Sunday, November 28, 2010

Femmes claires, hommes foncés (English post)

Averaged faces of 22 women and 22 men (White American subjects with no makeup). Female faces are lighter-skinned than male faces, while showing more contrast between facial skin and lips/eyes (see research by Richard Russell).

Les Presses de l’Université Laval has just published my book Femmes claires, hommes foncés. Les racines oubliées du colorisme.

An old Christian manuscript recounts the story of a man who went to live in a monastery with his infant son. As the boy became a young man, he began to see strange-looking beings in his dreams. One day, he ventured with his father into the outside world. On seeing some women, he exclaimed: “Father, those are the ones who came to see me during the night!”

We do not learn to recognize human faces. Nor do we learn to identify whether they are male or female. This type of image is processed by an innate mechanism that we inherit independently of other cognitive abilities. If this mechanism ceases to function following brain damage, the result is an odd syndrome: the patient may seem to be like everyone else but will not recognize a normally positioned face more easily than any other object, including an upside-down face.

This should be no surprise. If an object appears often enough in your visual field, while playing a significant enough role, you gain from recognizing it automatically (instead of having to learn its visual characteristics). Over time, natural selection gradually hardwires recognition of familiar objects, like human faces.

Which facial features are processed by this mechanism? There are notably the eyes and the mouth. There is also skin tone. Indeed, this feature seems to be key to distinguishing between male and female faces. When shown a facial photo, human subjects can tell its sexual identity even if the image is blurred and provides only the visual cue of skin tone. This cue has two aspects. First, women are paler than men, who are browner and ruddier because their skin is richer in melanin and blood. Second, women exhibit a higher luminous contrast between their facial skin and their lips and eyes.

This sex difference is universal, being larger if the population is medium in skin color and smaller if very fair or very dark. Just as universally, people tend to ritualize this difference in traditionalist societies, in general by lightening women’s skin even further through sun avoidance, wearing of protective clothing, use of cosmetics, and so on. In contrast, it gets little attention in modern societies, where skin color means ethnicity. Hence a common reproach I hear: “But that no longer matters nowadays. Skin color is about racism!”

Perhaps. But this sex difference used to matter. It once formed part of our very notions of femininity and masculinity. And whatever we may think, this past has shaped us.

Origins of male and female skin tones

Humans are born with very little skin pigment. This paleness is striking in dark-skinned populations, which consider it a mark of infancy. In Africa, a proud mother may invite her neighbors to see the ‘white’ person who has just arrived!

Skin darkens until just before puberty when this trend reverses in young girls, the result being a sex difference. This lightening of female skin after puberty coincides with the thickening of subcutaneous fat. It seems that both are part of the same process of sexual development, like other changes that happen at this life stage.

What use is a fairer skin for women? There are three hypotheses:

Infantile mimicry? A fairer color is one of several visual cues that identify the human infant including smooth hairless skin and a ‘baby face.’ These cues acquired the property of decreasing the observer’s aggressiveness and increasing his/her feelings of care. They were then retained by the adult female as a means to modify her partner’s behavior in the same way. Such mimicry exists in other primate species. To the degree that the sexual bond is stronger and longer-lasting, the adult female retains certain highly visible infant traits.

Hormonal side effect? Fairer skin, through a fortuitous interaction between pigmentation and the sex hormones, became a means to assess a potential partner’s fecundability. Remember that girls lighten in color after puberty. Later, women tend to darken during pregnancy and as they get older. They also darken slightly during the nonfertile phase of the menstrual cycle.

Means to prevent vitamin-D deficiency? Natural selection lightened women’s skin in order to increase vitamin-D production, thereby ensuring enough calcium and phosphorus during pregnancy and breastfeeding.

Whatever the initial cause, lighter skin also became a means to distinguish women from men, as long as this sex difference remained the main source of pigment variation.

Today, women’s lighter skin has become much less noticeable in an increasingly multiethnic context. This gender line has also been blurred with the entry of the suntanned look into women’s fashion. Thus, in a survey of university students, I found that only a quarter were aware of this sex difference.

Earlier meanings

Yet our ancestors were very much aware. Before their continent opened up to the world five centuries ago, Europeans described skin color with reference to the complexions they saw among themselves. ‘White,’ ‘brown,’ and ‘black’ corresponded to what we now call light, tan, and dark. Again contrary to current usage, these skin tones identified individuals rather than ethnic groups. A white was a fairer-skinned person and a black a darker-skinned one. This way of seeing things persists in family names that once referred to gradations in pigmentation within a single population, like Leblanc, Lebrun, and Lenoir among the French, White, Brown, and Black among the English, or Weiss, Braun, and Schwartz among the Germans.

This narrow spectrum was conducive to gendering of skin color. A woman had to have a fairer skin than average, i.e., ‘white’ in Europe or East Asia, ‘golden’ in South-East Asia, and ‘red’ in sub-Saharan Africa. Despite being normative for women, a fairer skin did not monopolize all erotic male desires. In old European folklore, some desires could target darker women, i.e., the nut-brown maid of the English, the braunes or schwarzbraunes Mädel of the Germans, the brune of the French, or the barna kislány of the Hungarians. This type of eroticism was ardent, but also stormy and short-lived.

Conversely, a man had to have a darker skin than average. There was nonetheless some ambivalence. A man was handsome if fair, but virile and strong if brown. In medieval England, the tenth token of a knight of ‘strong Corage’ required a ‘broun coloure in al the body’, a quality that many vaunted by adding ‘the brown’ to their names.

According to American psychologist Richard Russell, people still use skin color to identify gender, albeit unconsciously, by means of two visual cues: (1) the absolute luminosity of the skin and (2) the contrast between this luminosity and that of the lips and the eyes. Russell also argues that these cues have shaped the evolution of cosmetics in different culture areas and different time periods. There is a cross-cultural tendency for women to lighten facial color and to accentuate its contrast with lip and eye color.

Sexual attraction and other functions

Besides recognition of sexual identity, skin tone seems to be used for other mental tasks. First, there is sexual attraction, which nonetheless involves many other factors (personal history, social and physical context, nature of the sexual relationship, etc.). This task seems to be governed by the levels of the sex hormones, at least in women. This is what I found in presenting female participants with six pairs of facial photos, three being female and three male. Each pair of faces was identical, except for a slight difference in skin tone. The participant then had to choose the face she preferred. It turned out that preferences varied with the phase of the menstrual cycle, but only for male faces. The darker man was more strongly preferred if the participants were in the first two-thirds of their cycle (high ratio of estrogen to progesterone) than if they were in the last third (low ratio of estrogen to progesterone).

Other researchers have noted this cyclical effect for other secondary sexual characteristics, like face shape and body odor. The more the level of estrogen increases, the more the ‘masculine’ version is preferred.

Skin tone also seems to influence certain prejudices. This phenomenon has given rise to a large body of literature, the theme evidently being how children learn race prejudice. An exception would be the work of two American psychologists, Deborah Best and John Williams, who saw this prejudice as being constructed from a universal and early developing tendency to prefer fairer skin.

This was their finding in Japan and in Europe with young children who were unfamiliar with darker-colored ethnic groups. When the children were shown pictures of people or animals, they associated lighter skin with positive qualities (e.g., clean, pretty, nice) and darker skin with negative qualities (e.g., dirty, ugly, nasty). These associations did not seem to be learned. Their development did not follow a learning curve. Nor was there any correlation with the child’s IQ, as would be if they were learned.

But does all of this come down to fairer skin = positive qualities and darker skin = negative qualities? When through a translation error the children were given the word ‘robust,’ they associated this positive quality with darker skin. It seems that Best and Williams unconsciously chose feminine-sounding positive qualities and masculine-sounding negative ones.

Upcoming …

Last year, a Chinese research team demonstrated that the human face, as a visual object, is analyzed by a distinct mental mechanism. This was a major project with many participants but it proved in several months what had been suspected for several decades. With the same method, we could find out whether this mechanism analyzes skin tone. All that is needed is a research team ready to act.

Thirty years ago, such teams existed. Today, there are no longer any. In their time, Best and Williams attracted many collaborators and sources of funding. Then, in the 1990s, the team was dismantled and its members directed to more down-to-earth projects.

The same story played out elsewhere in North America. To a large degree, the reason was the shift to applied research. The watchword was to make research more ‘cost-effective’, more ‘targeted’ and more ‘realistic.’ This conservative discourse was taken up by other people, who saw a way to settle old scores and control the future …

It must be said that biological determinism bothers some people who see in it a fatalistic ideology, even an adversary of efforts to improve the human condition. In my opinion, such a view is exaggerated, but it was the one held by most decision makers in the 1980s and 1990s. The result? This paradigm has virtually disappeared from most social science departments. At times, this ‘departmental cleansing’ took place out in the open with much shouting and finger-pointing. But in general it happened quietly through attrition and reprioritizing.

North America has long claimed to lead the free world—a claim that is less and less true. If the social sciences are the ‘canary in the mine,’ we seem to be moving toward another model of society, one where people naively avoid difficult questions to avoid the troubles that come with them.


Tod said...

It's pretty conservative in the sense of leaving out much of your hypothesis on the origin of white skin. It seems progress is going to be agonizingly slow.

ItsTheWooo said...

I would point out those pix are deceptive as the lighter complexioned face has the beard stubble eliminated via the higher contrast. It also appears as if the facial features are softer and feminized because the dark cast of the strong nasal bridge and eye orbits are eliminated by the high contrast. The high contrast also makes the eye brows appear higher by thinning them out (as the stray hairs are eliminated by high contrast). All ina ll the high contrast does a lot more to feminize the face other than lighten the skin color - you have lost facial hair (both around the mouth and above the eyes) and you have lost visual cues as to the shape of the bone structures (the one with the deeper contrast has a pronounced nasal bridge and deep eye socket near the nose which are important testosterone cues).

I am not arguing whether or not light skin is a feminine gender/sex cue - I believe it is - however those pix only look so different because the high contrast is washing out important testosterone cues (facial hair stubble and strong facial features of nasal bridge/eye orbit/lower set brows appear higher etc).

Women often take pictures on high contrast for this reason - it washes out imperfections and all you are left with is a floating pair of eyes and a mouth which look beautiful of course (and hard features/large noses/ skin blemishes are washed out).

I would also argue that the lips on the left side are SLIGHTLY fuller, and this does not appear to be an optical illusion. The images have been manipulated in more ways than contrast. If you superimpose the dark face over the light face it is beyond apparent that this face has been manipulated to be more feminine (the upper lip plumps out very imperceptibly at the middle sides).

If you superimpose the dark face over the light face and play with the opacity it becomes OBVIOUS the main reason the light face looks more feminine is because high contrast washes out testosterone cues. In real life light skinned people this would not be a factor as shading and luminosity are unaffected by skin color (that is to say, people with light skin will have shadows cast by hard testosterone-treated facial features, it is ONLY in photoshop where the skin has been artificially lightened or darkened will we see an unnatural exaggeration or elimination of these features).

I agree skin color is one (minor) gender / sex cue ... but fraudulent reserach like this is not going to help the case. It is obvious to anyone with a basic understanding of photoshop or art why the left face appears more feminine and it has VERY LITTLE to do with skin color. Although, someone who is unfamiliar with art (i am an artist) or photoshop might be fooled.

Peter Frost said...


I wasn't trying to play safe by not discussing the evolution of European skin color in my book. It's really a separate question.

Its the Wooo2,

Those two pics are composite images of real-life men and women. There was no photoshopping, no 'manipulation,' and no 'fraud.' The pictures were taken under controlled conditions, using subjects with no makeup.

Yes, there are differences in lip size, face shape, and eye socket depth between the two pics. That is hardly evidence of manipulation. Perhaps you haven't noticed, but men and women differ in facial morphology.

I am familiar with Richard Russell's research, and I can assure you that he is a man of integrity with a long record of research work and publications from MIT and elsewhere.

On a final note, I don't normally get angry, but your accusations of 'fraud' make me sick.

Shame on you! Go find yourself a rock to crawl under.

Tod said...

ItsTheWooo2 obviously didn't bother actually reading this text.
Averaged faces of 22 women and 22 men (White American subjects with no makeup). Female faces are lighter-skinned than male faces, while showing more contrast between facial skin and lips/eyes (see research by Richard Russell).

He assumed the photos were being claimed to be an identical picture that had been digitally altered only with different skin tones which made an observer perceive it as male or female. As he has a good enough eye for detail to see they were not he proceeded to made a fool of himself.

I'll own up - I thought the same and didn't bother to read it properly either and made the same mistake (except I didn't realize the photos were actually unalike different because I don't have the same faith in my eye for detail).

And to make it worse that was after implying in my earlier comment that I understood all about the hard science research findings that have gone into the new book. ( Even though in fact I didn't).

Confession is good for the soul.

ItsTheWooo said...

Look, Peter, I'm sorry if you found what I said to be offensive but it doesn't make it untrue.

Why don't you get photoshop and superimpose the two images over each other like I said? IT IS OBVIOUS the dark face has features visible which the white face does not. This is exclusively a byproduct of bad photoshopping, it is not relevant to real life (i.e. a real life white skinned person you can see their beard stubble and you can see their eyebrows and you can see the hard ridges of the eye orbits/nasal bridge). The ONLY REASON the white face looks feminine is because of a piss poor photoshop job to artifricially lighten the face which washed out all testosterone driven features.

It's like you *so badly* want this idea to be true that you are completely willing to ingore big red flags that your "evidence" has problems.

Tod - so basically waht you are saying is that if you use photoshop to clearly make a picture more masculine (by exaggerating facial hair and the depth of boney prominences), then it is okay to attribute that difference to SKIN COLOR????


And if anyone has poor reading comprehension, it's you. The author meant that the contrast between dark lips and pale skin creates femininity (which i don't disagree with) however he did NOT mean to imply that his crappy research methods ended up making the dark face look masculine for reasons other than the gradient difference between the lips and the skin (the crappy photoshop job accentuated facial hair and boney prominences, which were washed out by the artificial lightening). THIS IS NOT A FACTOR IN REAL LIFE. In real life, facial hair and boney prominences are actually much more visible on white skin,only in photoshop (not the real world) is an artificially lightened face automatically given sexual reassignment. In the real world, lighter skin accentuates bone structures and facial hair and other masculine cues.
WOmen may prefer darker skin for other reasons, all the same, as slightly darker skin is a sign of high testosterone. Dark skin itself is NOT a sign of high testosterone, neither is white skin a sign of estrogen. You people are confusing ethnicity with sexual markers. There is a big difference between a woman with very black skin and a man with very white skin (1) women do not prefer very dark skin and men do not prefer very white skin (2) so I really don't see the point of constantly conflating ethnicity with sexual dimorphism.

Sideways said...

Itsthewoo, the reason you're not seeing signs of facial hair on the lighter picture is that women do not generally have noticeable stubble. Get it yet?

ItsTheWooo said...

Sideways said...

Itsthewoo, the reason you're not seeing signs of facial hair on the lighter picture is that women do not generally have noticeable stubble. Get it yet?

I'm really trying to stay unemotional here but the responses I"m getting are so obtuse and ridiculous.

You are using CIRCULAR logic. These are the facts:
1) The picture is a combination of 22 men and women's faces.
2) The light picture is PHOTOSHOPPED to be lighter. THe dark picture is PHOTOSHOPPED to be darker. Presumably nothing else is changed but tinkering with the contrast.
3) This photoshopping ARTIFICIALLY accentuated masculine features in the dark picture and dminished feminine features in the light picture.
4) this is not relevant in real life, as in real life light skinned people have more prominent masculine features as shadows and stubble are more noticable on lighter skin Only in photoshopped pictures which are artificially lightened will the face appear more feminine afterward because you have also lightened the color of the stubble and the color of the shadows making them less prominent.

I don't see why you are having a hard time understanding this.

Your circular logic of saing "the light picture doesn't have stubble because women don't have stubble, therefore light skinned people are feminine and this proves it"... it's proving nothing, and it is avoiding acknowledging the fact that the methods of research here are flawed (the photoshoped lightened picture looks significantly more feminine for reasons other than skin color, as the photoshoping has also washed out facial hair and shadows indicating boney prominences).

It would be like arguing an experiment to see if the sky is really red, used a camera which added a red tint to all the pictures... and I came a long and pointed out the results are not valid because the methods used are inappropriate... and you insisted "the results are valid because the sky is really red and so cameras will capture a red sky". You're just restating your original bias (that you think the sky is red / that you think women are white) without even ADDRESSING my point (the research method is flawed - the photoshoped picture has washed out beard stubble and shadows as well, which is not a factor in real life pale skinned people which actually have much more obvious stubble and more obvious shadows that show up better on slightly lighter skin tones than super dark ones)

Anonymous said...

ItsTheWooo2, it's a peer reviewed paper. You seriously think it has flaws that an interested amateur like yourself could spot? Pleeze.

Stephen said...

QUOTE"This way of seeing things persists in family names that once referred to gradations in pigmentation within a single population, like Leblanc, Lebrun, and Lenoir among the French, White, Brown, and Black among the English, or Weiss, Braun, and Schwartz among the Germans."

Seems more likely to me that these names are referring to hair color rather than skin color.