedness in culturally evolving systems and their consequences. What is Cultural Evolution? Cultural evolution is the idea that cultural change exhibits fundamental similarities to genetic evolution, such that we can pro-tably view and analyze cultural change as an evolutionary process. “Culture” here is de-ned broadly as any socially transmitted information that passes from person to person nongenetically, via imitation, language, teaching or other means of social learning. This includes what we colloquially label knowledge, beliefs, attitudes, norms, customs, skills, words, grammar, and institutions. If evolution is de-ned in abstract terms, as Darwin did (Darwin 1859), as comprising a system of variation, inheritance, and differential -tness (Lewontin 1970), then cultural change appears to meet the criteria of being an evolutionary process. Cultural traits (beliefs, ideas, attitudes, etc.) vary within a population; they are (by de-nition) inherited from person to person, not genetically but socially, via social learning; and not all traits are equally likely to persist over time, with some ideas or beliefs spreading at the expense of others (Mesoudi et al. 2004). This parallel was suggested by Darwin himself who, in The Descent of Man (Darwin 1871), pointed out the similarities between the evolution of species and of human languages. Despite initial interest in this parallel within the nascent social sciences of the late nineteenth century (Hodgson 2005), the idea fell from favor as the social and natural sciences diverged during the early twentieth century. A resurgence began in the 1970s with memetics (Dawkins 1976), evolutionary epistemology (Campbell 1974), and the mathematical modeling of culture using the methods of population genetics (CavalliSforza and Feldman 1981; Boyd and Richerson 1985). The latter, in particular, stimulated an ongoing empirical program that includes lab experiments, archaeological analyses, linguistics, history and nonhuman comparative work (Mesoudi 2011; 2017). As in many -elds, particularly relatively new ones, there are different approaches to cultural evolution that hold different assumptions. These include memetics (Blackmore 1999; Dennett 2017), cultural attraction or cultural epidemiology (Claidière et al. 2014; Sperber 1996), cultural phylogenetics (Gray and Watts 2017), and work derived from the aforementioned population genetic modeling (Boyd and Richerson 1985; CavalliSforza and Feldman 1981). One dimension on which they vary is the degree to which individuals are seen to be able to direct cultural dynamics by generating or transforming cultural traits nonrandomly, in particular directions or towards particular forms. This is also a reason others use to reject the entire cultural evolution approach, as explored in the next section. Randomness and Directionality in the Generation of Genetic and Cultural Variation It is a fundamental axiom of genetic evolutionary theory that novel genetic variants arise randomly with respect to function. Bene-cial mutations are no more likely to arise when they are needed (i.e., no more likely to subsequently confer -tness bene-ts to their bearer) than when they are not needed (Luria and Delbruck 1943). The primary directional component of genetic evolution is selection, which increases the frequency of variants that happen to confer a -tness advantage on their bearers (i.e., makes them more likely to survive and reproduce). The generation of variation, via mutation or recombination, is random, not directional. “Random” here does not mean that all genetic mutations are equally likely to occur, given the existence of developmental constraints that make some variants more likely to occur than others, and This content downloaded from 92.72.139.75 on Tue, 26 Jan 2021 16:19:44 UTC All use subject to https://about.jstor.org/terms BLIND OR DIRECTED? / 9 variation in mutation rates across loci. Rather, it means that the chance of a particular mutation occurring does not depend on whether or not that mutation is evolutionarily advantageous to its bearer. While there is some evidence for “adaptive mutation,” where the mutation rate increases during times of stress when bene!cial mutations are most needed (Foster 2004; Rosenberg 2001), this does not change the basic point that genetic mutation is undirected. Adaptive mutation, if it exists, increases the rate of random mutation, such that bene!cial and nonbene!cial mutations are both increasingly likely to occur, with no bias towards bene!cial mutations. Similarly, “facilitated variation” (Gerhart and Kirschner 2007) highlights the major constraints that conserved components of development and physiology place on the kind of phenotypic variation that genes can generate. While this may make phenotypic variation more likely to be adaptive, this is due to past selection of random genetic variation. The generation of novel cultural variation, however, seems far more directed and intentional. People strive to solve speci!c problems, invent useful or pro!table products, instigate political and social change that they believe will advance their chosen cause, and so on. This difference has frequently been used to argue against the claim that culture evolves. Pinker, for example, writes “Memes such as the theory of relativity are not the cumulative product of millions of random (undirected) mutations of some original idea, but each brain in the chain of production added huge dollops of value to the product in a nonrandom way” (Pinker, cited in Dennett 1995, p. 355). Similarly, Orr criticizes cultural evolution on the grounds that “new ideas—but not genes—are produced by a sort of directed mutation. Newton did not uncover the Fundamental Theorem of Calculus by conceiving millions of random ideas” (Orr 1996, p. 470). Hallpike writes ‘‘there is no signi!cant resemblance between the mutation, the basic source of variation in the Darwinian scheme of things, and social invention, which is purposeful, responsive, and can be diffused. Whereas biological variation can be treated as random, social variation is the product of particular societies and cultural traditions, and therefore far from random’’ (Hallpike 1986, p. 36, italics in original). Fracchia and Lewontin state that “[cultural v]ariations emerge not randomly, but as attempts by speci!c individuals and/or groups to solve speci!c social/cultural problems; and their origins are not unrelated to their fate” (Fracchia and Lewontin 2005, p. 21). Or this from Sternberg, in an entire article criticizing the blindvariation assumption of cultural evolution: “The blindvariation model is inadequate in accounting for all of human creativity, if only because the research on expertise . . . shows that humans do not blindly vary hundreds or even thousands of candidates for propagation—or mutations— for every one that succeeds” (Sternberg 1998, p. 171). This collection of quotes from psychologists, biologists, historians and anthropologists reveals the breadth of the criticism that culture cannot be said to evolve because the generation of cultural variation is not random or blind with respect to function. Is Cultural Variation Randomly Generated? While some approaches to cultural evolution explicitly incorporate nonrandom, or “guided” variation (see next section), some strongly imply that cultural variation is random with respect to se