Top-down approaches to primate personality offer insights into the evolution and diversity of personality traits within the primate order (Weiss, Reference Weiss, Vonk, Weiss and Kuczaj2017). Comparative approaches also allow us to address interspecific differences in endocrinological (Wilson, Guenther, Øverli, Seltmann, & Altschul, Reference Wilson, Guenther, Øverli, Seltmann and Altschul2019), genetic (von Borell, Weiss, & Penke, Reference von Borell, Weiss and Penke2019) and physical phenotypic correlates of behaviour (Kern, Robinson, Gass, Godwin, & Langerhans, Reference Kern, Robinson, Gass, Godwin and Langerhans2016). Such comparison is key to understanding variance in fitness outcomes such as stress responses, reproductivity and survival (Blaszczyk, Reference Blaszczyk2020). Within this framework, signal Footnote 1 strength is an important variable, since social signals are determinants of partner preference and combative encounters, which can affect fitness.
Amongst primates, faces are an important consideration for signalling, as they provide a wealth of information that can help inform social decisions. Primate faces are complex in terms of shape, colouration and the presence of facial hair, and all these components might be relevant for signalling (Waller, Kavanagh, Micheletta, Clark, & Whitehouse, Reference Waller, Kavanagh, Micheletta, Clark and Whitehouse2022). Dynamically, faces can express emotions and provide other information (Waller, Whitehouse, & Micheletta, Reference Waller, Whitehouse and Micheletta2017), which in primates can be used to mediate exchanges across the dominance hierarchy, such as using fear grins to show submissive (Maestripieri & Wallen, Reference Maestripieri and Wallen1997) and affiliative (Waller & Dunbar, Reference Waller and Dunbar2005) behaviour. Static faces can communicate information about fertility and fitness, such as reproductive-state-related changes in skin redness of rhesus macaques (Macaca mulatta), with male seasonal colouration changes linked to female mate preference (Dubuc, Allen, Maestripieri, & Higham, Reference Dubuc, Allen, Maestripieri and Higham2014). They have also been found to advertise dominance, via facial colouration in mandrills (Mandrillus sphinx) (Setchell, Smith, Wickings, & Knapp, Reference Setchell, Smith, Wickings and Knapp2008) and canine size in baboons (Papio sp.) (Galbany, Tung, Altmann, & Alberts, Reference Galbany, Tung, Altmann and Alberts2015). Thus, primate faces appear to contain multi-component information that is not only used in direct communication but also can indicate fitness to potential mates and competitors via sexually selected traits.
One topic that has been much debated in the human literature is the link between dominance and facial morphology. Starting with the work of Weston, Friday, Johnstone, and Schrenk (Reference Weston, Friday, Johnstone and Schrenk2004), who found an inverse relationship between canine height sexual dimorphism and bizygomatic width dimorphism amongst primates, the hypothesis emerged that in some primate species, dimorphism of facial width evolved as a result of intra-sexual selection by male–male competition (Weston et al., Reference Weston, Friday, Johnstone and Schrenk2004, Weston, Friday, & Liò, Reference Weston, Friday and Liò2007), whereby a wider zygomatic arch provides a combat advantage via skull strength (Lefevre et al., Reference Lefevre, Wilson, Morton, Brosnan, Paukner and Bates2014; Stirrat, Stulp, & Pollet, Reference Stirrat, Stulp and Pollet2012). This was supported by multiple human studies linking male facial width, typically measured as facial width-to-height ratio (fWHR, Fig. 1a) (Weston et al., Reference Weston, Friday and Liò2007), to dominance (mostly self-reported or inferred, but also estimated from military rank and economic game outcomes) (Geniole, Denson, Dixson, Carré, & McCormick, Reference Geniole, Denson, Dixson, Carré and McCormick2015; Polo et al., Reference Polo, Muñoz-Reyes, Valenzuela, Cid-Jofré, Ramírez-Herrera and Pavez2022) and aggression (Goetz et al., Reference Goetz, Shattuck, Miller, Campbell, Lozoya, Weisfeld and Carré2013; Haselhuhn, Ormiston, & Wong, Reference Haselhuhn, Ormiston and Wong2015; Stirrat et al., Reference Stirrat, Stulp and Pollet2012; Třebický, Havlíček, Roberts, Little, & Kleisner, Reference Třebický, Havlíček, Roberts, Little and Kleisner2013; Wen & Zheng, Reference Wen and Zheng2020; Zilioli et al., Reference Zilioli, Sell, Stirrat, Jagore, Vickerman and Watson2015). Wider faces are also perceived as more dominant (Alrajih & Ward, Reference Alrajih and Ward2014; Lefevre & Lewis, Reference Lefevre and Lewis2014; Mileva, Cowan, Cobey, Knowles, & Little, Reference Mileva, Cowan, Cobey, Knowles and Little2014), acting as cues or even signals of individual dominant/aggressive personality traits.
However, the results of studies of facial-dominance links in humans are often inconclusive and impeded by methodological issues. Firstly, several null findings are inconsistent with the proposed link between bizygomatic width and dominance (Özener, Reference Özener2012; Wang, Nair, Kouchaki, Zajac, & Zhao, Reference Wang, Nair, Kouchaki, Zajac and Zhao2019), including mixed findings for facial perceptions (Durkee & Ayers, Reference Durkee and Ayers2021) and sexual dimorphism of facial morphology (Kramer, Reference Kramer2017; Lefevre et al., Reference Lefevre, Lewis, Bates, Dzhelyova, Coetzee, Deary and Perrett2012; Penton-Voak et al., Reference Penton-Voak, Jones, Little, Baker, Tiddeman, Burt and Perrett2001; Summersby, Harris, Denson, & White, Reference Summersby, Harris, Denson and White2022; Wen & Zheng, Reference Wen and Zheng2020), as well as limited evidence linking facial width to combat success (Carré & McCormick, Reference Carré and McCormick2008; Stirrat et al., Reference Stirrat, Stulp and Pollet2012; Třebický et al., Reference Třebický, Havlíček, Roberts, Little and Kleisner2013; Zilioli et al., Reference Zilioli, Sell, Stirrat, Jagore, Vickerman and Watson2015). Secondly, the underlying mechanisms that could drive this relationship are poorly understood. There is little evidence for testosterone providing concurrent effects on craniofacial growth and behavioural traits (Bird et al., Reference Bird, Cid, Geniole, Welker, Zilioli, Maestripieri, Arnocky and Carré2016; Eisenbruch, Lukaszewski, Simmons, Arai, & Roney, Reference Eisenbruch, Lukaszewski, Simmons, Arai and Roney2017; Hodges-Simeon, Sobraske, Samore, Gurven, & Gaulin, Reference Hodges-Simeon, Sobraske, Samore, Gurven and Gaulin2016; Kordsmeyer, Freund, Pita, Jünger, & Penke, Reference Kordsmeyer, Freund, Pita, Jünger and Penke2019). Thirdly, the use of ratio measures can be problematic, leading to spurious correlations (Kronmal, Reference Kronmal1993). Fourthly, most research on facial dimorphism and links to behaviour or personality remains in the human domain. Expanding this to comparative work could help us understand which selection pressures might drive facial cues of dominance or other traits (Wilson, Weiss et al., Reference Wilson, Weiss, Lefevre, Ochiai, Matsuzawa, Inoue-Murayama, Freeman, Herrelko and Altschul2020).
In the remainder of this paper, we discuss studies conducted in nonhuman primates on this topic: focusing on static faces; discussing the potential selection pressures and underlying mechanisms; and making proposals for future research that could benefit a broader understanding of morphological cues to personality.
1. Facial morphology and behavioural correlates in nonhuman primates
Early studies on apes, focusing predominantly on cranial measurements, have revealed varying levels of sexual dimorphism. Gorillas (Gorilla gorilla, subspecies not specified) exhibit the strongest cranial sexual dimorphism amongst hominids (excluding bonobos, which were not tested) with low dimorphism in chimpanzee skull morphology (O’Higgins & Dryden, Reference O’Higgins and Dryden1993). Orangutans also display strong sexual dimorphism in cranial morphology, which begins in infancy, with males exhibiting a continuous growth of the zygomatic bone from early adolescence into adulthood (Hens, Reference Hens2005). Male orangutans also develop cheek flanges, secondary sexual characteristics that are linked to androgen levels (Marty et al., Reference Marty, van Noordwijk, Heistermann, Willems, Dunkel, Cadilek, Agil and Weingrill2015), and silverback gorillas develop sagittal crests of fatty and fibrous tissue (Breuer, Robbins, & Boesch, Reference Breuer, Robbins and Boesch2007). These sexually dimorphic features are correlated with their reproductive success (Banes, Galdikas, & Vigilant, Reference Banes, Galdikas and Vigilant2015; Breuer, Robbins, Boesch, & Robbins, Reference Breuer, Robbins, Boesch and Robbins2012).
More recent studies have focused on facial morphological links to behaviour, specifically to personality traits (individual behavioural characteristics consistent across time; capitalised further in the text, e.g., Dominance) and to characteristics of dominance hierarchies (relative ranking of individuals dependent on the outcomes of their present agonistic interactions; labelled further in the text with lowercase, e.g., dominance). Personality structures of nonhuman primates assessed by trait rating (specifically using the Hominoid Personality Questionnaire; Weiss, Reference Weiss, Vonk, Weiss and Kuczaj2017), unlike the human five-factor model (McCrae & Costa, Reference McCrae and Costa1987), include a separate Dominance-related personality dimension (labelled as Assertiveness or Dominance depending on the species, see Table 1). This separate Dominance-related personality dimension, indicating dominant and often aggressive social tendencies (except for Assertiveness in bonobos reflecting affiliative dominant tendencies; Martin, Staes, Weiss, Stevens, & Jaeggi, Reference Martin, Staes, Weiss, Stevens and Jaeggi2019), evolved as a result of the central role of dominance interactions in nonhuman primate social interactions (Weiss, Reference Weiss2022). Personality structures of some species, however, include more than one Dominance-related dimension. Similar to chimpanzees (Weiss et al., Reference Weiss, Feldblum, Altschul, Collins, Kamenya, Mjungu, Foerster, Gilby, Wilson and Pusey2023), rhesus macaque Assertiveness is related to dominance in social interactions (Kohn et al., Reference Kohn, Snyder-Mackler, Barreiro, Johnson, Tung and Wilson2016), whilst Confidence refers to confidence in the presence of environmental or social stressors (Adams et al., Reference Adams, Majolo, Ostner, Schülke, De Marco, Thierry, Engelhardt, Widdig, Gerald and Weiss2015). The characteristics of dominance hierarchies, assessed by observing outcomes of agonistic interactions, are usually expressed as rank (called “dominance status” in Altschul, Robinson, Coleman, Capitanio, & Wilson, Reference Altschul, Robinson, Coleman, Capitanio and Wilson2019; “alpha status” in Lefevre et al., Reference Lefevre, Wilson, Morton, Brosnan, Paukner and Bates2014; “agonistic dominance” in Martin et al., Reference Martin, Staes, Weiss, Stevens and Jaeggi2019).
Note. Facial morphology measures (see Fig. 1a for definition) were taken from facial photographs. Grey cells indicate the effect size for sexual dimorphism. Blue cells indicate effect size for behavioural measures. Values in bold indicate where coefficients were calculated from raw data. Rating of personality traits was conducted using the Hominoid Personality Questionnaire (Weiss, 2017). Results were only significant in individuals:
*Data represent 9 species of the Macaca genus.
**<8 years,
***>8 years,
****in females.
In brown capuchins (Sapajus apella), fWHR is sexually dimorphic, with higher ratios for males and alpha individuals, and correlates with the personality component Assertiveness (Lefevre et al., Reference Lefevre, Wilson, Morton, Brosnan, Paukner and Bates2014). Alpha males in particular are observed to have bulkier facial features, an effect that appears to be driven by direct access to females (Paukner et al., Reference Paukner, Slonecker and Wooddell2021). Higher fWHR was associated with despotic social style in the Macaca genus (Borgi & Majolo, Reference Borgi and Majolo2016) and with Assertiveness in immature rhesus macaques (Altschul et al., Reference Altschul, Robinson, Coleman, Capitanio and Wilson2019). fWHR has also been associated with Assertiveness (reflecting affiliative dominance) and rank (reflecting agonistic dominance) in bonobos (Pan paniscus) (Martin et al., Reference Martin, Staes, Weiss, Stevens and Jaeggi2019) and with Dominance in female chimpanzees (Pan troglodytes verus) (Wilson, Weiss et al., Reference Wilson, Weiss, Lefevre, Ochiai, Matsuzawa, Inoue-Murayama, Freeman, Herrelko and Altschul2020), yet for all these species (macaques, bonobos and chimpanzees) fWHR is not sexually dimorphic once accounting for sexual dimorphism in body size.
Several other measures, which have been found to be sexually dimorphic in humans (Penton-Voak et al., Reference Penton-Voak, Jones, Little, Baker, Tiddeman, Burt and Perrett2001), have also been assessed in nonhuman primates: face width/lower face height (Fig. 1a) has been linked to Assertiveness in brown capuchins and facial lower-height/full-height (Fig. 1a) to Neuroticism in brown capuchins (Wilson et al., Reference Wilson, Lefevre, Morton, Brosnan, Paukner and Bates2014) and to lower Confidence in immature rhesus macaques (Altschul et al., Reference Altschul, Robinson, Coleman, Capitanio and Wilson2019).
To summarise, personality, in particular Dominance-like traits, seems to have links to facial metrics in all species studied, but these relationships depend on age, sex and even subspecies (see Table 1), suggesting divergent selection pressures driving facial cues to Dominance. Together, these findings paint an interesting but varied picture of the role of facial morphology in primate behaviour, suggesting that human studies have only just scratched the surface of a potentially much broader field. The available studies on facial-personality links in primates, however, are not without limitations:
-
1. So far, the studied species represent a limited pool of the behavioural diversity found amongst primates, so potential socio-ecological drivers of interspecific differences cannot be untangled from phylogenetics.
-
2. Metric-based results may vary depending on whether measures are based on skull measurements or taken from photographs of living subjects (Kramer, Reference Kramer2017). Measures of skulls can be informative of sex-dependent growth trajectories independent of adiposity, but they do not account for facial characteristics that emerge in none-bone tissue, such as the secondary sexual characteristics found in gorillas and orangutans.
-
3. To date, most studies have focused on behavioural links to single, ratio-based measures (Fig. 1a).
2. The role of sexual selection in evolution of facial-personality links
Research linking the fWHR and facial lower-height/full-height to socially relevant Dominance-related personality traits (e.g., Dominance, Assertiveness, Neuroticism) suggests the potential role of static facial features as a cue in intraspecific communication (Table 1), which is long-lasting and available to many bystanders (compared to signalling through short-term facial expressions aimed at one or a few bystanders; Petersen & Higham, Reference Petersen and Higham2020). Sexual selection, including inter- and intra-sexual selection, has been hypothesised to generate and maintain facial-personality links (Lefevre et al., Reference Lefevre, Wilson, Morton, Brosnan, Paukner and Bates2014; Petersen & Higham, Reference Petersen and Higham2020; Wilson, Weiss et al., Reference Wilson, Weiss, Lefevre, Ochiai, Matsuzawa, Inoue-Murayama, Freeman, Herrelko and Altschul2020).
Intra-sexual selection acts on morphological or behavioural traits that increase an individual´s competitive ability. In primates (Table 1) including humans (e.g., Geniole et al., Reference Geniole, Denson, Dixson, Carré and McCormick2015; Haselhuhn et al., Reference Haselhuhn, Ormiston and Wong2015), fWHR and potentially other facial features might act as badges of status (Bergman & Sheehan, Reference Bergman and Sheehan2013; Petersen & Higham, Reference Petersen and Higham2020), reflecting status-seeking motivations, Dominance-related personality traits and fighting ability with the potential role of modulating social interactions. Indeed, evidence that faces of men are wider during early sexual maturity, compared with faces of older men, suggests that fWHR plays a role in intra-sexual competition (Summersby et al., Reference Summersby, Harris, Denson and White2022).
Inter-sexual selection acts on traits that improve individual reproductive success. For example, western lowland gorilla males with larger crests have a higher number of females (Caillaud, Levréro, Gatti, Ménard, & Raymond, Reference Caillaud, Levréro, Gatti, Ménard and Raymond2008) and better reproductive success (Breuer et al., Reference Breuer, Robbins, Boesch and Robbins2012). This suggests that certain sexually dimorphic cranial features are primarily under inter-sexual selection. In general, morphological traits reflecting individual quality either as a prospective partner or parent (direct benefits) or in terms of good genes (indirect benefits) are used as signals in mate choice. Experimental studies on birds and fish demonstrated that females consider their partner’s personality (e.g., Boldness or Exploration) as a signal of their quality in mate choice and that personality is often linked with variation in reproductive success (reviewed in Schuett, Tregenza, & Dall, Reference Schuett, Tregenza and Dall2010). Hence, any facial feature reflecting personality traits relevant to resource provision and paternal care (potentially Extraversion or Agreeableness) or protection from infanticide (potentially Assertiveness or Dominance) might be used as a cue by potential mates (Martin et al., Reference Martin, Staes, Weiss, Stevens and Jaeggi2019; Weston et al., Reference Weston, Friday, Johnstone and Schrenk2004). In primates, for example, facial symmetry (Fig. 1b) is perceived by partners as attractive and might be used as a cue to indirect benefits, such as good genes, better health and condition (Little, Paukner, Woodward, & Suomi, Reference Little, Paukner, Woodward and Suomi2012; Sefcek & King, Reference Sefcek and King2007; Waitt & Little, Reference Waitt and Little2006). Recent studies, however, found no significant association between attractiveness and facial symmetry in any of the tested human samples (Kleisner et al., Reference Kleisner, Kočnar, Tureček, Stella, Akoko, Třebický and Havlíček2017; Kočnar, Saribay, & Kleisner, Reference Kočnar, Saribay and Kleisner2019; Van Dongen & Gangestad, Reference Van Dongen and Gangestad2011). Hence, our understanding of face as a cue to personality and its role in mate choice in primates is limited.
3. Selection pressures driving facial-personality links
Unlike facial lower-height/full-height, which is associated with diverse personality traits across primates (Table 1) and thus might have diverged since the split of Catarrrhines and Platyrrhines, the link between fWHR and Dominance-related personality traits might predate this divergence. Yet, the pattern of associations between fWHR and Dominance-related personality traits in primates is not universal. Various species- or sex-specific selection pressures might explain these diverse patterns. First, provided that the face is a badge of status signalling personality characteristics relevant to choosing a partner or avoiding conflict escalation, it is expected to be found in primates living in large or dynamic groups, or where interactions with out-group members are frequent; in such scenarios, individual recognition is either limited or absent, and judgement based on first impression is crucial to fitness (Bergman & Sheehan, Reference Bergman and Sheehan2013; Grueter, Isler, & Dixson, Reference Grueter, Isler and Dixson2015). This might be further affected by sex-specific natal dispersal patterns – for example, facial-personality links might be more critical in the dispersing sex.
Second, the species-specific expression of dominance and a level of competitiveness might have affected the nature of personality traits linked to facial features and the strength of the correlation. In bonobos, in which social status is achieved by affiliative behaviour and coalitionary support, fWHR was more strongly related to Assertiveness (reflecting affiliative dominant tendencies) rather than rank (based on agonistic interactions) (Table 1). Contrastingly, macaque species with a despotic dominance style, for example, rhesus macaques, have larger fWHR than more tolerant species, for example, crested macaques (Macaca nigra) (Borgi & Majolo, Reference Borgi and Majolo2016). Humans are considered to have evolved from egalitarian societies (Boehm, Reference Boehm1999; Kaplan, Hooper, & Gurven, Reference Kaplan, Hooper and Gurven2009), as evidenced by their high social tolerance for out-group members, and reflected in their personality structure: compared to other primates, humans do not have a separate Dominance personality dimension (Weiss, Reference Weiss2022). These hierarchical differences could explain why links between fWHR and Dominance-related personality traits are relatively weaker than those in chimpanzees or capuchin monkeys (Table 1) (Wilson, Weiss et al., Reference Wilson, Weiss, Lefevre, Ochiai, Matsuzawa, Inoue-Murayama, Freeman, Herrelko and Altschul2020).
Third, the sex-specific patterns of facial-personality links might be attributed to the sex-specific dominance strategies and rank stability. Whilst in humans the fWHR-dominance link is found only in males, because human males exhibit dominant behaviour and aggression more than females (Archer, Reference Archer2004), in capuchins and bonobos, this link was found in both sexes (Table 1), as both sexes express similar levels of dominance (Gazes, Schrock, Leard, & Lutz, Reference Gazes, Schrock, Leard and Lutz2022; Stevens, Vervaecke, de Vries, & van Elsacker, Reference Stevens, Vervaecke, de Vries and van Elsacker2007). Female chimpanzees exhibiting relatively stable rank across the lifespan may rely on signalling by facial morphological features compared to males, who experience dynamic changes in rank across the lifespan and depend on signalling by facial expressions and aggressive behaviour (Wilson, Weiss et al., Reference Wilson, Weiss, Lefevre, Ochiai, Matsuzawa, Inoue-Murayama, Freeman, Herrelko and Altschul2020).
Finally, the mating system might act as another selection pressure. Faces of polygynous primates were rated by human raters as more masculine on a 6-point rating scale (based on presence of sexually dimorphic features, such as differences in pelage) than monogamous and promiscuous primates (A. Dixson, Dixson, & Anderson, Reference Dixson, Dixson and Anderson2005). A similar pattern is also expected in terms of facial-personality links. These explanations are of course not mutually exclusive, and several selection pressures might have acted simultaneously.
4. Proximate mechanisms underlying facial-personality link
Because the research on facial morphological cues to personality in nonhuman primates is relatively new, the direct causal links have not yet been studied. Hence, we must mainly rely on studies of the proximate mechanisms of human facial morphology. In this section, we will discuss two possible candidate mechanisms: neuroendocrine mechanisms and fluctuations during development.
Across vertebrate species, gonadal steroid hormones play a crucial role in bone growth (Gandelman, Simon, & McDermott, Reference Gandelman, Simon and McDermott1979; Juul, Reference Juul2001; Whitehouse et al., Reference Whitehouse, Gilani, Shafait, Mian, Tan, Maybery, Keelan, Hart, Handelsman, Goonawardene and Eastwood2015) and sex differentiation of behaviour (Berenbaum & Beltz, Reference Berenbaum and Beltz2011; Sisk & Zehr, Reference Sisk and Zehr2005; Thornton, Zehr, & Loose, Reference Thornton, Zehr and Loose2009). Since fWHR is linked to aggressive and dominant tendencies in primates including humans, testosterone exposure during pregnancy and puberty has been hypothesised to have critical organisational effects on morphology and neural structure, the latter causing behaviour differentiation. This was confirmed by a recent study on humans, documenting that newborns with higher umbilical cord testosterone levels had more masculine faces (measured as gender score) as adults (Whitehouse et al., Reference Whitehouse, Gilani, Shafait, Mian, Tan, Maybery, Keelan, Hart, Handelsman, Goonawardene and Eastwood2015). Surprisingly, this study did not find an association between prenatal testosterone levels and adult fWHR. In rhesus macaques, the experimental prenatal administration of testosterone masculinised juvenile and adult sexual behaviour (reviewed in Thornton et al., Reference Thornton, Zehr and Loose2009). The mediating effect of testosterone, however, may not be universal in primates. Instead, inter-species variability is expected due to the species- and sex-specific selection pressures. For example, the second-to-fourth digit ratio, a proxy of prenatal testosterone levels, was lower (higher prenatal testosterone levels) in polygynous primate species and species with high levels of intra-sexual competition compared to monogamous and polyandrous species and those with low levels of intra-sexual competition (Nelson & Shultz, Reference Nelson and Shultz2009).
The sex differentiation of morphology and behaviour established prenatally is further strengthened during puberty (Marečková et al., Reference Marečková, Weinbrand, Chakravarty, Lawrence, Aleong, Leonard, Perron, Pike, Richer, Veillette, Pausova and Paus2011). Adult circulating baseline or reactive testosterone levels have no (Bird et al., Reference Bird, Cid, Geniole, Welker, Zilioli, Maestripieri, Arnocky and Carré2016) or minimal effect (Lefevre, Lewis, Perrett, & Penke, Reference Lefevre, Lewis, Perrett and Penke2013) on fWHR. Regardless, the organisational effects of prenatal and pubertal testosterone exposure on behaviour and face morphology proved to be permanent and hence might also underlie the facial-personality links. Testosterone, however, might not be the only neuroendocrine mechanism underlying the facial-personality links. A study on lemurs, which are characterised by masculinisation of female genitalia and female dominance over males, concluded that the testosterone:oestrogen ratio rather than absolute prenatal testosterone levels promoted female masculinisation (Ostner, Heistermann, & Kappeler, Reference Ostner, Heistermann and Kappeler2003). The prenatal testosterone:oestrogen ratio has been documented to contribute to the behavioural sex differentiation also in humans (Mitsui et al., Reference Mitsui, Araki, Miyashita, Ito, Ikeno, Sasaki, Kitta, Moriya, Cho, Morioka, Kishi, Shinohara, Takeda and Nonomura2019). Finally, the effects of steroid hormones are often moderated by interaction with other mechanisms, such as cortisol (Carré & Archer, Reference Carré and Archer2018), growth hormone (Marečková et al., Reference Marečková, Weinbrand, Chakravarty, Lawrence, Aleong, Leonard, Perron, Pike, Richer, Veillette, Pausova and Paus2011), and androgen sensitivity caused by variation in a number of tri-nucleotide (CAG) repeats in the androgen receptor gene (Simmons & Roney, Reference Simmons and Roney2011).
Instability during critical phases of foetal and postnatal development caused by environmental (e.g., parasite load, resources quality and quantity, diseases) or genetic (e.g., homozygosity, inbreeding, mutation) stressors may result in fluctuating asymmetry (Caccavo, Lemos, Maroja, & Gonçalves, Reference Caccavo, Lemos, Maroja and Gonçalves2021; Parsons, Reference Parsons1992, but see also Lens, Dongen, Kark, & Matthysen, Reference Lens, Dongen, Kark and Matthysen2002). Fluctuating asymmetry refers to minor random deviations from perfect symmetry (Valen, Reference Valen1962) and can be expressed as individual variability in the asymmetry of craniofacial features (Caccavo et al., Reference Caccavo, Lemos, Maroja and Gonçalves2021; Fig. 1b) and other bilaterally symmetrical structures (Kucheravy, Waterman, & Roth, Reference Kucheravy, Waterman and Roth2022). Facial asymmetry in humans has been found to be associated with poor health, specifically number of respiratory diseases and their duration (Thornhill & Gangestad, Reference Thornhill and Gangestad2006, but see Rhodes et al., Reference Rhodes, Zebrowitz, Clark, Kalick, Hightower and McKay2001), and personality traits, such as Extraversion, Agreeableness, Neuroticism or Assertiveness (Borráz-León & Cerda-Molina, Reference Borráz-León and Cerda-Molina2015; Fink, Neave, Manning, & Grammer Reference Fink, Neave, Manning and Grammer2005; Holtzman, Augustine, & Senne, Reference Holtzman, Augustine and Senne2011; Pound, Penton-Voak, & Brown, Reference Pound, Penton-Voak and Brown2007), although the reported associations are usually weak or inconsistent (Hope et al., Reference Hope, Bates, Penke, Gow, Starr and Deary2011; Van Dongen & Gangestad, Reference Van Dongen and Gangestad2011) or show null associations with attractiveness (Kleisner et al., Reference Kleisner, Kočnar, Tureček, Stella, Akoko, Třebický and Havlíček2017; Kočnar et al., Reference Kočnar, Saribay and Kleisner2019). Moreover, stress might be theoretically dependent on the rank and personality of the mother as, for example, subordinate, less confident or more sociable females face a higher risk of injury or pathogen transmission (Robinson et al., Reference Robinson, Coleman, Capitanio, Gottlieb, Handel, Adams, Leach, Waran and Weiss2018). Investigating the potential links between facial symmetry and personality might be a promising avenue.
5. Does the primate face provide cues to personality?
Numerous studies reveal that humans perceive personality differences in conspecific faces (Kramer & Ward, Reference Kramer and Ward2010; Little & Perrett, Reference Little and Perrett2007), including dynamic features (Kavanagh, Whitehouse, & Waller, Reference Kavanagh, Whitehouse and Waller2022), and even in other species (Clark, Butler, Ritchie, & Maréchal, Reference Clark, Butler, Ritchie and Maréchal2020; Kramer & Ward, Reference Kramer and Ward2012). In parallel, there is a large body of comparative research on how people and nonhuman primates look at faces, from what features they attend to (Dahl, Wallraven, Bülthoff, & Logothetis, Reference Dahl, Wallraven, Bülthoff and Logothetis2009; Gothard, Erickson, & Amaral, Reference Gothard, Erickson and Amaral2004; Kano, Call, & Tomonaga, Reference Kano, Call and Tomonaga2012), to holistic processing (Carp et al., Reference Carp, Santistevan, Machado, Whitaker, Aguilar and Bliss-Moreau2022; Parr, Reference Parr2011; Wilson, Kade et al., Reference Wilson, Kade, Moeller, Treue, Kagan and Fischer2020). Yet, little is known about whether and how nonhuman primates perceive different facial cues to personality traits.
Studying how individuals perceive information encoded in facial features of conspecifics is key to understanding whether personality traits can be read from faces. The simplest way to test this is using the looking time paradigm, which assesses gaze responses to two different stimuli (Wilson, Bethell, & Nawroth, Reference Wilson, Bethell and Nawroth2023). Numerous studies have used this approach to examine whether primates differentiate between images of faces that depict varying social information (Wilson et al., Reference Wilson, Bethell and Nawroth2023; Winters, Dubuc, & Higham, Reference Winters, Dubuc and Higham2015), such as differences in fitness (indicated by skin redness) (Waitt et al., Reference Waitt, Little, Wolfensohn, Honess, Brown, Buchanan-Smith and Perrett2003), age (Almeling, Hammerschmidt, Sennhenn-Reulen, Freund, & Fischer, Reference Almeling, Hammerschmidt, Sennhenn-Reulen, Freund and Fischer2016) and emotion expression (Pritsch, Telkemeyer, Mühlenbeck, & Liebal, Reference Pritsch, Telkemeyer, Mühlenbeck and Liebal2017). Such studies suggest that being able to attend to socially relevant information is a potentially universal feature amongst primates.
Yet, to what extent different species advertise socially relevant traits in facial features, or use this information to inform social interactions, is largely unknown. One study found that long-tailed (Macaca fascicularis) and rhesus macaques can distinguish between human faces based on features depicting Trustworthiness (Costa et al., Reference Costa, Gomez, Barat, Lio, Duhamel and Sirigu2018). Moreover, both monkey and human participants exhibited reduced-looking durations to faces with higher fWHRs. Male rhesus macaques also demonstrate rank-related preferences, by sacrificing fluid to view images of high-status, but not low-status monkeys (Deaner, Khera, & Platt, Reference Deaner, Khera and Platt2005). However, an attempt to determine whether female rhesus macaques visually distinguish facial “masculinity” from varying face morphology of males revealed borderline results (Rosenfield et al., Reference Rosenfield, Semple, Georgiev, Maestripieri, Higham and Dubuc2019). Similarly, assessment of whether brown capuchins differentiate between faces of varying facial width found that subjects did not differ in their approach latency to a life-sized capuchin model with manipulated facial image depicting lower or higher fWHR (Wilson, Gartner, D’Eath, Buchanan-Smith, & Morton, Reference Wilson, Gartner, D’Eath, Buchanan-Smith and Morton2018). However, it was unclear whether their response was because they did not use fWHR as a cue to Assertiveness or as a result of viewing static stimuli rather than live conspecifics.
Alongside behavioural paradigms, studies from neuroscience can also address this question. For example, in response to superior compared with inferior co-players, human participants show increased activation in the occipital/parietal cortex, ventral striatum, parahippocampal cortex and dorsal lateral prefrontal cortex (Zink et al., Reference Zink, Tong, Chen, Bassett, Stein and Meyer-Lindenberg2008). This overlaps with findings of responses to social cues of status, such as direct gaze and brow position (Marsh, Blair, Jones, Soliman, & Blair, Reference Marsh, Blair, Jones, Soliman and Blair2009). Notably, men with higher fWHRs had increased amygdala activation in response to images of angry facial expressions (Carré, Murphy, & Hariri, Reference Carré, Murphy and Hariri2013). Lesions to the amygdala in rhesus macaques reduced affiliative tendencies in response to threats from conspecifics compared with pre-lesion baseline (Machado & Bachevalier, Reference Machado and Bachevalier2006). This raises the possibility that emotional reactivity could explain the relationship between facial features and dominance. Given the likely differential expression of such behaviour between species and social organisations, this could also explain the wide variance in reported findings to date. Whilst these studies examine response to behavioural, or perceived behavioural information, approaches that use EEG or fMRI could be useful in understanding social perceptions of static facial morphology.
It is clear that more research into the perception of personality and facial morphology in nonhuman primates is needed. Whilst the looking time paradigm can provide a baseline for understanding whether a study group can distinguish such information, this approach can have limited interpretation in understanding the social value of these features (Morton et al., Reference Morton, Brosnan, Prétôt, Buchanan-Smith, O’Sullivan, Stocker, D’Mello and Wilson2016; Wilson et al., Reference Wilson, Bethell and Nawroth2023). This approach should therefore be complemented not only by a neuroscience perspective but also by studies that try to understand social outcomes of sexually dimorphic facial features in natural settings.
6. Future directions
In this manuscript, we have drawn on literature addressing the behavioural and neuroendocrinological correlates of facial morphology within an evolutionary framework. By doing so, we hope to have highlighted the benefits of comparative research for understanding how, and why, the primate face can cue personality traits. Here we highlight some specific steps that would be of benefit to the field:
Firstly, we should move away from single measures such as the fWHR and measures based on ratios (due to spurious correlations; Kronmal, Reference Kronmal1993) and consider facial morphology holistically (Dixson, Reference Dixson2018; Kleisner, Pokorný, & Saribay, Reference Kleisner, Pokorný and Saribay2019). This would require taking multiple measures (e.g., sexual shape dimorphism, averageness, Kočnar et al., Reference Kočnar, Saribay and Kleisner2019) based on the configuration of facial landmarks delineating facial shape, ideally from both skulls and from flesh (or non-invasive proxies of either) (Fig. 1c). Moreover, diverse facial ornamentation (such as the presence of ear or head tufts, beards or other facial hair) evolved across the primate taxa (Petersen & Higham, Reference Petersen and Higham2020). Despite extensive research suggesting the potential role of human facial hair in signalling dominance (Dixson & Vasey, Reference Dixson and Vasey2012, but see also Kowal et al., Reference Kowal, Sorokowski, Żelaźniewicz, Nowak, Orzechowski, Żurek, Żurek and Nawrat2021), the link between shape and size variation of primate facial hair features and behavioural variation has not been studied yet. Future studies thus should explore the potential facial hair-personality links and their selection pressures (Dixson et al., Reference Dixson, Dixson and Anderson2005; Petersen & Higham, Reference Petersen and Higham2020).
Secondly, we need to better understand the proximate mechanisms underlying facial-personality links in primates. Such research could benefit from in-depth consideration of developmental neuroendocrine mechanisms, as well as other possible latent variables such as emotional reactivity. Laboratory colonies of primates represent a viable research opportunity due to their experience in neuroendocrine, genetic and developmental research. Later, expanding the research beyond fWHR measures and beyond the laboratory to a broad spectrum of primate species with varying social and mating strategies and relatedness levels would help shed light on proximate as well as ultimate facial-personality links in primates, including humans.
Thirdly, we can utilise a comparative understanding of the role of faces in social cues or signals, by drawing comparisons amongst phylogenies of varying socio-ecology. Specifically, more studies on personality links to reproductive success (Brent et al., Reference Brent, Semple, MacLarnon, Ruiz-Lambides, Gonzalez-Martinez and Platt2014; Masilkova, Boukal, Ash, Buchanan-Smith, & Konečná, Reference Masilkova, Boukal, Ash, Buchanan-Smith and Konečná2022; Weiss et al., Reference Weiss, Feldblum, Altschul, Collins, Kamenya, Mjungu, Foerster, Gilby, Wilson and Pusey2023) and the importance of personality in mate choice are needed. Studies that not only incorporate populations from western, educated, industrialised, rich and democratic populations but also people from indigenous communities could better aid the understanding of such selection pressures in humans.
To conclude: does the primate face cue personality? The answer appears to be yes, but we are only just beginning to understand how and why.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/pen.2023.5.
Acknowledgements
We are grateful to Karel Kleisner for his insightful comments on the earlier versions of the manuscript.
Competing interests
The authors have nothing to disclose.