Both in more hierarchical & patriarchal communities, 6 in less ones, men who were ranked as having more marital conflict with their spouses had higher testosterone than men in more harmonious relationships

Sharing and caring: Testosterone, fathering, and generosity among BaYaka foragers of the Congo Basin. Lee T. Gettler, Sheina Lew-Levy, Mallika S. Sarma, Valchy Miegakanda & Adam H. Boyette. Scientific Reports volume 10, Article number: 15422, September 22 2020. https://www.nature.com/articles/s41598-020-70958-3

Abstract: Humans are rare among mammals in exhibiting paternal care and the capacity for broad hyper-cooperation, which were likely critical to the evolutionary emergence of human life history. In humans and other species, testosterone is often a mediator of life history trade-offs between mating/competition and parenting. There is also evidence that lower testosterone men may often engage in greater prosocial behavior compared to higher testosterone men. Given the evolutionary importance of paternal care and heightened cooperation to human life history, human fathers’ testosterone may be linked to these two behavioral domains, but they have not been studied together. We conducted research among highly egalitarian Congolese BaYaka foragers and compared them with their more hierarchical Bondongo fisher-farmer neighbors. Testing whether BaYaka men’s testosterone was linked to locally-valued fathering roles, we found that fathers who were seen as better community sharers had lower testosterone than less generous men. BaYaka fathers who were better providers also tended to have lower testosterone. In both BaYaka and Bondongo communities, men in marriages with greater conflict had higher testosterone. The current findings from BaYaka fathers point to testosterone as a psychobiological correlate of cooperative behavior under ecological conditions with evolutionarily-relevant features in which mutual aid and sharing of resources help ensure survival and community health.

Discussion

A major goal of these analyses was to test for relationships between T and locally-defined measures of fathering quality in a small, egalitarian society. There has been limited past research on the biology of fatherhood in forager societies in which men are often committed to caring for their immediate families but are also contributors to the broader community’s pooled resources and shared engagement in collective caregiving36,38,39,52. We found that BaYaka fathers who were seen as better community sharers and those in less conflicted marriages, respectively, had lower T than their peers. Below, we contextualize our findings for BaYaka fathers and comparisons with their fisher-farmer Bondongo neighbors within theoretical frameworks and empirical research related to the psychobiology of family life, cooperative behavior, and competition/risk taking.

BaYaka fathering rankings and T

Consistent with our predictions, we found that BaYaka fathers who were ranked by their peers as being more generous sharers with the broader community had lower T than men who were seen as poorer sharers. Gettler has argued that one potential evolutionary and psychobiological implication of some men experiencing declines in T when they became committed fathers is that this might enhance their ability to cultivate social capital within their communities through greater cooperative and generous behavior9. This proposed neuroendocrine role of reduced paternal T fits within broader conceptual models that emphasize the importance of cooperation, reciprocal altruism, and empathy to the evolution of human life history13,14,26,32. Our finding for BaYaka fathers’ sharing is consistent with this framework. This result also had a meaningful, medium-level effect size (standardized B = -0.52) that well exceeds the effect sizes from recent meta-analyses on human T and fathering and past meta-analyses examining the positive associations between T, competition, and dominance behaviors7,53,54. To our knowledge no other research has specifically explored the relationship between fathers’ T and cooperative behavior, beyond nuclear family childcare roles9.

However, relevant research outside of family life has found that men with lower T commonly show greater generosity and empathic capacities10,11,12. Generally consistent with these patterns, a small number of recent studies have also found that lower T individuals have more social support, feel socially closer to others, and generate new friendships within social networks55,56,57. In contrast, experimental studies involving T administration have shown that men with elevated T will also circumstantially behave generously if such behavior helps increase social status58,59. In BaYaka communities, men are viewed positively for their generosity and likely gain benefits from community sharing (e.g. through reciprocity)60,61. Yet, in this egalitarian setting there are strong cultural values and practices geared towards reducing status hierarchy (see Methods)30,33,62. Thus, we suggest that our finding linking lower T and better sharing for BaYaka men aligns with social dynamics within their cultural context as well as past psychobiological research on T and generosity.

In our full models, BaYaka men who were ranked as better providers also tended to have lower T than men who were seen as poorer providers. This finding was not statistically significant and the effect was smaller (standardized B = -0.21), compared to the results for Share. This may reflect the strong positive relationship between men’s scores for Share and Provider and is potentially consistent with the idea that community members are basing BaYaka men’s rankings as providers at least partially on men’s sharing of acquired resources63. In addition, in multiple societies that still engage in foraging as a part of their routine subsistence, men who are better hunters often achieve some semblance of higher social status, despite egalitarian social norms, and hunting ability and reputation have been linked to higher fitness45. These competition- and status-related aspects of hunting in such societies could theoretically be linked to elevated T6,7,16. Indeed, among Amerindian Tsimane forager-horticulturalists men who achieved kills during hunts showed short-term increases in T, potentially reflecting these status-competition psychobiological effects37.

Meanwhile, in BaYaka communities, men who were skilled, successful elephant hunters (ntuma) are remembered historically as having had higher social standing in their communities. However, there were no ntuma in the community at the time of our data collection. Elephant hunting is now illegal and severely reduced in practice today and the importance of men’s hunting as a pathway to status is potentially attenuated in contemporary BaYaka communities. In comparison, local healers (nganga) and BaYaka council members remain prestigious positions within the contemporary community, and help foster cooperation through ceremony and conflict resolution64,65. Finally, we note that hunting is only one component of men’s provisioning in this setting, and our Provider ranking captures this broader breadth of subsistence activities, according to the participants’ characterizations30. Collectively, these issues help to highlight the relative dearth of studies on social neuroendocrine function among contemporary forager societies, limiting our ability to disentangle the importance of variation in status, cooperation and generosity, and paternal care in different societies36,38,39,52.

We also did not find a significant relationship between BaYaka men’s rankings as teachers and their T after adjustment for covariates. BaYaka fathers’ teaching involves direct interaction using a variety of behaviors, such as instruction and opportunity scaffolding66,67,68. Fathers have also been observed to do more teaching with very young children, whereas less of this “vertical transmission” from parents occurs for older children67,69. As direct caregivers, BaYaka fathers are often warm, nurturing, and patient33. This parenting style may be beneficial during teaching of young children and this type of nurturant direct care, especially with young children, is likely to be linked to lower paternal T, based on theory and past research elsewhere1,6,17,18. Thus the lack of a significant association and the relatively smaller effect size (standardized B = − 0.24) between men’s T and their teaching ran somewhat counter to our predictions. We do note that while teaching emerged as a locally-valued domain of fathering and may be significant for children developmentally70, BaYaka fathers may spend relatively less time engaged in teaching of their offspring, compared to other direct paternal care. For example, while Aka fathers were the second most important adult teachers of infants, their contribution was still much lower than those of mothers (12% vs. 59% of observed teaching). However, fathers’ teaching nearly exceeded all other adults combined (15%)66. Recent work by Lew-Levy and colleagues has shown that older BaYaka children and adolescents commonly accompany adults (not necessarily their parents) to forage71, and those trips afford opportunities for learning and teaching67,69. In future work, we hope to include further observational data on specific domains of men’s daytime direct care as well as family cosleeping, as these have been linked to lower or declining paternal T in multiple other settings20,22,23,24,25,72,73.

BaYaka and Bondongo fathering and T

Among BaYaka and Bondongo participants, there was a consistent emphasis that “good” fathers work to reduce conflict with their wives, as negative interactions between parents can lead to poor outcomes for children30,51,74. Despite this similar recognition of the importance of positive marital functioning in both societies, the two cultures have differing perceptions of the place of spousal conflict in day-to-day life, which align with broader cultural values in each community. Specifically, BaYaka are relatively gender egalitarian and value individual autonomy. They disapprove of conflict between wives and husbands and particularly have cultural mechanisms for attenuating and avoiding control of one partner over the other. Meanwhile, Bondongo communities are more hierarchical and patriarchal, thus men are generally higher in status and power than women, and disputes are seen as a part of normal day-to-day marital functioning30,51,74.

Based on these cultural differences, we predicted that Bondongo marital conflict would be more strongly related to men’s T than among BaYaka fathers. This prediction was not supported. Rather, we found a main effect for Dispute, such that men in both societies who were ranked as having more marital conflict with their spouses had higher T than men in more harmonious relationships. We also note that in the Dispute model focusing solely on BaYaka men we observed a complementary, stand alone finding relating greater marital conflict to higher T. These main effects potentially reflect that cultural variation in norms regarding the acceptability of marital conflict may have little effect on the actual frequency of such conflict. Our findings also align with findings elsewhere from a range of socio-ecological settings. For example, U.S. men with higher T reported that they felt less satisfied and committed to their relationships and had more marital conflict49,75, and in a large decade-long longitudinal U.S. study higher T men had greater risk of divorce48. Similarly, in a large longitudinal analysis from the Philippines, men with greater T functioning were more likely to experience relationship dissolution over a five-year period76. It is somewhat difficult to compare effect sizes across these varied studies. However, the bivariate correlation between marital conflict and men’s T in our combined sample (r = 0.33) is similar in size to the findings from Edelstein and colleagues’ (2014) research linking higher T to lower relationship satisfaction, investment, and commitment in the U.S. (rs = -0.29 to -0.36)49. In total, our findings add additional cross-cultural support to a growing body of literature linking higher T to poorer relationship functioning and outcomes in very different societies.

Finally, we did not find that the relationship between men’s Provider scores and T were significantly different between the two cultural groups. However, we do note that the slope of the lines relating men’s Provider rankings and T were in the opposite direction for men in the two communities (Bondongo: positive; BaYaka: negative). This reflects our past findings showing that Bondongo fathers who were rated as better providers had higher T than their peers77, and an opposite non-significant pattern linking lower T to higher Provider scores for BaYaka fathers in the present analyses. These patterns hint at the potential importance of cultural variation in social norms and complements foundational36 and recent39 anthropological work that similarly explored cross-cultural variation in fathers’ T.

Muller et al. (2009) found that Hadza forager fathers in Tanzania had lower T than non-fathers while among their Datoga pastoralist neighbors there was no significant difference for T between fathers and non-fathers. Building on this work, Alvarado et al. (2019) compared Hadza and Datoga men with Qom transitional foragers of Argentina. Relative to young Datoga fathers, Hadza and Qom men with children had lower T during their reproductive primes. The authors suggested that Datoga men’s elevated T as young fathers is particularly linked to the cultural practice of polygyny (i.e. involving competition) and the relative lack of routine contact between men and their families, due to men’s subsistence. Meanwhile, Qom and Hadza fathers more routinely engage in proximate interactions with their families and are generally serially monogamous39. In the present study, we need to be restrained in over-interpreting non-significant results. Yet, along this past research and other relevant work19,20,78, we hope our findings can help bring further attention to the potential importance of operationalizing cultural norms and practices in studies of social neuroendocrine function.

Limitations

There are limitations to the present study that merit attention. As we have discussed in past work from this study, our sample sizes of fathers were relatively small compared to some studies of paternal psychobiology in industrialized settings in more highly populated societies72,79,80,81,82. However, we also note that our pooled analyses (n = 45) and BaYaka sample size (n = 29) compare favorably to other recent work in this area25,72, especially research in similar societies39. That said, small sample sizes limit statistical power, as may have been the case in our moderation analyses predicting T from men’s provider rankings (Provider × ethnicity), and can also contribute to inflated effect sizes for statistically significant results83. While there has been substantial growth in the study of paternal psychobiology, much of the emerging research is in the U.S., Europe, and similar settings53,54. The present study makes an important, complementary contribution by focusing on these questions in two small-scale, subsistence-level societies, which differ politically, economically, and culturally from one another and from most prior study samples in this research area17. In that vein, as we have discussed in our past work from this site, smaller sample sizes are a research design trade-off that result from working at a highly remote field site with participants residing in small communities77. To that end, for each of the two communities, our sample of fathers represents ~ 90–100% of the eligible men in the village at the time of data collection. Moreover, to help attenuate sample size concerns, we collected repeated samples across participants and also used data analytical techniques that maximized the information from these repeated observations.

In addition, the BaYaka do not record their calendar ages, thus we estimated an approximate age based on a procedure from Diekmann and colleagues and with their assistance84. In validating their method, Diekmann et al. found the calculations were reliable within a year of known ages (median: 4 months; mean: 11 months) for another forager society, giving us confidence in the BaYaka calculated ages84. Still, because of inter-correlations between BaYaka age, T, and fathering rankings, the reliability of these age calculations could be potentially concerning in terms of adjusting our models for age. For BaYaka men, their calculated ages and T were qualitatively more strongly correlated (Rho = -0.46) than for the Bondongo (Rho = -0.34), who do record their ages. While populations can vary in age-related declines in T39,85, we suggest this is one helpful indicator of the validity of the calculated ages for BaYaka men. Moreover, the BaYaka age-T-rankings inter-correlations could pose potential issues for multi-collinearity for the independent variables in our regression models. Following each of our regression models, we calculated variance inflation factors (VIF) for the predictors. Although conventions can vary, it is common to use VIF of > 10 as an indicator of concerning multi-collinearity86, and the calculated values for the present analyses were well below this threshold, as we reported in the Results.

Finally, there was a minor unintended difference in the handling of the saliva samples our team collected from the participants in the two communities. During both field seasons (see Methods), we froze the saliva samples on site in portable liquid nitrogen dewars, and they remained frozen until shipment. The Bondongo samples were kept frozen throughout their transport to the U.S. During shipment of the BaYaka samples, we encountered a logistical problem, which resulted in the saliva samples going through a freeze–thaw cycle while in transit. All the BaYaka samples were exposed to identical conditions, and salivary T is generally robust to limited freeze–thaw cycles and short-term exposure to ambient temperature87,88, which attenuates concerns over this issue.