A key adaptive challenge humans face is the ability to make decisions under uncertainty, where learning is at best incomplete because of inherent complexity relative to information-processing capacities, an insufficient time span for learning, or nonstationarity of variability (Turner, Smaldino, Moya, & Jones, Reference Turner, Smaldino, Moya and Jones2022). The general principle for the management of uncertainty is robustness (Kay & King, Reference Kay and King2020): Solutions cannot be optimized for specific conditions, but need to be good enough for a wide range of conditions. The time frame for the emergence of the genus Homo coincided with a particularly variable period in Earth's history in a region with remarkably high-frequency variability (Antón, Potts, & Aiello, Reference Antón, Potts and Aiello2014). Moreover, this variability was nonstationary (Levin, Reference Levin2015). Learning fully about environmental conditions was largely not possible because of this nonstationary, high-amplitude, fine-grain variability. This suggests that adaptive decision-making under uncertainty is a fundamental aspect of being human. The development of culture is clearly an essential feature of the success of the genus Homo, and H. sapiens in particular. Johnson et al.'s identification of narrative as a tool for dealing with radical uncertainty is a major conceptual contribution and has big implications for debates in the field of cultural evolution.

As it is shared knowledge, culture requires transmission. What distinguishes culture as an adaptive strategy is that it is shared and its advantages stem from relaxing the need to learn decisions anew across individuals and through time. What Sterelny (Reference Sterelny2017) calls the “California” school of cultural evolution emphasizes high-fidelity transmission of cultural elements effectively through copying with errors (e.g., Boyd & Richerson, Reference Boyd and Richerson1985; Cavalli-Sforza & Feldman, Reference Cavalli-Sforza and Feldman1981).

However, the organ that permits the development of culture, the brain, is not a passive copying machine; rather it is an information processor. In a world of infinite distractibility, elements of cultural knowledge must be attended to and meanings must be ascribed. As a result, it is likely that people transform elements of culture as they learn, process, and transmit them. This is the fundamental argument of Sperber (Reference Sperber1996) and subsequent work in the tradition of cultural attractor theory, what Sterelny (Reference Sterelny2017) calls the “Parisian” school. The focus of the Parisians is on cultural attractors, which attempt to account for the fidelity and persistence of cultural ideas without the imitation of experts by novices that the California school presumes. While the idea that learners actively transform cultural information is appealing, the Parisian approach remains weak on formal mechanisms of transmission and is missing the strong modeling tradition of the California school (Sterelny, Reference Sterelny2001, Reference Sterelny2017). Without this, cultural attractor models lack explanatory power. Sterelny (Reference Sterelny2001, p. 848) suggests that the Parisian models “represent the consequences of transmission biases but they explain nothing about the sources of those biases.” Narrative provides this.

While ideas may not replicate, we suggest that stories do. This makes narrative a very important type of cultural attractor. Narrative transforms ideas and generates emergent qualities of meaning, causal explanation, etc., but narratives are also uniquely transmittable. Stories are highly transmissible, even in the presence of substantial noise or long transmission chains (Mesoudi, Whiten, & Dunbar, Reference Mesoudi, Whiten and Dunbar2006). By facilitating transmission, narrative also functions as a means to distribute cognition beyond the single individual. Narrative improves recall and interest in information, aiding that crucial limiting resource, attention (Glaser, Garsoffky, & Schwan, Reference Glaser, Garsoffky and Schwan2009). By facilitating attention, retention, and transmission, narrative helps accomplish the crucial elements of culture, namely, extrasomatic information storage and processing.

Narrative includes a number of features that make it useful for encoding and transmitting information under conditions of uncertainty. Narrative provides (1) a mechanism for transmitting complex information that goes well beyond imitation, (2) causal structure, thereby facilitating exploration, and (3) coherence, which ensures robustness. Currie and Sterelny (Reference Currie and Sterelny2017, p. 19) note, “As our information about the causal background is enriched, coherence becomes an increasingly important, increasingly demanding constraint.” Checks on coherence ensure a degree of fidelity in transmission and provide the foundation for selection to work.

As Johnson et al. note, narratives are causal explanations. Their causal structure, representable as a directed graph (as in their Fig. 3), means that the elements of narratives are relational. Through their relationship, as indicated by an arc in the directed graph, two elements cease to be independent of each other. Their transmission, and selection on them, will be correlated. Correlation greatly increases the complexity of both population dynamics (e.g., Tuljapurkar, Reference Tuljapurkar1990) and the response to selection (e.g., Lande, Reference Lande1979). Very little work in cultural evolution has addressed the complications of indirect selection, but some initial forays have been made (Yeh, Fogarty, & Kandler, Reference Yeh, Fogarty and Kandler2019).

Narratives, as transmissible cultural attractors, complicate selection on cultural elements. In addition to the relational nature of the elements of a narrative, highly compelling narratives attract associated concepts through semantic association, metaphor, and connotation (as noted in sect. 5.1.3). If we want to understand the dynamics of selection on particular cultural traits, we need to consider them as multivariate traits because of the emergence of covariance between elements both within the narrative and between the narrative itself and associated narratives. This covariance means that cultural traits will respond to both direct selection, as with standard cultural-evolutionary models, but also to indirect selection on covarying cultural traits. Figure 1 provides a demonstration of two simple narratives that provide a causal structure for two outcomes (Z and Z’), with common element across the two narratives of Y. Not only will the elements within a narrative be correlated from the perspective of selection acting on their frequency, the two narratives themselves will be correlated.

Figure 1

Two simple narratives, represented as directed graphs, that provide a causal structure for two outcomes (Z and Z') with a shared element, Y.

This has many implications. One important implication, particularly for the early selective advantage of narrative transmission, is that we expect narrative structures to be more robust to mistakes about specific elements. Evolutionary robustness is the persistence of an organismal trait under perturbations (Wagner, Reference Wagner2013). Of course, narratives as cultural attractors can shape meaning in potentially harmful ways as well. A high-coherence narrative can capture otherwise orthogonal information or can be resistant to updating as contradictory evidence accumulates.