تقارن، میانگین و قابلیت اندازه در جذابیت صورت زنان

Female facial attractiveness was investigated by comparing the ratings made by male judges with the metric characteristics of female faces. Three kinds of facial characteristics were considered: facial symmetry, averageness, and size of individual features. The results suggested that female face attractiveness is greater when the face is symmetrical, is close to the average, and has certain features (e.g., large eyes, prominent cheekbones, thick lips, thin eyebrows, and a small nose and chin). Nevertheless, the detrimental effect of asymmetry appears to result solely from the fact that an asymmetrical face is a face that deviates from the norm. In addition, a factor analysis indicated that averageness best accounts for female attractiveness, but certain specific features can also be enhancing.

Attractiveness is an important element of social life. A large body of research in social psychology has shown that attractive persons enjoy many advantages that unattractive persons do not have (for reviews, see Berscheid and Walster, 1974 and Bull and Rumsey, 1988; for a meta-analysis, see Eagly, Ashmore, Makhijani, & Longo, 1991). This well-known stereotype, described by social psychologists as the What is beautiful is good prototype ( Dion, Berscheid, & Walster, 1972), mainly applies to the formation of first impressions, but it can also extend into less superficial interactions (e.g., Frieze, Olson, & Russell, 1991). Until recently, beauty was considered to lie in the eyes of the beholder (for a historical background, see Bruce & Young, 1998). In that sense, a person’s physical appearance is not the main aspect of his/her attractiveness. It is determined instead by the beholder, who probably takes personality into account in addition to physical appearance. Above all, it is believed that each beholder has personal likings that may differ from those of others. Thus, a face may be attractive for some people but unattractive for others. However, some recent studies have reported substantial agreement, not only within the sexes but also across sexes, ethnic groups, and ages (e.g., Cunningham, Roberts, Barbee, Druen, & Wu, 1995; Langlois, Ritter, Roggman, & Vaughn, 1991; Langlois & Roggman, 1990). Moreover, some facial characteristics have been shown to be factors of attractiveness, including closeness to the average (e.g., Langlois & Roggman, 1990), symmetry (e.g., Thornhill & Gangestad, 1993), and the physical characteristics of individual facial features (e.g., Cunningham, 1986). From an evolutionary standpoint, the fact that humans share views about which features are attractive suggests that there are species-typical psychological adaptations that have evolved because of a preference for healthy and fertile mates ( Symons, 1979 and Thornhill and Gangestad, 1999). Working on the closeness-to-average factor, (Langlois & Roggman, 1990; see also Grammer & Thornhill, 1994) created composite computer faces by averaging photographs of women or men. The composite faces were assumed to be increasingly closer to the average as the number of faces that went into generating them rose. The authors found that by increasing the number of faces averaged into the composite, they also increased their attractiveness ratings. This observation was found for both female and male faces. Later, this finding was extended to other cultures (Langlois et al., 1994 and Pollard, 1995). The caricature versus anti-caricature technique (see Rhodes, Brennan, & Carey, 1987) has also been used to study the relationship between averageness and attractiveness. A caricatured face is a face in which the distinctive features in a set of faces are exaggerated by increasing the distance between the features and their average positions. Decreasing this distance makes the resulting face closer to the average, thereby creating anti-caricatured faces. In studies by Rhodes and Tremewan (1996) and Deffenbacher, Vetter, Johanson, and O’Toole (1998), anti-caricatured faces were deemed more attractive than normal faces, themselves being rated as more attractive than caricatured faces. Thus, it appears that moving a face closer to the average increases its attractiveness. This is consistent with a previous report in which typical faces (more difficult to recognize) were judged more attractive than distinctive ones (more easily recognized; Light, Hollander, & Kayra-Stuart, 1981). Symons (1979) suggested that averageness is associated with a good phenotypic condition because of a stabilizing effect of natural selection on facial features. More recently, Thornhill and Gangestad, 1993 and Thornhill and Gangestad, 1999 suggested that a preference for average traits may have evolved because averageness denotes genetic heterozygosity, which could signal an outbreed mate or provide genetic diversity in defense against parasites. On the other hand, some studies also indicate that, even if an average face is attractive, the most attractive faces are not close to the average. Shepherd and Ellis (1973), for example, did not obtain the linear relationship that Light et al. (1981) did between facial attractiveness and memorability, the latter dimension being linked to distinctiveness. They reported a U-shaped relationship. This suggests that distinctive faces comprise both the least and the most attractive faces. Perrett, May, and Yoshikawa (1994) provided support for this possibility using the caricature technique. They showed that the average face among a set of sixty women was less attractive than the average face of the fifteen most attractive women in the set. Moreover, caricaturing of the latter composite face increased its attractiveness. In conclusion, an average face is attractive, but the most attractive faces are not average ones. A face’s symmetry has also been reported as a factor of its attractiveness. Some studies indicate that symmetrical female and male faces are seen as more attractive than asymmetrical ones, and that having a symmetrical face is an advantage in sexual competition (Gangestad et al., 1994, Grammer and Thornhill, 1994, Mealey et al., 1999 and Thornhill and Gangestad, 1993). In the same way, many studies have suggested that facial symmetry reflects the person’s phenotypic and genetic condition (e.g., Waynforth, 1998) as well as his/her psychological state (e.g., Shackelford & Larsen, 1997). So, like closeness to the average, symmetry has been proposed as a “certificate of health” in a mating perspective. However, the role of symmetry has been questioned. Some authors have concluded that symmetry has little influence on attractiveness (Farkas, 1994 and Jones and Hill, 1993). Others have found no impact at all (Langlois et al., 1994 and Shackelford and Larsen, 1997). Langlois et al. (1994) suggested that only highly asymmetrical faces are unattractive; and finally, in three studies, asymmetrical normal faces were found to be more attractive than symmetrical composite versions of those same faces (Kowner, 1996, Samuels et al., 1994 and Swaddle and Cuthill, 1995). However, the techniques used in the latter studies have been criticized (Rhodes et al., 1998 and Thornhill and Gangestad, 1999). Kowner (1996) and Samuels et al. (1994) used composite faces created by aligning a hemi-face with its mirror image. This technique may create abnormalities in facial feature size, such as a large nose. Swaddle and Cuthill (1995) averaged normal full faces with their mirror images, but they did not control facial expression. Using the same technique as Swaddle and Cuthill (1995) but with neutral expressions only, Rhodes et al. (1998) found that chimerical symmetrical faces were more attractive than their original normal faces. As a whole, then, it seems that symmetrical faces are more attractive than asymmetrical ones. The third approach used to study facial attractiveness consists in measuring certain facial features or manipulating feature size. For female faces, Cunningham (1986) (see also Keating, 1985) showed that women are more attractive when they have big eyes, a small nose, thin jaws, prominent cheekbones, a large mouth, high eyebrows, and so forth. These observations were later confirmed by Cunningham et al. (1995), who extended them to faces and judges from other cultures. Cunningham, 1986 and Cunningham et al., 1995 proposed the “Multiple Fitness Model,” which states that attractiveness depends on multiple dimensions, with each feature representing a different aspect of mate value. They distinguished five dimensions, each related to specific facial features. Three of the dimensions are biological and age-related (neonatal, sexual maturity, and senescence) and two are influenced by more personal and social variables (facial expression and grooming). Neonatal features are thought to suggest desirable qualities of youthful vivaciousness, open-mindedness, and agreeableness. Sexual maturity features indicate postpubescent status and may suggest strength, dominance, status, and competency. They may also indicate an effective immune system, healthy resistance to parasites, and proper functional adaptation to the environment. Senescence features may convey non-competitive dominance. Expressive features suggest happiness, congeniality, non-threatening interest, social approachability, excitement, and arousal, whereas grooming features indicate successful adaptation, group membership, and status (see Cunningham et al., 1995). The attractiveness of a face depends on the harmonious presence of these different aspects (except for the features of senescence). Cunningham et al. (1995) proposed that an attractive face has neonatal features in the center of the face and sexual maturity features at the periphery. In contrast to the symmetry or closeness-to-average hypotheses, this model explains how two faces that seem attractive can nonetheless be distinctive. According to the literature, an attractive woman has a symmetrical, close-to-average face with large eyes, a small nose, and so forth. However, at this time, no authors appear to have jointly studied the role of each of these characteristics. Rhodes, Sumich, and Byatt (1999) showed that an average face is not deemed attractive solely on the basis of its symmetry, but they did not include feature size in their studies. Many questions remain unanswered. Notably, what is the role of each of these characteristics? Is a woman more attractive with a symmetrical face, an average face, or large eyes? Some ensuing questions also emerge: What is the relationship between facial symmetry and facial averageness? One can assume that an average face is also symmetrical, but a symmetrical face is not necessarily close to the average. For example, one may have perfectly symmetrical eyes that are lower in the face than in other faces. Another question concerns the averageness and size of some features. It seems contradictory to find that a face is more attractive with both close-to-average features (e.g., Langlois & Roggman, 1990) and some larger (or thinner) features (e.g., Cunningham, 1986). For example, is it more important to have large eyes or average eyes? The answer is probably that it all depends on the feature in question. Having close-to-average features implies having no unattractive features, which would make the face attractive. In this case, deviations from the average would enhance or decrease attractiveness according to (i) which feature is the deviating one, and (ii) the direction of the deviation. This suggestion is in line with the studies above which showed that close-to-average faces are attractive, but the most attractive faces are not close to average (e.g., Perrett et al., 1994). In order to study the respective roles of symmetry, averageness, and feature size in the facial attractiveness of women, we parametrized these characteristics on a set of female faces. Attractiveness ratings of these faces were also collected. A first goal of the experiment was to confirm that symmetry, averageness, or feature size influence facial attractiveness. So, correlations were computed between attractiveness ratings and measures of symmetry, averageness, and feature size factors. According to the litterature, each of these factors should be correlated with ratings. A second goal was to study the weight of each of these factors. Partial correlations and factor analyses were computed to test if they had an independent influence on attractiveness ratings and if they had the same weight.