Cultural evolution – The mystery of production
Alberto Acerbi’s book is not just an impressive and timely summary of our knowledge of cultural transmission in the digital world (that in itself would be a good reason to read it). It also replaces issues of digital transmission in the broader perspective of theories of cultural evolution. In the process it provides us with an excellent entry point into the field. Students and interested outsiders could do much worse than following this guide – perhaps to be used in combination with How traditions live and die (Morin, 2016) and the more recent Not born yesterday (Mercier, 2020). Acerbi’s book is also a pleasure to read, which adds to the many reasons to acquire this volume (should I also mention it is reasonably priced?)
One central theme of the book, and one reason why it will or at least should be of great interest outside our field, is that it tackles the many myths and clichés that plague any discussion of digital information, and of social media in particular, described as a frightful post-truth world of echo chambers, fake news, paranoia and junk culture. Acerbi proposes a much more sober appreciation of this new world of digital culture, one that is based on empirical studies and plausible psychology rather than lurid fantasies and scaremongering.
Instead of singing a pean to the many qualities of the book, I should focus on two points that may be crucial to developing the study of cultural diffusion. One is the fairly familiar point that cultural evolution without a rich psychology is not a great proposition. The other point is that models of cultural transmission are overly focused on consumption instead of production.
The idea of modeling cultural evolution properly started with Culture and the evolutionary process (Boyd & Richerson, 1985) a very impressive exploration of the many ways in which importing tools from population genetics could clarify issues of “culture”, that is, population level patterns of information transmission. The book provided a fairly exhaustive survey of the various ways in which the selection of information could affect cultural patterns, including though drift, guided transmission, direct bias, etc. This was foundational and, as is required with many foundational work of this kind, Boyd and Richerson had to start with simplifying assumptions, idealizations that would make modeling possible. One of these idealizations was the assumption of domain-generality, that is, the notion that different patterns of ‘culture’ emerge because of combination of parameters that would apply to different domains of culture in the same way. Obviously, one could at a later stage complete these models with specific content biases (Boyd & Richerson, 2005).
Over the following decades, there occurred a bifurcation in the field, whereby some people kept working on the assumption of domain generality, but applied them to fairly limited domains of human cultures, while others took on board the diversity of domain-specific psychological systems that psychologists were uncovering, but without proposing much by way of formal models of their domains (Acerbi & Mesoudi, 2015). It is only recently that cultural evolution has started to include the rich domain-specific psychology in formal approaches to diffusion, see, e.g., Kirby, Griffiths, & Smith (2014); Morin (2013, 2018) and many others, including of course Acerbi himself (Acerbi, Kendal, & Tehrani, 2017; Acerbi & Parisi, 2006; Ruck, Bentley, Acerbi, Garnett, & Hruschka, 2017).
That is an important development because, for a long time, the assumption of domain-generality hampered our understanding of cultural patterns in many domains. As Acerbi points out: “Cinderella is still with us, not because of a general-purpose mechanism supporting high fidelity transmission but because the reconstruction of stories is not random […we observe,] not random errors […] but biased transformations, an essential and constructive part of cultural transmission” (Acerbi, 2019, p. 162)
The same goes for such domain-general features as prestige, which, against the domain-general assumption, works… only sometimes. Highly specific assumptions about persons are more important to cultural diffusion than the domain-general notion that the sources of information are “prestigious”.
Another assumption in most models of cultural evolution is that cultural patterns are, first and foremost, to be explained in terms of consumption. Alberto Acerbi writes, “cultural evolutionists think of social learners as information scroungers, as opposed to individual learners, who are information producers” (p. 12). That is true of both domain-general perspectives like dual-inheritance theories, but also to some degree of the alternative, epidemiological or “cultural attractor” models.
But we should indeed consider production, as Acerbi mentions at the end of his Précis. I do not see the lack of production models as a flaw in the book – the state of reflection and empirical findings in this domain is much too uncertain, so that speculative considerations about the motivation to produce culture would not have added much to his argument.
But such considerations are allowed in a blog post… so let us dive in. Acerbi remarks that “people everywhere seem to be happy to produce content […] for apparently no gain” (p. 12), a point most clearly illustrated by Wikipedia contributors. The same point applies to customer reviews on shopping sites and to many other domains of digital communication.
Why do people bother? Direct material rewards are certainly not the prime motivation. Is reputation the main motivation? That is not as clear as we would wish, as many Wikipedia contributors, for instance, remain anonymous. Acerbi formulates the more plausible hypothesis, that a motivation to contribute may be the outcome of an evolutionary mismatch between the conditions under which we evolved as communicators, and those of the modern world (long before digital media in fact) (p. 15).
Now, I would say that Acerbi’s explanation is probably on the right track, but also probably insufficient. (Which is no criticism, since that was not his aim in the first place). In all human communities people constantly produce information for others. It may be of varying epistemic quality. Telling people of your experience with this or that vegetable may be useful, while explaining people’s sickness as the result of witchcraft does not actually contribute to people’s welfare. But in all these cases, what would be the motivation?
The reputation + mismatch account, although possibly true, remains at the level of proximate mechanisms. We assume that the mismatch occurs because people mistakenly have the intuition that online contributions are recognized by receivers, and they want to be seen as contributors to knowledge. But why do they have that motivation?
We should consider this in evolutionary terms . If there is a motivation to broadcast information, what fitness benefits does it bring? Using Hamilton’s classification of initiator-receiver dynamics (Hamilton, 1964), we could imagine that producers of information may be selfish (enhancing their fitness at the expense of others), altruistic (enhancing others’ fitness more than their own), or mutualistic (for both parties’ benefit). In some cases the answer is clear. The cult leader who persuades his followers to provide him with sex and money is exploiting them. Or, the parents who transmit knowledge to their offspring are probably increasing their own inclusive fitness. But these are rather exceptional cases. What happens much more often is that people just broadcast all manners of ideas they entertain about the world, for the benefit of anyone who cares to listen.
This process may constitute an instance of mutualistic cooperation. Consider this. In a species that depends so much on information from others, having reliable sources of information is certainly a fitness factor, and the motivation to seek out such sources and pay attention to them may be enhanced by natural selection. Indeed, that capacity to identify good sources is an important component of our epistemic vigilance, especially in domains for which we have little direct access to information (Bonaccio & Dalal, 2006; Mercier & Sperber, 2011).
If humans readily identified good sources and tried to keep such individuals in their social environment, this would result in a dynamic whereby it is fitness-enhancing not just to have good sources in one’s environment but also to be a good source, as other people value you, a process somewhat similar to building friendship (Tooby & Cosmides, 1996).
Another possibly relevant dynamic is that of prestige in the selection of sources, a variation on the kind of dynamic described in dual inheritance models (Boyd & Richerson, 1985, pp. 241-280). In some domains of high uncertainty, one criterion that a source is of high quality may be that others seem to take that source seriously, and act on that person’s advice. This could create information cascades and result in bandwagon effects, whereby people assume some individual is a good source because others follow his or her advice, and so forth.
Going further into speculation, the motivation to be seen as a good source may incentivize the production of junk culture. Sadly, junk culture does exist. True, moral panics about fake news and post-truth echo chambers are based on wild exaggerations, as Acerbi nicely demonstrates throughout his book. Still, there is a large production of information that is of low epistemic value, such as rumors and conspiracy theories. That is the case, partly, because of consumer preferences (p. 200). But what are the producers’ motivations?
One possibility is that, while the detection of good sources is roughly accurate (otherwise our epistemic vigilance mechanisms would not be adaptive), it may not be perfect. There may be domains in which it makes sense to accept information of not-so-great epistemic value. One such context would be information about potential threats. We often operate on error-management principles, as we tend to err on the side of caution when the cost of misses is greater than that of false alarms (Haselton & Buss, 2000). So it would make sense to pay attention to a source that warns us of potential danger – Acerbi does mention empirical studies that seem to support that hypothesis. The “good source” dynamics may enhance the reputation of a Cassandra, even though her warnings may be (at least in part) false alarms – and mostly because people who follow the recommended precautions will never find out if the threat was real. That would explain why people who think they have found out about potential threats are so eager to tell us, and so eager to monitor our acceptance of that information, as they may have the intuition that providing such information will make them good sources in our estimation.
These informal remarks suggest how Alberto Acerbi’s book is so valuable. It does not just provide the best account so far of the field of cultural evolution (in digital but also in traditional media) but also makes it possible to envisage new perspectives for the field – some still in the firmly speculative range. So, again, read it.
Acerbi, A. (2019). Cultural Evolution in the Digital Age. Oxford, UK: Oxford University Press.
Acerbi, A., Kendal, J., & Tehrani, J. J. (2017). Cultural complexity and demography: The case of folktales. Evolution and Human Behavior, 38(4), 474-480.
Acerbi, A., & Mesoudi, A. (2015). If we are all cultural Darwinians what’s the fuss about? Clarifying recent disagreements in the field of cultural evolution. Biology & philosophy, 30(4), 481-503.
Acerbi, A., & Parisi, D. (2006). Cultural transmission between and within generations. Journal of Artificial Societies and Social Simulation, 9(1).
Bonaccio, S., & Dalal, R. S. (2006). Advice taking and decision-making: An integrative literature review, and implications for the organizational sciences. Organizational Behavior and Human Decision Processes, 101, 127-151.
Boyd, R., & Richerson, P. J. (1985). Culture and the evolutionary process. Chicago: University of Chicago Press.
Boyd, R., & Richerson, P. J. (2005). The origin and evolution of cultures. Oxford ; New York: Oxford University Press.
Hamilton, W. D. (1964). The genetical evolution of social behaviour I and II. Journal of theoretical biology, 7, 1-16 and 17-52.
Haselton, M. G., & Buss, D. M. (2000). Error management theory: A new perspective on biases in cross-sex mind reading. Journal of Personality and Social Psychology, 78, 81-91.
Kirby, S., Griffiths, T., & Smith, K. (2014). Iterated learning and the evolution of language. Current opinion in neurobiology, 28, 108-114.
Mercier, H. (2020). Not Born Yesterday: The Science of Who We Trust and What We Believe. Princeton, NJ: Princeton University Press.
Mercier, H., & Sperber, D. (2011). Why do humans reason? Arguments for an argumentative theory. Behavioral and Brain Sciences, 34, 57-74. doi:10.1017/s0140525x10000968
Morin, O. (2013). How portraits turned their eyes upon us: visual preferences and demographic change in cultural evolution. Evolution and Human Behavior, 34, 222-229.
Morin, O. (2016). How traditions live and die. Oxford: Oxford University Press.
Morin, O. (2018). Spontaneous emergence of legibility in writing systems: the case of orientation anisotropy. Cognitive Science, 42(2), 664-677.
Tooby, J., & Cosmides, L. (1996). Friendship and the banker’s paradox: Other pathways to the evolution of adaptations for altruism. In W. G. Runciman, J. M. Smith, & et al. (Eds.), Evolution of social behaviour patterns in primates and man. (pp. 119-143). Oxford, England UK: Oxford University Press.
 Most of the ideas sketched here were developed in conversations with Nicolas Baumard and Jean-Baptiste André, who of course are only partly to blame for any errors and ambiguities in this formulation.