Placing morality in the context of evolutionary theory

From an evolutionary theory perspective, theories of human morality draw primarily from research on group-level selection (Smith, Reference Smith2020). Under this paradigm, MFT views morality as a complex function of evolved psychological mechanisms and their genetic, cognitive, and cultural correlates that increase cooperation and group fitness, i.e. the ability for extended members of a family, clan, tribe, or society to function better (reproductively fitter) than another family, clan, tribe, or society (Curry, Reference Curry2016; Curry, Mullins, & Whitehouse, Reference Curry, Mullins and Whitehouse2019; Henrich & Henrich, Reference Henrich and Henrich2006). Common components of these mechanisms include empathy, the basic emotions (e.g. fear), the moral emotions (e.g. guilt and shame), cultural norms and taboos, and reward and punishment (e.g. punitive deterrents). Importantly, the functional fitness gained from such mechanisms refers to incremental gains throughout evolutionary history and does not necessarily have anything to say about how beneficial these evolved psychological and cultural mechanisms are in the contexts of modern-day societies (Gross, Reference Gross1998; Richerson & Boyd, 2006; Tooby & Cosmides, Reference Tooby and Cosmides1990).

Evolutionary theory also suggests certain sex differences in moral intuitions, and these differences might be primarily derived from differences in reproductive strategies. For example, sexual strategies theory suggests that males are biased towards short-term promiscuous sexual strategies whereas females are biased towards long-term and more committed sexual strategies (Buss, Reference Buss2016), and different moral intuitions will be more or less aligned with these strategies. For example, empathy and loyalty are consistent with long-term strategies whereas autonomy is more consistent with a short-term strategy.

Moral foundations theory therefore attempts at an evolution-based explanation for many of the moral phenomena demonstrated in experimental studies and everyday life, e.g. the tendency to rapidly judge others’ behaviours as right or wrong (good or bad), the increased emotional salience of moral violations when compared with social norm violations, and the persistence with which people will maintain their moral beliefs even in the face of evidence against their rational concerns, a.k.a., moral dumbfounding (McHugh et al., Reference McHugh, McGann, Igou and Kinsella2017). Moral foundations theory researchers have systematised the testing of these theoretical foundations of moral concern using five specific criteria: (1) evidence of stimulating a common concern in third-party normative judgements; (2) evidence of automatic affective stimulation; (3) evidence of widespread cultural distribution; (4) evidence of some form of innate preparedness; and (5) evidence that an evolutionary model demonstrates an adaptive advantage. For more elaboration on the development of these criteria and examples for how they are met, see Graham et al. (Reference Graham, Iyer, Nosek, Haidt, Koleva and Ditto2011, Reference Graham, Haidt, Koleva, Motyl, Iyer, Wojcik and Ditto2012).

The six most widely accepted and studied foundations of MFT are care, justice, loyalty, authority, sanctity and autonomy. However, more recent studies of adults have shown that models that split the care foundation into animal physical harm, human physical harm and human emotional harm exhibit better model fit statistics (Clifford et al., Reference Clifford, Iyengar, Cabeza and Sinnott-Armstrong2015). Prior studies of adolescents have assumed a latent factor structure of moral intuitions based on the original six factors mentioned above (Bespalov et al., Reference Bespalov, Prudnikova, Nyamdorj and Vlasov2017; Cingel & Krcmar, Reference Cingel and Krcmar2020; Joeckel et al., Reference Joeckel, Bowman and Dogruel2012, Reference Joeckel, Bowman and Dogruel2013; Tamborini, Reference Tamborini2011; Valkenburg & Peter, Reference Valkenburg and Peter2013). However, in more recent studies designed specifically to assess the latent factor structure of adolescents, only three of these foundations found psychometric support consistent with MFT: autonomy, loyalty and animal physical harm (Bretl, Reference Bretl2020). Therefore, the present study explored the theoretical potential of considering these three foundations as primary components of adolescent moral development by testing for psychometric validity and comparing their latent factor mean differences based on age and sex.

Autonomy, loyalty and empathy in developmental context

For this study, we explored the possibility of considering the three factors exhibiting prior psychometric validity with adolescent populations (autonomy, loyalty and animal physical harm) as a triad of primary latent factor constructs relevant to adolescent moral development. Ratings of animal physical harm were considered a proxy for empathy because links between empathy development and judgements of animal cruelty have empirical support (Eisenberg-Berg & Mussen, Reference Eisenberg-Berg and Mussen1978; McPhedran, Reference McPhedran2009; Taylor & Signal, Reference Taylor and Signal2005). Indeed, the animal physical harm vignettes used in the present study were especially well suited to assess an empathy construct because they theoretically allowed for the bypassing of intentionality and culpability processing. The important role of intentionality and culpability processing in moral judgements and the ability to bypass such processing has been demonstrated experimentally as well (Alicke, Reference Alicke2000; Clark et al., Reference Clark, Chen and Ditto2015; Cushman, Reference Cushman2008; Davies & Rogers, Reference Davies and Rogers2009).

Autonomy is often considered the goal of adolescent development, and thousands of research papers attest to the importance of this construct across cultural settings (Smetana, Reference Smetana2017; Soenens et al., Reference Soenens, Vansteenkiste and Van Petegem2017; Zimmer-Gembeck et al., Reference Zimmer-Gembeck, Collins and Adams2003). Previous research using autonomy violation vignettes to assess moral judgements has demonstrated that judgements of autonomy violations seem to be especially sensitive to the degree of kinship between actors. For example, moral violation ratings seem to be particularly sensitive to autonomy violations between immediate family members – and even more precisely sensitive to mother/daughter relationships, thus suggesting possible evolutionary correlates (Curtin et al., Reference Curtin, Barrett, Bolyanatz, Crittenden, Fessler, Fitzpatrick and Henrich2020; Lieberman et al., Reference Lieberman, Tooby and Cosmides2007).

Loyalty can be seen as the antithesis to autonomy development, as loyalty involves the balancing of commitments to others with commitments to self. Often described as ‘the vexing virtue’, there is good reason for adolescents to be especially sensitive to loyalty violations (e.g. betrayal) as they navigate increasingly meaningful peer relationships and at the same time seek to increase their independence (autonomy).

Unfortunately, little research to date has studied these developmental constructs under an MFT-based paradigm. Nonetheless, what can be gleaned from the scant literature suggests that the latent factor constructs of autonomy, loyalty and empathy as a triad of primary constructs related to adolescent moral development provide a promising starting point for future research investigating innate correlates of moral development during the transition from childhood to adulthood.

Age differences and hormonal influence

Hormonal influence on cognition and brain development has been generally understudied despite overwhelming evidence that there is a lot of meaningful sex-hormone-dependent brain development that takes place (Cahill, Reference Cahill2018). Indeed, pubertal onset is when many sex-based brain structure and function differences emerge (Goddings et al., Reference Goddings, Beltz, Peper, Crone and Braams2019). Of relevance to general social processing and affectively mediated moral intuitions, a common theme of U-shaped and inverted U-shaped patterns of affective reactivity during adolescent development have been shown to be influenced by sex hormones. For example, inverted U-shaped peaks in amygdala and hippocampus responses to fearful facial expressions and sex differences in social- and self-evaluative processing brain regions have been found. However, testosterone levels in females exhibited an inverse pattern (U-shaped) in these neural responses correlated with pubertal stage (Vijayakumar et al., Reference Vijayakumar, Pfeifer, Flournoy, Hernandez and Dapretto2019). All of this is to suggest the possibility of meaningful correlations between pubertal sex hormones and affectively mediated moral intuitions.

More specifically, previous research suggests that early adolescence is a critical period for the maturation of brain areas that are involved in both socio-cognitive and affective processing, and these developmental changes may lead to fluctuating empathy levels during early adolescence (Blakemore & Choudhury, Reference Blakemore and Choudhury2006). Interestingly, previous empirical findings suggest a temporal decline in general empathy during early adolescence (Goering & Mrug, Reference Goering and Mrug2021; Van der Graaff et al., Reference Van der Graaff, Branje, De Wied, Hawk, Van Lier and Meeus2014). Concurrently, early adolescence is a critical period for the nascent development of autonomy from parents as youth begin to detach from caregivers and spend more time with their peers (Meeus, Reference Meeus2016). In turn, adolescents’ feelings of loyalty, specifically to other adults, may also decrease – which often manifests as parent–child conflict and increased feelings of adults as out-group ‘others’ during early adolescence (Marceau et al., Reference Marceau, Ram and Susman2015). Thus, we might expect to see higher sensitivity to violations of autonomy and lower sensitivity to violations of loyalty corresponding with pubertal onset; however, it should be noted that this would only hold true for moral violations committed by adults.

On average, the onset of gonadarche is about age 11 for females and about age 13 for males, and the activation of the growth hormone axis (a primary consequence of puberty) is about age 12.5 for females and about age 13.8 for males (Figure 1) (Hansen et al., Reference Hansen, Bretl and Amini2019). Therefore, we hypothesised that sensitivity to violations of autonomy would be particularly pronounced in 11–12-year-old females and in 13–14-year-old males owing to the onset of pubertal processes. We also hypothesised a decrease in loyalty sensitivity in these same groups owing to the fact that the moral violation vignettes in the current study involve adult perpetrators.

Figure 1. CFS-3 latent factor mean comparisons between age and sex groups.

Sex differences

Sex differences in empathy, autonomy and loyalty during early adolescence have been reported in prior research. In general, early adolescent females have higher levels of empathy compared with their male counterparts, and males experience a stronger temporal decline in empathy during early adolescence (Van der Graaff et al., Reference Van der Graaff, Branje, De Wied, Hawk, Van Lier and Meeus2014). Autonomy development has been specifically linked to the onset of puberty, and therefore females may develop concerns for autonomy earlier than their male counterparts as a result of having earlier pubertal timing (Branje, Reference Branje2018; Negriff & Susman, Reference Negriff and Susman2011), and indeed, research in this area has demonstrated that females have higher levels of autonomy and self-reliance compared with males during early adolescence (Steinberg & Silverberg, Reference Steinberg and Silverberg1986).

A meta-analysis on sex differences in friendship expectations found that females, in general, put greater emphasis on peer relationships, operationalised as feelings of loyalty, compared with males (Hall, Reference Hall2011). There is also evidence that males and females exhibit different degrees of betrayal sensitivity, with females showing greater sensitivity (Atlas & Ingram, Reference Atlas and Ingram1998; Burton et al., Reference Burton, Halpern-Felsher, Rankin, Rehm and Humphreys2011; Feldman & Cauffman, Reference Feldman and Cauffman1999; Gobin & Freyd, Reference Gobin and Freyd2009; Keng et al., Reference Keng, Noorahman, Drabu and Chu2019; Leets & Sunwolf, Reference Leets and Sunwolf2005; Singer & Doornenbal, Reference Singer and Doornenbal2006). However, relatively few studies have examined sex differences in loyalty expectations in early adolescence outside the realm of traumatic betrayal (Blakemore & Choudhury, Reference Blakemore and Choudhury2006).

Current study

This study sought to confirm the psychometric validity of these three latent factors as assessed by the 3-Factor Character Foundations Survey (CFS-3). The study then sought to determine if there exist any significant differences between males and females during the developmental period of early adolescence and between two distinct age groups (11–12-year-old and 13–14-year-old). Based on prior research, it was hypothesised that compared with males, females would have a higher latent factor mean rating in the animal physical harm factor (which serves as a proxy for empathy), a higher latent factor mean in the autonomy factor, and a higher latent factor mean in the loyalty factor. It was also hypothesised that 11–12-year-old females and 13–14-year-old males (when compared with same sex, different age groups) would have higher ratings in the autonomy factor but lower ratings in the loyalty factor owing to pubertal timing. In short, we hypothesised that there would be significant differences based on age and sex, yet owing to the lack of prior developmental research under an MFT-based paradigm and no direct measure of pubertal stage, our primary goals were exploratory in nature.