The symbolic behaviors collectively referred to as “art and ritual” (e.g., storytelling, visual art, song, dance, ceremony) emerged tens of thousands of years ago, when all humans lived in hunter-gatherer economies. Thus, hypotheses regarding the origin, design, and/or function (adaptive or otherwise) of these behaviors must comprehend the recurrent conditions of forager life that motivated and constrained their form and content. One important motivator was the need for information: making a living by hunting, fishing, and/or gathering entails the acquisition and application of large sets of contingent, local ecological knowledge (Tooby & DeVore, 1987; Turner, Ignace, & Ignace, 2000). These data sets cannot be acquired solely through personal experience by the time individuals must begin fending for themselves in early adulthood. Social learning mitigates this problem by reducing the time and energy costs of information gathering, avoiding the dangers inherent in experiential learning, and bypassing the need to wait around for natural learning opportunities to occur (Boyd, Richerson, & Henrich, 2011; Tooby & Cosmides, 2001:21). This in turn increases the rate of knowledge acquisition and, thereby, the amount of knowledge that can be accumulated. However, accelerated acquisition rates do not ensure successful replication of local knowledge bases over time, which presents a major obstacle to the emergence of cumulative culture: to be applied effectively, information must be input, retrieved, and transmitted accurately (Tomasello, Kruger, & Ratner, 1993). In oral cultures, these processes are constrained by the limits of human attentional and memory systems, making communal knowledge bases vulnerable to data corruption and loss. Thus, cumulative culture requires the development of data management systems that support high-fidelity encoding, recall, and communication of local knowledge. Art and ritual solve these problems by providing mnemonic frameworks for encoding and recalling knowledge, rules for faithfully copying it, and regular occasions for refreshing and transmitting it. This essay focuses on one of these systems: storytelling.

The proposition that storytelling emerged as a data management system rests on the assumption that stories communicate generalizable knowledge, or information that can be applied beyond the immediate context. In this light, storytelling can also be seen as a form of teaching. This claim may seem implausible given that forager oral traditions commonly include information that, to Western eyes, appears to be fictitious. A Pumé tale about the origin of toads and fire is a case in point:

Kumañí created the toad, who is not a common toad but Kiberome. “This is how I will create you,” Kumañí said, making the toad jump so she could go and steal fire. “Go and steal some embers,” Kumañí said. Toad replied, “All right, let me go.” She went running to where the … [Pumé] were warming themselves by the fire. Quickly she swallowed the burning embers and left. Then she vomited and told them: “Here you have fire.” (Wilbert & Simoneau, 1990a:36).

Forager myths and legends include anthropomorphic beings, shape-shifters, and other supernatural phenomena that appear to belie the claim that narrative is used to store and transmit practical knowledge. At best, this claim begs the question of how audiences sift fact from fiction during storytelling events.

Five key concepts are critical to understanding storytelling as a data management system. The first is teaching, which occurs when an expert (i.e., an individual in possession of generalizable knowledge) modifies its behavior in the presence of a novice (i.e., an individual lacking the knowledge in question), such that the expert incurs a cost (or receives no immediate benefit), and the novice benefits by acquiring skills or knowledge it would not otherwise acquire or would not acquire as efficiently (Caro & Hauser, 1992). The second is narrative, a representational format used to organize information about temporally and causally linked events and goal-directed actions. In literary study, narrative is understood as the structure of a story, or the elements that cause an utterance to be perceived as a story. A story, in turn, is an oral or written utterance that has a narrative structure. Indeed, this structure is so integral to story-ness that narrative is commonly used as a synonym for story. Storytelling is the sharing of stories with conspecifics. Thus, storytelling is a form of cooperation involving information exchange. Although often characterized by Western science as religious narrative, myth is more accurately characterized as “narrative . . . regarded as important and significant even when it is not directly religious or actively linked with religious ritual” (Berndt & Berndt, 1994:1). From an emic perspective, myths are stories that are regarded as sacred because of their cultural significance, and are actively curated by the community.

Storytelling meets the criteria for teaching outlined by Caro and Hauser (1992). In this case, the “expert” is the storyteller, whose “expertise” consists of a corpus of tales and the knowledge encoded within them. The “modification of behavior” is the act of telling a story, and the “novices” are audience members who are unfamiliar or only partly familiar with the story. The cost incurred by the expert is the time and energy it takes to tell the story. The novice benefits from the expert's behavior by acquiring the knowledge encoded in the story and the mnemonic framework for retaining it. This argument rests on the assumption that stories encode practical knowledge, returning us to the question of how and what stories teach.

The answer lies in narrative, the framework on which stories are built. Significantly, cognitive psychologists and literary scholars agree on the key components of this structure: character, actions and events, setting, sequence, and conflict and resolution (Scalise Sugiyama, 2005). Collectively, these components represent goal-directed action. Characters–which are almost always human or anthropomorphic–represent agents. Actions represent goal-directed behaviors performed by agents, and events represent non-goal directed happenings (e.g., thunderstorms, tsunamis) that affect character actions. Setting represents a set of environmental constraints and affordances that delimit the kinds of actions and events that can occur, while sequence represents the temporal and causal relationships between actions and events. Finally, conflict represents obstacles (i.e., problems) that impede goal-directed action, and resolution represents strategies deployed to surmount obstacles. As this description illustrates, the purview of narrative is agency. Thus, narrative is almost certainly an adaptation that emerged as a component of the agency-detection system (Scalise Sugiyama, 2017a). Humans co-opt this framework to recreate actual experiences (e.g., travelogues, histories) and invent novel ones (e.g., legends, plans), which are experienced as stories.

In sum, stories—both fiction and non-fiction–are simulations of experience, or what Tooby and Cosmides (2001) describe as invented learning opportunities that simulate relevant features of naturally-occurring learning opportunities (also known as developmental niche construction; see, e.g., Sterelny, 2011). As Pinker writes, characters in story worlds “do exactly what our intelligence allows us to do in the real world. We watch what happens to them and mentally take notes on the outcomes of the strategies and tactics they use in pursuing their goals” (Pinker, 1997:541). Thus, when storytellers share narratives, they are providing the audience with vicarious experience (e.g., Mar & Oatley, 2008; Scalise Sugiyama, 2001, Scalise Sugiyama, 2005).

The fitness benefit of engaging in virtual experiences is summed up by the old adage, “Forewarned is forearmed.” Predator evasion is a case in point: learning about predator behavior and habitat prior to actual encounters can mean the difference between life and death. In the Amazon, for example, “even though a 4-year-old Shuar child might not be able to fend off an attacking anaconda in a direct encounter, knowledge that anacondas tend to hunt near riverbanks could have major fitness benefits for the child” (Barrett, 2005:442). A Yąnomamö tale demonstrates that stories can and do simulate this scenario:

“Older brother … don't go over there! An anaconda appears there regularly!” In vain he warned his brother; the latter insisted on going … fishing. His younger brother repeated: “Don't go; you'll be killed at the water's edge”… The man left in spite of his brother's warnings. When he reached the riverbank he … caught several fish as wide as a hand ...Meanwhile, an enormous anaconda stood erect in the river, supporting itself on its backbone ... In the place where you were fishing in order to satisfy your desire to eat fish, Kuyere, you were killed in the water. (Wilbert & Simoneau, 1990b:448).

This story encodes information that is instrumental to avoiding Kuyere's fate: it identifies an animal known as an anaconda whose habitat includes riverbanks, notes that this animal attacks and kills humans, and documents a kill. In so doing, it also encodes a warning: these areas should be avoided if possible, or approached with caution.

This story also graphically illustrates the importance of high-fidelity transmission: if a story isn't copied accurately, the critical information it encodes might be lost. The problem is, successful replication is easier said than done. In oral cultures, social learning is sharply constrained by the limits of human attentional, perceptual, and memory systems. Also, there are multiple points of vulnerability across the learning process: information can be corrupted during encoding, retrieval, and/or transmission. These constraints present a formidable obstacle to the emergence of cumulative culture. However, developmental niche construction plays a role here as well: humans can use their innovative and instrumental intelligence to invent strategies and techniques that increase the likelihood of faithful copying, and then build these into artificial learning environments.

This essay makes the case that storytelling emerged in ancestral human populations as a solution to the problems of preserving and transmitting communal knowledge bases from generation to generation. To this end, it reviews evidence from forager and other small-scale oral cultures that (a) storytelling is a form of teaching that encodes local ecological knowledge, (b) multiple strategies are used to maintain fidelity from transmission event to transmission event, and (c) fictional stories are marked in ways that enable listeners to separate fact from fantasy.