ANTHROPOLOGICAL REVIEW • Vol. 70, 45-79 (2007)

Facial attractiveness: General patterns of facial preferences

Krzysztof Kościński

Institute of Anthropology, Adam Mickiewicz University, Umultowska 89, 61-614 Poznań, Poland; E-mail: [email protected]

ABSTRACT This review covers universal patterns in facial preferences. Facial attractive- ness has fascinated thinkers since antiquity, but has been the subject of intense scientific study for only the last quarter of a century. Many facial features contribute to facial attrac- tiveness: Averageness and symmetry are preferred by males and females, probably because they signal genetic quality and developmental stability. Men prefer highly feminized female faces because they reflect high estrogen levels and low testosterone levels. This indicates that the woman is reproductively healthy. Women, on the other hand, prefer a moderate level of male facial masculinity, since facial masculinity that is too pronounced signals high level of testosterone and, thereby, a poorly developed pro-family personality. In women, facial hair is detrimental to facial attractiveness. In men, the effect is not consistent. Faces with a clear complexion are attractive to both men and women. Men prefer light and smooth skin in women. Positive facial expressions also enhance facial attractiveness. Many factors, in par- ticular skin condition and facial proportions, affect perceived age, which is an important component of facial attractiveness. Men in particular strongly prefer youthful-looking female faces. Facial preferences enable an individual to recognize reproductively fit mates. There- fore, facial preferences are adaptive, although non-adaptive mechanisms related to general brain function also play a role. KEY WORDS: physical attractiveness, facial esthetics, human face, beauty

Everything is in the face Cicero Facial attractiveness is something that attractiveness has proven to have such is intuitively perceived rather than meas- serious interpersonal and social conse- urable with instruments. Until recently, quences that science cannot exclude it facial attractiveness was more a subject from its realm of study. Moreover, scien- of interest for artists and philosophers tific research on physical attractiveness is than for scientists. However, physical justified because it is connected to the

© 2007 Polish Anthropological Society DOI: 10.2478/v10044-008-0001-9 46 K. Kościński features of the perceiver and the person body proportions which is now referred perceived. to as the classical system, or the canon. Physical attractiveness is assessed on Among the many criteria for the ideal the basis of different sensory data, e.g., face were that the length of the ears voice, scent, appearance. The face is should be equal to the length of the nose especially important, because it has many and that the distance between the eyes morphological elements, is the main should be equal to the width of the nose. channel of interpersonal communication Ancient sculptors, beginning with Poly- and, in most cultures, is clearly visible. kleitos and Phidias, also believed that It is impossible to precisely define the relationship between many pairs of physical and facial attractiveness. None- measurements should be equal to the theless, an attractive face attracts the golden ratio (φ), which is approximately perceiver, the perceiver can make a equal to 1.618 (the ratio of A to B con- judgment about it, tends to look at it, and forms to the golden ratio if (A+B)/A = wants to make contact with the owner. A/B). Usually, but by no means always, the During the Renaissance, this system contact desired is erotic in nature. was transformed and incorporated by The author wishes to comprehensively Albrecht Dürer and Leonardo da Vinci review the current state of knowledge on into the neoclassical system. Since then, the attractiveness of the human face. In the it has been widely used in the arts. How- first of two parts, this review focuses on ever, research showed that the neoclassi- general, universal patterns of facial prefer- cal system described neither the typical ence. Individual variation in preferences, face nor the ideally beautiful face of course, contribute substantially to at- [FARKAS et al. 1987b, FARKAS 1994, tractiveness judgments, but will be presen- BAKER and WOODS 2001]. Nowadays, ted separately, together with other issues most researchers have abandoned the not covered here. For this review, the fol- geometrical approach to FacA since they lowing abbreviations will be used to avoid no longer consider it correct nor even repetition: PA for physical attractiveness, scientific at all. Nevertheless, this view is and FacA for facial attractiveness. still propounded by some orthodontists and plastic surgeons [JEFFERSON 1996, History of research 2004; MARQUARDT 2002]. Psychologists turned their interest to Theoretical foundations PA in the 1960s, after two myths about it were disproved. The first of these was History of research on FacA has re- that the perception of attractiveness is vealed three phases that differ in theo- almost exclusively a matter of private retical approach. During the first period, taste and that there are no universal stan- facial beauty was understood in a ma- dards of beauty. This was supported by thematical sense, that is, as specific pro- rigorous ethnographic research [FORD portions between elements of the face. and BEACH 1951]. If this were true, there This approach was proposed in the fifth would be no basis for conducting scien- century BC by the sculptor Polykleitos, tific research on PA. This myth was dis- who formulated the system of human proved by studies involving many thou- General patterns of facial preferences 47 sands of raters in the United Kingdom sociobiology and the development of [ILIFFE 1960] and the United States sexual selection theory. In the 1970s, [UDRY 1965] which showed wide agree- some arguments that PA is connected ment amongst the raters regarding their with reproductive ability, age, health and assessment of FacA (r ≈ 0.5). conspicuous sexual traits were put forth The second myth was that PA is of lit- [EIBL-EIBESFELT 1970, GUTHRIE 1976, tle social significance. This myth was SYMONS 1979]. disproved by WALSTER et al. [1966] who Further theoretical considerations, of- found that, after a blind-date, a person ten based on the results of animal re- would want to date the other person search, have led to the formulation of again solely on the basis of PA. It ap- many testable hypotheses on FacA, in- peared that the PA of a person influences cluding: (1) that an attractive face has many of the traits attributed to that per- average proportions, conspicuous sexual son, and thus, how that person is treated traits, a smooth skin, and symmetry; (2) by others. That is, perceivers appear to that all these traits are signs of the good apply the stereotype of “what is beautiful health and genetic quality of the owner; is good” [DION et al. 1972]. and (3) that the facial features that are In the 1970s and 1980s, these discov- perceived as attractive as well as the eries inspired hundreds of experiments preferences for these features have been on: (1) the consistency of various groups shaped during the course of biological of raters in assessing PA; (2) the relation- evolution. Research based on the biologi- ships between PA and perceived or actual cal approach to FacA increased during psychological traits; (3) the difference the 1990s. between how physically attractive and physically unattractive people are treated. Research methods Although these studies revealed many interesting phenomena, they could not For decades, the most common method explain why one person is considered to of studying FacA was based on the as- be attractive and another not, and why sessment of faces in photographs by ex- something such as PA exists in the first perimental subjects acting as judges. place. More recently, faces to be evaluated are One of the co-authors of the theory of frequently presented to the judges as evolution, WALLACE [1889], connected computerized images. In some cases, PA with biological fitness. Curiously, judges were instructed to watch video DARWIN [1871] did not share this point films of the faces to be evaluated (e.g., of view. Instead, he proposed that mating RUBENSTEIN [2005]), or told to evaluate preferences in humans and other animals the faces of people in real life (e.g., are based on esthetics. KNIFFIN and WILSON [2004]), or instruc- In the following decades, the biological ted to imagine a person with specified and evolutionary approach to PA was traits, to assess attractiveness [WAGAT- abandoned because of its connotations SUMA and KLEINKE 1979, KOWNER with racism, colonialism and nazism. It 1998]. generated renewed interest in the 1970s In the majority of studies, both the with the founding of the science of sample of judges and that of stimulus 48 K. Kościński persons were students. In some studies, relationship between FacA and the spa- the faces represented were of beauty tial locations of anthropometric points contest competitors [CUNNINGHAM [FERRARIO et al. 1995]. 1986], models [JONES 1995, NGUYEN Newer methods of describing facial and TURLEY 1998, PETTIJOHN and shape include: (1) principal component JUNGEBERG 2004, YEHEZKEL and analysis, which describes a face based on TURLEY 2004], and actresses [FERRARIO a number of mathematically calculated et al. 1995, PETTIJOHN and TESSER factors related to facial appearance in a 1999], or in works of art [COSTA and global, rather than local, way [JOHNSTON CORAZZA 2006]. The judges were some- et al. 2003, EISENTHAL et al. 2006, times orthodontists or maxillofacial sur- VALENZANO et al. 2006]; and (2) geome- geons (e.g., MAPLE et al. [2005]), or tric morphometric methodology, which is pedophiles and rapists [MARCUS and based on the location coordinates of es- CUNNINGHAM 2003]. sential points, or landmarks, instead of on Examples of recent developments in the distances between them [FINK et al. this research are: (1) assessment of FacA 2005a, VALENZANO et al. 2006]. by judges over the internet (e.g., In order to modify one specific feature http://www.faceresearch.org), (2) analy- of a face on a photo while leaving the ses of the relationships between FacA other features intact, researchers have and sex hormone levels [PENTON-VOAK retouched photos [CARELLO et al. 1989] and CHEN 2004, LAW SMITH et al. 2006, and computer images [LAENG et al. RONEY et al. 2006], or genetic makeup 2007], drawn stubble on models [THORNHILL et al. 2003; ROBERTS et al. [MUSCARELLA and CUNNINGHAM 1996], 2005a, b]. or used drawings of faces with definite In the past, researchers seeking corre- features in place of photos [BERRY and lations between FacA and facial traits MCARTHUR 1986, REED and BLUNK sometimes used whimsical or ambiguous 1990]. Judges have been allowed to por- terminology (e.g., “Roman”, “hawk” or tray their conception of the most attrac- “pug” noses in WAGATSUMA and tive face by using sticks to form the face KLEINKE [1979]). In the middle of the [HERSHON and GIDDON 1980], moving 1980s, researchers began to use the fa- disks to denote eyes and lips [FRĄC- cialometric method, in which specified KIEWICZ 2001], or even employing a facial features such as distances and an- system of hinged telescopic pipes gles were measured with calipers on liv- [SERGL et al. 1998]. In the mid-1980s, ing people or photographs [MCARTHUR researchers started to use facial draw- and APATOW 1983/1984; BERRY and ings produced by superimposing trans- MCARTHUR 1985, 1986; KEATING 1985; parent sheets with specific variations of CUNNINGHAM 1986; KEATING and BAI each facial part (or an analogous com- 1986; FARKAS et al. 1987a; MCARTHUR puterized technique) [MCARTHUR and and BERRY 1987; STRZAŁKO and APATOW 1983/1984, KEATING 1985, KASZYCKA 1988, 1992]. Nowadays, VON FAUSS 1988]. these features are measured with the aid An important approach to studies on of special computer software programs. FacA is the merging of two or more Few studies, however, have analyzed the facial photographs. In the nineteenth General patterns of facial preferences 49 century, Herbert Spencer superimposed PERRETT 1992, JOHNSTON and FRANK- semi-transparent sheets with facial LIN 1993, PERRETT et al. 1994, KUJAWA photos, and Francis Galton developed and STRZAŁKO 1998]. The techniques a step-wise method of exposing a plate themselves have been described by ROW- of different facial photos [GALTON LAND and PERRETT [1995] and TIDDE- 1878]. Over one hundred years later, MAN et al. [2001]. By judiciously com- LANGLOIS and ROGGMAN [1990] de- bining these techniques, the researcher veloped an analogous computerized can test many hypotheses on FacA. method. A turning point in studies on FacA was Theories of facial attractiveness the development of computer warping and morphing. Warping is an elastic Adaptation-oriented theories change of facial shape (Fig. 1). The proc- ess resembles localized pulling on the There are two types of theories to ex- surface of a balloon with a face drawn on plain why some faces are considered it. Morphing is a process of creating a attractive and others not: adaptation- face intermediate in shape and color be- oriented theories and non-adaptation- tween two other faces. The intermediate oriented theories. The adaptation-orien- shape is obtained by the process of ted theories state that facial preferences warping. The first studies on FacA based are adaptations, that is, they arose by on these techniques appeared at the be- natural selection, and are therefore ad- ginning of the 1990s [BENSON and vantageous to their owner. Natural selec-

(A)

→

(B)

+=

Fig. 1. Warping and morphing of a face. (A) Warping deforms a face in an elastic way. (B) Morphing blends two faces into an image with intermediate shape and color. 50 K. Kościński tion favored these preferences because an studies on the heritability of preferences individual possessing them had higher for facial features have yet been con- reproductive success. Bonding or mating ducted even though heritability of mating with a partner who is infertile, unhealthy preferences has been proved in animals or genetically loaded is disadvantageous [JENNIONS and PETRIE 1997]. to the individual. On the other hand, However, for the adaptive view of having a partner who is phenotypically facial attractiveness to be true, some and genotypically healthy is advanta- conditions need not necessarily be met: geous. An individual X has a high mate (1) People need not be conscious of value (MV) for an individual Y if mating the adaptive value of their preferences. with individual X is more profitable for (2) Individuals possessing preferred traits individual Y than mating with a random need not be healthier in the broadest partner. The benefits may be direct or sense of the word, nor have a higher sur- indirect. Direct benefits include invest- vival rate than others. Specifically, high ments by the partner in the interested quality individuals may invest so much individual and their common offspring. energy in the development and the Indirect benefits include passing of good maintenance of traits preferred by the genes from the partner to the common opposite sex that they have less energy offspring. for the development and the maintenance The main problem in mating is distin- of other biological traits, such as immu- guishing an individual with a high MV nity to disease [GETTY 2002]. (3) The from an individual with a low MV based preferences prevailing in contemporary solely on sensory input (Fig. 2). This is westernized populations may differ from made even more difficult by the fact that those that drove sexual selection in hu- the prospective partner has a reproductive mans in the evolutionary past (because of interest in signaling a high MV, even changes in the environment). The above when this is not the case. Therefore, there phenomena make it difficult to test the are reasons to cheat. There is also a risk hypothesis that facial preferences are of being cheated, and consequently, of adaptive. losing out in terms of reproductive suc- cess. According to the concept of honest Non-adaptation-oriented theories signaling, the co-evolution of signals and preferences for them leads to a situation Non-adaptation-oriented theories state where signals are reliable cues to the that preferences for specific facial fea- quality of the sender (see GANGESTAD tures are by-products of phenomena that and SCHEYD [2005]). have little to do with the assessment of A prerequisite for the evolution of the attractiveness. Therefore, these prefer- face and preferences for facial features is ences are not adaptations and do not lead that they are in part genetically deter- to the choice of a high-MV partner. Non- mined. The coefficient of heritability ran- adaptive preferences may arise in several ges from 0.34 to 0.65 for various facial ways: features [SUSANNE 1977] and amounts Overgeneralization. Assume that it is 0.64 for female FacA [MCGOVERN et al. advantageous for an individual to re- 1996]. To the best of my knowledge, no spond with reaction R to stimulus S and GENOTYPE transmitted genes PROSPECTIVE genes quality MATE genes complementarity fitness of common offspring

visible cues to PHENOTYPE the will and possibilities parental to invest and to genes investments fitness of the the will to invest quality and interested complementarity individual possibilities to invest pro-mate investments

e b n ld tio ou la the assessment by the interested individual decision on mating h re S or c a

Fig. 2. The evolutionary view of the mechanisms of mate choice. An appropriate partner is one that has a high a mate value. This ensures that the offspring will be biologically fit. Physical appearance and other sensory cues have evolved to facilitate the decision process. An appropriate choice can be made only after evaluating reliable indicators of high genotypic and phenotypic quality. 52 K. Kościński that the individual possesses an evolu- Mere exposure effect. This phenome- tionarily-shaped mechanism of perform- non depends on a change in attitude to- ing the reaction to the stimulus. There ward a stimulus to being more positive may also exist a stimulus T that is similar by merely observing the stimulus (see to stimulus S to the degree that the indi- ZAJONC [2001]). vidual reacts to stimulus T with reaction All of the above mechanisms are based R, even though the reaction in this case is on distinct modes of sensory organ and not advantageous. Therefore, performing nervous system function that unadap- reaction R in response to stimulus T is tively biases preferences toward some not adaptive but the by-product of the kinds of stimuli. This is the reason that adaptation to stimulus S. In this case, the these preferences are called sensory or individual is said to overgeneralize perceptual biases. A quite different and stimulus S (see ZEBROWITZ and RHODES frequently non-adaptive mechanism by [2004]). For example, female zebra which preferences are formed is social finches prefer males with red beaks. This learning, most often, mimicking. This is an adaptation because a red beak is a phenomenon is common in humans. At sign of health. However, they also prefer the population level, it takes the form of males with red slips experimentally at- short-lived, constantly changing fashions. tached to their legs – this is an overgen- eralization (see JONES [1996b]). Determinants of attractiveness Overdiscrimination. If it is assumed that it is advantageous for an individual Age to react with reaction R to stimulus S but disadvantageous to react this way to A child’s face has the following fea- stimulus T, a well-fitting individual will tures: high forehead, large, round eyes, distinguish stimuli S from stimulus T and bluish sclera, high eyebrows, thick, red perform reaction R only in response to lips, short, wide, concave nose, wide stimulus S. In many conditions such as cheeks, short ears, and light, soft skin this, an individual reacts especially [BERRY and MCARTHUR 1985, 1986; strongly with reaction R to the stimulus JONES 1996b; ETCOFF 1999]. In chil- when those elements that differentiate dren, the lower part of head is smaller stimulus S from stimulus T are exagger- than in adults [MARK et al. 1988]. In ated. This phenomenon is called overdis- adolescence, a boy’s face changes much crimination or super-normal reaction to more than does a girl’s face. Above all, a super-normal stimulus and is wide- the jaw and superciliary arches become spread in humans and other animals (see larger. For this reason, a child’s face is ENQUIST and ARAK [1993]). more similar to a female face than that The fluency of data processing. Not of a male. In adulthood and old age, the all stimuli are analyzed by the brain with most apparent facial change is wrinkling equal speed and efficiency. The attrac- [MARK et al. 1980]. Other changes also tiveness of some stimuli may be due occur, including: (1) descent of the eye- to the ease with which they are proces- brows over the orbital ridges, which sed by the brain (see HALBERSTADT makes the eyes appear smaller; (2) in- [2006]). crease in size of cartilaginous structures General patterns of facial preferences 53 such as the ears and the nose; (3) thin- and TIBERGHIEN 2004, MESKO and ning of the vermilion [JONES 1996b]. BERECZKEI 2004]. FacA in females is Age is an important criterion for mate also positively correlated with facial choice. Men in particular strongly prefer youthfulness as perceived by judges younger females [BUSS 1999]. Even (KORTHASE and TRENHOLME [1982], four-month-old infants look longer at TATARUNAITE et al. [2005], but see also the faces of children than at the faces of BERRY [1991a] and ZEBROWITZ and adults [MCCALL and KENNEDY 1980]. MONTEPARE [1992]). JONES [1995] also Preferences for youth are reflected in the showed that models have baby-like facial assessment of FacA of people at various proportions. The features that decrease ages, at least as far as female faces are perceived age and increase FacA in fe- concerned. According to MATHES et al. males are most of all big eyes, short nose, [1985], women’s FacA gradually de- small chin, thick lips, but also narrow creases with age, both for men and jaw, thin eyebrows, and wide-set eyes. women judges, although the effect was Besides having many baby-like traits, an weaker when the judge was a woman, or attractive woman should also possess when the judge was over 40 years old. some adult traits, in particular pro- On the other hand, for males, age is not nounced cheekbones and narrow cheeks generally related to FacA. In both males [CUNNINGHAM 1986, CUNNINGHAM et and females, the relative FacA of a par- al. 1995, BAUDOUIN and TIBERGHIEN ticular individual in comparison with his 2004]. A female face in which all of the peers remains constant from childhood proportions are childlike is perceived to middle adulthood [ALLEY 1993, simply as childish and not as attractive ZEBROWITZ et al. 1993, TATARUNAITE [GRAMMER et al. 2003]. et al. 2005]. The positive correlation between According to ZEBROWITZ and MON- babyfacedness and FacA in women was TEPARE [1992], there is a correlation confirmed by many studies carried out between FacA and perceived age only in on faces which age-related features were infants. HILDEBRANDT and FITZGERALD modified [MCARTHUR and APATOW [1979] determined that cute infants are 1983/1984, VON FAUSS 1986, RIEDL characterized by large eyes and pupils, 1990, JOHNSTON and FRANKLIN 1993, and a large forehead. HÜCKSTEDT PERRETT et al. 1994, JONES 1995]. The [1965] found that women prefer babyish traits having the greatest effect on FacA proportions to such a degree that they were a small lower part of the face (nar- consider children with pathologically row jaw, short chin, and lips close to the enlarged crania to be attractive because nose), large eyes and thick lips. this is an exaggerated baby-like trait. JOHNSTON and FRANKLIN [1993] found Studies of real faces consistently show that the most attractive versions of adult that FacA in females is positively corre- women’s faces were those having pro- lated with the measured baby-likeness of portions typical of 11-14 year-old girls. the facial proportions [CUNNINGHAM Real faces studies were less consistent 1986, FARKAS et al. 1987, FARKAS 1994, for men than they were for women. CUNNINGHAM et al. 1995, JONES 1995, JONES [1995] and ZEBROWITZ and MEHRABIAN and BLUM 1997, BAUDOUIN MONTEPARE [1992] found no correlation 54 K. Kościński between babyfacedness in males and of children she can rear) is very limited. FacA. BERRY and MCARTHUR [1985] A woman cannot give birth more fre- reported that some childlike features, quently than once a year. She usually including large, round eyes, high eye- gives birth to a single child. Furthermore, brows and a small chin, increase FacA in her reproductive potential decreases with males. CUNNINGHAM et al. [1990] found age, and ceases when she reaches the age that an attractive male face is a specific of about 50. Therefore, from a reproduc- combination of child-like features, such tive point of view, it is better for a man to as large eyes, and mature features, such bond with a young woman than with an as a prominent chin. The results of studies old one. If the bond is expected to be conducted on modified faces are consis- short-lived, the criterion for partner tent in that the most attractive male face choice should be the probability that the has medium values for age-related traits woman will conceive, which peaks at (MCARTHUR and APATOW [1983/1984], around the age of 25 years. If the bond is RIEDL [1990], JONES [1995], but see also expected to be long-lasting, the criterion VON FAUSS [1988]). for partner choice should be the repro- Both studies based on real as well as on ductive potential of the woman (i.e., the modified faces gave similar, compatible number of children she can still bear), results: (1) in children, at least in infants, which peaks just after puberty. For these FacA is positively correlated with con- reasons, during the course of evolution, spicuous baby-like features, (2) there is a men developed a preference for young substantial preference for baby-faced women, more precisely for apparent, women, yet also for two adult features, though not necessarily authentic, signals i.e., pronounced cheekbones and narrow of female youthfulness. Therefore, male cheeks; (3) there is a weak correlation preference for young-looking women is between FacA and facial youthfulness in an adaptation, and a preference for very men (FacA is highest in men with me- young-looking women is an overdis- dium or slightly baby-like values for age- crimination. In order for male prefer- related traits). ences not to be directed toward females In many animals, young individuals that are too young, they have to be de- have special features to signal their age. pendent on maturity cues such as pro- These traits trigger caring instincts in nounced cheekbones, narrow cheeks, and their parents and inhibit aggressive be- also a mature, feminine body shape. havior in all adults. One cue to young age Reproductive capacity is far higher in in mammals is a relatively large cranium males than in females. It declines with with large eyes. This kind of age- age far more slowly than in women. signaling and its social effects are also Therefore, a woman’s reproductive suc- found in humans. Children who poorly cess depends relatively little on the age exhibit their young age get less care from of her partner, as long as he is neither a their parents and are maltreated more child or an old man. In old age, a man’s frequently [MCCABE 1984]. reproductive efficiency decreases sub- Female reproductive capacity (ex- stantially. His health, physical fitness pressed as the number of pregnancies she and, frequently, social status, deterio- can successfully undergo and the number rates. The probability that he will die General patterns of facial preferences 55 increases, and the probability that he will coordinates. In spite of various meth- still be around to carry out his duties to odological drawbacks, their results were his wife and children decreases. There- unambiguous: (1) in almost every case, fore, there are good reasons why a wo- the composite face was more attractive man would choose not to mate with an than the original pair of faces; (2) higher old man. On the other hand, a very young order composites, that is, composites man is usually psychologically immature, produced from other composites, were emotionally unstable and irresponsible. even more attractive; (3) the effect ap- He has low social status and possesses plied equally to both sexes; and (4) the few resources. The theoretically best FacA of a composite did not depend on partner for a woman is middle-aged, the FacAs of the original faces. The which is why male faces signaling this authors concluded that the averageness age are the most attractive to women. At is the essence facial attractiveness. the same time, BORKAN and NORRIS These results were extensively criti- [1980] reported that young-looking old cized, however. ALLEY and CUNNING- men are healthier and have a lower HAM [1991] disputed that the most beau- biological age than their peers. Therefore, tiful faces are average in shape. After all, youthful appearance in old men is a hon- it had recently been established that FacA est cue of their biological quality. (at least in females) is increased by de- partures from averageness toward baby- Averageness facedness, and not by averageness itself. Critics [ALLEY and CUNNINGHAM 1991, The common belief that exceptionally BENSON and PERRETT 1992] suggested attractive faces have some special fea- that the high FacAs recorded for the tures responsible for their beauty was composites were not due to average pro- challenged at the end of nineteenth portions, but to smooth, clean, flawless century by GALTON [1878] who found skin, symmetry, and youthful appear- that an experimentally averaged face ance, or that the reason lay in the fact that (obtained by exposing a plate step-wise the images were imperfectly merged, to different facial photographs) pro- which blurred the face and enlarged fa- duced an image of above-average at- cial elements such as the eyes and lips. tractiveness, even when the original Langlois and Roggman attempted to faces belonged to people who were not refute the criticism with theoretical ar- at all attractive. Over 100 years later, guments and further experiments [LANG- Judith Langlois and Lori Roggman LOIS et al. 1994]. But they succeeded transferred the technique of face su- only partially. They qualified their con- perimposition from plate to computer. clusion by stating that while averageness This resulted in the most illuminating is not decisive in determining FacA, it is paper on FacA to date [LANGLOIS and the most important factor of FacA. ROGGMAN 1990]. The researchers in- With warping, the negative effects of putted gray-scale images of faces in a simply superimposing faces on one an- computer, and then collapsed them other can be avoided. The results of pairwise by calculating the arithmetical studies using warping were clear, and means for pixel brightness at the same the conclusions drawn from them can 56 K. Kościński be summarized as follows: (1) Warping In her meta-analysis, RHODES [2006] a face toward averageness increases found a strong correlation between FacA FacA; (2) Caricaturing, or warping in and averageness (r = 0.52). The correla- the opposite direction, decreases FacA; tion held true for both male and female (3) Averaging facial shape improves faces and was higher in manipulated FacA more than averaging (i.e., smooth- faces (r = 0.67) than real faces ing) of facial texture in both male and (r = 0.40). For real faces, the correlation female faces; (4) Averageness increases was much higher when facial aver- FacA in both frontal and profile views ageness was assessed by independent (frontal: BENSON and PERRETT [1992], judges (r = 0.47) than when measured RHODES and TREMEWAN [1996], KUJA- objectively (r = 0.09). This proves that WA and STRZAŁKO [1998], O’TOOLE et the methods used to measure aver- al. [1999], RHODES et al. [1999], LITTLE ageness were unsatisfactory. and HANCOCK [2002], VALENTINE et al. Adaptation-oriented explanations are [2004]; profile: SPYROPOULOS and HALA- based on the assumption that individuals ZONETIS [2001], PEARSON and ADAMSON with average faces have a higher biologi- [2004], VALENTINE et al. [2004], VALEN- cal quality than others. They therefore ZANO et al. [2006]). have a high MV for a potential mate. If From studies in which the co-variables so, the preference for facial averageness were controlled, it was concluded that is adaptive and may have been shaped by averaging face shape increases FacA in evolution. The relationship between av- and of itself and not by merely smooth- erageness and biological fitness may ing the skin [LITTLE and HANCOCK have evolved by natural or sexual selec- 2002, RHODES and TREMEWAN 1996], tion: by lowering the perceived age [O’TOOLE (1) The performance efficiency in a et al. 1999], by improving symmetry given individual depends on the size and [BAUDOUIN and TIBERGHIEN 2004, shape of various parts of his body, in- RHODES et al. 1999, VALENTINE et al. cluding the components of his face. 2004], or by giving the face a positive Therefore, natural selection adjusted expression [RHODES et al. 1999]. Most facial proportions to be optimal for the studies conducted on real faces confirm environment our ancestors lived in. Natu- that FacA is correlated with the level of ral selection continues to act in a stabi- averageness [LIGHT et al. 1981; FARKAS lizing manner. Nonetheless, in the faces et al. 1987; STRZAŁKO and KASZYCKA of some individuals, there are departures 1988, 1992; FARKAS 1994; RHODES and in one direction or another from the op- TREMEWAN 1996; RHODES et al. 1999, timal proportions. Therefore, individuals 2005b; GRAMMER et al. 2002; BAU- with average facial proportions are more DOUIN and TIBERGHIEN 2004; but see biologically fit than individuals with also POLLARD 1999, and JONES 1995]. atypical facial proportions. EDLER et al. [2006] found that the level (2) Since the association of facial aver- of improvement in FacA after cosmetic ageness with biological quality arose as surgery is positively correlated with the the result of natural selection, a prefer- degree to which the proportions of the ence for mates with average faces should face were altered towards averageness. also have evolved. This preference for General patterns of facial preferences 57 average faces is an additional selective dency. In that case, the preference for pressure for facial averageness, and this average faces would not just be as a re- is known as sexual selection. In the sult of a specific adaptation that evolved population as a whole, the correlation to enable individuals to find high-quality between facial averageness and biologi- mates. The most popular non-adaptation- cal quality will persist because average oriented explanations for the preference facial features can develop only in indi- for average stimuli are those known as viduals with a high biological quality. Processing Fluency and Familiarity: Factors such as mutations, malnutrition (1) Processing Fluency. Typical faces and infections lower biological quality are more similar to neutral facial proto- and disturb the normal development the types than atypical faces. Therefore, they face so that it is atypical in some re- are more quickly and accurately proc- spects. essed [HALBERSTADT 2006]. This makes To be able to determine how far the them more pleasant to look at, and there- shape of a face departs from the average, fore, more attractive. an observer has to possess a neutral (2) Familiarity. Through mere expo- model of the face, or prototype, with sure, familiar objects are regarded as which he can compare the faces he sees. more attractive than unfamiliar objects. More precisely, he needs not only one LANGLOIS and ROGGMAN [1990] found prototype, but several prototypes, or sev- that average faces were assessed as being eral subtypes of a single prototype. This more familiar. This is understandable as allows him to recognize and judge the most of the faces we see during our lives faces of men, women, and children, etc. have proportions close to average, and The credibility of adaptation-oriented relatively few of the faces we see are explanations was challenged by the dis- very atypical. For these reasons, it was covery that people prefer not only aver- hypothesized that average faces are at- age faces, but also average-looking ani- tractive simply because they are more mals and objects such as dogs and cars familiar or perceived as such. [HALBERSTADT and RHODES 2000, Unfortunately, studies on the role of 2003; HALBERSTADT 2006]. They also processing fluency and familiarity in preferred arrangements of abstract ele- determining the attractiveness of average ments such as dots that were similar to faces are not consistent [RUBENSTEIN et arrangements to which they had been al. 1999; RHODES et al. 2001a, 2005a; previously exposed during the course of HALBERSTADT 2006]. Therefore, we the experiment [WINKIELMAN et al. cannot easily resolve the question of 2006]. Furthermore, the correlations whether preference for average faces is a between attractiveness and averageness specific, non-adaptive example of a gen- for non-facial categories are approxi- eral preference for averageness in any mately as strong as they are for faces. category, whether the general preference This suggests that the human brain tends is an overgeneralization of the adaptive to prefer stimuli that are typical for the preference for average faces, or whether category to which they belong. The pref- the preference for average faces depends erence for average faces is only therefore on both non-adaptative and adaptive a specific example of this general ten- mechanisms. 58 K. Kościński

Facial sexual dimorphism KOEHLER et al. [2004] found a positive correlation between FacA and perceived In small children, facial dimorphism is facial femininity in women. Many re- minimal. Substantial differences only searchers have used morphing to con- appear in puberty, when facial features struct sets of female faces that repre- undergo more radical changes (in both sented various mixtures of the typical size and proportions) in males and fe- male face and the typical female face. males. These changes are more marked Some of the images included in the sets in males than in females. Adult female were hyper-feminine faces with exagger- faces are therefore more similar to chil- ated feminine features. Most male sub- dren’s faces than male faces. The differ- jects considered the hyper-feminine faces ences between male and female faces to be even more attractive than a 100% correspond largely to the differences typical female face [PERRETT et al. 1998, between adults’ and children’s faces. RHODES et al. 2000, JOHNSTON et al. Compared to the female face, the male 2001, DEBRUINE et al. in press]. How- face is large, angular and convex in pro- ever, highly exaggerated feminine facial file view. It also has small, deep-set features reduces FacA because they make eyes, prominent superciliary arches, a woman appear dominating [JOHNSTON thick, bushy eyebrows that are low-set, et al. 2001]. In many studies, male sub- less prominent cheekbones, a wide, pro- jects preferred female faces with thick truding nose, thin lips, a wide mouth, and prominent lips [GIDDON 1995, big chin and a wide jaw [ETCOFF 1999]. BAUDOUIN and TIBERGHIEN 2004, In both males and females, high testos- TURKKAHRAMAN and GOKALP 2004, terone levels during the fetal stage or SCOTT et al. 2006]. BISSON and GROB- during puberty magnify masculine facial BELAAR [2004] found that, compared to traits [PENTON-VOAK and CHEN 2004, ordinary women, female models had FINK et al. 2005a]. In women, estrogen thicker upper and lower lips having dou- levels during puberty determine fat ble the surface area and a more distinct deposition throughout the body. In the Cupid’s bow. The meta-analysis carried face, adipose tissue develops in the lips, out by RHODES [2006] showed that the which makes them look full, protruding, correlation between FacA and facial and even curled up [ETCOFF 1999]. In femininity in women is strong for both relation to overall facial size, women real and modified faces (r = 0.64). have relatively thicker lips than children, In studies on the correlation between while children have thicker lips than FacA and facial masculinity in men, the men. In this sense, the thick lips of the results were inconsistent: (1) A negative adult female face are a hyper-childlike correlation between FacA and facial feature. On the other hand, in terms of masculinity was found in studies using the prominence of the cheekbones, the digitally modified faces [PERRETT et al. width of the cheeks, and the width of the 1998, RHODES et al. 2000, LITTLE et al. nose, adult male faces are more similar 2001, LITTLE and HANCOCK 2002], real to children’s faces than are female faces faces [BERRY and MCARTHUR 1985], [CUNNINGHAM 1986; CUNNINGHAM and drawings [MCARTHUR and APATOW et al. 1990, 1995; KOEHLER et al. 2004]. 1983/1984]. (2) A positive relationship General patterns of facial preferences 59 between FacA and facial masculinity was This reduced the perceived age for the found in studies using digitally modified modified faces [LITTLE and HANCOCK faces [JOHNSTON et al. 2001, SCARB- 2002]. This, in turn, may affect the cor- ROUGH and JOHNSTON 2005, DEBRUINE relation between FacA and facial mascu- et al. 2006], real faces [CUNNINGHAM et linity [RHODES 2006]. One way to avoid al. 1990, GRAMMER and THORNHILL these problems is to use warping to 1994, SCHEIB et al. 1999, PENTON-VOAK modify real faces, and not the composites et al. 2001, NEAVE et al. 2003, KOEHLER made from them. In two studies that used et al. 2004, FINK et al. 2007], and draw- this approach [KEATING and DOYLE ings [KEATING 1985]. (3) Two studies 2002, SWADDLE and REIERSON 2002], on real faces found no correlation be- the highest values for FacA were coinci- tween FacA and facial masculinity dent with moderate values for facial mas- [RHODES et al. 2003, WAYNFORTH et al. culinity. If these results are accurate, it is 2005]. (4) In two studies using digitally easy to explain (1) why no correlation modified faces, there was a curvilinear between FacA and facial masculinity was relationship between FacA and facial detected in some studies (because not all masculinity, with the highest values for statistical methods can identify parabolic FacA coincident with moderate values relationships), and (2) why positive or for facial masculinity [KEATING and negative correlations were found in some DOYLE 2002, SWADDLE and REIERSON studies (because many small-effect ran- 2002]. dom impacts can move the FacA peak in In her meta-analysis, RHODES [2006] either direction along the face masculin- showed that there was a negative corre- ity axis). Therefore, studies to date sug- lation between FacA and facial mascu- gest that women prefer facial shapes that linity in studies using digitally modified are near the masculine average. How- faces (r = -0.47) but a positive correla- ever, in studies on real faces, women tion in studies conducted on real faces preferred facial shapes which were above (r = 0.35). Therefore, the results obtained the masculine average. This is probably partly depend on the experimental because masculinity of male facial shape method used. is correlated with some facial features In studies on real faces, being correla- that determine FacA (e.g., symmetry or tional ones, the possible influence of co- skin quality) [SCHEIB et al. 1999]. variables such as complexion, skin tex- Since female faces and children’s faces ture, facial expression and stubble were have similar features, men prefer highly not controlled for. Therefore, the positive feminized female faces for the same rea- correlations observed in these studies do sons that they prefer baby-faced women. not necessarily indicate that there is a Evolutionary pressure for youthful ap- causal relationship between facial mas- pearance applies exclusively, or at least culinity and FacA. For most studies using mainly, to women. Therefore, the differ- digitally modified faces, one of the ences between male and female faces preparative stages involves digital modi- correspond to the differences between fication of original faces to produce fa- adults’ and children’s faces. cial composites which have visibly Conspicuously reddened lips are under smoother complexions than the originals. the control of estrogen, and are probably 60 K. Kościński a signal of fecundity. It is more difficult VOAK and CHEN [2004], WAYNFORTH to explain why high estrogen levels are et al. [2005]). Therefore the curvilinear manifested in this particular way. It has relationship between FacA and facial been suggested that conspicuously red masculinity in male faces may be ex- lips resemble the female genitals in shape plained as a compromise made by and color, and thereby sexually arouse women between biological quality and men. Therefore, they might have been good character. On the other hand, subject to sexual selection during the a highly masculinized male face, being course of evolution [MORRIS 1967]. On a signal of physical strength [FINK et al. the other hand, JONES [1996a] suspected 2007] and dominant, aggressive person- that conspicuous red lips signal that the ality [KEATING 1985, CUNNINGHAM et woman does not suffer from anemia or al. 1990], allows the man to successfully infections, and are therefore a reliable compete with other men, to reach a high signal of the woman’s current health and social status, and to have better access to fecundity. more and better women. Therefore, qual- Men with highly masculinized faces ity-dependent facial masculinity in males are perceived to be, and actually are of may have evolved initially through intra- high biological quality, including good sexual selection. Only later, after these genes and health in the broadest sense of traits became honest cues to male quality, the word [RHODES et al. 2003, ZEBRO- did masculinized male faces become WITZ and RHODES 2004, THORNHILL and attractive to women. Inter-sexual selec- GANGESTAD 2006, FINK et al. 2007]. tion then began to influence the further However, they are also perceived as evolution of the male face. having poor pro-family attitude (willing- Some male facial features might have ness to invest time and resources in both evolved through natural selection rather the woman and the common offspring) than intra-sexual or inter-sexual selec- and as being dominant, aggressive, and tion. For example, men have larger even asocial [KEATING 1985, KEATING noses and mouths than women because and BAI 1986, BERRY 1991b, PER- our male ancestors were more frequently RETT et al. 1998, JOHNSTON et al. 2001, required to subject themselves to pro- KEATING and DOYLE 2002]. The above longed physical effort which demanded correlations can be attributed to the ac- intensive breathing. Similarly, thick tion of testosterone. High testosterone eyebrows and prominent superciliary levels are a reliable indicator of high arches in men have evolved to protect biological quality and determine the high their eyes from the sun and sweat. Some level of facial masculinity [PENTON- studies on preferences that are depend- VOAK and CHEN 2004, RONEY et al. ent on contingencies provide strong 2006]. Men with high testosterone levels evidence that the preference for facial have less successful relationships with masculinity in male faces is adaptive. their women and children, are more For example, women prefer more mas- likely to cheat on their partners, be single culinized male faces while in the fertile than involved in a relationship, and ex- phase of their menstrual cycle than the hibit antisocial behavior (see PENTON- infertile phase. This is because they are VOAK and PERRETT [2001], PENTON- more likely to be seeking a man with General patterns of facial preferences 61 good genes when the probability of con- ANNs is also responsible for the evolu- ception is high (e.g., PENTON-VOAK tion of dimorphism in the human face et al. [1999]). Female preferences for and for the evolution of preferences for facial masculinity in male faces may dimorphic facial features. (Note that the also be a by-product of age preference, mechanism presented above can increase as is the case with male preference for the already existing dimorphism but can- facial femininity in females. BOOTH- not initiate it.) ROYD et al. [2005] found that perceived GHIRLANDA et al. [2002] claimed that age in male faces was correlated with they observed overdiscrimination in chick- perceived masculinity. Furthermore, wo- ens (that is, in natural neural networks) men who preferred older men also pre- which had been taught to distinguish ferred men with more facial masculinity. between male and female faces and con- On the basis of their experiments with cluded that preferences for highly dimor- artificial neural networks (ANNs), phic facial features are non-adaptive. ENQUIST and ARAK [1993] concluded Further research is therefore needed. that preferences for conspicuous sexual However, the adaptation-oriented and features evolved as a by-product of the non-adaptation-oriented explanations are need to distinguish between males and not necessarily mutually exclusive. Both females. An ANN (representing an indi- adaptive and non-adaptive mechanisms vidual of a given sex) that has been might have played a role in the evolution taught to react (i.e., to mate with) to of preferences for facial dimorphism. a stimulus of type A (the opposite sex individuals) and ignore a stimulus of type Symmetry B (the same sex individuals) reacts espe- cially strongly to a stimulus of type A in There are several kinds of asymmetry which the differences between types A of biological structures: and B are exaggerated (that is, individu- Directional asymmetry (DA): Differ- als with exaggerated sexual features). ences of values of a trait between the This is an example of overdiscrimination, sides are systematic, and the mean value and, over many simulated generations, of the difference in the population differs results in an increasing preference for from zero significantly. That is, in most highly conspicuous traits in the opposite individuals, the value on one side (left or sex (stimuli of type A) as well as in- right) is larger than on the other. For creasing sexual dimorphism (differences example, in humans the left kidney is between stimuli of type A and B). This located a few centimeters higher than the co-evolution of preferences and stimuli is right one. SIMMONS et al. [2004] found an example of so called runaway selec- DA for most of measured facial traits tion. After a time, it is counteracted by (23/35 in men, 28/35 in women). the negative effects of exaggerated sexual Anti-symmetry (AS): Values of a trait traits which include high production strongly differ between the sides but the costs, dysfunction due to deformed or- larger side in some individuals is the left gans, and overall reduced survival rates. one, while in the others the right one. ENQUIST and ARAK [1993] proposed that Claw-size in fiddler crabs is a good ex- the mechanism observed in studies on ample of this. 62 K. Kościński

Fluctuating asymmetry (FA): Dif- on the left side of the face. This means ferences in values of a trait between that the right side of the face is more sides are random, while the mean value attractive than the left side. The authors of the difference in the population may suspected that this was due to the ob- be zero (i.e., no DA) or non-zero (i.e., scuring of attractiveness cues by facial DA exists). expression which is more strongly ex- DA and AS have functional signifi- pressed on the left side of the face cance, or are the by-products of function- [KOWNER 1996]. ally significant DA or AS in another trait. Numerous studies have been carried Thus, they are determined by processes out on the relationship between FacA and controlled by the organism. On the other the level of facial asymmetry because the hand, FA is caused by random deviations level of asymmetry is believed to be an from the normal and is not controlled by indicator of developmental stability and, the organism. Two methods are used to therefore, biological quality. The method estimate the level of facial asymmetry: employed has a considerable influence Measurements. With this method es- on the result obtained. sential points (landmarks) are first lo- (1) In many of the studies performed cated on a facial photograph. For paired on real faces a positive correlation be- points such as the outer corners of the tween FacA and facial symmetry was eyes, the distance of each point from found. Some of these studies were based other paired or unpaired points is re- on facial measurements [GRAMMER and corded. The difference between the dis- THORNHILL 1994, RIKOWSKI and GRAM- tance on the left and that on the right MER 1999, HUME and MONTGOMERIE indicates the level of asymmetry. For 2001, JONES et al. 2001, PENTON-VOAK unpaired points such as the tip of the et al. 2001, BAUDOUIN and TIBERGHIEN nose, the deviation from the vertical axis 2004]. Others were based on direct is recorded. [RHODES et al. 1998, 2001c; KOEHLER et Assessment by judges (perceived al. 2004; SIMMONS et al. 2004] or indi- asymmetry). Judges may assess facial rect assessment of asymmetry [MEALEY asymmetry either directly or indirectly. et al. 1999, PENTON-VOAK et al. 2001]. In direct assessment, the judges simply FARKAS [1994] and JONES [1996b] estimate the level of asymmetry they see found little support for positive correla- in a face. In indirect assessment, two tion between FacA and facial asymme- chimeras are made for each face. A chi- try. Only SHACKELFORD and LARSEN mera is an image that consists of one side [1997] found a negative correlation. of the original face combined with its CUNNINGHAM [1986] and LANGLOIS et mirror image. One chimera is based on al. [1994] found no correlation between the left side of the face and the other FacA and facial symmetry. based on the right side of the face. Judges (2) A chimera is constructed from the then assess how similar the two chimeras image of only one side of the face and is are to each other. therefore perfectly symmetrical by defi- ZAIDEL et al. [1995] found that the nition. Nevertheless, FacA was lower chimera based on the right side of a face in chimeras than in real faces [LANG- was more attractive than chimera based LOIS et al. 1994; KOWNER 1996, 1997; General patterns of facial preferences 63

RHODES et al. 1999a; NOOR and EVANS less attractive than the faces from which 2003]. they had been constructed. In images (3) A symmetrical image can be con- created by morphing, skin blemishes may structed by morphing a face with its mir- be duplicated, which decreases FacA. ror image. In two studies, FacA was This is believed to be the reason why higher in the symmetrical face than the SWADDLE and CUTHILL [1995] found a original face [RHODES et al. 1998, negative correlation between FacA and 2001b]. SWADDLE and CUTHILL [1995], facial symmetry. On the other hand, however, found that FacA was lower in RHODES et al. [1998, 2001a] circum- the symmetrical face than in the original vented this problem by graphically re- face. moving all facial blemishes before (4) Warping allows construction of a morphing, and found a positive correla- symmetrical image without changing tion between FacA and facial symmetry. skin texture. The texture of the images to Warping allows modification the level of be compared may be derived from the asymmetry without any serious side- original face (rendering it asymmetrical effects. In all studies using warped im- to some degree) or from a composite ages, there was a positive correlation image (when it remains smooth and between FacA and facial symmetry. symmetrical). Warping can also be used Furthermore, preference for symmetry is to exaggerate asymmetry in the original so strong that judges consider symmetri- face. In all studies using warping, there cal versions of their own or their ac- was a positive correlation between FacA quaintances’ faces to be more attractive and facial symmetry [RHODES et al. 1998, than the originals [LITTLE and JONES 1999a, 2001a; PERRETT et al. 1999; 2003]. LITTLE and JONES 2003]. SIMMONS et al. [2004] found that di- In most studies on real faces, there was rectly perceived asymmetry was corre- a positive correlation between FacA and lated with the measured FA but not with facial symmetry. The few studies in DA. This means that perceived symmetry which there was not a clear positive cor- is not based on DA, probably because relation found were mostly conducted people are accustomed to DA and there- some time ago when methodologies were fore ignore it when evaluating facial not as refined. In the process of con- symmetry. In her meta-analysis, RHODES structing chimeras, the face is bisected [2006] found that for real faces there was exactly along the vertical axis of the face. a weak positive correlation between However, because few faces are perfectly FacA and facial symmetry (r = 0.23). For symmetrical, the vertical axis of the face morphed and warped faces, there was a rarely goes through the middle of every moderate positive correlation (r = 0.43). structure which is located in the center of For chimeras, there was a strong negative the face. Therefore, chimeras may have correlation (r = -0.62). This is true for some structures unnaturally large or both male and female faces and for both small, and they are considered by judges male and female judges. The correlation to be strange and atypical RHODES et al. between FacA and facial symmetry was [1999a]. This is why, in all studies using somewhat stronger when facial symme- chimeras, the chimeras were found to be try was assessed by judges (r = 0.30) 64 K. Kościński than when it was determined by meas- alcohol consumption and smoking by urement (r = 0.19). parents during pregnancy) disturb the In the 1990s, results of studies using symmetry of an animal’s body and a chimeras or poorly constructed morphed human body and face [JONES 1996b, images supported the hypothesis that ETCOFF 1999, ŻĄDZIŃSKA 2003]. It is facial asymmetry reduces FacA only believed that a low level of FA indicates when it is extreme or pathological, and (i) that the individual is able to develop that at normal levels asymmetry actually symmetric structures that are normal for increases FacA. Normal faces are slightly the species even despite adverse envi- asymmetric because of DA: for example, ronmental conditions, and (ii) the accu- facial expressions are more pronounced mulation of environmental stresses to on the left side of a face and vocal which the individual has been exposed movements on the right. Recent studies, during the course of his lifetime [THORN- however, proved that the correlation HILL and MØLLER 1997]. between FacA and facial symmetry is Facial asymmetry may be so severe as positive. The studies also suggested that to impair normal facial function. For judges either do not notice or even ignore example, an asymmetric jaw can reduce DA when evaluating FacA [RHODES et chewing efficiency or even cause the al. 1998, SIMMONS et al. 2004]. head to lean to one side, which can even If facial asymmetry is correlated with lead to scoliosis [JEFFERSON 1996]. An some other facial feature, a preference for asymmetric nose can impede breathing that feature may be falsely interpreted as which can reduce stamina [MANNING a preference for facial symmetry. Several and PICKUP 1998]. Therefore, natural features are possible covariates: age selection favors symmetrical faces, and, [FINK et al. 2005b], facial averageness in turn, the preference for symmetrical [JONES 1996b], sexual dimorphism faces should appear in the opposite sex. [SCHEIB et al. 1999], and skin condition Consequently, individuals with symmet- [JONES et al. 2004b]. However, facial rical faces have an advantage not only in symmetry also increases FacA in faces terms of anatomical function but also in that are digitally modified without terms of reproductive success. changing the level of masculinity or Sexual selection for symmetrical without changing skin texture, which faces should have produced a positive may provide cues to age and health. correlation between the level of facial Furthermore, RHODES et al. [1999a,b] symmetry and biological quality of the found that facial symmetry and facial individual (because only high-quality averageness contributed independently to individuals are able to develop a highly FacA. Therefore, facial symmetry in and symmetrical face). Indeed, a high de- of itself affects FacA even if it is corre- gree of facial asymmetry is positively lated with other determinants of FacA correlated with many somatic and men- such as age, averageness, sexual dimor- tal disorders [THORNHILL and MØLLER phism and skin condition. 1997]. However, the correlation between Many genetic (inbreeding, mutation) low levels of asymmetry and biologi- and environmental factors (parasitic, cal quality is not high (see RHODES malnutrition, environmental pollution, [2006]). General patterns of facial preferences 65

It has been proposed that the preference oriented rather than the non-adaptation- for facial symmetry is only a specific oriented explanation of the preference for example of the preference for symmetry facial asymmetry. in general, and is therefore not an adap- Therefore, support exists for both for tation which allows an individual to rec- the adaptation-oriented and the non-adap- ognize potential partners with high bio- tation-oriented explanations for prefer- logical quality. Indeed, common objects ence for facial symmetry. However, both and patterns are perceived as more pleas- explanations are not necessarily mutually ant if they are symmetrical (see LITTLE exclusive and both adaptive and non- and JONES [2003], CARDENAS and adaptive mechanisms may contribute to HARRIS [2006]). Explanations for the the existence of the preference. generality of preference for symmetric forms were attempted in terms of proc- Other determinants of facial essing fluency. But it may also be ex- attractiveness plained by the fact that our bodies as a whole are vertically symmetrical (this The preference for light skin is present would be an example of overgeneraliza- in most human populations [VAN DEN tion of the mere exposure effect for verti- BERGHE and FROST [1986]. However, cal symmetry) [LITTLE and JONES 2003]. the preference within a population is for Non-adaptive preferences for symme- skin tones that are lighter for that popu- try were also observed in ANNs. lation and not the lightest possible skin ENQUIST and ARAK [1994] found that tones. In most cultures, the preference for when an ANN was learning to recognize light skin in women is stronger than the an asymmetric stimulus presented from preference for light skin in men [VAN different rotational angles, it was devel- DEN BERGHE and FROST 1986]. There is oping a preference for a symmetric ver- also a general preference for light hair in sion of the stimulus. JOHNSTONE [1994] women and for dark hair in men [FEIN- trained an ANN to recognize bilaterally MAN and GILL 1978, JONES 1996b]. symmetrical objects by presenting the Probably, several mechanisms contribute ANN with images, all of which were to the preference for skin color: randomly asymmetric to either the left or In hunter-gatherer populations, a wo- the right. After training, the ANN was man is pregnant a large part of the time. most responsive to perfectly symmetric Her calcium requirement increases be- images, even though it had never en- cause of the developing fetus. Intestinal countered them before. JANSSON et al. calcium absorption is enhanced by vita- [2002] and SWADDLE et al. [2004] con- min D. The synthesis of vitamin D de- ducted the same experiment on birds (i.e., pends on capacity of the skin to allow the natural neural networks) and their results UV component of sunlight to penetrate. were consistent with those of Johnstone. UV penetration is more efficient when On the other hand, LITTLE and JONES the skin is light in color. This would ex- [2006] found that the intensity of the plain why natural selection favored preference for facial symmetry is not lighter skin in women more so than in correlated with the ability to detect facial men. On the other hand, UV light is asymmetry. This supports the adaptation- harmful (it destroys folic acid). Adults 66 K. Kościński are exposed to more UV light than chil- terms of the common genetic determina- dren are. Therefore, skin color is darker tion of skin color in both sexes [FROST in adults. Because the biologically opti- 2006]. In turn, the parallel evolutionary mal skin color is lighter in women than changes in skin and hair color and the men, the preference for light skin color preferences for them can be explained in should be stronger in men than women terms of the common genetic determina- [RUSSELL 2003]. tion of skin and hair color. Because skin Light skin became a trait associated color is lighter in women than men, the with childishness and femininity and men contrast between skin color and eye and prefer women with feminine and childish lip color is greater in women than men. traits. This gave rise to the preference for RUSSELL [2003] showed that digitally markedly lighter skin and hair color increasing the contrast by darkening the [FROST 2006]. Estrogen lightens the skin eyes and lips increases FacA in female whereas progesterone darkens it. There- faces but decreases FacA in male faces. fore, skin color in women is darker dur- Little is known about preferences for ing pregnancy and lighter around the eye color. FROST [2006] proposed that time of ovulation [FROST 1988]. This is eye color and hair color are determined more cogently explained by sexual se- by frequency-dependent selection. This lection rather than natural selection be- means that rare, exceptional colors are cause natural selection would dictate that preferred. LITTLE et al. [2003] found that skin color in females be lighter during both men and women preferred the same pregnancy when the calcium requirement eye color and hair color that their oppo- is particularly high. site-sex parent possessed. LAENG et al. In western societies, there is currently a [2007] found a particularly strong prefer- fad for tanned skin [JONES 1996b, FINK ence of blue-eyed men for blue-eyed et al. 2001]. It is believed that this is women, which is probably an adaptation because tanned skin is a marker of high for detection of non-paternity. socio-economic status (SES) because Healthy hair is shiny and strong only individuals with a high SES possess [ETCOFF 1999]. It is a reliable indicator the time for tanning. The preference for of overall reproductive and non- skin color may therefore be conditioned reproductive health in women [HINSZ et by the preference for a partner with a al. 2001]. If a woman has healthy hair, it high SES. Light skin is also a better clue makes sense for her to display it, that is, to health than dark skin. Many disorders, to wear the hair long and loose. Accord- including anemia, cyanosis, jaundice and ing to this supposition, HINSZ et al. infection, are easier to detect in an indi- [2001] found a positive relationship be- vidual with lighter skin. Therefore, it tween hair quality and hair length in makes sense not to choose a partner with women (age being controlled for). Using dark skin, because they may be conceal- images to which hairstyles were digitally ing cues to poor biological quality added, MESKO and BERECZKEI [2004] [ETCOFF 1999]. showed that FacA was correlated with The parallel evolutionary changes in hair length in women and this was espe- skin color and preferences for skin color cially true if the hair was worn loose. in women and men can be explained in Hair length and hair quality are nega- General patterns of facial preferences 67 tively correlated with old age, regardless The results of studies for the correlation of health status [HINSZ et al. 2001]. Hair between FacA and facial hair in men are quality also decreases with every succes- conflicting. Some studies found a posi- sive pregnancy (see HINSZ et al. [2001]). tive correlation [PELLEGRINI 1973, PAN- In summary, hair length and hair quality CER and MEINDL 1978, HATFIELD 1986, in women reflect general health, repro- REED and BLUNK 1990, HELLSTRÖM and ductive health, age and the number of TEKLE 1994] while others found a nega- past pregnancies. tive correlation [FEINMAN and GILL Little is known about women’s prefer- 1977, CUNNINGHAM et al. 1990, WO- ences for hair length and hair quality in GALTER and HOSIE 1991, MUSCARELLA men. PANCER and MEINDL [1978] found and CUNNINGHAM 1996]. Nevertheless, that women prefer men with long hair. there is general agreement that men with However, this preference is most proba- beards are perceived as older, more ma- bly subject to cultural factors and ture, manly, dominant, aggressive and changes over time. MUSCARELLA and threatening [CUNNINGHAM et al. 1990, CUNNINGHAM [1996] photographed men REED and BLUNK 1990, WOGALTER and wearing wigs ranging from full hair to HOSIE 1991, BARBER 1995, MUSCA- complete baldness. Both receding hair- RELLA and CUNNINGHAM 1996]. Be- lines and baldness decreased FacA and cause a beard accentuates the lower part perceived aggressiveness. On the other of the face, men with beards are per- hand, they increased perceived age and ceived as more masculine, just as are social maturity. The authors proposed men with large jaws and chins. The effect that baldness is an adaptive process and of facial hair on FacA may be similar to serves as an indicator of old age and re- the effect of facial proportions in males duced reproductive drive. This reduces on their FacA. Facial hair indicates high competitive and aggressive behavior in biological quality and social status1, and younger males. Thus, an old man can a weak will to invest in his mate and invest in his offspring and grandchildren offspring. Therefore, in regard to facial without interference from younger males. hair, women have to compromise when In women, on the other hand, intra-sexual choosing a partner. competition is weaker than in men and With age, facial skin becomes matt, does not usually involve physical vio- limp, and wrinkled, the cheeks cave in lence [BUSS 1999]. This explains why and the facial bones become more con- baldness evolved only in men. spicuous [ETCOFF 1999]. The same traits Facial hair seems not to serve a pur- may occur in younger individuals. Skin pose as far as health and survival are condition is therefore a cue to age and concerned [BARBER 1995]. It occurs only health, so it is hardly surprising that it in adults, and almost exclusively in affects FacA. Many skin changes that are males. The development of facial hair is under control of the male sex hormones. 1 Of course, each man decides for himself whether to wear a beard or not. However, the presence of the beard is also A lack of facial hair is an indicator of a signal of his ability and readiness to compete with other youthfulness and femininity. In women, men who test the honesty of the signal. MUELLER and the presence of facial hair decreases MAZUR [1997] showed that military men whose physiog- nomy dishonestly signaled their physical, social and FacA [ETCOFF 1999, FINK et al. 2001]. intellectual abilities, received the worst treatment. 68 K. Kościński clearly symptomatic of disease processes TURKKAHRAMAN and GOKALP 2004, (e.g., acne, eruptions, warts, moles, cysts, HONN et al. 2005, MAPLE et al. 2005]. tumors and lesions) reduce FacA There are no symmetric elements in fa- [MORRIS 1967]. Some of these changes, cial profiles, so symmetry cannot effect such as inflammation and acne, are in- FacA in profile views. Therefore the duced by relatively high levels of andro- positive effect of shape averaging on gens. Therefore, these changes reduce FacA can be explained only in terms of FacA in women more so than in men averageness and not symmetry. [ETCOFF 1999, FINK and NEAVE 2005]. In profile view, as in frontal view, the Images with perfectly clear and smooth differences between men’s and women’s skin (constructed by digitally averaging faces correspond to the differences be- skin texture) possess a higher FacA than tween adults’ and children’s faces. In the originals [BENSON and PERRETT women, the profile has a smaller jaw, a 1992, O’TOOLE et al. 1999, LITTLE and less protruding nose, and larger, more HANCOCK 2002]. Men prefer female protruding lips. FacA is higher in women faces with smooth and slightly reddened with less protruding chins and noses, skin [FINK et al. 2001]. FINK et al. more protruding lips, a forwardly posi- [2006] found that the more heteroge- tioned jaw angle, and a generally flatter nously distributed visible skin color was contour than in men [FERRARIO et al. in female faces (a cue to an old age), the 1995, POLK et al. 1995, SERGL et al. lower was the FacA. In studies on male 1998, PEARSON and ADAMSON 2004, faces, JONES et al. [2004a,b] found a TURKKAHRAMAN and GOKALP 2004, correlation between FacA and health MATOULA and PANCHERZ 2006]. How- assessed either on the basis of the whole ever, there is at best a weak preference face or on the basis of a section of cheek for high dimorphism in male faces skin. In summary, there is a preference [SWADDLE and REIERSON 2002]. There for skin that indicates good health (in is also a preference for childlike features both sexes), and, in females, youthfulness in women, such as a slightly up-turned or and low androgens levels. concave nose [SERGL et al. 1998, CHOE In faces that were equalized in terms of et al. 2004, PEARSON and ADAMSON skin texture, FacA positively correlates 2004, VALENZANO et al. 2006]. with facial profile averageness [VALEN- In a study using morphed and warped ZANO et al. 2006]. FacA is higher in a images, SPYROPOULOS and HALAZO- composite image constructed by digitally NETIS [2001] showed that FacA is higher averaging profiles than in the original in profiles in which the skin had been faces [SPYROPOULOS and HALAZONETIS smoothed by averaging. 2001, PEARSON and ADAMSON 2004]. Generally, facial expressions signaling FacA is also higher in profiles which interest and kindliness increase FacA in have been warped toward the average both males and females [RHODES et al. shape [VALENTINE et al. 2004]. In sev- 1999a,b]. RONEY et al. [2006] showed eral studies in which the protrusion of the that kindliness and the liking of children maxilla and jaw were digitally manipula- perceived in a male face affected FacA ted, FacA was highest in faces that were independently of each other. Smiling in- typically orthognathic [GIDDON et al. 1996, creases FacA in both males and females General patterns of facial preferences 69

[REIS et al. 1990, RHODES et al. 1999a, tations. This indicates the need for more TATARUNAITE et al. 2005]. FacA is research carried out under conditions that highest for smiles that show a lot of are closer to natural conditions. white teeth but that do not show the gums or the buccal corridors [MOORE Conclusions et al. 2005, ONG et al. 2006, PAREKH et al. 2006]. The human face consists of many In facial photographs, enlarging the elements and features, most of which pupils increases FacA [ETCOFF 1999]. contribute to facial attractiveness. Most This is because in normal light dilated researchers have focused on aver- pupils indicate interest. The direction of ageness, symmetry and dimorphism, and the gaze is also important. FacA is high- relatively few have focused on head est when the gaze is directed at the ob- hair, facial hair, and skin, hair and eye server and not off to the side [JONES et color. Only rarely have studies on the al. 2006]. FacA is also higher in indi- effect of skin condition and facial ex- viduals who shift their gaze toward the pression been carried out. Although age observer than in individuals who stop is not a facial feature in and of itself, it looking at the observer [MASON et al. has a significant effect on how a person 2005]. JONES et al. [2006] found that a is evaluated in social and erotic terms. smile increases FacA if the individual is Therefore many previous studies have looking at the observer because the ob- focused on how age and age-related server interprets the smile as a reward. changes determine FacA. On the other hand, a smile decreases In both men and women, FacA is FacA if the individual is looking else- increased by symmetry, averageness, where because the observer feels disap- a smooth and clear complexion, and pointment. a pleasant expression. This is true irre- RUBENSTEIN [2005] instructed judges spective of whether the judges are of the to evaluate FacA and facial emotional opposite or the same sex. FacA in fe- expressiveness after watching ten-second- males is improved by all of the typically long film clips. Each clip presented a feminine features: prominent lips, a small person with a neutral expression reading lower part of the face, small supra-orbital a book. The judges were also instructed ridges, light skin, light hair, and an ab- to evaluate single-frame images from sence of facial hair. On the other hand, each clip. The correlation between as- male faces with a moderate level of mas- sessments of dynamic and static facial culinity are considered more attractive views was surprisingly low (r = 0.2). The than faces with a high level of masculin- correlation between FacA and facial ex- ity. Although skin condition is the main pressiveness was high for film clips indicator of a person’s age, all of the (r = 0.48), but low for photographs factors mentioned above also play a role. (r = 0.11). This proves that the determi- Facial cues to young age are important nants of FacA differ depending on determinants of FacA in women. whether moving images or still images The most strongly correlated factors are being evaluated. Studies based on with FacA are facial femininity in static facial images therefore have limi- women and averageness in both sexes 70 K. Kościński

(for each trait, r exceeds 0.5). Facial References symmetry itself seems to be a relatively weak determinant of FacA. The weight ALLEY T.R., 1993, The developmental stability of facial attractiveness: New longitudinal data and of possible other determinants of FacA is a review, Merrill-Palmer Q., 39, 265-278 not yet recognized. The strong correla- ALLEY T.R., M.R. CUNNINGHAM, 1991, Averaged tions between FacA and some facial faces are attractive, but very attractive faces are morphological features shows that the not average, Psychol. Sci., 2, 123-125 search for the physical basis of facial BAKER B.W., M.G. WOODS, 2001, The role of the divine proportion in the esthetic improvement of beauty has validity. There is some evi- patients undergoing combined orthodon- dence that preferences for specific facial tic/orthognathic surgical treatment, Int. J. Adult traits are evolutionary adaptations that Orthod. Orthognath. Surg., 16, 108-120 facilitate the choice of an appropriate BARBER N., 1995, The evolutionary psychology of mate or lifetime partner. However, non- physical attractiveness: Sexual selection and human morphology, Ethol. Sociobiol., 16, 395- adaptive mechanisms are also responsible 424 for the existence of these preferences. BAUDOUIN J.Y., G. TIBERGHIEN, 2004, Symmetry, This paper has focused on only some of averageness, and feature size in the facial the many aspects of FacA. It has briefly attractiveness of women, Acta Psychol., 117, surveyed the history of research on FacA 313-332 BENSON P., D. PERRETT, 1992, Face to face with and the main theories that have been the perfect image, New Sci., 133, 32-35 proposed to explain FacA. It has re- BENSON P.J., D.I. PERRETT, 1993, Extracting viewed and discussed in detail the results prototypical facial images from exemplars, of studies on general patterns in facial Perception, 22, 257-262 preferences. However, there are many BERRY D.S., 1991a, Accuracy in social perception: Contributions of facial and vocal information, issues that have not been covered here: J. Pers. Soc. Psychol., 61, 298-307 (1) Variation in facial preferences be- BERRY D.S., 1991b, Attractive faces are not all tween populations, within populations created equal: Joint effects of facial babyishness and within individuals, (2) The relation- and attractiveness on social perception, Pers. ship between FacA and biological quality Soc. Psychol. Bull., 17, 523-531 BERRY D.S., L.Z. MCARTHUR, 1985, Some com- and mate value of the face’s owner, ponents and consequences of a babyface, (3) The various neural, physiological and J. Pers. Soc. Psych., 48, 312-323 behavioral responses to attractive faces, BERRY D.S., L.Z. MCARTHUR, 1986, Perceiving (4) The social and biological conse- character in faces: The impact of age-related craniofacial changes on social perception, Psy- quences of perceiving some people chol. Bull., 100, 3-18 as attractive, and others as unattractive, BISSON M., A. GROBBELAAR, 2004, The esthetic (5) The selection pressures by which properties of lips: A comparison of models and evolution has shaped facial preferences in nonmodels, Angle Orthod., 74, 162-166 humans. All of these aspects will be dis- BOOTHROYD L.G., B.C. JONES, D.M. BURT, R.E. cussed in the author’s forthcoming paper CORNWELL, A.C. LITTLE, B.P. TIDDEMAN, D.I. PERRETT, 2005, Facial masculinity is related to on FacA. perceived age but not perceived health, Evol. Hum. Behav., 26, 417-431 Acknowledgements BORKAN G.A., A.H. NORRIS, 1980, Assessment of biological age using a profile of physical pa- I thank the two anonymous reviewers rameters, J. Gerontol., 35, 177-184 BUSS D., 1999, Evolutionary psychology: The new for their helpful comments. science of the mind, Allyn & Bacon, Boston General patterns of facial preferences 71

CARDENAS R.A., L.J. HARRIS, 2006, Symmetrical EIBL-EIBESFELDT I., 1970, Ethology: The biology decorations enhance the attractiveness of faces of behavior, Holt, Rinehart and Winston, New and abstract designs, Evol. Hum. Behav., 27, York 1-18 EISENTHAL Y., G. DROR, E. RUPPIN, 2006, Facial CARELLO C., A. GROSOFSKY, R.E. SHAW, J.B. attractiveness: Beauty and the machine, Neural PITTENGER, L.S. MARK, 1989, Attractiveness of Comput., 18, 119-142 facial profiles is a function of distance from ar- ENQUIST M., A. ARAK, 1993, Selection of exag- chetype, Ecol. Psychol., 1, 227-251 gerated male traits by female aesthetic senses, CHOE K.S., A.P. SCLAFANI, J.A. LITNER, G.P. YU, Nature, 361, 446-448 T. ROMO, 2004, The Korean American woman’s ENQUIST M., A. ARAK, 1994, Symmetry, beauty face. Anthropometric measurements and quan- and evolution, Nature, 372, 169-172 titative analysis of facial aesthetics, Arch. Facial ETCOFF N., 1999, Survival of the prettiest: The Plast. Surg., 6, 244-252 science of beauty, Anchor Books, New York COSTA M., L. CORAZZA, 2006, Aesthetic phenom- FARKAS L.G., 1994, Anthropometry of the attrac- ena as supernormal stimuli: The case of eye, lip, tive North American Caucasian face, [in:] An- and lower-face size and roundness in artistic thropometry of the head and face, L.G. Farkas portraits, Perception, 35, 229-246 (ed.), Raven Press, New York, pp. 159-179 CUNNINGHAM M.R., 1986, Measuring the physical FARKAS L., I. MUNRO, J. KOLAR, 1987a, Linear in physical attractiveness: Quasi-experiments on proportions in above- and below-average the sociobiology of female facial beauty, J. Pers. women’s faces, [in:] Anthropometric facial pro- Soc. Psychol., 50, 925-935 portions in medicine, L.G. Farkas, I.R. Munro CUNNINGHAM M.R., A.P. BARBEE, C.L. PIKE, (eds.), Charles C. Thomas, Springfield, pp. 119- 1990, What do women want? Facialmetric as- 130 sessment of multiple motives in the perception of FARKAS L., I. MUNRO, J. KOLAR, 1987b, The male facial physical attractiveness, J. Pers. Soc. validity of neoclassical facial proportion can- Psychol., 59, 61-72 ons, [in:] Anthropometric facial proportions in CUNNINGHAM M.R., A.R. ROBERTS, A.P. BARBEE, medicine, L.G. Farkas, I.R. Munro (eds.), Char- P.B. DRUEN, C.H. WU, 1995, “Their ideas of les C. Thomas, Springfield, pp. 57-66 beauty are, on the whole, the same as ours”: FEINMAN S., G.W. GILL, 1977, Female’s responses Consistency and variability in the cross-cultural to male beardedness, Percept. Mot. Skills, 58, perception of female physical attractiveness, 533-534 J. Pers. Soc. Psychol., 68, 261-279 FEINMAN S., G.W. GILL, 1978, Sex differences in DARWIN C., 1871, The descent of man, and selec- physical attractiveness preferences, J. Soc. Psy- tions in relation to sex, John Murray, London chol., 105, 43-52 DEBRUINE L.M., B.C. JONES, A.C. LITTLE, L.G. FERRARIO V.F., C. SFORZA, C.E. POGGIO, G. TAR- BOOTHROYD, D.I. PERRETT, I.S. PENTON-VOAK, TAGLIA, 1995, Facial morphometry of television P.A. COOPER, L. PENKE, D. FEINBERG, B.P. actresses compared with normal women, J. Oral TIDDEMAN, 2006, Correlated preferences for Maxillofac. Surg., 53, 1008-1014 facial masculinity and ideal or actual partner’s FINK B., K. GRAMMER, P.J. MATTS, 2006, Visible masculinity, Proc. R. Soc. Lond. B., 273, 1355- skin color distribution plays a role in the per- 1360 ception of age, attractiveness, and health in fe- DEBRUINE L.M., B.C. JONES, L. UNGER, A.C. male faces, Evol. Hum. Behav., 27, 433-442 LITTLE, D.R. FEINBERG, in press, Dissociating FINK B., K. GRAMMER, P. MITTEROECKER, averageness and attractiveness: Attractive faces P. GUNZ, K. SCHAEFER, F.L. BOOKSTEIN, J.T. are not always average, J. Exp. Psychol. Hum. MANNING, 2005a, Second to fourth digit ratio Percept. Perform. and face shape, Proc. R. Soc. Lond. B., 272, DION K.K., E. BERSCHEID, E. WALSTER, 1972, 1995-2001 What is beautiful is good, J. Pers. Soc. Psychol., FINK B., K. GRAMMER, R. THORNHILL, 2001, 24, 285-290 Human (Homo sapiens) facial attractiveness in EDLER R., P. AGARWAL, D. WERTHEIM, D. GREEN- relation to skin texture and color, J. Comp. Psy- HILL, 2006, The use of anthropometric propor- chol., 115, 92-99 tion indices in the measurement of facial attrac- FINK B., N. NEAVE, 2005, The biology of facial tiveness, Eur. J. Orthod., 28, 274-281 beauty, Int. J. Cosmet. Sci., 27, 317-325 72 K. Kościński

FINK B., N. NEAVE, J.T. MANNING, K. GRAMMER, HALBERSTADT J., 2006, The generality and ulti- 2005b, Facial symmetry and the ‘big-five’ per- mate origins of the attractiveness of prototypes, sonality factors, Pers. Indiv. Differ., 39, 523-529 Pers. Soc. Psychol. Rev., 10, 166-183 FINK B., N. NEAVE, H. SEYDEL, 2007, Male facial HALBERSTADT J., G. RHODES, 2000, The attrac- appearance signals physical strength to women, tiveness of nonface averages, Psychol. Sci., 11, Am. J. Hum. Biol., 19, 82-87 285-289 FORD C.S., BEACH F.A., 1951, Patterns of sexual HALBERSTADT J., G. RHODES, 2003, It’s not just behavior, Harper, New York average faces that are attractive: Computer- FRĄCKIEWICZ W., 2001, The aesthetics the eyes and manipulated averageness makes birds, fish, and mouth position in a three-point face schema, automobiles attractive, Psychon. Bull. Rev., 10, Prz. Antropol. – Anthropol. Rev., 64, 93-100 149-156 FROST P., 1988, Human skin color: A possible HATFIELD E., S. SPRECHER, 1986, Mirror, mirror: relationship between its sexual dimorphism and The importance of looks in everyday life, State its social perception, Perspect. Biol. Med., 32, University of New York Press, Albany, New 38-58 York FROST P., 2006, European hair and eye color: A HELLSTRÖM Å., J. TEKLE, 1994, Person perception case of frequency-dependent sexual selection?, through facial photographs: Effects of glasses, Evol. Hum. Behav., 27, 85-103 hair, and beard, on judgments of occupation GALTON F., 1878, Composite portraits, Nature, 18, and personal qualities, Eur. J. Soc. Psychol., 24, 97-100 693-705 GANGESTAD S.W., G.J. SCHEYD, 2005, The evolu- HERSHON L.E., D.B. GIDDON, 1980, Determinants tion of human physical attractiveness, Ann. Rev. of facial profile self-perception, Am. J. Orthod., Anthropol., 34, 523-548 78, 279-295 GETTY T., 2002, Signaling health versus parasites, HILDEBRANDT K.A., H.E. FITZGERALD, 1979, Am. Nat., 159, 363-371 Facial feature determinants of perceived infant GHIRLANDA S., L. JANSSON, M. ENQUIST, 2002, attractiveness, Infant Behav. Dev., 2, 329-339 Chickens prefer beautiful humans, Hum. Nat., HINSZ V.B., D.C. MATZ, R.A. PATIENCE, 2001, 13, 383-389 Does women’s hair signal reproductive poten- GIDDON D.B., 1995, Orthodontic applications of psychological and perceptual studies of facial tial, J. Exp. Soc. Psychol., 37, 166-172 esthetics, Semin. Orthod., 1, 82-93 HONN M., K. DIETZ, A. GODT, G. GOZ, 2005, Perceived relative attractiveness of facial pro- GIDDON D.B., D.L. BERNIER, C.A. EVANS, J.A. KINCHEN, 1996, Comparison of two computer files with varying degrees of skeletal anomalies, animated imaging programs for quantifying fa- J. Orofac. Orthop., 66, 187-196 cial profile preference, Percept. Mot. Skills, 82, HÜCKSTEDT B., 1965, Experimentelle Unter- 1251-1264 suchungen zum "Kindchenschema", Zeitschrift GRAMMER K., B. FINK, A. JUETTE, G. RONZAL, für experimentelle und angewandte Psychologie, R. THORNHILL, 2002, Female faces and bodies: 12, 421-450 N-dimensional feature space and attractiveness, HUME D.K., R. MONTGOMERIE, 2001, Facial [in:] Facial attractiveness: Evolutionary, cogni- attractiveness signals different aspects of tive, and social perspectives, G. Rhodes, L.A. “quality” in women and men, Evol. Hum. Be- Zebrowitz (eds.), Ablex Publishing, Westport, hav., 22, 93-112 pp. 91-125 ILIFFE A.H., 1960, A study of preferences in femi- GRAMMER K., B. FINK, A.P. MØLLER, R. THORN- nine beauty, Br. J. Psychol., 51, 267-273 HILL, 2003, Darwinian aesthetics: Sexual selec- JANSSON L., B. FORKMAN, M. ENQUIST, 2002, tion and the biology of beauty, Biol. Rev., 78, Experimental evidence of receiver bias for sym- 385-407 metry, Anim. Behav., 63, 617-621 GRAMMER K., R. THORNHILL, 1994, Human (Homo JEFFERSON Y., 1996, Skeletal types: Key to unrav- sapiens) facial attractiveness and sexual selec- eling the mystery of facial beauty and its bio- tion: The role of symmetry and averageness, logic significance, J. Gen. Orthod., 7, 7-25 J. Comp. Psychol., 108, 233-242 JEFFERSON Y., 2004, Facial beauty – Establishing GUTHRIE R.D., 1976, Body hot spots: The anatomy a universal standard, Int. J. Orthod., 15, 9-22 of human social organs and behavior, Van JENNIONS M.D., M. PETRIE, 1997, Variation in Nostrand Reinhold, New York mate choice and mating preferences: A review General patterns of facial preferences 73

of causes and consequences, Biol. Rev., 72, KNIFFIN K.M., D.S. WILSON, 2004, The effect of 283-327 nonphysical traits on the perception of physical JOHNSTON V.S., M. FRANKLIN, 1993, Is beauty in attractiveness, Evol. Hum. Behav., 25, 88-101 the eye of the beholder, Ethol. Sociobiol., 14, KOEHLER N., L.W. SIMMONS, G. RHODES, 183-199 M. PETERS, 2004, The relationship between sex- JOHNSTON V.S., R. HAGEL, M. FRANKLIN, B. FINK, ual dimorphism in human faces and fluctuating K. GRAMMER, 2001, Male facial attractiveness – asymmetry, Proc. R. Soc. Lond. B., (Suppl.) evidence for hormone-mediated adaptive design, 271, S233-S236 Evol. Hum. Behav., 22, 251-267 KORTHASE K.M., I. TRENHOLME, 1982, Perceived JOHNSTON V.S., C.J. SOLOMON, S.J. GIBSON, A. age and perceived physical attractiveness, Per- PALLARES-BEJARANO, 2003, Human facial cept. Mot. Skills, 54, 1251-1258 beauty: Current theories and methodologies, KOWNER R., 1996, Facial asymmetry and attrac- Arch. Facial Plast. Surg., 5, 371-377 tiveness judgment in developmental perspective, JOHNSTONE R.A., 1994, Female preference for J. Exp. Psychol. Hum. Percept. Perform., 22, symmetrical males as a by-product of selection 662-675 for mate recognition, Nature, 372, 172-175 KOWNER R., 1997, The perception and attribution JONES B.C., L.M. DEBRUINE, A.C. LITTLE, C.A. of facial asymmetry in normal adults, Psychol. CONWAY, D.R. FEINBERG, 2006, Integrating Rec., 47, 371-384 gaze direction and expression in preferences for KOWNER R., 1998, Effects of social deviance labels attractive faces, Psychol. Sci., 17, 588-591 on judgements of facial attractiveness: A com- JONES B.C., A.C. LITTLE, D.M. BURT, D.I. parison of labelling procedures using Japanese PERRETT, 2004a, When facial attractiveness is raters, Int. J. Psychol., 33, 1-16 only skin deep, Perception, 33, 569-576 KUJAWA B., J. STRZAŁKO, 1998, Standard of JONES B.C., A.C. LITTLE, D.R. FEINBERG, I.S. physical attractiveness, Prz. Antropol. - Anthro- PENTON-VOAK, B.P. TIDDEMAN, D.I. PERRETT, pol. Rev., 61, 31-48 2004b, The relationship between shape symme- LAENG B., R. MATHISEN, J.A. JOHNSEN, 2007, Why try and perceived skin condition in male facial do blue-eyed men prefer women with the same attractiveness, Evol. Hum. Behav., 25, 24-30 eye color?, Behav. Ecol. Sociobiol., 61, 371-384 JONES B.C., A.C. LITTLE, I.S. PENTON-VOAK, B.P. LANGLOIS J.H., L.A. ROGGMAN, 1990, Attractive TIDDEMAN, D.M. BURT, D.I. PERRETT, 2001, faces are only average, Psychol. Sci., 1, 115- Facial symmetry and judgements of apparent 121 health. Support for a "good genes" explanation LANGLOIS J.H., L.A. ROGGMAN, L. MUSSELMAN, of the attractiveness-symmetry relationship, 1994, What is average and what is not average Evol. Hum. Behav., 22, 417-429 about attractive faces, Psychol. Sci., 5, 214-220 JONES D., 1995, Sexual selection, physical attrac- LAW SMITH M.J., D.I. PERRETT, B.C. JONES, R.E. tiveness, and facial neoteny: Cross-cultural evi- CORNWELL, F.R. MOORE, D.R. FEINBERG, L.G. dence and implications, Curr. Anthropol., 36, BOOTHROYD, S.J. DURRANI, M.R. STIRRAT, S. 723-748 WHITEN, R.M. PITMAN, S.G. HILLIER, 2006, Fa- JONES D., 1996a, An evolutionary perspective on physical attractiveness, Evol. Anthropol., 5, 97- cial appearance is a cue to oestrogen levels, 109 Proc. R. Soc. Lond. B., 273, 135-140 LIGHT L.L., S. HOLLANDER, F. KAYRA-STUART, JONES D., 1996b, Physical attractiveness and the theory of sexual selection, Museum of Anthro- 1981, Why attractive people are harder to re- pology, University of Michigan, Ann Arbor member, Pers. Soc. Psychol. Bull., 7, 269-276 ITTLE URT ENTON OAK KEATING C.F., 1985, Gender and the physiognomy L A.C., D.M. B , I.S. P -V , D.I. of dominance and attractiveness, Soc. Psychol. PERRETT, 2001, Self-perceived attractiveness Q., 48, 61-70 influences human female preferences for sexual KEATING C.F., D.L. BAI, 1986, Children’s attribu- dimorphism and symmetry in male faces, Proc. tions of social dominance from facial cues, R. Soc. Lond. B., 268, 39-44 Child Dev., 57, 1269-1276 LITTLE A.C., P.J.B. HANCOCK, 2002, The role of KEATING C.F., J. DOYLE, 2002, The faces of desir- masculinity and distinctiveness in judgments of able mates and dates contain mixed social status human male facial attractiveness, Br. J. Psy- cues, J. Exp. Soc. Psychol., 38, 414-424 chol., 93, 451-464 74 K. Kościński

LITTLE A.C., B.C. JONES, 2003, Evidence against MCARTHUR L.Z., D.S. BERRY, 1987, Cross- perceptual bias views for symmetry preferences cultural agreement in perceptions of babyfaced in human faces, Proc. R. Soc. Lond. B., 270, adults, J. Cross. Cult. Psychol., 18, 165-192 1759-1763 MCCABE V., 1984, Abstract perceptual informa- LITTLE A.C., B.C. JONES, 2006, Attraction inde- tion for age level: A risk factor for maltreat- pendent of detection suggests special mecha- ment?, Child Dev., 55, 267-276 nisms for symmetry preferences in human face MCCALL R.B., C.B. KENNEDY, 1980, Attention of perception, Proc. R. Soc. Lond. B., 273, 3093- 4-month infants to discrepancy and babyishness, 3099 J. Exp. Child. Psych., 29, 189-201 LITTLE A.C., I.S. PENTON-VOAK, D.M. BURT, D.I. MCGOVERN R.J., M.C. NEALE, K.S. KENDLER, PERRETT, 2003, Investigating an imprinting-like 1996, The independence of physical attractive- phenomenon in humans: Partners and opposite- ness and symptoms of depression in a female sex parents have similar hair and eye colour, twin population, J. Psychol., 130, 209-219 Evol. Hum. Behav., 24, 43-51 MEALEY L., R. BRIDGESTOCK, G.C. TOWNSEND, MANNING J.T., L.J. PICKUP, 1998, Symmetry and 1999, Symmetry and perceived facial attractive- performance in middle distance runners, Int. ness: A monozygotic co-twin comparison, J. J. Sports Med., 19, 205-209 Pers. Soc. Psychol., 76, 151-158 EHRABIAN LUM MAPLE J.R., K.W.L. VIG, F.M. BECK, P.E. M A., J.S. B , 1997, Physical ap- LARSEN, S. SHANKER, 2005, A comparison of pearance, attractiveness, and the mediating role providers’ and consumers’ perceptions of facial- of emotions, Curr. Psychol., 16, 20-42 profile attractiveness, Am. J. Orthod. Dentofa- MESKO N., T. BERECZKEI, 2004, Hairstyle as an cial Orthop., 128, 690-696 adaptive means of displaying phenotypic qual- ity, Hum. Nat., 15, 251-270 MARCUS D.K., M.R. CUNNINGHAM, 2003, Do child MONIN B., 2003, The warm glow heuristic: When molesters have aberrant perceptions of adult liking leads to familiarity, J. Pers. Soc. Psychol., female facial attractiveness?, J. Appl. Soc. Psy- 85, 1035-1048 chol., 33, 499-512 MOORE T., K.A. SOUTHARD, J.S. CASKO, F. QIAN, MARK L., R.E. SHAW, J.B. PITTENGER, 1988, T.E. SOUTHARD, 2005, Buccal corridors and Natural constraints, scales of analysis, and in- smile esthetics, Am. J. Orthod. Dentofacial Or- formation for the perception of growing faces, thop., 127, 208-213 [in:] Social and applied aspects of perceiving MORRIS D., 1967, The naked ape, Cape Ltd, Lon- faces, T.R. Alley (ed.), Lawrence Erlbaum don Associates, Inc, Hillsdale, NJ, pp. 11-49 MUELLER U., A. MAZUR, 1997, Facial dominance MARK L.S., J.B. PITTENGER, H. HINES, C. CAREL- in Homo sapiens as honest signaling of male LO, R.E. SHAW, J.T. TODD, 1980, Wrinkling and quality, Behav. Ecol., 8, 569-579 head shape as coordinated sources of age-level MUSCARELLA F., M.R. CUNNINGHAM, 1996, The information, Percept. Psychophys., 27, 117-124 evolutionary significance and social perception MARQUARDT S.R., 2002, Dr. Stephen R. on the of male pattern baldness and facial hair, Ethol. Golden Decagon and human facial beauty. In- Sociobiol., 17, 99-117 terview by Dr. Gottlieb, J. Clin. Orthod., 36, NEAVE N., S. LAING, B. FINK, J.T. MANNING, 339-347 2003, Second to fourth digit ratio, testosterone MASON M.F., E.P. TATKOW, C.N. MACRAE, 2005, and perceived male dominance, Proc. R. Soc. The look of love: Gaze shifts and person per- Lond. B., 270, 2167-2172 ception, Psychol. Sci., 16, 236-239 NGUYEN D.D., P.K. TURLEY, 1998, Changes in the MATHES E.W., S.M. BRENNAN, P.M. HAUGEN, Caucasian male facial profile as depicted in H.B. RICE, 1985, Ratings of physical attractive- fashion magazines during the twentieth century, ness as a function of age, J. Soc. Psychol., 125, Am. J. Orthod. Dentofacial Orthop., 114, 208- 157-168 217 MATOULA S., H. PANCHERZ, 2006, Skeletofacial NOOR F., D.C. EVANS, 2003, The effect of facial morphology of attractive and nonattractive symmetry on perceptions of personality and at- faces, Angle Orthod., 76, 204-210 tractiveness, J. Res. Pers., 37, 339-347 MCARTHUR L.Z., K. APATOW, 1983/1984, Impres- ONG E., R.A. BROWN, S. RICHMOND, 2006, Peer sions of baby-faced adults, Soc. Cogn., 2, 315- assessment of dental attractiveness, Am. J. Or- 342 thod. Dentofacial Orthop., 130, 163-169 General patterns of facial preferences 75

O’TOOLE A.J., T. PRICE, T. VETTER, J.C. PETTIJOHN T.F., A. TESSER, 1999, Popularity in BARTLETT, V. BLANZ, 1999, 3D shape and 2D environmental context: Facial feature assess- surface textures of human faces: The role of av- ment of American movie actresses, Media Psy- erages in attractiveness and age, Image. Vis. chol., 1, 229-247 Comput., 18, 9-19 POLK M.S., A.G. FARMAN, J.A. YANCEY, L.R. PANCER S.M., J.R. MEINDL, 1978, Length of hair GHOLSTON, B.E. JOHNSON, 1995, Soft tissue and beardedness as determinants of personality profile: A survey of African-American prefer- impressions, Percept. Mot. Skills, 46, 1328-1330 ence, Am. J. Orthod. Dentofacial Orthop., 108, PAREKH S.M., H.W. FIELDS, M. BECK, S. 90-101 ROSENSTIEL, 2006, Attractiveness of variations POLLARD J., J. SHEPHERD, J. SHEPHERD, 1999, in the smile arc and buccal corridor space as Average faces are average faces, Curr. Psychol., judged by orthodontists and laymen, Angle Or- 18, 98-103 thod., 76, 557-563 REED J.A., E.M. BLUNK, 1990, The influence of PEARSON D.C., P.A. ADAMSON, 2004, The ideal facial hair on impression formation, Soc. Behav. nasal profile: Rhinoplasty patients vs the gen- Pers., 18, 169-176 eral public, Arch. Facial Plast. Surg., 6, 257-262 REIS H.T., I.M. WILSON, C. MONESTERE, S. PELLEGRINI R.J., 1973, Impressions of the male BERNSTEIN, K. CLARK, E. SEIDL, M. FRANCO, E. personality as a function of beardedness, Psy- GIOIOSO, L. FREEMAN, K. RADOANE, 1990, chology, 10, 29-33 What is smiling is beautiful and good, Eur. J. PENTON-VOAK I.S., J.Y. CHEN, 2004, High sali- Soc. Psychol., 20, 259-267 vary testosterone is linked to masculine male RHODES G., 2006, The evolutionary psychology of facial appearance in humans, Evol. Hum. Be- facial beauty, Annu. Rev. Psychol., 57, 199-226 hav., 25, 229-241 RHODES G., J. CHAN, L.A. ZEBROWITZ, L.W. PENTON-VOAK I.S., B.C. JONES, A.C. LITTLE, SIMMONS, 2003, Does sexual dimorphism in S. BAKER, B. TIDDEMAN, D.M. BURT, D.I. human faces signal health?, Proc. R. Soc. Lond. PERRETT, 2001, Symmetry, sexual dimorphism in facial proportions and male facial attractive- B., (Suppl.) 270, S93-S95 ness, Proc. R. Soc. Lond. B., 268, 1617-1623 RHODES G., J. HALBERSTADT, G. BRAJKOVICH, PENTON-VOAK I.S., D.I. PERRETT, 2001, Male 2001a, Generalization of mere exposure effects facial attractiveness: Perceived personality and to averaged composite faces, Soc. Cogn., 19, shifting female preferences for male traits 57-70 across the menstrual cycle, Adv. Study Behav., RHODES G., J. HALBERSTADT, L. JEFFERY, R. 30, 219-259 PALERMO, 2005a, The attractiveness of average PENTON-VOAK I.S., D.I. PERRETT, D.L. CASTLES, faces is not a generalized mere exposure effect, T. KOBAYASHI, D.M. BURT, L.K. MURRAY, R. Soc. Cogn., 23, 205-217 MINAMISAWA, 1999, Menstrual cycle alters face RHODES G., C. HICKFORD, L. JEFFERY, 2000, Sex- preference, Nature, 399, 741-742 typicality and attractiveness: Are supermale and PERRETT D.I., D.M. BURT, I.S. PENTON-VOAK, K.J. superfemale faces super-attractive?, Br. J. Psy- LEE, D.A. ROWLAND, R. EDWARDS, 1999, Sym- chol., 91, 125-140 metry and human facial attractiveness, Evol. RHODES G., K. LEE, R. PALERMO, M. WEISS, S. Hum. Behav., 20, 295-307 YOSHIKAWA, P. CLISSA, T. WILLIAMS, M. PERRETT D.I., K.J. LEE, I. PENTON-VOAK, D. PETERS, C. WINKLER, L. JEFFERY, 2005b, At- ROWLAND, S. YOSHIKAWA, D.M. BURT, S.P. tractiveness of own-race, other-race, and mixed- HENZI, D.L. CASTLES, S. AKAMATSU, 1998, Ef- race faces, Perception, 34, 319-340 fects of sexual dimorphism on facial attractive- RHODES G., F. PROFFITT, J.M. GRADY, A. SUMICH, ness, Nature, 394, 884-887 1998, Facial symmetry and the perception of PERRETT D.I., K.A. MAY, S. YOSHIKAWA, 1994, beauty, Psychon. Bull. Rev., 5, 659-669 Facial shape and judgements of female attrac- RHODES G., J. ROBERTS, L.W. SIMMONS, 1999a, tiveness, Nature, 368, 239-242 Reflections on symmetry and attractiveness, PETTIJOHN T.F., B.J. JUNGEBERG, 2004, Playboy Psychol. Evol. Gend., 1, 279-295 playmate curves: Changes in facial and body RHODES G., A. SUMICH, G. BYATT, 1999b, Are feature preferences across social and economic average facial configurations attractive only conditions, Pers. Soc. Psychol. Bull., 30, 1186- because of their symmetry?, Psychol. Sci., 10, 1197 52-58 76 K. Kościński

RHODES G., T. TREMEWAN, 1996, Averageness, of good genes, Proc. R. Soc. Lond. B., 266, exaggeration and facial attractiveness, Psychol. 1913-1917 Sci., 7, 105-110 SCOTT C.R., M.S. GOONEWARDENE, K. MURRAY, RHODES G., S. YOSHIKAWA, A. CLARK, K. LEE, 2006, Influence of lips on the perception of mal- R. MCKAY, S. AKAMATSU, 2001b, Attractive- occlusion, Am. J. Orthod. Dentofacial Orthop., ness of facial averageness and symmetry in non- 130, 152-162 Western cultures: In search of biologically SERGL H.G., A. ZENTNER, G. KRAUSE, 1998, An based standards of beauty, Perception, 30, 611- experimental study of the esthetic effect of facial 625 profiles, J. Orofac. Orthop., 59, 116-126 RHODES G., L.A. ZEBROWITZ, A. CLARK, S.M. SHACKELFORD T.K., R.J. LARSEN, 1997, Facial KALICK, A. HIGHTOWER, R. MCKAY, 2001c, Do asymmetry as an indicator of psychological, facial averageness and symmetry signal health?, emotional, and physiological distress, J. Pers. Evol. Hum. Behav., 22, 31-46 Soc. Psychol., 72, 456-466 RIEDL B.I.M., 1990, Morphologisch-metrische SIMMONS L.W., G. RHODES, M. PETERS, N. Merkmale des männlichen und weiblichen Part- KOEHLER, 2004, Are human preferences for fa- nerleitbildes in ihrer Bedeutung für die Wahl cial symmetry focused on signals of develop- des Ehegatten, Homo, 41, 72-85 mental instability?, Behav. Ecol., 15, 864-871 RIKOWSKI A., K. GRAMMER, 1999, Human body SPYROPOULOS M.N., D.J. HALAZONETIS, 2001, odour, symmetry and attractiveness, Proc. Significance of the soft tissue profile on facial R. Soc. Lond. B., 266, 869-874 esthetics, Am. J. Orthod. Dentofacial Orthop., ROBERTS S.C., A.C. LITTLE, L.M. GOSLING, B.C. 119, 464-471 JONES, D.I. PERRETT, V. CARTER, M. PETRIE, STRZAŁKO J., K. KASZYCKA, 1988, Atrakcyjność 2005a, MHC-assortative facial preferences in fizyczna - obiektywne i subiektywne składowe humans, Biol. Lett., 1, 400-403 oceny, Prz. Antropol., 54, 7-17 ROBERTS S.C., A.C. LITTLE, L.M. GOSLING, D.I. STRZAŁKO J., K.A. KASZYCKA, 1992, Physical PERRETT, V. CARTER, B.C. JONES, I. PENTON- attractiveness: Interpersonal and intrapersonal VOAK, M. PETRIE, 2005b, MHC-heterozygosity variability of assessments, Soc. Biol., 39, 170- and human facial attractiveness, Evol. Hum. 176 Behav., 26, 213-226 SUSANNE C., 1977, Heritability of anthropological RONEY J.R., K.N. HANSON, K.M. DURANTE, characters, Hum. Biol., 49, 573-580 D. MAESTRIPIERI, 2006, Reading men’s faces: SWADDLE J.P., J.P.K. CHE, R.E. CLELLAND, 2004, Women’s mate attractiveness judgments track Symmetry preference as a cognitive by-product men’s testosterone and interest in infants, Proc. in starlings, Behaviour, 141, 469-478 R. Soc. Lond. B., 273, 2169-2175 SWADDLE J.P., I.C. CUTHILL, 1995, Asymmetry and ROWLAND D.A., D.I. PERRETT, 1995, Manipulating human facial attractiveness: Symmetry may not facial appearance through shape and color, always be beautiful, Proc. R. Soc. Lond. B., 261, IEEE Comput. Graph. Appl., 15, 70-76 111-116 RUBENSTEIN A.J., 2005, Variation in perceived SWADDLE J.P., G.W. REIERSON, 2002, Testoster- attractiveness: Differences between dynamic one increases perceived dominance but not at- and static faces, Psychol. Sci., 16, 759-762 tractiveness in human males, Proc. R. Soc. RUBENSTEIN A.J., L. KALAKANIS, J.H. LANGLOIS, Lond. B., 269, 2285-2289 1999, Infant preferences for attractive faces: A SYMONS D., 1979, The evolution of human sexual- cognitive explanation, Dev. Psychol., 35, 848- ity, University Press, Oxford 855 TATARUNAITE E., R. PLAYLE, K. HOOD, W. SHAW, RUSSELL R., 2003, Sex, beauty, and the relative S. RICHMOND, 2005, Facial attractiveness: A luminance of facial features, Perception, 32, longitudinal study, Am. J. Orthod. Dentofacial 1093-1107 Orthop., 127, 676-682 SCARBROUGH P.S., V.S. JOHNSTON, 2005, Individ- THORNHILL R., S.W. GANGESTAD, 2006, Facial ual differences in women’s facial preferences as sexual dimorphism, developmental stability, and a function of digit ratio and mental rotation susceptibility to disease in men and women, ability, Evol. Hum. Behav., 26, 509-526 Evol. Hum. Behav., 27, 131-144 SCHEIB J.E., S.W. GANGESTAD, R. THORNHILL, THORNHILL R., S.W. GANGESTAD, R. MILLER, 1999, Facial attractiveness, symmetry, and cues G. SCHEYD, J.K. MCCULLOUGH, M. FRANKLIN, General patterns of facial preferences 77

2003, Major histocompatibility genes, symmetry WAYNFORTH D., S. DELWADIA, M. CAMM, 2005, and body scent attractiveness in men and The influence of women’s mating strategies on women, Behav. Ecol., 14, 668-678 preference for masculine facial architecture, THORNHILL R., A.P. MØLLER, 1997, Developmen- Evol. Hum. Behav., 26, 409-416 tal stability, disease and medicine, Biol. Rev., WINKIELMAN P., J. HALBERSTADT, T. FAZENDEIRO, 72, 497-548 S. CATTY, 2006, Prototypes are attractive be- TIDDEMAN B., M. BURT, D. PERRETT, 2001, cause they are easy on the mind, Psychol. Sci., Prototyping and transforming facial textures for 17, 799-806 perception research, IEEE Comput. Graph. WOGALTER M.S., J.A. HOSIE, 1991, Effects of Appl., 21, 42-50 cranial and facial hair on perceptions of age TÜRKKAHRAMAN H., H. GÖKALP, 2004, Facial and person, J. Soc. Psychol., 131, 589-591 profile preferences among various layers of YEHEZKEL S., P.K. TURLEY, 2004, Changes in the Turkish population, Angle Orthod., 74, 640-647 African American female profile as depicted in UDRY J.R., 1965, Structural correlates of feminine fashion magazines during the 20th century, Am. beauty preferences in Britain and the United J. Orthod. Dentofacial Orthop., 125, 407-417 States: A comparison, Sociol. Soc. Res., 49, ZAJONC R.B., 2001, Mere exposure: A gateway 330-342 to the subliminal, Curr. Dir. Psychol. Sci., 10, VALENTINE T., S. DARLING, M. DONNELLY, 2004, 224-228 Why are average faces attractive? The effect of ZAIDEL D.W., A.C. CHEN, C. GERMAN, 1995, She view and averageness on the attractiveness of is not a beauty even when she smiles: Possible female faces, Psychon. Bull. Rev., 11, 482-487 evolutionary basis for a relationship between VALENZANO D.R., A. MENNUCCI, G. TARTARELLI, facial attractiveness and hemispheric speciali- A. CELLERINO, 2006, Shape analysis of female facial attractiveness, Vis. Res., 46, 1282-1291 zation, Neuropsychologia, 33, 649-655 VAN DEN BERGHE P.L., P. FROST, 1986, Skin color ZEBROWITZ L.A., J.M. MONTEPARE, 1992, Im- preference, sexual dimorphism, and sexual se- pressions of babyfaced individuals across the lection: A case of gene-culture co-evolution?, life span, Dev. Psychol., 28, 1143-1152 Ethnic Racial Stud., 9, 87-113 ZEBROWITZ L.A., K. OLSON, K. HOFFMAN, 1993, VON FAUSS R., 1988, Zur Bedeutung des Gesichtes Stability of babyfacedness and attractiveness für die Partnerwahl, Homo, 37, 188-201 across the life span, J. Pers. Soc. Psychol., 64, WAGATSUMA E., C.L. KLEINKE, 1979, Ratings of 453-466 facial beauty by Asian-American and Caucasian ZEBROWITZ L.A., G. RHODES, 2004, Sensitivity to females, J. Soc. Psychol., 109, 299-300 “bad genes” and the anomalous face overgen- WALLACE A.R., 1889, Darwinism, (2nd edition), eralization effect: Cue validity, cue utilization, Macmillan, London and accuracy in judging intelligence and health, WALSTER E., V. ARONSON, D. ABRAHAMS, L. J. Nonverbal Behav., 28, 167-185 ROTTMAN, 1966, The importance of physical ŻĄDZIŃSKA E., 2003, Fluctuating asymmetry of attractiveness in dating behavior, J. Pers. Soc. some head structures and its possible causes, Psychol., 4, 508-516 Prz. Antropol. – Anthropol. Rev., 66, 39-54

Streszczenie

Praca omawia ogólnoludzkie wzorce postrzegania atrakcyjności twarzy. Atrakcyjność twa- rzy jest tematem rozważań od tysięcy lat, jednak dopiero kilkadziesiąt lat temu stała się przedmiotem badań naukowych. Już w starożytnej Grecji istniały koncepcje łączące piękno ludzkiej twarzy z tzw. złotą proporcją jej odcinków (tzn. stosunek ich długości miał być równy 1,618) lub równą długością określonych jej wymiarów. Poglądy te pochodziły z roz- ważań geometrycznych, nic więc dziwnego, że nie znalazły uzasadnienia we współczesnych badaniach nad postrzeganiem atrakcyjności twarzy. Dzisiejsze, ewolucyjne podejście do zagadnienia atrakcyjności twarzy zakłada, że dostrze- ganie piękna lub brzydoty twarzy jest adaptacją osoby postrzegającej. Adaptacja ta polega na 78 K. Kościński pozytywnym odbiorze tych twarzy, których wygląd sugeruje (sygnalizuje) obecność pożąda- nych cech u ich właścicieli, np. zdrowia lub „dobrych genów”. Istnieją także alternatywne, nieadaptacyjne próby wyjaśnienia fenomenu atrakcyjności fizycznej, odwołujące się do ogólnych mechanizmów funkcjonowania mózgu. Badania empiryczne pokazują, że ludzkie preferencje w stosunku do pewnych twarzy są wynikiem działania zarówno mechanizmów adaptacyjnych jak i nieadaptacyjnych. Wyróżniono wiele czynników, które wpływają na atrakcyjność twarzy. Jednym z ważniej- szych jest wiek, wyraźnie związany z cechami uznawanymi za atrakcyjne. Okazało się na przykład, że mężczyźni wykazują tak silną preferencję dla oznak młodości na twarzy kobiet, że za najatrakcyjniejsze uważają proporcje twarzy typowe dla 11-14-letnich dziewcząt. Oprócz dziecięcych proporcji większości cech , atrakcyjna kobieta powinna jednak posiadać też cechy świadczące o jej dojrzałości. Takie preferencje u mężczyzn zapewne zostały ukształtowane ewolucyjnie, w związku z wysokim potencjałem reprodukcyjnym oraz wyso- ką zdolnością rozrodczą młodych kobiet. Z kolei kobiety nie wykazują wyraźnych preferen- cji dla oznak wieku na twarzach mężczyzn, choć nisko oceniają obie skrajności: twarze bardzo młode oraz twarze starców. Istnieją doniesienia, że u dzieci (a przynajmniej u nie- mowląt) młody wygląd podnosi atrakcyjność twarzy. Już w XIX w. zauważono, że twarze o przeciętnych proporcjach cechują się ponadprze- ciętną atrakcyjnością. Nowoczesne techniki komputerowej obróbki obrazu twarzy (tzw. warping i morfing) pokazały, że korzystny wpływ przeciętności proporcji na atrakcyjność twarzy ma miejsce zarówno dla widoku z przodu jak i z profilu, dla twarzy obu płci oraz według sędziów obu płci. Do dziś nie rozstrzygnięto czy preferencja dla przeciętnych twarzy jest adaptacją pozwalającą wybrać dobrego partnera (przeciętność proporcji miałaby tu być oznaką zdrowia somatycznego i genetycznego) czy też skutkiem sposobu funkcjonowania mózgu (atrakcyjność form przeciętnych znaleziono dla wielu innych typów bodźców, np. psów, samochodów itd.). Różnice pomiędzy twarzą mężczyzny a twarzą kobiety w dużym stopniu pokrywają się z różnicami pomiędzy twarzą osoby dorosłej a twarzą dziecka. Nic więc dziwnego, że męż- czyźni za atrakcyjniejsze uznają silnie sfeminizowane twarze kobiet. Twarze takie wydają się młodsze, co sugeruje wysoki potencjał reprodukcyjny, a ponadto sygnalizują też wysoki poziom estrogenu, a zatem zdrowie reprodukcyjne. Kobiety preferują twarze mężczyzn o umiarkowanym stopniu maskulinizacji. Powiązania atrakcyjności twarzy mężczyzny z jej stopniem maskulinizacji są bardziej złożone: maskulinizacja oznacza wysoki poziom testo- steronu, a to, z jednej strony sygnalizuje posiadanie przez mężczyznę „dobrych genów”, a z drugiej może oznaczać zmniejszoną wierność mężczyzny i jego chęć inwestowania w potomstwo. W kilku pracach z lat 1990. wykazano ujemny związek między symetrią twarzy a jej atrakcyjnością, jednak wkrótce okazało się, że były to artefakty wynikające z zastosowania wadliwych metod cyfrowej obróbki twarzy. Nowsze badania, przeprowadzone z użyciem techniki warpingu, wykazały, że wzrost symetrii twarzy zwiększa jej atrakcyjność. Istnieje hipoteza, że symetryczna twarz, podobnie jak wszelkie inne symetryczne obiekty, jest uwa- żana za atrakcyjną dlatego, że neuronowa obróbka obiektów symetrycznych jest szybka i mniej zawodna niż obiektów niesymetrycznych. Bardziej prawdopodobne wydaje się jednak wyjaśnienie adaptacyjne: asymetria twarzy jest oznaką niezdolności organizmu do precyzyjnego wytwarzania struktur fenotypowych na podstawie genotypu. Wybór twarzy symetrycznej oznacza więc wybór osobnika o wysokiej stabilności rozwojowej, a zatem o „dobrych genach”. General patterns of facial preferences 79

Generalnie, mężczyźni preferują jasny odcień skóry u kobiet, co wydaje się być formą pre- ferencji w stosunku do młodego wieku (ponieważ dzieci mają jaśniejszą skórę niż dorośli) oraz wysokiego poziomu estrogenu (estrogen rozjaśnia skórę). Preferencje mężczyzn ze względu na kolor włosów i oczu są urozmaicone, co próbuje się tłumaczyć działaniem dobo- ru zależnego od częstości. Mężczyźni są wrażliwi na jakość włosów, która jest wskaźnikiem ogólnego zdrowia kobiety. Owłosienie twarzy (zarost) jest cechą typowo męską, dlatego jego obecność u kobiety obniża jej atrakcyjność. Broda u mężczyzn jest odbierana przez kobiety równie dwuznacznie jak silnie zmaskulinizowana twarz, dlatego kobiety, na ogół, pozytyw- nie oceniają obecność cienia po zgolonej brodzie (oznaka męskości), ale już nie obecność brody. Czysta (tzn. wolna od brodawek itp.) i zdrowo wyglądająca skóra pozytywnie wpływa na atrakcyjność twarzy obu płci. Ponadto u kobiet atrakcyjność twarzy jest obniżana przez oznaki zaawansowania wieku (zmarszczki) lub względnie wysokiego poziomu androgenów (np. trądzik młodzieńczy). Wyraz twarzy sugerujący pozytywne nastawienie podnosi atrak- cyjność tej twarzy. Znaczenie ma tu uśmiech, rozszerzone źrenice, wzrok skierowany na patrzącego, a także „odczytywana” z twarzy uprzejmość i lubienie dzieci. Podsumowując, wyniki badań dowodzą, że istnieją pewne ogólne wzorce preferencji w stosunku do twarzy, które w dużej mierze są kryteriami rozpoznawania wartościowych, z reprodukcyjnego punktu widzenia, partnerów. Preferencje w stosunku do twarzy mają zatem charakter adaptacji, choć w niektórych przypadkach istotną rolę mogą odgrywać nie-adaptacyjne mechanizmy związane z ogólnymi zasadami funkcjonowania mózgu.