Physiognomy is the ancient art and now science of deducing someones personality and character from their appearance, especially from the features of the face, possibly including expression and movement patterns across a brief amount of time.
Physiognomy has gone out of fashion in mainstream scientific research due to its pseudoscientific history. Physiognomy is commonly compared to forms of quackery such as phrenology (the belief that one could determine a person's personality and propensity to engage in criminal activity by evaluating the bumps on their head). With the machine learning hype, the field has regained some recognition, e.g., machine learning was used to detect people's individual's behavioral tendencies via facial analysis; however, it remains unclear whether the software can produce above-chance results based on facial or contextual cues.
There have also been many scientific studies linking physical appearance with certain personality/character traits above chance, though most researchers shy away from using the term 'physiognomy' explicitly. More tangible results are described below, but none of them have effect sizes large enough that one could assume them to be consistently accurate in actual lived experience. However, proponents of modern physiognomy argue that making decisions based on even a small correlation is superior to doing so based on no information. The chances of making a correct prediction increase as certain traits tend to stack on top of each other.
Regardless of the validity of physiognomic judgments in general, people frequently make very severe and swift judgments based on stereotypes and perceived physiognomic characteristics. As there appear to be genuine associations between appearance and actual behavior (however weak these associations are), the natural human tendency to 'judge a book by its cover' likely has adaptive value. This tendency helped ancestral man avoid potentially dangerous situations (or choose particularly suitable sexual partners or allies) in his evolutionary past. These cognitive heuristics based on superficial appearance would likely have lead to higher reproductive success for those that acted on them for long enough that this judgmental tendency was bred into the human gene-pool over time.
Self-fulfilling prophesy[edit | edit source]
It is argued that since people are generally treated based on how they look, people learn to behave as they look through social reinforcement, thus creating a feedback loop. This is especially true for people with exotic looks. For example, Anton LaVey, founder of the Satanic Church, always had an extremely evil, unconventionally attractive looking face. Anton (real name Howard) was always treated as being a literal handsome devil. Anton enjoyed this treatment and decided to mold his personality to his looks.
Still, others may not be so accepting of how people treat them based on facial features. A good example is baby-faced men. While they are commonly perceived as more trustworthy and warm, they are also perceived as less competent and dominant. However, there is evidence that baby-faced men are more likely to excel in school and commit crimes as adolescents, in stark contrast to how people generally expect them to behave. This suggests that baby-faced men may be attempting to over-compensate for the infantilizing treatment they often receive from others by attempting to prove their dominance or competence.
Another example would be a pretty boy that may not enjoy being treated as such; thus he may attempt to become more muscular in an attempt to be treated as, or seen to be a gigachad.
Modern physiognomy research[edit | edit source]
In recent years, many studies have concluded that facial features are correlated with perceived and actual personality and character traits. The following is a quick summary of some of these findings:
Criminal/violent physiognomy[edit | edit source]
- A meta-analysis of 19 studies found a weak but statistically significant correlation between fWHR (a measure of the broadness of the face) and aggression, ranging from r = .09 for field and archival studies to r = .21 for studies conducted in research labs. Another study in 2016 found weak but significant correlations between various psychopathic traits and fWHR (r = .12 for the whole sample and r = .27 for a sample of prison inmates). This greater aggression is posssibly mediated by an association between fWHR and higher levels of pre-natal and pubertal testosterone exposure. However, these findings may be confounded by the fact that fWHR is positively associated with greater body size, which is moderately positively associated itself with the propensity to engage in physically violent behavior. Thus, any future studies that propose a link between fWHR and physically aggression would do well to control for BMI and muscularity.
- Studies conducted in 2013 found that people were able to accurately predict the outcomes of fights based on facial features, above chance. The fighters with faces rated as more aggressive were more likely to win their bouts, but they were also confounded by weight; thus, it only held for heavyweight fighters. The facial features associated with aggressiveness were an overall broader face, broader chin, darker eyebrows and horizontally narrowed eyes.
- Holtzman (2011) created a series of prototypical faces corresponding to each of the traits of the dark triad, using the photos of 81 study participants, who completed self-report inventories designed to measure the levels of the dark triad traits. The participants were also evaluated in regard to their level of dark triad traits by their peers. It was found that observers could (above chance) correctly distinguish between high and low morphs of the various "dark traits," thus lending some evidence to the idea that these traits are correlated with a certain facial structure. This correlation was explained by several hypotheses, the facial traits and the dark triad being co-evolved, the facial traits influencing people's self-perception and thus behavior, or that individuals are possibly conditioned to behave in a way 'congruent' with their facial structure by peers, through constant social reinforcement.
- A Chinese study examined the association between fWHR and domestic violence in 144 individuals of both sexes (Wen & Zheng, 2020) found a medium effect size (d = .67) for greater fWHR (as measured to the midbrow) and the men's likelihood of being involved in a domestic violence incident in the past. No such association was found for the female subjects, though greater fWHR was associated with certain aspects of interpersonal dominance. The effect size was much more substantial for fWHR as measured to the midbrow compared to measuring it to the eyelid. However, both were statistically significant and fairly large compared to the usual effects one finds in the social sciences. This discrepancy in effect sizes for the two alternative forms of measuring of fWHR may suggest the midbrow measurement is more accurate in terms of discerning the effects this particular facial metric has on behavior. This discrepancy in the predictive validity for various measures of fWHR is something to keep in mind when interpreting the results of such research. 
- A Chinese study claimed to be able to tell whether someone is a criminal based on machine learning, but the technique turned out to detect smiling instead.
- Stillman et al. (2010) had people rate the estimated propensity of violent behavior of a group of convicted violent and non-violent sexual offenders (N = 87) after a brief (2s) exposure to a static photograph of the offenders. It was found that participants were able to determine whether the offenders were violent or not above chance (d =.44), with no sex differences in these judgements' accuracy. However, women perceived a significantly higher level of threat from the men's photos than men (d =.38, though the significance was borderline marginal). An analysis of the individual target related factors that determined people's judgements of the men's violence proneness found that there was a mix of valid (predictive above chance) and invalid (not predictive or even deceptive cues) cues involved in rater's perceptions. The valid cues were generally markers of overall masculinity and robustness, such as facial masculinity and perceived strength, with age being negatively correlated with violent offending. Physical attractiveness, sadness, and smiles were poor cues of violent behavior, with the associations between these cues and actual offending being weak or non-existent. Interestingly, the deceptive cues included better grooming and general displays of positive affect (happiness), which were believed to be negatively associated with violent behavior. In reality, the actual link between these things and the presence of violent offending was non-existent. The strongest deceptive cues were certain displays of negative affect, such as anger and disgust, which contributed substantially to participants negative evaluations of the men in question, despite the link between these traits (as expressed in the static photos used in the study) and violent offending being non-significant.
- Hoskin & Ellis found evidence for significant associations between several indices of both pre-natal and post-natal androgen exposure and criminal behavior. They examined the correlations between several proposed markers of pre-natal testosterone exposure such as 2D:4D ratio (the ratio of the index finger to the ring finger, which is sexually dimorphic, the ratio being typically lower in men, meaning their ring fingers tend to be longer than their index fingers) and anogenital distance (the length of one's perineum, which is strongly sexually dimorphic; it has been proposed to be a possible marker to use in order to predict the risk of an infant boys future potential likelihood of experiencing reproductive disorders.)
The association between other sexually dimorphic male traits and offending behavior was examined, such as bodily masculinity, masculine behavior, physical strength, and height.
Such associations were generally in the expected direction, though of the pre-natal markers, only AGD was significantly associated with violent tendencies in men (the association was stronger when not controlling for sex). Interestingly, a higher 2D:4D ratio (more feminine) was weakly, but significantly, associated with non-violent delinquency among women.
The authors had no explanation for this finding, though it is in line with the insights from certain philosophers of a misogynist tendency that assert that feminity is associated with provocative and disruptive behavior.
It could also be due to the fact that certain forms of delinquent behavior, such as shop-lifting, are sexually dimorphic in a direction that favors females. The effects discovered for specific masculine traits and criminality were most potent for muscularity, strength, and masculine behavior in men. This could be due to the obvious explanation that such men are more formidable and capable of violence. It could be that men with a general disposition towards criminality deliberately enhance their physical capabilities to carry out violent assaults (reverse causation) more successfully, though the link between masculine behavior and criminality suggests some innate link here as personality is generally both highly heritable and quite stable over time, though more behaviorally masculine men may also have a higher drive to attain muscularity. Of course, the tendency to behave violently is itself a behaviorally masculine trait to some extent, so these are likely overlapping constructs to a degree. It also could be that these traits are associated with higher levels of systemic androgens in the long term, with such long-term exposure potentially being necessary for androgens to alter brain function in a manner that is associated with a greater propensity for aggression and delinquency.
Homosexual physiognomy[edit | edit source]
- Skorska et al. (2015) used a computer modeling program that examined the facial metrics of N = 390 male and female subjects of varying sexual orientations. It was found that lesbians had 'marginally more masculine facial shapes,' upturned noses, puckered mouths, and smaller foreheads than heterosexual women. Homosexual men tended to have more sloped foreheads, convex cheeks, and smaller noses (with nose size being typically considered a masculine feature, perhaps due to larger noses generally indicating greater prenatal androgen exposure, when controlling for ethnicity)  compared to heterosexual men. Therefore, the researchers concluded that facial structure was associated with sexuality in both men and women. This is likely due to prenatal hormonal exposure playing an important role in determining later sexual preferences, while also partially determining facial bone development.
- Wang & Kosinski (2017) used a deep neural network that, analyzing 35,326 'selfie' images, correctly determined homosexuality in 81% of cases for men, and in 74% of cases for women. The decisions made by the neural network were compared to human judges, who could distinguish a man's homosexuality in 61% of cases and women in 54% of cases (slightly above chance). However, this study has been heavily criticized for being confounded by differences in facial expression, grooming, clothing, camera angle, and other contextual factors unrelated to facial structure.
Intelligence physiognomy[edit | edit source]
- A twin study in 2017 found a weak but significant relationship between wider IPD (Interpupillary distance) and actual measured IQ. An earlier study found that people were able to gauge measured IQ accurately from a photograph, but this only held in the case of men's IQ, not women's.
- There is evidence that there is some validity to the popular 'four-eyes nerd' stereotype that people that wear glasses are more likely to be highly intelligent. Specifically, myopia (short-sightedness) has been found to be positively linearly linked to intelligence:
Rosner & Belkin (1987) examined the medical data of 157,748 Israeli military conscripts aged 17-19 and their intelligence test scores. They found a linear association with intelligence and myopia (controlling for education), with the lowest scoring cohort (IQ ≤ 80) having rates of myopia much lower than the general population (8% vs. 15.8%). Conversely, the highest-scoring cohort (IQ ≥ 128) had significantly higher incidences of myopia (27.3% vs. 15.8%). The authors also found a positive association between myopia and attained education level, with intelligence test performance being controlled for in this analysis.
It is not entirely clear how this positive relationship between intelligence relates to the different lower-order factors that derive from general intelligence in the standard model of general intelligence (i.e., non-verbal vs. verbal intelligence). Some studies have found no association between myopia and non-verbal intelligence, indicating that the link between intelligence and myopia may be environmental in nature, i.e. myopic people read more or reading puts strain on the eyes that can induce myopia; however, they did not control for reading frequency or preference for near-work. Saw et al. (2004) found a large positive link between myopia and non-verbal intelligence in a large sample of Singapore Chinese children (aged 8-12, N = 1204), controlling for books read a week and preference for near work (multivariate OR for myopia among the highest quartile of intelligence: 2.4, 95% CI 1.7-3.4, as compared to the lowest quartile of intelligence, the reference group).
In order to determine the root cause(s) of the observed link between myopia and intelligence, Williams et al. (2017) conducted a study on a large number of twin pairs aged 14-18) who were administered a verbal and non-verbal intelligence test at age 16. The genotypic, lifestyle, and medical data of this cohort were pooled and subjected to extensive statistical analysis. The conclusion was that the phenotypic link between myopia and intelligence was largely a result of pleiotropic (several genes influencing the development of the same trait) genetic influences and that this indicates that certain aspect of brain and eye growth may be determined by the pleiotropic effects of several genes that may be co-inherited, as both traits are strongly polygenetic (the result of additive effects of several genes).
Mental health physiognomy[edit | edit source]
- Certain behavioral conditions may result in particular facial expressions. For example, very intense people might bite their lips, crunch their teeth, chew their nails and engage in other nervous tics. Autists are often described to have an empty, monotonous facial expression. Some mental conditions result in a tilted mouth. Socially excluded individuals may develop tics as nobody is there to correct their behavior, or such behavior may be induced by the greater stress such people often face. People are, in fact, able to infer other's mental conditions above chance level based on photos. Autists are often recognized as less approachable and attractive within seconds.
Infidelity physiognomy[edit | edit source]
- Foo et al. (2019) found that women relied on men's attractiveness, facial masculinity, and overall perceived trustworthiness to attempt to determine their likelihood of committing infidelity, however of these traits only facial masculinity predicted actual infidelity related behaviors above chance. This effect however, was weak, with only 14.1–18.0% of raters being able to accurately gauge men's infidelity related behaviors above chance. There was no such effect found for women, despite men being able to judge infidelity related behaviors above chance in other men. Earlier research, found that men could only make above accuracy judgements of infidelity when presented with two women, one of whom had committed infidelity several times in the past compared to who had not.
Leadership physiognomy[edit | edit source]
- A study in 2017 found that fWHR influences social status, with Popes and CEOs typically having higher than average fWHRs. This association between fWHR and social dominance is due to either broad-faced leaders being more effective and socially dominant, or due to their being perceived to be so.
Personality and physiognomy[edit | edit source]
- Several studies have found a link between a less acute slant to the forehead and impulsivity. Apolo et al. (2018) measured the angle of forehead slant of a largely male sample (80.2%) of traffic offenders. They found that a more heavily slanted forehead correlated positively with several psychometric measures of impulsivity. A follow up study by the same author (Apolo, 2020) examined a pre-screened sample of people who had been subjected to neuropsychiatric testing to exclude those with hyperactivity disorders, attentional deficits or a family history of mental illness (unlike the aforementioned study which was made up of a sample of offenders, who would be expected to be more impulsive compared to a general population sample). Apolo found multiple weak to moderate (.3-.6) significant positive correlations between the degree of forehead slant and several psychometric measures of general impulsivity, sensation-seeking and positive urgency (the tendency to engage in risky behavior more often when one is experiencing positive affect). The study also found a positive relationship between cortical thickness in certain regions of the frontal and parietal lobes and impulsive behavior. This suggests that differences in the development in different brain regions that affect behavior can be observable by examining the shape of a person's skull, a finding reminiscent of the 'pseudo-science' of phrenology, which also proposed a link between skull shape and behavior. As forehead slant is sexually dimorphic (with men generally having a more slanted forehead than women due to having a more pronounced brow ridge), it could also be that this link is due to a more slanted forehead being reflective of higher levels of androgen exposure during key developmental periods, which has also been found to be linked to a higher risk of engaging in impulsive and violent behavior.
Life history theory and physiognomy[edit | edit source]
There is some evidence that several physical traits are related to a person's life history strategy (whether someone is adapted to a life-fast die young ecology, or is geared towards higher long term resource acquisition and investment in offspring). Thus these physical life-history traits may also be somewhat reflective of their personality.
For instance, tattooing is associated with impulsiveness and promiscuity. Heywood et al. (2012) found that tattooing was associated with drug use, promiscuity, and lower SES, despite changes in the societal acceptance of tattooing. Other studies have replicated this finding, and also found that various forms of body modification such as piercings are also associated with greater levels of promiscuity, particularly in women. In terms of (mostly) unalterable traits, there is some (very tentative) evidence that women with larger breasts are generally more promiscuous, impulsive, and extroverted.
These traits are likely reflective of a fast life history strategy, focused on monopolizing as many mating opportunities as possible (for men), and maximally sexually advertising oneself to desirable mates (for women).
Other bodily traits are possibly linked to a slower life history strategy. For instance, the study previously mentioned suggests that women with smaller breasts may generally be more introverted, chaste, and generally exhibit more self-control than larger breasted women. Wiggins et al. (1969) found that men who preferred smaller breasts were also generally more restrained, introverted, and less sexually successful. It has been argued that this could be because smaller breasted women may exhibit a slower life history strategy (being less focused on producing attractive ornaments). Men pursuing a similar life history strategy could be drawn to this through a process of assortative mating (preference for similar traits to oneself in a sexual partner).
Villainous physiognomy[edit | edit source]
A team of Psychologists from the University of Warwick conducted a study which showed downward pointing triangles are perceived as threatening as well as negative faces in a crowd. Dr. Blagrove says, "If we look at cartoon characters, the classic baddie will often be drawn with the evil eyebrows that come to a downward point in the middle. This could go some way to explain why we associate the downward pointing triangle with negative faces. These shapes correspond with our facial features, and we are unconsciously making that link." This explains why the evil eyebrows and pointy chin activate our innate threat instincts. A cross-sectional study done in 2017 on the Dermatologic Features of Classic Movie Villains shows that the top 10 villains display a significantly higher incidence of dermatologic findings than the top 10 heroes (60% vs. 0%), which include alopecia, periorbital, hyperpigmentation, deep rhytides on the face, scars on the face, Verruca Vulgaris on the face, and rhinophyma. Showing that people readily associate signs of unhealthy and or damaged skin with an evil, villainous nature.
In men, a large cut at just the right angle over the face seems to improve attractiveness. A study done in 2008 by a team of researchers at the University of Liverpool proves that women indeed find battle-scarred men more attractive for short term relationships. This may be due to women perceiving such men as brave, dominant 'bad boys' moreso than men without facial scarring. Deliberate scarification as a form of signaling bravery and stoicism is common in many non-industrialized cultures around the world. It was also popular in some Western societies in the past, for instance, it was common among aristocrats in 19th century Germany, where these scars became associated with the aristocratic military caste due to them often being members of university fencing clubs. In this milieu, using face masks was viewed as a sign of timidity and weakness.
Facial trust physiognomy[edit | edit source]
Someone's facial appearance plays a significant role in how trustworthy they are perceived to be. For instance, attractive people are seen as warmer and more trustworthy than unattractive people due to the halo effect. However, attractiveness is not the only facial characteristic people use to make a snap judgment of trustworthiness.
For example, it has been found that wider-faced men are seen as less trustworthy, which means people are less willing to cooperate with them. This type of bias can sometimes result in dire outcomes for men judged as less trustworthy based on their facial appearance. For instance, Wilson & O'Rule (2015) found that men with less trustworthy appearing faces were more likely to receive hard prison sentences and even the death penalty.
DeBruine (2002) found that people also tend to trust those who bear a greater facial resemblance to themselves, possibly due to kin selection based evolutionary mechanisms and perceived relatedness.
See Also[edit | edit source]
[edit | edit source]
Review of some the research on physiognomy:https://www.youtube.com/watch?v=43zciPFbTpc
References[edit | edit source]
- Zebrowitz, Andreoletti, Collins, Lee, & Blumenthal, 1998 https://link.springer.com/article/10.1007/s10508-011-9821-6
- Fishbein, M. Progress in Social Psychology: Volume 1