Occupation of the foraging niche requires considerable stamina, strength, coordination, skill, and – significantly – knowledge, which require a long time to acquire (Kaplan et al., 2000, 2007; Gurven et al., 2006). This costly period of development is supported by intensive parental (Lancaster and Lancaster, 1983) and alloparental (Hrdy, 2005, 2009) investment. Although human life history theory commonly characterizes this investment in terms of food resources and care (Lancaster and Lancaster, 1983; Kaplan et al., 2000, 2007), ethnographic research on the transmission of foraging skills suggests that the flow of resources from old-to-young includes information as well. Although certain skills are acquired and/or refined through play (Bruner et al., 1976; Boulton and Smith, 1992; Steen and Owens, 2001; Bock, 2002; Pellegrini and Bjorkland, 2004; Göncü and Gaskins, 2007; Pellegrini, 2011), ethnographic evidence suggests that children acquire much of their knowledge from their parents (Hewlett and Cavalli-Sforza, 1986; Ohmagari and Berkes, 1997; Shennan and Steele, 1999). This research is paralleled by extensive evidence that cognitive systems dedicated to social learning begin emerging in early infancy (for a review, see Carey, 2009) and that social learning is more highly developed in humans than in any other species (Byrne, 1995; Tomasello et al., 2005). Although an incipient form of this strategy – obtaining access to food resources by exploiting the expertise of conspecifics – is seen in some non-human primates (Stammbach, 1988; Russon, 1997), humans are unmatched in their highly elaborated ability to access the expertise of conspecifics in order to predict and manipulate both the physical and social environment.

Intergenerational information transfer raises the question: in what format(s) is this information transmitted? Humans are distinguished by two highly developed social learning complexes: transmission of physical skills (e.g., knapping stone, making fire) via demonstration and imitation, and transmission of factual knowledge via language (Gibson, 1999). Our highly elaborated ability to share knowledge gives us access to a much greater range of resources than our ape cousins, and enables us to surmount hazardous terrain and vast distances in pursuit of them (Gibson, 1999). Information sharing also “reduces uncertainty and increases efficiency by reducing search costs in hunting” (Hames, 1990, p. 93; see also Heffley, 1981). Specifically, information sharing increases foragers’ “sample size of resource distributions and characteristics” and provides “information on prey and patches that the foragers have never exploited” by allowing access to information that “covers long time spans and the experiences of hundreds of individuals” as well as “[e]xtremely rare events unlikely ever to be experienced by a single forager” (Kaplan and Hill, 1992, p. 196). One way that foragers share information is by “telling the hunt” (Blurton Jones and Konner, 1976) – that is, by recounting and listening to others recount their hunting experiences. Among the Ache, for example, “each man usually reports to the others in considerable detail concerning every game item that he encountered that day, and the outcome of the encounter” (Kaplan and Hill, 1992, p. 196; see also Leacock, 1954, p. 14; Laughlin, 1968, p. 308; Nelson, 1969, p. 374; Biesele, 1978, p. 940; Ridington, 1988, p. 47). Although hunting stories are ostensibly exchanged among adults, children certainly overhear them. As Kaplan and Hill (1992) note, an important source of foraging information is older individuals, and most Ache “children have detailed knowledge of resource characteristics and capture techniques before ever foraging themselves” (196). Although genetic relationships between the storyteller and audience members are rarely specified, storytelling in small-scale societies appears to be characterized by a vertical transmission pattern, with elders telling stories to an audience of adults and children (Lowie, 1918; Opler, 1940; McIlwraith, 1948; Jacobs and Jacobs, 1959; Biesele, 1976; Tonkinson, 1978; Ridington, 1988; Sobel and Bettles, 2000). Occasionally, a general pattern of story transmission is identified: for example, in Jicarilla culture, “Those who ‘give’ the stories to the children stand usually in a grandparental relationship to them” (Opler, 1938, p. xii). Although these observations are anecdotal, a general pattern of old-to-young transmission in an extended family setting can be discerned, which is in alignment with the proposal that parents and alloparents invest knowledge as well as food and care in offspring.

The costs of acquiring information at first hand can be considerable (Kaplan and Hill, 1992), and investment of information in offspring is but one way of defraying these costs. Information may be shared with collateral kin, thereby enhancing the donor's inclusive fitness (Scalise Sugiyama, 2005; see also Sibly, 1999). Information may be also be bartered in social exchanges: recipients might be motivated to reciprocate in the future, saving the donor future information acquisition costs (Kaplan and Hill, 1992; Scalise Sugiyama, 2005). Yąnomamö peccary hunting illustrates this point:

Gossip networks are another example of reciprocal information sharing: the price of inclusion is willingness to share gossip with fellow members (Bott, 1971; Capp, 2003). The use of gossip to manage one's own reputation and manipulate the reputations of others (Emler, 2001) illustrates another potential benefit of information sharing: influencing the beliefs and behavior of conspecifics in ways that serve the transmitter's interests (Scalise Sugiyama, 1996). Finally, visiting other groups for the purpose of acquiring information may provide access to potential mates (Steward, 1938; Kaplan and Hill, 1992; MacDonald and Hewlett, 1999). The proposition that parents invest knowledge in offspring is presented not as an alternative but as a complement to these hypotheses. The claim is not that humans only share information with offspring but that, given the value of information and the potentially high cost of acquiring it at first hand, one would expect parental investment to include knowledge as well as food and protection, and that storytelling is an important means of parent-to-offspring knowledge transfer.

This argument assumes that storytelling is an ancient part of H. sapiens life. For the same reason that the finely crafted parietal and portable art of Upper Paleolithic Europe is not evidence of a sudden change in human cognitive capacities (McBrearty and Brooks, 2000; Sterelny, 2011), narrative is unlikely to be a recent development. The earliest hard evidence of non-verbal symbolic behavior (shell beads and engraved ochre artifacts found in Blombos Cave, South Africa) dates to 75 kya (Henshilwood et al., 2002; d'Errico et al., 2005), but several lines of evidence indicate that the capacities requisite to these behaviors were in place approximately 300 kya (McBrearty and Brooks, 2000). Similarly, storytelling is dependent on cognitive capacities – e.g., shared attention, theory of mind, episodic memory – that evolved prior to the H. sapiens diaspora. Consider the inherent verbal nature of story transmission: because non-verbal media are highly inefficient and imprecise narrative devices (Scalise Sugiyama, 2005), storytelling requires language, which almost certainly evolved before our species migrated out of Africa 100 kya (Falk, 1984; Pinker, 1994; Dunbar, 1996; Holloway, 1996; Locke and Bogin, 2006). Thus, it is likely that human foragers living tens of thousands of years ago were as capable of telling stories as modern foragers. The tales of our ancestors are lost to us, but the oral traditions of historically documented foraging peoples offer a window on our storied past. Although these peoples are not facsimiles of ancestral humans, their lifeway is similar in that they depend largely on foraging for their subsistence, live in small natural-fertility populations, and lack telecommunication, motorized transport, and western medicine (Marlowe, 2005). The thematic similarities between modern forager oral traditions indicate that we would see a similar overlap in content between modern and ancient forager oral traditions.

This paper reviews several lines of evidence in support of the hypothesis that parental investment includes knowledge and that, in forager societies, storytelling is an important means of parent-to-offspring knowledge transfer. Part 2 reviews evidence that acquisition of foraging knowledge takes many years and that this knowledge is transmitted primarily from parent to child. Part 3 reviews evidence that the human mind contains adaptations specific to social learning, and discusses the developmental trajectory of these structures in relation to prolonged juvenility. Part 4 reviews cross-cultural patterns in forager folklore content vis-à-vis key information demands of the foraging niche, and Part 5 suggests possible tests of the proposed hypothesis.

Humans are characterized by a long lifespan, large brain relative to body size, and long period of juvenile dependency. The evolution of these traits is the subject of vigorous debate, centered on the question of whether learning is a consequence or a cause of prolonged juvenility (Bock and Sellen, 2002; Hrdy, 2005; Konner, 2010). The former view sees prolonged juvenility as the by-product of selection for larger body size: the benefits of larger body size outweigh the costs of the longer growing period they require (Charnov, 1993; Charnov and Berrigan, 1993). A complementary view, the grandmother hypothesis, posits that the evolution of longevity was driven by the provisioning of descendants by post-reproductive women (Hawkes et al., 1989, 1997, 1998; O'Connell et al., 1999). The latter view, formulated by Kaplan and colleagues as the embodied capital hypothesis, posits that our long lifespan and large brain co-evolved through the exploitation of nutrient-dense resources using complex extraction techniques, the acquisition of which requires a long growth and learning period supported by intensive parental investment (Lancaster and Lancaster, 1983; Bogin, 1997, 1999; Kaplan et al., 2000, 2007). On this view, the costs of investment in growth and development early in the lifespan are offset by the benefits they yield later in the lifespan. In functional terms, embodied capital includes a range of physical and mental attributes, including strength, speed, stamina, immune function, skill, and knowledge. A related argument, the social intelligence hypothesis, posits that encephalization was driven by the demands of predicting and manipulating the behavior of conspecifics, necessitating an extended juvenile period for the development of complex social skills and the cognitive capacities that scaffold them (Byrne and Whiten, 1988; Dunbar, 1992, 1996, 2007; Whiten and Byrne, 1997; Flinn et al., 2005). The chief bone of contention between the latter two hypotheses is whether the selection pressures driving encephalization were social or ecological. However, this argument dissipates in light of the role that sharing and cooperation play in hunter-gatherer life (Reader and Laland, 2002; Kaplan et al., 2007; MacDonald, 2007a,b). Humans are social foragers: many of the resource-extraction and foraging-risk-management strategies deployed by foragers rely heavily on intraspecific cooperation (e.g., Tonkinson, 1978; Wiessner, 1982; Berkes et al., 1994), including social learning (Buckley and Steele, 2002; MacDonald, 2007a,b). Cooperative hunting and/or fishing, for example, are practiced across a wide range of cultures (Downs, 1966; Nelson, 1969; Lee, 1979; Hill and Hawkes, 1983; Yost and Kelley, 1983; Hewlett, 1991; Chagnon, 1997; Sugiyama and Chacon, 2000; Brody, 2002), and successful implementation of such strategies can depend on information sharing (e.g., Heffley, 1981; Hames, 1990). Thus, the exploitation of difficult-to-acquire resources – and concomitant tasks such as social exchange, mate acquisition, and child care – require complex manipulations of both the physical and social environment (Locke and Bogin, 2006).

Although different researchers emphasize different areas of growth, there is widespread agreement that successful occupation of the foraging niche requires a high degree of strength, coordination, skill, and/or knowledge; that these attributes require a significant period of time to acquire; and that this costly period of development is supported by intensive parental – and, some argue, alloparental (Hawkes et al., 1997, 1998; Hrdy, 1999, 2005, 2009) – investment. This investment is often characterized in terms of food resources and care (e.g., Lancaster and Lancaster, 1983; Kaplan et al., 2000, 2007). However, both theory and evidence suggest that information transfer is an important component of parental investment. All else equal, individuals who effectively communicated accurate information to close kin would have increased the representation of their genes in subsequent generations. The potential fitness payoffs of sharing information with immature offspring would have been especially large because, due to their small size and inexperience, juveniles are more vulnerable to environmental hazards. 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, p. 442). Proactive provisioning of reliable information would have increased offspring survival rates and, hence, parental fitness (Fitch, 2004). As Locke and Bogin (2006) note, “Even a small amount of vocal–verbal behavior would have facilitated warnings and instruction” (274), a claim supported by evidence that the mind contains mechanisms specialized for reasoning about precautions (Fiddick et al., 2000). As language increased in complexity, it exponentially increased the information sharing capacities of humans; however, it also added a new set of selection pressures. By providing an additional means of acquiring economic and social knowledge, language imposed a new learning requirement on ancestral humans: mastery of “skills associated with the social, pragmatic, and performative uses of language” – not only for the purposes of warning and instruction, but also for “deceiving, mollifying, negotiating, and persuading” (Locke and Bogin, 2006, pp. 264–265). In oral cultures, these ends are achieved through the use of such forms as gossip, public speaking, and – significantly – narrative (Scalise Sugiyama, 1996; Locke and Bogin, 2006).

Ethnographic research on the transmission of foraging skills supports the hypothesis that that the flow of resources from parents to children includes information. In a now classic study, Hewlett and Cavalli-Sforza (1986) asked 72 Aka adults, adolescents, and children whether they had each of 50 daily life skills (e.g., shelter construction, net hunting, food gathering, food preparation, infant care, sharing). For each skill, informants were asked if there were any person(s) who had shown them how to perform that skill. Of the 40 adults in the sample, for all skills lumped together, the proportion of individuals who reported being self-taught was 0.9%. Significantly, parents were the primary teachers: for all traits combined, 80.7% of adults reported learning from a parent. Percentages were even higher for key foraging skills: food gathering (89.3%); net hunting (84.5%); sharing (83.9%); infant care (85.6%). A similar study of indigenous knowledge and bush skills asked James Bay Cree women in the villages of Moose Factory and Peawanuck whether they had a given skill, who their major teacher was, and how old they were when they learned the skill. Parents were by far the most frequent source of knowledge: 61% of Moose Factory participants and 72% of Peawanuck participants named one or both parents as their main teachers (Ohmagari and Berkes, 1997). Although their study of the transmission of traditional food procurement techniques examined peasants, Ruddle and Chesterfield (1977) also found that parents were the primary teachers. Shennan and Steele (1999) expressly argue that skill transmission is a form of parental investment. In a survey of over a dozen cultures of varying complexity, they found that transmission of craft skills was most often from parent to child. There are some noteworthy exceptions to this rule: among the Hadza, for example, boys “almost never go hunting with their fathers” (Marlowe, 2010, p. 157), learning instead through a combination of practice and observation of older boys.

Several lines of evidence suggest that foraging knowledge takes considerable time to acquire and refine. Hewlett and Cavalli-Sforza (1986) found that, by age 10, the majority of Aka children have acquired most of the skills they need to survive in the forest. However, certain key skills are not mastered until the mid-teens. At age 10, 13% of girls and 0.0% of boys report having mating skills, whereas at age 15, 100% of girls and 94% of boys report having these skills. At age 10, 69% of girls and 36% of boys report having infant care skills, whereas at age 15, 84% of girls and 70% of boys report having these skills. During this same period, boys also increase their net hunting and other hunting skills: at age 10, 75% have net hunting and 28% have other hunting skills, while at age 15, 96% have net hunting and 63% have other hunting skills. This latter finding is important given our species’ dietary dependence on nutrient-dense foods to support the high metabolic cost of brain growth and maintenance (Leonard et al., 2007), and the importance of hunting in obtaining these foods. Other studies report similar ages for indigenous skill acquisition. For example, among the James Bay Cree, the mean age of transmission was 14 in Moose factory and 11 in Peawanuck; older Cree women reported that they had learned the same skills by the age of 13–15, and had mastered all bush skills before the age of marriage (Ohmagari and Berkes, 1997, pp. 209–210). These findings are consistent with ethnographic research indicating that Cree children had the skills requisite to surviving in the bush by the time they reached their mid-to-late teens (Rogers and Rogers, 1960; Flannery, 1962; Blythe et al., 1985). Ruddle and Chesterfield (1977, p. 104) report similar findings for Venezuelan peasants. Observations of single-skill acquisition, too, suggest that mastery takes years of practice. For example, Mer children begin spearfishing as soon as they begin walking, yet it is not until between the ages of 10 and 14 that those “children that choose to invest in spearfishing practice reach the same efficiency as the most practiced adult” (Bliege Bird and Bird, 2002, p. 262). Waorani boys can consistently hit leaf and fruit targets with a blowgun at age 5, but do not become proficient until their late teens (Yost and Kelley, 1983).

The embodied capital hypothesis is supported by comparative evidence that primate foraging is cognitively demanding (Menzel, 1997) and ethnographic evidence that successful exploitation of the foraging niche requires extensive zoological, botanical, and geographical knowledge (Laughlin, 1968; Blurton Jones and Konner, 1976). For example, the Aka exploit 63 plant species, 20 insect species, honey from 8 species of bees, and 28 species of game, the latter of which requires knowledge of several capture techniques (e.g., spear, net, crossbow, snare, trap; Hewlett, 1991). Similarly, the Ache utilize a minimum of 78 mammal species, 21 reptile and amphibian species, 150 bird species, and 14 fish species (Kaplan et al., 2000, p. 171). Like hunting, gathering requires extensive knowledge. Among the Washo, for example, efficient gathering “requires a vast fund of knowledge about the growth cycle of dozens of plant species, [and] an understanding of the effect of weather on growth and knowledge of soils and growing conditions” (Downs, 1966). Among the !Kung, “knowledge of plant identification, growth, ripeness, and location… is extremely complex” and “!Kung women are highly skilled at distinguishing useful from non-useful or dangerous plants and at finding and bringing home sufficient quantities of the best food species available” (Lee, 1984, p. 37). The use of food storage to survive predictable annual shortfalls creates additional information demands, such as “boning, drying, salting, and smoking as well as the use of specialized tools” (Kramer, 2005b, p. 227). Hunting and gathering also require extensive topographical and wayfinding knowledge: many foraging groups utilize more than 1000 km² in the course of a year (Kelly, 1995, pp. 112–115). The James Bay Cree are a case in point: “a knowledge of fish distributions and behavior, as well as the ability to navigate the coast are prerequisites to successful fishing” (Ohmagari and Berkes, 1997, p. 200, emphasis added). Indeed, traditional Cree and Inuit knowledge of seasonal cycles, rivers, currents, sea ice, and the food web is so extensive that it has been used to aid in assessments of sustainable resource development in the Hudson Bay region (McDonald et al., 1997).

Perhaps the most impressive display of forager knowledge is seen in tracking. From the angle at which urine has landed on snow, a Beaver hunter can tell whether the moose he is stalking is a cow or a bull; from the distances between hoof marks, he can infer the animal's speed; and from the texture and shape of a hoof mark or droppings, he can tell how long ago an animal passed through an area (Brody, 2002).!Xo trackers were able to tell that “the spoor leaving the sleeping place had been made early that morning and was therefore relatively fresh. The spoor then followed a straight course, indicating that the animal was on its way to a specific destination” (Liebenberg, 1990, p. 80). When they came upon several sets of footprints all in the same area, one tracker determined “that these footprints all belonged to the same animal, but were made during the previous days. He explained that the particular area was the feeding ground of that particular wildebeest. Since it was, by that time, about midday, it could be expected that the wildebeest may be resting in the shade in the near vicinity” (Liebenberg, 1990, p. 80).

Ethnographic evidence suggests that forager education begins early. As noted above, Mer children begin spearfishing as toddlers (Bliege Bird and Bird, 2002) and Waorani boys are using blowguns by age 5 (Yost and Kelley, 1983). Similarly, Hadza boys are given their first bows at age 3 (Marlowe, 2010). Cree children began learning as soon as they could walk and started helping with work around camp (Long, 1978). Mardudjara children began accumulating foraging knowledge at very early age (Tonkinson, 1978), and most Ache “children have detailed knowledge of resource characteristics and capture techniques before ever foraging themselves” (Kaplan and Hill, 1992, p. 196). Chagnon (1997) observes that Yanomamö children are “accomplished naturalists” (130) at an early age – for example, that most 12-year-old boys can name 20 species of bees, give anatomical or behavioral reasons for their distinctions, and tell which ones produce the best honey. Similarly, “any 12-year-old [!Kung] boy can accurately reproduce in the sand the prints of a dozen species” (Lee, 1984, p. 47). However, an incident reported by Helena Valero (a Brazilian woman who was captured by the Yąnomamö) suggests that, although children's foraging knowledge may be considerable, it is not exhaustive. During Valero's captivity, a little girl died and another became ill from eating the eggs of a poisonous toad. An older woman scolded the sick girl: “So you don't know that the eggs of that toad are poisonous? When I was with the Namoeteri five children died after eating the eggs of that toad…. We had gone to pick mumu, and the children, who stayed behind with only the old women, had cooked the toad's eggs” (Biocca, 1970, p. 86). Unfortunately, these observations are anecdotal and therefore inconclusive. Because there are so few systematic studies of forager knowledge transmission, it is not clear how or when children master this body of knowledge.

In some habitats, the age at which children begin acquiring foraging skills is limited by local ecology. Some skills can't be taught until the child has sufficient stamina to travel long distances, sufficient patience to endure discomfort, and/or sufficient strength to manage heavy loads or animals. For example, although acquisition of foraging knowledge began in early childhood, Mardudjara men did not take their sons hunting with them until they were mature enough to travel long distances and endure the tension of stalking game without detection (Tonkinson, 1978). One Dunne-za informant reports that kids start to go on hunts “After they're about the same size Joe, you know [age 10]. Just enough to think, you know. To talk, and think. Walk” (Ridington, 1988, p. 170). Netsilingmiut boys did not begin to learn breathing-hole seal hunting until between the ages of 12 and 14 (Rasmussen, 1931), presumably because of the strength required to haul in a seal and the endurance required to lie still on the ice in frigid temperatures for hours at a time. In some environments, foraging is simply too dangerous and/or difficult for young children, which constrains their opportunities for acquiring certain skills until they are older.

Blurton Jones et al.’s (1994) comparison of !Kung and Hadza children's foraging illustrates this point. The Kalahari Desert is characterized by high temperatures, little water, predators, and thorny plants. Hunting and gathering excursions may take adults far from camp, and the days are hot 8 months out of the year. Due to the flat featureless terrain and thick brush, a child who falls behind or wanders off is likely to get lost, which is very dangerous: children are much more vulnerable to heat exhaustion, cold, starvation, and predation than adults. Not surprisingly, !Kung parents regard children as a hindrance to hunting and gathering. Children over age 3 rarely accompany women on gathering forays: a woman can't carry an older child who gets tired, because she is usually carrying an infant or toddler, in addition to whatever she has gathered. These problems are exacerbated in the dry season, when there are no water sources en route to gathering sites. Although there is no direct research on !Kung children's mastery of foraging skills, children's contribution to subsistence gives some indication of when this occurs. According toDraper (1976),/Du/da !Kung children make virtually no economic contribution to subsistence: boys do not begin serious hunting until age 16 or older, and girls do not begin regular food, water, and wood gathering until around age 14 (see also Draper and Cashdan, 1988). !Kung marriage practices suggest that gathering skills are not mastered for several years. Traditionally, !Kung girls were married between the age of 12 and 16 but, due to the practice of bride service, effectively remained at home until they were older. One reason the !Kung give for the practice of bride service is that “a girl of 12 or 14 was simply too young to leave her parents” (Lee, 1984). The fact that/Du/da !Kung girls don't begin gathering until age 14 suggests that they continue to live with their parents after marriage because their foraging knowledge is incomplete.

In contrast to the Kalahari, the Hadza environment has ample shade, salient landmarks, panoramic vistas, and sufficient dry season water sources. These features greatly reduce the danger of heat exhaustion and the chances of getting lost. Accordingly, Hadza children are allowed to play and forage on their own, and may spend several hours a day away from camp (Blurton Jones et al., 1994). Similarly, Martu children begin foraging on their own without parental supervision at a very young age (Bird and Bliege Bird, 2005). Bird and Bliege Bird found that the winyjikity goanna return rates for younger (age 5–7) Martu children were not significantly different from those of older (age 12–14) children, and conclude that, “Beyond about 5 years, age alone has little effect on children's hunting success in rocky outcrops” (142). Although this research appears to indicate that foraging competence is swiftly achieved, the range of skills examined in these studies is quite narrow. For example, Bird and Bliege Bird (2005) only examined the harvesting of one wild resource (i.e., hunting winyjikity goannas) in one location (rocky outcrops). The exploitation of other resources and ecological zones may require a much longer learning period. Also, participants drove to field camps and in some cases to foraging locations, which removed the route selection and navigation tasks normally associated with hunting and gathering. Topographical information – e.g., the distribution of resources across the landscape, their season(s) of availability, their location relative to travel routes – is unlikely to be acquired by hunting in rocky outcrops close to camp, or to be mastered by age 5. Aboriginal children are able to capture small game and gather grubs and fruit, but they remain dependent on their parents for foods with higher capture and processing costs (Gould, 1980). A Martu woman's memories of her childhood foraging experiences illustrates this point:

As Konner (2010) astutely observes, the informant's references to mother feeding all the kids and father being a good hunter “show that the children did not really fend for themselves” (645). In sum, in low-risk environments with abundant easy-to-harvest foods, children may make considerable contributions to their subsistence (Bliege Bird and Bird, 2002; Kramer, 2002, 2005a,b; Tucker and Young, 2005). However, children's foraging return rates for single or a small subset of resources are not a proxy for breadth and depth of foraging knowledge.

Turning this observation on its head, Tucker and Young (2005) argue that children's contributions to their caloric intake are not an accurate measure of their maximum foraging potential, which makes it difficult to pinpoint the terminus of juvenile dependency. This point is made in the context of Mikea children's specialization in a high-quality, easily extracted resource: the ovy tuber. Tucker and Young suggest that Mikea children could provide most of their caloric needs, but choose not to. However, the Mikea forest is an exceptionally safe foraging habitat for children; thus, Mikea children do not face the information demands associated with avoiding dangerous animals, exposure, heat exhaustion, and other environmental hazards. Moreover, although ovy is “the most important foraged food source” (Tucker and Young, 2005, p. 152), the authors identify at least 23 other wild foods included in the Mikea diet, begging the question of when Mikea children master the skills needed to harvest them. High foraging return rates (measured in calories) have been documented among Hadza children as well. However, most of these calories are acquired on trips to distant berry patches with adults (Hawkes et al., 1995, p. 694); thus, these children are largely spared the tasks of route finding, predator detection, and remembering patch location(s).

Bird and Bliege Bird (2005) add another turn of the screw by noting that, because it is an indirect measure, foraging efficiency does not tell us exactly how long it takes children to acquire a given foraging skill. While true, this observation misses the larger point that successful exploitation of the foraging niche requires a “basket of competencies” (Bock, 2002, p. 168) that are interdependent and additive. A given competence may be mastered early and/or quickly, but its successful deployment may require recruitment of other competencies that do not come online or are not mastered until later in life. As Blurton Jones and Marlowe (2002) note, “The difference between being a good archer and being a successful hunter may be very large; hunting benefits from extensive knowledge and requires several skills” (229).

Measuring children's contributions to their subsistence in calories is also problematic in that it does not indicate whether children's foraging returns meet their nutritional needs. For example, Hadza children's mean in-patch acquisition rates for two types of berries (tafabe and ondishibe) are 2223 and 964 cal/h, respectively – about 50–70% of adult rates (Hawkes et al., 1995). Based on this finding, the authors argue that Hadza children “could earn their own daily consumption requirements in less than 3 h once they reached the patch” (693). However, no amount of berries will provide sufficient fat or protein for a growing child, nor is it likely to provide all necessary vitamins and minerals. Another Hadza study estimates that, “If children spent just 2 h per day foraging for baobab, or (for older boys with access to an axe) honey, they would acquire around 800–1000 cal, almost half the calories they need…. Thus it is quite possible that Hadza children from the age of about 5 years old provide half of their food by their own efforts” (Blurton Jones et al., 1989, p. 380). While high foraging return rates among dependent children may contribute to shorter inter-birth intervals for mothers, a child who provides half of his/her daily calories is by no means economically independent, and a diet of baobab and/or honey cannot supply the fatty acids requisite to brain growth at this stage of life (Leonard et al., 2007). On this point, the authors observe that Hadza boys age 10 and older can and do shoot birds and small animals, “but this was not a significant part of their diet” (Blurton Jones et al., 1989, p. 379). However, the authors only followed children during the dry season; there may be seasonal variation in Hadza boys’ hunting and in Hadza children's gathering returns in general.

Although researchers disagree over whether the comparatively low productivity of forager children is due to insufficient size/strength or insufficient knowledge, research indicates that both factors are involved. Both within and across habitats, economic tasks vary in the degree of strength, skill and knowledge they demand. This variation may be parsed in terms of easy-to-harvest collected resources (e.g., fallen fruit) versus difficult-to-harvest extracted resources (e.g., large game; Kaplan et al., 2000). In studies that find high foraging return rates for children, the targeted resource(s) tends to fall into the former category (e.g., Hawkes et al., 1995; Bliege Bird and Bird, 2002; Bird and Bliege Bird, 2005; Tucker and Young, 2005). Moreover, as noted above, these high rates tend to occur either in exceptionally safe, easily navigable habitats (e.g., Blurton Jones et al., 1994; Bliege Bird and Bird, 2002; Bird and Bliege Bird, 2005; Tucker and Young, 2005) or when children forage in the company of adults (Hawkes et al., 1995) – conditions that spare children the tasks of remembering resource locations, route finding, and/or avoiding environmental hazards, and thus belie the “ease” with which these resources are harvested.

More importantly, humans do not live by collected resources alone. Harvesting extracted resources typically involves a suite of skills, some of which may be acquired relatively quickly (e.g., marksmanship), and others of which may take years to master (e.g., locating and successfully capturing game; Blurton Jones and Marlowe, 2002; Walker et al., 2002; Gurven et al., 2006). As Bock (2002) argues, for any given task, two aspects of technology potentially come into play: the strength needed to effect it and the knowledge and/or skill needed to deploy it. For example, picking up nuts may seem simple, but carrying them takes strength and finding them takes knowledge. A !Kung woman may walk 16 km per day carrying a full day's harvest and a child. She needs to know where foods are to be found, and in what season they are edible; how to keep oriented in the bush (5- and 6-year-old !Kung children cannot do this; Blurton Jones et al., 1994); and how to read animal tracks. The latter skill is integral to avoiding dangerous animals (e.g., lions, leopards, elephants, snakes) while gathering. On this point, !Kung women are able to tell whether a given snake species is dangerous or not, and how it reacts to the presence of humans. While gathering, !Kung women also make note of animal tracks and report recent game movements to the men when they return to camp. Once gathered, foods may require processing, which may take both strength and skill. For example, although a certain amount of strength and coordination are needed to crack mongongo nuts effectively, peak efficiency takes “some years to acquire after adults stop growing” (Bock, 2002), suggesting that knowledge affects efficiency. Thus, gathering nuts is not a simple matter of being able to carry a heavy load.

Bock's (2005) study of fishing among San and Bantu boys and girls in the Okavango Delta illustrates this point by teasing out the relationships between size, experience, and children's foraging returns. His findings support the claim that childhood is a period of both strength and knowledge acquisition. He found that, for girls’ basket fishing, arm strength does not significantly affect the quantity of fish caught. Variation in fishing returns is due to age (i.e., experience, knowledge) alone: older girls catch more fish. In contrast, arm strength, not age, is a predictor of size (mean weight) of fish caught: stronger girls catch bigger fish. This is because catching larger fish requires a larger basket and the strength to move the basket through deep water. Bock concludes that age (experience) is important in knowing where to find fish, whereas strength is integral to catching larger fish. For boys’ canoe fishing, Bock found that arm strength, not age, is predictor of quantity of fish caught: stronger boys catch more fish. This may be related to the ability to pole a canoe close to the reed beds, which requires poling a long distance through deeper water. In contrast, age, not arm strength, is a predictor of size (mean weight) of fish caught: older boys catch bigger fish. Bock concludes that stronger boys can pole a canoe close to the reeds, whereas more experienced boys may have greater knowledge of where to find big fish and/or how to use hook and line.

In sum, in order to understand why prolonged juvenility evolved in humans, it is critical to determine when children master the full spectrum of skills it takes to be a competent forager. In so doing, it is important to include not only subsistence skills per se (e.g., arrow shooting, fish jigging, tuber digging), but skills that scaffold foraging (e.g., wayfinding, tool manufacture, predator avoidance), including skills that are integral to navigating the social world (e.g., sharing, mating, infant care). Research indicates that, at least in some habitats, it takes many years to reach peak foraging efficiency (Ohtsuka, 1989; Kaplan et al., 2000, 2003; Kaplan and Bock, 2001; Bock, 2002; Walker et al., 2002; Gurven et al., 2006), and juvenile consumption exceeds production until the late teens or older (Kaplan and Robson, 2002; Robinson et al., 2008). However, as noted above, foraging efficiency is an indirect measure and does not tell us exactly how long it takes to learn these skills. With regard to the question of whether and to what degree parents invest knowledge in their children, it would be useful to know how much forager children learn on their own, and how much they learn from adults. Unfortunately, there is no data on these groups comparable to Hewlett and Cavalli-Sforza's (1986), showing the age at which children master a representative sample of foraging skills and how they acquire them. Despite this lack, cross-cultural ethnographic evidence suggests that successful occupation of the foraging niche requires extensive knowledge, that at least some of this knowledge is transmitted from parents to children, and that mastery of the “basket of competencies” is not attained until the mid-teens – more or less the length of the extended juvenile period.

The question at the heart of the prolonged juvenility debate is not whether children learn during childhood, but whether learning is a cause or consequence of the extended juvenile period (Bock and Sellen, 2002; Hrdy, 2005; Konner, 2010). Although presented as alternative processes, they are most likely complementary. Human infants are born when the brain is only 25% of its adult size, and are incapable of supplying the energy needed to complete the growth process. Moreover, neural tissue is metabolically expensive to grow and operate, commandeering 87% of resting metabolic rate in infancy and 20–25% in adulthood (Leonard et al., 2007). Thus, brain expansion could not have been selected for in the absence of either, increased provisioning/protection or compensatory benefits. On this view, extended provisioning of offspring and encephalization are tandem processes, each driving the other in a 2-million-year-long feedback loop. In the course of this process, some parts of the brain were enlarged more than others. These areas include the cerebellum, neocortex, basal ganglia, and hippocampus (Stephan et al., 1981), all of which are integral to learning processes. Thus, it appears that brain expansion was driven, at least in part, by selection for increased learning capacities. This hypothesis is supported by extensive research in developmental psychology showing early onset of a host of domain-specific cognitive systems that guide attention to relevant stimuli and organize informational inputs (for a review, see Carey, 2009).

A sizeable portion of this cognitive equipage appears to be dedicated to mechanisms for social learning: to a greater degree than other species, humans appear to be designed to acquire information from conspecifics (Byrne, 1995; Tomasello, 1999; Tomasello et al., 2005). Perhaps the most compelling evidence in support of this claim is shared attention, which appears to be unique to humans (Baron-Cohen, 2005) and emerges predictably near the end of the first year (Scaife and Bruner, 1975; Butterworth and Cochran, 1980; Leslie, 1994; Baron-Cohen, 1995, 2005; Carpenter et al., 1998b) – well before infants have mastered the art of walking. This is what we would expect if social learning were integral to occupation of the foraging niche and if learning took a long time. If humans could quickly accumulate all knowledge requisite to success in their ecological niche, there would be no need for the structures that subserve learning to develop early.

In addition to shared attention, learning from others requires shared intentionality: the ability to attribute a goal to an individual, to understand how the individual's actions subserve his/her goal, and to make the extrapolation, “When I have the same goal I can use the same means” (Tomasello et al., 2005, p. 680). This capacity emerges by about 14 months, and is neatly illustrated in Meltzoff's (1995) study of imitation in 18-month-old infants. Subjects were presented with two types of demonstrations. In the first condition, infants watched an adult perform actions on objects, whereas in the second condition, infants watched an adult try but fail to perform actions on objects (e.g., try but fail to pull apart two parts of an object). Both groups reproduced the target actions equally well. Tellingly, in Condition 2, infants appeared to understand what the adult was trying (i.e., intended) to do and performed that action instead of mimicking the adult's actual behavior. In a related study, Carpenter et al. (1998a) examined 14- to 18-month-old infants’ imitation of accidental versus intentional actions. Infants watched as an adult performed two-action sequences on objects. One action was marked vocally as intentional (“There!”), and one action was marked as accidental (“Woops!”). Infants were then given the opportunity to make the result occur themselves. Infants imitated twice as many intentional actions as accidental ones regardless of the order in which they saw them. Gergely et al. (2002) report an even more sophisticated understanding of the relationship between intention and action. In this study, 14-month-old infants watched as an actor used his head to turn on a light in either a hands-free or a hands-occupied condition. When given the opportunity to turn on the light themselves, subjects who saw the hands-free condition used the head method more often than infants who saw the hands-occupied condition. These results suggest that infants in the hands-occupied condition assumed that the actor wouldn't have used his head if his hands had been free. Conversely, infants in the hands-free condition appear to have assumed that the actor had a reason for using his head instead of his hands. These studies suggest that infants infer the goal of an actor, and use this representation to replicate that goal.

Shared intentionality is also integral to language learning (Tomasello et al., 2005) and thus to the acquisition of knowledge via linguistic communication. As Baldwin et al.’s (1996) “dawnoo” experiment illustrates, language acquisition is dependent on the ability to understand what a conspecific intends (“means”) by an utterance. In this study, toddlers (19–20 months) were shown a novel object while a speaker simultaneously produced a novel label (e.g., “A dawnoo!”). In the first condition, the speaker was seated next to the infant and looking at the novel object when the novel label was vocalized; in the second condition, the speaker was seated out of sight of the infant. Infants in the first condition readily chose the correct object over an equally salient distractor object when they were later asked to find the dawnoo. In contrast, infants in the second condition were just as likely to choose the distractor object as the correct object. Baldwin's (1993) team concluded that toddlers are capable of understanding that speakers intend to talk about objects and of using this understanding to acquire new information – in this case, to learn new object–word associations.

Pinpointing the developmental onset of the mechanisms involved in social learning is important because, as noted above, their early emergence suggests that social learning is integral to occupation of the foraging niche. It also suggests that assembly of these mechanisms – including information input – takes a lengthy period of time. However, this is only a supposition unless we know when development of these capacities is complete: to understand the relationship between learning and the evolution of prolonged childhood, we need to understand how long learning takes. Thus, it is critical to identify not only the age of onset of learning systems, but also the age of maturity. At this point, the amount of time required for complete assembly of these mechanisms is largely unknown – a lack that parallels the paucity of ethnographic data on the time frame for acquisition of essential knowledge sets among foragers.

To illustrate this point, consider three knowledge sets critical to success in the foraging niche: information about the social environment, animal behavior/characteristics, and topography. Acquisition of social information is subserved, in part, by the mindreading system. Research indicates that mindreading involves several different capacities, with different developmental trajectories (e.g., Leslie, 1994; Baron-Cohen, 2005). For example: the ability to represent affective states emerges at around 3 months (Walker, 1982); the understanding that an entity with eyes can perceive emerges between birth and 9 months (Baron-Cohen, 2005); the ability to represent goal-directed action emerges between 5 and 7 months (Woodward, 1998; Csibra et al., 1999; Hamlin et al., 2007); shared attention emerges by 12 months (Scaife and Bruner, 1975; Butterworth and Cochran, 1980; Carpenter et al., 1998b); and the understanding that seeing-leads-to-knowing emerges by 15 months (O'Neill, 1996; Onishi and Baillargeon, 2005). Pretense, too, appears to develop in stages: pretend-play emerges between 18 and 24 months, but the understanding of pretense (i.e., recognizing when others are engaging in it) emerges by 15 months (Onishi et al., 2007). Thus, the mindreading system is additive, involving the successive assembly and integration of multiple components over time. This assembly is assumed to be complete when children can understand that others can have beliefs that are different from theirs and that these beliefs can be false (i.e., false belief). However, mindreading development does not end with the ability to understand the behavior of others in terms of epistemic states. The mindreading system cannot be considered complete until it can be used to construct a reliable, integrated “social map of the persons, relationships, motives, interactions, emotions, and intentions” in the individual's social environment (Cosmides and Tooby, 1992, p. 163). Just as a spatial map represents the relative positions of a given set of entities, a social map represents the motives, emotions, and intentions of individuals in relation to one another. In order to make such a map, a person must be able to (1) note and remember the actions performed by a specific individual; (2) determine which of the individual's actions subserve which of the individual's goals (i.e., infer the individual's plans for attaining his/her goals); (3) integrate this representation with representations of the goals and actions of the other individuals in his/her social world; (4) identify points of conflict between the goals of these individuals; and (5) track the outcomes of these conflicts (Scalise Sugiyama, 2009). In short, mapping one's social world requires the ability to track interactions between multiple lines of goal-directed action. The question of when this capacity develops is an open one.

On this point, children's acquisition of narrative competence is illustrative. Story plots trace the conflicts between characters’ goals and the outcomes of those conflicts (Scalise Sugiyama, 2005, 2009). Thus, narrative processing entails tracking interactions between multiple lines of goal-directed action – i.e., the construction of a social map of the story world. Interestingly, children's early narrative is organized by a central theme rather than a plot (Sutton-Smith, 1986). After age 3 or 4, story focus shifts from a central theme to a central character, but conflict – a central goal and opposition to that goal – is still absent. Central characters with a clear conflict do not appear regularly until somewhere between age 5 and 7. Children “can state a conflict before they can develop and resolve it” and do not master the latter ability until as late as age 10 (Sutton-Smith, 1986, p. 82). Thus, plot structure consists of several components, which children appear to master in stages: (1) organization around a central character, (2) with a clear goal, and (3) obstacles to the attainment of that goal (including goals of other characters). Of course, children's acquisition of narrative competence cannot tell us when the ability to track multiple lines of goal-directed action emerges, as development of this ability may precede development of the linguistic skills requisite to expressing it. However, children's narrative development shows that mapping the social world requires the ability not only to attribute mental states to others, but to chart the mental states of multiple individuals as they converge and/or collide with one another in an endless series of moves and countermoves. Little is known about the design or ontogeny of the cognitive structures involved in this task. Additionally, because social environments vary across locales, mapping them requires local information inputs. Yup'ik girls’ storyknifing (oral narrative accompanied by drawings sketched in mud with a knife) suggests that storytelling may be a source of these inputs. Among other topics, participants “tell about the people they saw during the day and where they fit into the kinship system. Often, one girl will have information that corrects or supports another's account. The stories may explain why a member of the community is known by a particular nickname, to which family a woman belonged before her marriage, or how certain people are related to others. The stories are a means by which the girls can work together to construct the kinship system of the community” (de Marrais et al., 1994, p. 202). Participants in de Marrais et al.’s (1994) study ranged in age from 9 to 12, but it is not known when children acquire a comprehensive map of their social world.

Another key foraging knowledge set is information about dangerous animals and how to avoid them. Evidence suggests that at least some of the cognitive mechanisms that scaffold this task begin emerging early in development. For example, humans appear to have evolved intuitive ontologies (Spelke et al., 1992; Wellman and Gelman, 1992; Carey, 2009) – cognitive structures that categorize entities in the world by the properties they have in common. These ontological categories are invaluable for making valid inferences. For example, putting jaguar in the “agent” category activates the inferences “eats,” “periodically experiences hunger” and “is capable of goal-directed action,” which have obvious fitness benefits for the individual making them. A number of studies suggest that the agency system contains intentional schemas – inference systems activated under specific conditions – which are used to predict the behavior of agents (Gergely et al., 1995; Csibra et al., 1999), including animal predators (Csibra et al., 2003; Barrett, 2005).

However, finer-grained predictions of animal behavior require knowledge acquisition as well as inference (Barrett, 2005). Boyer's (2001) “template” model of knowledge acquisition illustrates the interdependence of conceptual primitives (e.g., innate ontological categories) and environmental input (e.g., learning). Boyer proposes that the acquisition of knowledge about entities in the world is guided by cognitive structures – templates – that specify the general attributes of the category to which that entity belongs (e.g., person, animal, natural object, tool). Thus, the animal template specifies the general attributes of animal agents (e.g., eats, reproduces, occupies a specific habitat), and the particulars of a given species’ attributes are fleshed out as relevant information becomes available over the individual's lifespan. Initially, a child may only know a species’ name – e.g., “jaguar.” Then, as additional information is acquired, it is plugged into the appropriate slot. For example, the information that jaguars attack and eat humans would be added to the slot for “diet.” Thus, while ontological categories and at least some intentional schemas (Gergely et al., 1995; Csibra et al., 1999, 2003) emerge early and reliably in development, full assembly and calibration of these systems requires informational inputs from the local environment. The age by which this occurs is unknown. Nor do we know the degree to which information inputs are provided through verbal transmission. Answers to these questions are critical to determining when learning is complete.

A third critical foraging knowledge set is wayfinding information. Navigation is integral to nearly every facet of forager life: locating food, water, and shelter; avoiding predators, dangerous conspecifics, and inanimate hazards; and mate acquisition (Silverman and Choi, 2005). This raises the question: at what age is this capacity sufficiently developed for an individual to cognitively map the distribution of resources and hazards across the landscape and travel long distances through this landscape without getting lost? Psychological research indicates that children begin to navigate by landmark around age 2, but do not use orientation (dead reckoning) strategies until age 8 (Anooshian and Young, 1981; Scholnick et al., 1990; Blades and Medlicott, 1992). Once again, however, we must not confuse onset with maturity. Navigating by landmark requires informational input (i.e., knowledge of landmarks in one's local environment) and, given that many forager groups maintain knowledge of more than 1000 km² (Kelly, 1995, pp. 112–115), this information may take years to acquire. Given the wide range of latitudes occupied by our species, dead reckoning mechanisms may similarly require local environmental inputs for assembly and calibration (i.e., to local and seasonal variation in the sun's zenith, hours of daylight, etc.).

Testing the claim that engaging in storytelling develops skills and knowledge sets that are useful later in life is a difficult proposition. This is due to the universality of the behavior in question: if all humans are exposed to storytelling, how are we to determine whether exposure to storytelling enhances knowledge or whether humans do indeed draw on information contained in stories to solve real-world problems? An ideal test would involve a within-culture comparison of performances on specific foraging tasks between two sets of individuals, one raised with storytelling and one raised without. For obvious ethical and practical reasons, such a test is untenable. However, a newspaper article about the tsunami of December 26, 2004 (Mason, 2005) suggests an alternative experimental paradigm. According to the article, only 7 of the 75,000 inhabitants of Simeulue island died, because the overwhelming majority remembered the stories their grandparents had told them about the “semong” (tsunami) of 1907 and fled to higher ground. As one informant reported, “After the earthquake, I looked for the water to suck out…. I remember the story of the ‘semong’ and I ran to the hill.” A test similar to this natural experiment can be envisioned in which performance on a given task is compared between subjects who are told a story that contains information useful to performing the task and subjects given the equivalent information in non-narrative form (or who are told a story unrelated to the performance of the task). Neuroimaging provides another means of testing the hypothesis that storytelling is used to acquire information instrumental to constructing future scenarios. Are the brain regions activated by planning tasks also activated during narrative processing? Are the same brain regions activated by episodic memory tasks also activated during narrative processing?

We can also make predictions about cross-cultural patterns in story content. As the tsunami article suggests, we would expect stories to contain information about phenomena that occur at intervals longer than the human lifespan (e.g., calamitous earthquakes/tsunamis, severe droughts, 100-year floods). Another prediction is that stories should contain information that is difficult or dangerous to acquire at first hand – for example, information regarding environmental hazards (e.g., toxic substances, dangerous animals, getting lost, inclement weather). On this point, Bock and Johnson (2004) argue that play is an important means of learning when there are constraints on learning through first-hand experience. Conversely, because some skills (e.g., shooting an arrow, cracking nuts, weaving a basket) are best acquired through observation and practice, we would not expect stories to contain detailed explanations of how do these things. However, we might expect them to contain information relevant to locating, selecting, and/or processing materials used in these tasks.

To test the hypothesis that parents invest knowledge in their offspring via storytelling, the most critical and obvious data needed is information on who tells stories to whom. Although studies of daily life skill transmission sometimes include “non-practical” skills such as dance, pottery decoration, and bead making (Hewlett and Cavalli-Sforza, 1986; Shennan and Steele, 1999), they do not include storytelling. Anecdotal observations in the ethnographic and folklore literature indicate that folklore and mythology are transmitted vertically, but the degree to which this transmission is from parent (or alloparent) to child is unknown.

Another important avenue of inquiry is the age at which children become “accomplished naturalists.” If it is true that “the years between the five-to-seven shift and puberty are a human adaptation for a large amount of cultural acquisition” (Konner, 2010, p. 637), we would expect older children to be more knowledgeable than younger children. This could be tested by administering a questionnaire to different age groups to measure proficiency in various knowledge sets (e.g., animal behavior, spatial and temporal distribution of resources). Data could also be collected on the age at which children begin foraging for various resources, then compared with foraging knowledge data to see whether (and how much) knowledge acquisition occurs before children begin foraging on their own. These data would provide a basis for estimating how much knowledge is acquired through social learning. Finally, if it is true that stories are an important source of knowledge in forager cultures, then another knowledge set foragers must master is the oral tradition itself. Anecdotal evidence suggests that this knowledge is not mastered until relatively late in life. Among the !Kung, for example, young and even middle-aged storytellers are rare, and it is “the addition of knowledge, not secret knowledge but a large collection of items which takes a long time to accumulate, which results in competent storytelling. It is generally the case that by the time a person is a zhu n!a (‘old person’) he will have accumulated a good deal of this knowledge” (Biesele, 1976, p. 308). Quantitative data on how many different stories each group member knows is lacking, but would not be prohibitively difficult to collect.

In arguing that storytelling is a means of social learning, I am not arguing that storytelling is the only means of social learning. Rather, storytelling is one of several means by which humans invest in cognitive capital. Nor am I arguing that every information byte in a given story is useful information. Rather, my claim is that forager oral tradition includes information that is potentially useful in solving problems recurrently encountered in forager life. Whether a given piece of information proves to be useful will vary from individual to individual, depending on his/her attributes, constraints, and the events that befall him/her.

If prolonged juvenility evolved, in part, to support learning, then humans may have been under selection to invest in cognitive capital. A case in point is the enjoyment humans get out of listening to stories, which may be rooted in an evolved interest in acquiring information beyond what is immediately necessary. Such an interest has been documented for animal behavior: for example, !Kung men “sometimes observe animals more than is necessary for the purpose of the hunt in which they are involved” (Blurton Jones and Konner, 1976, pp. 337–338; see also Nelson, 1969). This seemingly superfluous interest is hypothesized to be a risk-buffering strategy – a means of acquiring knowledge about animal behavior that might be useful in the future (Blurton Jones and Konner, 1976; Mithen, 1990). On this logic, we would expect humans to have an evolved interest in acquiring information useful not only for hunting, but for gathering, wayfinding, mating, parenting – in short, any evolutionarily recurrent task with inherent risks that can be buffered by knowledge. Humans may find stories interesting because they simulate the whole of human experience – human beings, their actions, and environmental constraints on those actions – and therefore present information concerning a variety of task domains.

As noted above, information exchange “and the effects of incomplete information are important issues particularly relevant to human foragers, who rely extensively on learning and communication to assess resource distribution, abundance, and profitability” (Kaplan and Hill, 1992, p. 186). If information is so critical to success in the foraging niche, we would expect foragers to exhibit concern regarding its reliability. This appears to be the case. The !Kung, for example, distinguish between directly observed behavior, behavior deduced from tracks, hearsay (behavior reported by another person), and behavior they think could happen, and are hesitant to accept as fact behavior they have not witnessed themselves (Blurton Jones and Konner, 1976). This attitude is widespread in oral cultures (Rogoff, 2003, p. 40), and is evinced in storytelling as well. For example,

Across forager oral tradition, a common refrain is a statement to the effect that the narrator has seen the object or event in question. For example, in an Apache variant of the running rocks tale, the informant asserts: “Those rocks are over near Taos still. There is a black mark on them even now where Coyote defecated on them…. The story says they are tied together, but when I saw them they were joined together” (Opler, 1938, p. 336). Similarly, a Yanomamö storyteller claims, “I have seen it [the Kaithë u River] with my own eyes, at the source of the Hwara u, in that direction. The old people with whom I was traveling in the highlands showed it to me and said: ‘There's the Kaithë u! That's the river that Omamë caused to burst forth by piercing the ground!’” (Wilbert and Simoneau, 1990, p. 406).

This attention to the source of information is paralleled by an interest in keeping story content consistent across time and space. Among the Klamath and Modoc, for example, only adults were allowed to narrate myths, and if a narrator deviated from the standard text, listeners could interrupt and debate until the correct version was decided upon. For this purpose, the Klamath preferred that two individuals well-versed in the myth were present to prevent transmission errors on the part of the narrator (Sobel and Bettles, 2000). Among the Dene, “repetition of the tale by the same author serves to fix it in the memory of the listeners, and… the ability of the young people to repeat the story accurately, and honestly (without help or hints), would be tested later by their elders” (de Laguna, 1995, pp. 74–75). These measures are what we would expect if oral tradition were a means of storing and transmitting valuable information – that is, if humans’ “most effective technology… is one that can be carried around in their minds” (Ridington, 1988, p. 73).

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.