The sample

Intensive field work was conducted in two communities, comprising 264 and 410 individuals, of which 145 and 208 were children (or people <16 years of age). Both communities differed in their proximity to the village of Bantu-speaking neighbors- the Nzime- and in the type of school. The first village is settled on the prolongation of the Nzime farmer’s village and Baka children have the opportunity to attend a public national school, together with Nzime children. In contrast, the second village is located at approximately 2 km from closest Nzime neighbor village and has a private school managed by a local institution promoting schooling among the Baka.

The sample for this research included all the children between 5 and 16 years of age willing to participate. The upper limit was fixed at 16 years of age because at this age, the Baka generally start a separate household and are thus considered adults. Although evidence exists that the transmission of local knowledge starts earlier [31], the lower limit was fixed at 5 years of age because younger children were generally too shy or too unreliable to answer interview questions.

Data collection

Over the whole period of data collection, the two first authors lived in the two selected Baka communities, following Baka socio-cultural norms and participating on the daily life of neighboring households, e.g., while washing clothes, cooking, taking care of children, accompanying them on fishing, hunting and gathering expeditions and to their forest camps and agricultural plots. Participant observation was conducted among adults and children and with as many households as possible.

Data analysis

It is worth highlighting that the performance of any of the listed activities might contribute to the acquisition of local ecological knowledge. Indeed, as all the activities a child performs are embedded in the Baka’s culture and environment, they can potentially contribute to local ecological knowledge acquisition. However, given the content of the different activities listed, we assume that some activities are more likely to improve the acquisition of LEK than others. For example, activities directly related to subsistence, in addition to be developed in Baka social and environmental context, require the child to use her skills and put her knowledge into practice. This largely contrasts with other activities, such as listening to modern music, where the child would be less likely to involve ecological knowledge and skills necessary to subsistence activities.

We aggregated data on children’s daily activities by computing the frequency with which each of the 15 categories of activities was reported by each child. To avoid the potential biases derived from the overrepresentation of children who were interviewed more than others, we randomly selected one observation per child (n = 102). Frequency was coded as 1 (if the activity was performed) or 0 (otherwise), independently of whether the activity was performed more than once during the 24-h recall (i.e., if a child reported hunting with traps and hunting with bows and arrows we only coded hunting as 1).

Using census data, we analyzed Baka children’s engagement in different activities by sex and age. Age categories were determined drawing on bibliographic references [33, 34] and our own interviews with Baka adults and consisted in i) middle childhood (>5–9 = <years-old), ii) pre-adolescence (<9–13 = < years-old), and iii) adolescence (<13–16 > years-old). We also compared involvement in the different activities from children in the two studied villages. Then, since our data were collected during two different seasons - the major dry season (from beginning of February to mid March) and the minor rainy season (from mid March to the end of June) - we also analyzed seasonal variation.

To assess differences in children’s involvement in daily activities by sex, village and season we calculated the difference of proportions [35]. For that aim, we first calculated the relative proportion of each cluster of activities (i.e., the number of observations reported by children in each cluster of activities from the total amount of observations reported) and for each of the groups of analysis (i.e., girls/boys, village closest/furthest to neighbor’s village, and major dry/minor rainy season). We then ran two-sample test of proportions at the level of 95 % of confidence, to evaluate whether spotted differences were statistically significant [35]. We tested for significant differences across age categories by running Fisher’s exact tests for each cluster of activities. In a final analysis, we used multivariate logistic regression models to analyze the association between sex, age category, seasonality, and village of residency with each of the 15 types of activities.

Children’s main daily activities

Among the subsistence related activities, household maintenance was predominant. Children mentioned having conducted household maintenance in almost 95 % of the interviews. Other common activities in this cluster included hunting and gathering, reported in 45 and 49 % of the interviews. Agricultural work was reported in 31 % of the interviews, fishing in almost 26 %, and handicraft in 12 %.

Among the non-subsistence related activities, playing appears as the most frequently mentioned activity, reported in almost 59 % of the interviews. Overall, playing is the second most frequently cited activity, after household maintenance.

Finally, among the recently introduced activities, listening to music was reported in 45 % of the interviews, whereas attending school was only reported in 22 % of the interviews.

Gendered involvement in daily activities

Statistically significant differences also exist in the frequency with which boys and girls perform hunting and fishing activities. Hunting appeared as a male orientated activity, with 70 % of the boys but only 18 % of the girls reporting hunting (Table 3). In contrast, fishing is a girl-oriented activity, with girls reporting fishing twice as often as boys (35 vs.15 %). In the same line, but to a lower extent, gathering tends to be more a girl- than boy-oriented activity (cited in 57 % of the interviews with girls and in 41 % of the interviews with boys). From all the activities in the cluster of subsistence-related activities, the only one that seems to be reported with similar frequency between boys and girls is agricultural work.

Girls and boys seem to engage with the same frequency in activities indirectly related to subsistence, with one exception: girls report performing traditional songs, tales, and dances more frequently than boys (12 vs. 6 %; Table 2). Regarding recently introduced activities, the only statistically significant difference between girls and boys relates to playing football, a male-oriented activity in the study area (23 % for boys vs. 2 % for girls) (Table 3).

Variation among age-sex categories

Given the magnitude of the gendered differences found, we kept the sample of boys and girls separated to analyze children’s involvement in daily activities by age-categories (Fig. 1a). Overall, the analysis shows differences in time investment as girls and boys move from middle childhood to adolescence. Among the activities related to subsistence, both girls and boys perform more hunting activities but play less frequently as they grow up. In contrast, they spend more time performing non-traditional activities, such as socializing, listening to music and drinking alcohol. As girls move into adolescence, they invest less time in fishing, (27 % of adolescent girls’ interviews vs. 50 % of middle childhood girls) (Fig. 1). In the same line, adolescent girls invest even less time in hunting than middle childhood girls do (13 vs. 20 %). Contrarily adolescent girls are more frequently involved in agriculture (47 vs. 40 %) and in gathering (67 vs. 50 %) than middle childhood girls.

Although the differences on the frequencies of activities between age categories appear higher among girls, we only found statistically significant differences in the sample of boys (Fig. 1b). Thus, as boys enter adolescence, they become less frequently involved in hunting (reported in 44 % of the interviews with adolescent boys vs. 80 % of the interviews with middle childhood boys). Differently to girls, as they grow up, boys are less frequently involved in household maintenance activities. Also in contrast to the pattern found among girls, adolescent boys allocate less time than younger boys to gathering (38 % for adolescents vs. 67 % for middle childhood boys).

Differences between villages

Seasonal variation

Some activities indirectly related to subsistence show seasonal differences. Playing was reported more frequently during the minor rainy than during the dry season (62 vs. 54 %), a seasonal distribution that probably relates to the higher frequency of subsistence activities conducted during the dry season. Interestingly, children also seem to perform more frequently maintenance activities during the dry than during the minor rainy season (32 vs. 20 %). Finally, whereas children tend to perform recently introduced activities with the same frequency during both seasons, alcohol drinking is almost three-fold more frequent during the dry than during the rainy season (20 vs. 7 %).

Correlates of children’s involvement in daily activities

Multivariate models also confirm the importance of age in explaining children’s involvement in daily activities. Thus, older children engage less frequently in household maintenance (odd ratio = 0.72. p < 0.1) and hunting (odd ratio = 0.86. p < 0.1). Age is also negatively associated to play (odd ratio = 0.83. p < 0.01), but positively associated to alcohol drinking (odd ratio = 2.06. p < 0.01), socializing (odd ratio = 1.63. p < 0.5) and, to a lower extent, listening to music (odd ratio = 1.13. p < 0.1).

Finally, multivariate regressions also show that there are some differences between both villages but not between seasons. Thus, children from the village closest to Nzime’s village tend to be less engaged in agricultural work (odd ratio = 0.32. p < 0.1), play (odd ratio = 0.35. p < 0.5), traditional songs and dances (odd ratio = 0.45. p < 0.1), and school attendance (odd ratio = 0.34. p < 0.1) than children from the other village. In relation to the season, the only statistically significant associations found were a lower frequency of hunting during the dry season (odd ratio = 0.2. p <0.01) and a higher frequency of play during the dry season (odd ratio = 2.34. p < 0.1).

It is worth noting that most multiple logistic regressions explain a relatively small fraction of the variation found (between 3 and 45 %), which suggest that other variables not accounted for in our model do affect children’s involvement in daily activities. The two models with higher predictive power are the model for hunting, which explained almost 30 % of the variation found, and the model for alcohol drinking (45 %).

Variations in Baka childhood activities

A main finding of our work is that, irrespectively of their sex and age category, most Baka children engage in household maintenance, a finding that has also been reported among other small-scale societies [11, 13]. Thus, from the earliest age, Baka children are expected to participate in household chores, for example helping their mothers with tasks such as collecting firewood, fetching water, and taking care of younger children. But, differently to what has been reported in farmer societies [36], Baka children are not expected to participate in income generating activities (although some of them do, especially after reaching adolescence). While such situation that has led some researchers to coin the notion of ‘children in paradise’ [37], Baka children do frequently engage in productive subsistence activities (i.e., hunting, gathering, fishing, and agricultural labor). Our ethnographic observations suggest that Baka children perform these subsistence activities mostly out of enjoyment, especially as these activities are often embedded in games [38]. However, it should also be noticed that such activities seem to provide an important part of children nutritional intake during parental absences (for similar results in other settings see [39–41]). Thus, from early age, Baka children hunt birds or rodents and gather sub-spontaneous tubers, all products which typically are immediately cooked and eaten by the children themselves.

An important finding of our work relates to gendered differences of Baka children’s daily activities, a finding that dovetails with other studies both in farmer [37, 42] and hunter-gatherer [13] societies. The finding, however contrast with at least one study reporting few differences in the activities performed by girls and boys among the Aka, another hunter-gatherer group from the Congo Basin [16]. Baka children do tend to reproduce adult’s same-sex activities. Thus, as Baka women, Baka girls are more involved in children caretaking, cooking, agricultural work and fishing than Baka boys. Similarly, as Baka men, Baka boys are more often involved in hunting than Baka girls.

In addition to gendered differences in frequency of engagement in certain activities, our results also suggest that there are additional differences in the way activities are practiced. For example, girls occasionally hunt. But they only hunt little mammals using their hands, the machete or, during adolescence, unearthing game with smoke. Boys, however, not only hunt more frequently, but they also use a broader diversity of techniques, such as bow and arrows, slingshot, spear and snares. Contrarily, fishing is more frequent among girls, who typically engage in collective fishing expeditions, in which a group of girls and women elevate dams in shallow rivers and extract the water to catch the fishes with their hands. Differently, although boys also fish, they are more likely to use poles or ichtyotoxics, techniques practiced generally alone or in small groups (for a similar finding see [Díaz-Reviriego I. et al., under review]).

It is worth noting, however, that while some activities are clearly gender-oriented, there are not strict gender exclusions in the performance of most activities. Thus girls and boys, as women and men, occasionally perform activities most commonly performed by people from the opposite sex. The flexibility in activity performance, beyond standard gender roles, is a common, but seldom noted distinction of hunter-gatherers versus farmers [43].

The study of children’s involvement in daily activities also shows that preferred activities change as children grow up. As Aka children [11], Baka children tend to spend less time playing and more time in productive and specifically in income generating activities (i.e., agricultural wage labor or commercial hunting and gathering) as they move into adolescence.

In sum, consistent with the Whitings’ [10] predictions, the descriptive analysis of Baka children’s daily activities suggests that such activities are largely shaped by their specific cultural settings, although the sex and the age of the child are important factors that pattern children’s involvement in activities.

Knowledge acquisition through daily activities

We devote the second part of the discussion to analyze how the frequency of performance of different activities might shape LEK acquisition and to describe how such acquisition varies according to the age and the sex of the children.

First, as mentioned, subsistence related activities are predominant during Baka childhood. Additionally, most of the activities Baka children perform through all their childhood occur in their natural environment. We argue that the performance of such activities might directly contribute to local knowledge acquisition, and more specifically to the acquisition of local ecological knowledge. For example, during their youngest childhood, boys spend a considerable amount of time hunting and girls invest time fishing, allowing them to embody hunting and fishing knowledge. Time involved in both activities decreases once they become adolescents; however as knowledge is already embodied, adolescents are able to practice these activities even if they do so more infrequently.

Thus, an important aspect to consider when discussing LEK acquisition and the performance of daily activities relates to the variation, across the lifespan, in the use of techniques and practices of different complexity. Overall, the number of practices and the complexity of tools children use during their daily activities increases with age. Take the case of hunting. From the earliest age, children play various hunting-related games, such as shooting wheels or throwing spears to easy-to-target objects and animals. Then, during middle childhood they start to use popular hunting tools, such as the , or small replicas of the common snare with cable used by adults (called ). Differently, adolescents prefer hunting with spears or the collective hunting of small mammals using smoke. The case of gathering of wild edibles is similar. Even young children gather tubers of spontaneous agricultural plants, such as (Ipomoea batatas), and (Xanthosoma mafaffa), which typically grow around Baka villages. Differently, adolescents gather the leaves of (Gnetum africanum), an important component of the household consumption, or other forest products, such as mbalaka (Pentaclethra macrophylla), (Baillonella toxisperma), or payo (Irvingia excelsa) which can be sold in local markets. Although gathering does not require many tools and techniques (except for some specific products such as honey and yams), effective gathering requires the acquisition of knowledge related to observation, the capacity to identify wild edibles, and the ability to navigate the landscape, abilities that also evolve across the lifespan. In sum, the analysis of children’s activities suggests that LEK acquisition seems to follow a ‘multi-stage learning model’, according to which children would first acquire basic knowledge and abilities that would allow them to progressively acquire more complex skills and knowledge [44].

One more point requires attention. Although the predominance of activities related to household maintenance might apparently underscore our argument of the importance of subsistence-related activities for LEK acquisition, we argue that the performance of these activities are key to obtain the cultural bases of adult’s livelihood and, therefore for LEK acquisition. Indeed, activities such as fetching water, collecting firewood, cooking, sharing meals, or even taking care of younger siblings, are considered by parents as key elements in Baka children’s learning process. Additionally, it is worth noticing that while conducting such household chores, especially those that take place in the forest, children also engage in other activities such as hunting birds, fishing, or gathering mushrooms. Adults clearly know that those tasks let children learn and practice on their own, alone or in groups of peers, skills that they would later need.

Our final point relates to the engagement in activities not traditionally performed by the Baka, such as schooling or listening to recorded music in bars [24]. Consistent with previous studies, both among the Baka [45, 46] and among other hunter-gatherer groups [13, 47], we found that school attendance was very limited, but existent. Thus, for the school year 2012–2013, 54 % of the schooled-aged children were registered at school. However, school attendance was low, irregular, and decreased as the school year advanced. Reasons for this low attendance do not differ from those highlighted by Kamei [46] (i.e., teachers’ low level of commitment) and largely reflect the lack of fit between the national educative system and Baka’s livelihood. Our data also show that new leisure activities, i.e., listening to recorded music, were increasingly common among Baka children. From our ethnographic experience, we know that Baka children generally listen to African popular music during the evening when they might also drink alcohol. These activities, generally very common among all the Baka, are now included in the standard use of time of adolescents and young adults, as they seem to have become the new way of socializing and even of potentially finding a partner [48].

As such activities are new in Baka repertoire, we have no way to assess how their performance might affect LEK acquisition. Studies focusing on the impact of schooling on LEK have found that schooling might have a negative impact on LEK acquisition, unless schooling is adapted to the local social-environmental context [47]. One could speculate that the case would be similar for the Baka. The work presented here, however, brings to the light that not only schooling, but also non-traditional leisure activities might impact the process of LEK acquisition. For example, activities that are mostly conducted at night, such as listening to recorded music in bars, might displace other cultural activities such as the performance of tales, songs and traditional dances, many of which begin as night falls [49]. Interestingly, none of the previous studies on the erosion of LEK have addressed the impact of children’s involvement in new leisure activities in LEK acquisition. Such neglect is worrisome, especially if-as results from our study show-children devote more time to modern forms of leisure than to schooling.