The bio-logic of facial geometry

Abstract

Perrett et al. reply — Our computer graphic manipulations of the ‘geometrical’ differences between male and female face shapes generate stimuli that embody the ‘psychological’ meaning of masculinity and femininity¹. Meyer and Quong question whether our stimuli accurately imitate the ‘biological’ effect of sex hormones on facial structure, arguing that this should be defined from same-sex individuals whose levels of sex hormones differ.

Main

As visible sexual dimorphism in the human face increases rapidly at puberty when sex steroids influence tissue growth, our working assumption is that the geometrical differences between male and female faces parallel the differences between individuals with high and low androgen levels. Masculinity in mammals is testosterone dependent, whereas femininity represents a more neutral developmental state². Individuals with an XY chromosome pattern and complete androgen insensitivity syndrome have a female appearance. It is unlikely that geometrically ‘feminizing’ a face produces facial attributes associated with higher than average levels of androgens.

Meyer and Quong suggest that our subjects may be attracted to cues of youth rather than femininity in faces. Cues to femininity and youth are inextricably related, however, because of sex differences in the amount of facial growth at puberty. Pubescent male faces grow a great deal, whereas female faces change less and therefore remain relatively child-like. As we have shown¹, manipulations based on geometric sex differences in face shape alone affect apparent age. Certainly, face stimuli derived from males with measured high and low levels of testosterone would reflect androgenic effects on target tissues more accurately than our stimuli. Nevertheless, high-testosterone composite faces will carry more ‘mature’ features and may still appear older than low-testosterone composites.

Evolutionary theories do not predict that females should favour youthful-looking males; instead, they predict that cues to high male status, such as dominance, will be considered attractive³. Our masculinized male faces were rated as more dominant than feminized or average facial stimuli. Facial dominance is a predictor of male status in at least some human hierarchies⁴ and yet subjects in our experiments preferred subdominant male faces. It is unclear whether this is due to a preference for youthfulness or a preference for other characteristics attributed to femininity (such as apparent warmth, kindness or a willingness to invest in progeny). What is apparent is that explanations of male attractiveness based on female sensory biases⁵ or honest Zahavian handicaps fail to predict the preference for slightly feminized male facial characteristics.

Our investigation indicates that females select mates on characteristics other than, or in addition to, dominance.