Hadza Hunter-gatherers Disagree on Perceptions of Moral Character

Smith, Kristopher M., and Coren L. Apicella. 2019. “Hadza Hunter-gatherers Disagree on Perceptions of Moral Character.” PsyArXiv. March 29. doi:10.31234/osf.io/er4jb

Abstract: To the extent that moral character is grounded in stable and observable truths, there should exist agreement between people in their judgements of others’ character. In Western populations, this agreement is found. We examine whether this is universal in Hadza hunter-gatherers of Tanzania. Ninety-four judges ranked their campmates on global character and relevant character traits for a total of 824 observations. Judges disagreed on rankings of global character, generosity, and honesty, but agreed more on hard work and hunting ability. Individual rankings on specific traits predicted character evaluations. There was agreement between judges on the extent to which generosity and hard work related to character. These findings suggest that Hadza have shared beliefs about what traits constitute character, but disagree on which of their campmates exhibit these traits. We discuss these findings in light of other research suggesting that stable moral dispositions may not be universal.

---
Evaluation of moral character is an important component of person perception (Goodwin,
2015; Goodwin, Piazza, & Rozin, 2014). When learning about a new person, we seek
information about whether they are trustworthy (Brambilla, Rusconi, Sacchi, & Cherubini,
2011). When identifying features and traits most relevant to identity, people consider morality to
be an essential component (Strohminger & Nichols, 2014). And when considering what
attributes a partner should have in different types of relationships, morally relevant features, such
as trustworthiness, are most important (Cottrell, Neuberg, & Li, 2007). This makes sense because
a person’s character is used to infer their intentions toward us and whether they would help or
hinder our goals (Landy, Piazza, & Goodwin, 2016). Indeed, people use information about moral
character to decide who to interact and cooperate with (Everett, Faber, Savulescu, & Crockett,
2018; Martin & Cushman, 2015; van der Lee, Ellemers, Scheepers, & Rutjens, 2017).
Despite the importance of moral character in person perception, some have argued that
character does not exist and that people do not have stable moral dispositions  (Doris, 2002;
Harman, 2003). Social psychologists and philosophers have used classic findings from social
psychology, such as the bystander effect (Darley & Latané, 1968; Latané & Darley, 1968) or the
good Samaritan experiment (Darley & Batson, 1973), to argue against the existence of moral
character and that moral behavior is determined wholly by the situation (Doris, 2002; Harman,
2003). One way to determine whether people behave similarly across situations is to examine
agreement between independent observers. Because different observers are likely to interact with
the target in different situations, if they agree in their evaluations, it then suggests there is a
stable disposition that is being observed (Kenrick & Funder, 1988). So, if independent observers
have similar perceptions of targets’ moral character, then it provides some evidence for the
existence of moral character.
People generally agree on who does, and does not, have moral character. Self-report and
informant ratings of morally-relevant traits, such as honesty or guilt-proneness, moderately
correlate (Cohen, Panter, Turan, Morse, & Kim, 2013). Independent observers also agree on
global evaluations of moral character, as well as specific moral traits and trait profiles. (Helzer et
al., 2014). And people agree on morally relevant traits displayed by respected cultural figures,
even across the US political divide (Frimer, Biesanz, Walker, & MacKinlay, 2013). Again, this
agreement is used as evidence that moral character exists.
Like much of behavioral and social science research, samples in studies of moral
psychology have largely been drawn from Western, Educated, Industrialized, Rich, and
Democratic—or WEIRD—societies (Ellemers, van der Toorn, Paunov, & van Leeuwen, 2019;
Henrich, Heine, & Norenzayan, 2010). Despite this, the importance of moral character in identity
and person perception is theorized to be universal (Strohminger, Knobe, & Newman, 2017).
Most often, when moral psychology is examined in other cultures, the emphasis is on the content
of moral norms and the shared or unique prescriptions and prohibitions across cultures (Curry,
Mullins, & Whitehouse, 2019; Haidt, Koller, & Dias, 1993; Purzycki et al., 2018). Yet, when
research has looked at processes in moral judgments, important differences have been found. For
example, whether a wrong is done intentionally is an important distinction in moral judgments
among Western populations, presumably because it reveals information about moral character
(Landy & Uhlmann, 2018). However, unintentional violations are judged as wrong as intentional
violations in some cultures, including the Hadza and South Pacific islanders (Barrett et al., 2016;
McNamara, Willard, Norenzayan, & Henrich, 2019). To our knowledge, no research has been
conducted on perceptions of moral character in small-scale societies.
There are reasons to suspect important differences in moral character and its perception
in small-scale societies. First, there is some evidence for less personality variation in non
WEIRD societies. For example, personality traits in the Tsimané forager-horticulturalists of
Bolivia do not cluster into five distinct factors, but rather two, and there is less variation within
those factors compared to Western samples (Gurven, von Rueden, Massenkoff, Kaplan, & Vie,
2013). And in fact, across 55 nations, populations with fewer economic opportunities to
specialize have less variation in personality traits (Lukaszewski, Gurven, von Rueden, &
Schmitt, 2017). To the extent that there is a relationship between personality traits and moral
character (Thalmayer, Saucier, Srivastava, Flournoy, & Costello, 2019), we might then similarly
expect less variation in morally-relevant character traits. Second, there is no evidence for
generous dispositions in small-scale societies. In longitudinal data among the Hadza,
contributions to a public good game were not predicted by previous contributions, but rather the
contributions of an individual’s campmates (Smith, Larroucau, Mabulla, & Apicella, 2018).
Here, strong, local norms governing generosity may be reducing individual variation in morally
relevant behavior leading to a lack of agreement on perceptions of moral character.
In the current study, we examine perceptions of moral character among the Hadza of
Tanzania, one of the last remaining hunter-gatherer groups in the world. The Hadza are an ideal
population because they live in small groups of known individuals where behavior is observable,
and because of their harsh environment, knowing who is moral would be seemingly important.
We examine agreement on these perceptions in two ways. First, do Hadza agree on who has
moral character? And second, do Hadza agree on what traits contribute to global moral
character? To answer these questions, we ask the Hadza to rank their campmates on moral
character, as well as specific traits of hard work, generosity, and honesty. We examine the
consensus within each camp on rank orderings for each trait to answer the first question. We
examine the relationship between the specific traits and global character rankings and the
variation between Hadza on the importance of the specific traits in determining global character
to answer the second question.


Discussion

 In WEIRD societies, people evaluate the moral character of others and use those
perceptions to decide with whom to interact. Underscoring the importance of character in these
populations, independent observers agree on how moral others are (Helzer et al., 2014). But is
this universal? To answer this, we asked if Hadza hunter-gatherers agree on who is moral and
what traits make someone moral. The Hadza disagree on which of their campmates have a good
heart, are generous, and are honest, and agree more on which campmates are hard working
(effort) and produce the most food (hunting ability). At the level of the population, hard work,
generosity, and honesty contribute to global character; however, there is variation between
Hadza judges on how much honesty contributes to global character, though judges agree more on
how much hard work and generosity contribute to character. Overall, these results suggest that
Hadza use some of the same criteria—hard work and generosity—for evaluating moral character,
but disagree on who displays those traits, leading to disagreement on global character
perceptions. 
 Agreement between independent observers on ratings about a trait is taken as evidence
for that trait existing because raters are likely observing the same behaviors despite being in
different situations (Kenrick & Funder, 1988). The disagreement between Hadza judges on
character traits suggests that Hadza do not have moral dispositions. However, disagreement does
not definitively rule out the existence of moral character. For example, the Hadza may have been
unwilling to make assessments about their campmates’ character, though notably we do see
agreement on hunting ability, which is highly valued in the Hadza. Or there could be
disagreement because there are not many opportunities to display moral behavior; however, it
should be easy to observe moral behavior because they live together in small groups and depend
on each other for survival.
 One alternative interpretation of the data is that the Hadza can agree on moral character,
and in fact they do have moral dispositions, but that our measure is unreliable and cannot detect
agreement. A good measure measuring a phenomenon that does not exist and a bad measure
measuring a phenomenon that does exist will produce the same result: noise. However, we argue
there are two reasons to suspect that our measure would be reliable enough to detect agreement
on moral character if it existed. First, we were able to detect moderate relationships between the
specific character traits and moral character, indicating reliability was not so low as to be unable
to detect any effects. Second, we did find moderate agreement on hard work and hunting ability.
And in fact, given what we know about the noisy relationship between hunting returns and
hunting reputation (Stibbard-Hawkes et al., 2018), the fact that we were able to detect agreement
suggests low reliability can not fully explain the disagreement in perceptions of moral character. 
 It may seem that hunting ability would be easily observable, but in the anthropological
literature, this is notoriously difficult to measure, and because of this hunting reputation is
criticized as a measure of hunting success (Hill & Kintigh, 2009). First, hunting ability is rarely
directly observed, as most hunting happens alone. And second, there is high variance in hunting
returns, in which men return to camp with nothing on most days, but occasionally (about 3% of
days) bring in large game (Hawkes, O’Connell, & Blurton-Jones, 1991). In fact, for
anthropologists to reliably estimate hunting ability using hunting returns, they need 200 to 600
days of observations (Hill & Kintigh, 2009). Despite this, in our study and others (Stibbard
Hawkes et al., 2018), the Hadza are able to agree on who the best hunters are, and hunting
reputation does relate to proxies of actual hunting ability, such as strength, accuracy, and
ecological knowledge (Apicella, 2014; Stibbard-Hawkes et al., 2018). This suggests that if there
are moral dispositions among the Hadza, the signal is much weaker than that of hunting ability,
which is itself a noisy signal (Stibbard-Hawkes, 2019). And in fact, if it is this hard to detect
moral dispositions, it then raises the question of whether the Hadza can reliably determine
character enough to provide useful social information.
 Data measuring morally-relevant behavior, such as generosity, further suggest a lack of
moral dispositions in the Hadza and other non-WEIRD populations. In a longitudinal study, a
Hadza’s previous generosity in an economic game did not predict their subsequent contributions,
and instead the only significant predictor was how much his or her campmates contribute (Smith
et al., 2018). And in a small study (n = 12) of the Tsimané of Bolivia, generosity in a dictator
game in one year did not predict generosity in a later year (Gurven, 2014).  
 These results further support recent research finding that character and moral reputation
do not play a role in Hadza campmate preferences. When asked who they prefer to live with,
Hadza do not choose the most generous people, whether generosity is measured using an
economic game (Apicella, Marlowe, Fowler, & Christakis, 2012) or via reputation (Smith &
Apicella, 2019). Rather, Hadza prefer to live with better hunters (Smith & Apicella, 2019; Wood,
2006). If moral behavior changes across time and situations as our results here suggest, then
choosing campmates based on their current behavior is useless. Instead, traits related to
productivity, such as being a hard worker or a good hunter, may become more important in
campmate preferences (Barclay, 2016); if everyone is expected to share because of strong norms,
such as in the Hadza, then choosing productive campmates is more important. And in fact, a
preference for productive partners may influence friendships in Western societies. People prefer
partners in economic games and are more generous to partners who are perceived to be more
productive, even though it is irrelevant to the game (Eisenbruch, Grillot, Maestripieri, & Roney,
2016; Eisenbruch & Roney, 2017). The effect of a productivity preference in various
relationships may be a fruitful area for future research. 
 Throughout the latter half of the 20th century, the situationist paradigm in social
psychology casted doubt on the existence of moral character. However, more recent research in
moral psychology has argued that moral character does in fact exist (Fleeson, Furr,
Jayawickreme, Meindl, & Helzer, 2014). In Western societies, people agree on who is moral
(Helzer et al., 2014), and perceptions of moral character play an important role in social
cognition (Goodwin, 2015; Landy & Uhlmann, 2018). Our results here question the universality
of moral character and its centrality in social life, and highlights the importance of cross-cultural
research using underrepresented samples. By conducting research with populations in a variety
of socio-ecologies, we can better understand the variation in our moral psychology.