Identifying personality from the static, nonexpressive face in humans and chimpanzees: evidence of a shared system for signaling personality

Abstract

Many aspects of personality are honestly signaled on the human face, as shown by accurate identification of personality traits from static images of unknown faces with neutral expressions. Here, we examined the evolutionary history of this signal system. In four studies, we found that untrained human observers reliably discriminated characteristics related to extraversion solely from nonexpressive facial images of chimpanzees (Pan troglodytes). In chimpanzees, as in humans, there is therefore information in the static, nonexpressive face that signals aspects of an individual's personality. We suggest that this performance is best explained by shared personality structure and signaling in the two species.

Introduction

Personality traits describe the stable, context-general behavioral biases of an individual organism. Factor-analytical approaches have identified a small number of human personality traits, with three-factor (Eysenck & Eysenck, 1985) and five-factor (Goldberg, 1993) models being the most used. Human personality as defined by these models has not only cultural but also biological bases. Behavioral genetics studies estimate heritability coefficients for individual traits to be 0.40–0.60 (Bouchard & Loehlin, 2001), and human models show cross-cultural and even cross-species generalization of personality factors (Gosling & John, 1999). In particular, humans and chimpanzees demonstrate similar, although not identical, factor structures. The most important distinction is that the factor-analytical approach identifies an additional (King & Figueredo, 1997), highly heritable (Weiss, King, & Figueredo, 2000) dominance-related factor present in chimpanzees but not in humans.

Many socially relevant traits can be accurately identified in humans solely from visible cues in the static, nonexpressive face, including personality (Kramer & Ward, 2010, Little & Perrett, 2007), sociosexuality (Boothroyd, Jones, Burt, DeBruine, & Perrett, 2008), trustworthiness (Stirrat & Perrett, 2010), and aggression (Carré, McCormick, & Mondloch, 2009). Interpreting these results within animal signaling theory suggests a close association between the facial morphology and behavior of the signal “sender” and the cognitive processes for understanding the signal in the “receiver” (Maynard-Smith & Harper, 2003). Here, we examined the evolutionary history of this signal system. We reasoned that if the face were part of an evolved signal system, then humans and chimpanzees might share aspects of this system. If so, some facial morphology signals should be expressed and understood between species.

The possibility of a shared signal system is plausible in part due to evidence from comparative studies showing similarities in face processing for the two species, including homologous specialized brain regions (Parr, Hecht, Barks, Preuss, & Votaw, 2009), cross-species identification of relatedness (Alvergne et al., 2009), sensitivity to facial configurations (Parr, Heintz, & Akamagwuna, 2006), and homologies in expression (Parr, Waller, Vick, & Bard, 2007). However, a shared signal system of personality would require a variety of other physical and psychological homologies, including behavioral biases as reflected by aspects of personality structure, facial morphology, and the cognitive means for correctly interpreting and using these signals from the face.

To test the possibility of a shared signal system, we measured the ability of humans to understand signals from the static chimpanzee face related to extraversion. Our focus on extraversion was motivated by previous findings. First, in the human face, the signal for extraversion is strong and apparent in both individual (Penton-Voak, Pound, Little, & Perrett, 2006) and composite (Kramer & Ward, 2010) faces. Second, personality characteristics related to human extraversion, such as individual differences in dominance, sociability, and activity levels, are widespread in nonhuman animals (Gosling & John, 1999), including other primates (e.g., chimpanzees; King & Figueredo, 1997), other mammals (e.g., hyenas; Gosling, 1998), fish (e.g., guppies; Budaev, 1997), and even invertebrates (e.g., octopuses; Mather & Anderson, 1993). Finally, the characteristic of “dominant” is encompassed by the trait of “extraversion” in human taxonomy (Goldberg, 1990) and is a particularly robust measure in chimpanzees, demonstrating the single highest factor weighting of any characteristic in the chimpanzee personality model (King & Figueredo, 1997) and both reliability (Freeman & Gosling, 2010) and external validity (Pederson, King, & Landau, 2005) in predicting individual behavior. We carried out a series of four studies in order to investigate accuracy in identifying characteristics relating to extraversion from static chimpanzee faces.