The perception of facial expressions

A team at Glasgow University in Scotland published research in 2009 in the journal Current Biology on differences in the interpretation of facial expressions by different racial groups [1]. The research suggests that Whites [2] and East Asians differ significantly in their mode of scrutiny of faces and their success in identifying emotions from facial expressions.

Whites concentrate their attention on the eyes and the mouth equally, while East Asians concentrate largely on the eyes. The consequence is that the latter have difficulty in distinguishing expressions which have a similarity around the eyes. Whites, who use two reference areas, are significantly more adept at correctly identifying such expressions. The difference in the mode of scanning faces used by the two groups translates into a difference in the emoticons used by Whites and East Asians. Whites use representations of the mouth for happy and for sad; East Asians use representations of the eyes for happy (^.^) and for sad (;_;).

The research involved White and East Asian subjects [3] viewing still images of faces whose emotions were classified using the Facial Action Coding System (FACS), which categorises emotion depending on the muscles used. Those wishing for the full technical details of the study can find them here.

The subjects were shown both White and East Asian photographs, with expressions classified as Happiness, Surprise, Fear, Disgust, Anger, and Sadness, as well as Neutral, with 'Same Race' and 'Other Race' pictures, all in accord with FACS criteria.

Whites correctly identified expressions all the time, but East Asians confused Fear with Surprise and Disgust with Anger. The Sadness, Happiness, and Neutral images produced similar results amongst both White and East Asian subjects. Let us try to winkle out why this might be.

The nature of the mis-identified emotions

There are two obvious differences between the expressions which are and are not correctly identified. First, all the emotions involved in incorrect identifications are in some manner unpleasant emotions, while happiness and sadness contain one pleasant and one unpleasant emotion.

Fear & Surprise and Disgust & Anger are pairs which have some tangential similarity. Indeed, they may be experienced at the same time or at least in rapid sequence, giving the impression of emotions being mixed. Probably because they are emotionally cousins to one another, their facial representations have similarities, for example, we raise our eyebrows and widen our eyes for both Fear and Surprise.

Contrariwise, Happiness and Sadness are two diametrically opposed emotions. They have no tangential similarity and their facial expressions are perceived as discrete and presumably more easily recognised.

The second major point of difference is the response they produce in others. Fear, Surprise, Disgust, and Anger are all what one might call active emotions. When we experience them we do so in an energetic way, for it is impossible to feel any of these emotions without being physical aroused because to experience them will result in a rush of adrenaline. Conversely, Sadness and Happiness, although they may be experienced in an energetic way, for example in ecstasy or violent grief, can also be experienced in a physically quiet manner.

It is also arguable that the Sadness or Happiness of others, unless we are significantly emotionally attached to the person, does not evoke as strong a response in an observer as Fear, Surprise, Disgust, and Anger do, regardless of how well or little the person displaying the expressions is known to the observer. The reason for this is easy to see: Fear, Surprise, Disgust, and Anger all offer a potential threat, whether that be experiencing something unpleasant (Disgust), concern about whether there is something to worry about of which we have yet to be aware (Fear, Surprise) or the fear of someone indicating they are in a state to do you harm (Anger).

Masking their emotions?

The bias in the East Asian mis-identifications

There was a pattern to the East Asian mis-identifications. They showed a bias towards the softer, less threatening emotions. Faced with a choice between Fear and Surprise they chose Surprise and between Disgust and Anger, Disgust.

The researchers attribute this tendency amongst East Asians to select less threatening emotions to be culturally determined. This may be the case, but it would be rash to accept it as self-evident. East Asians may choose less threatening emotions when they misidentify expressions simply because their mode of scanning the face makes one type of emotion easier to identify than another. Alternatively, and more interestingly, East Asians could be genetically slanted towards selecting less threatening emotions. Unless you argue that personality is not subject to any genetic influence [4], you cannot say that the genes which control personality are not subject to natural selection. If the form of a society favours the quiescent personality, then those with the genes which tend towards such personalities will be favoured. There is evidence that there are innate behavioural differences between racial types and the reported differences in facial perception between Whites and East Asians seem to correlate with this.

A quarter of a century ago E.O. Wilson reported on studies by D.G. Freedman (1974, 1979) on new born infants which...
"...demonstrated marked racial differences in locomotion, posture, muscular tone and emotional response of newborn infants that cannot reasonably be explained as the result of training or even conditioning within the womb. Chinese-American newborns, for example, tend to be less changeable, less easily perturbed by noise and movement, better able to adjust to new stimuli and discomfort, and quicker to calm themselves than Caucasian-American infants."
(P.274 Sociobiology: The Abridged Edition, abridged version) 
More recently Professor Philip Rushton addressed the subject:
"Temperamental differences, measured objectively by activity recorders attached to arms and legs, show up in babies. African babies are more active sooner and develop earlier than White babies who, in turn, are more active than East Asian babies. Motor behaviour is a highly stable individual difference variable. Even among Whites, activity level measured during free play shows highly significant negative correlations with IQ: more restrained children average higher intellects.
Solving The African IQ Conundrum: "Winning Personality" Masks Low Scores
In my October 2009 American Renaissance article, Why Have East Asians Not Dominated? I addressed the failure of East Asians to become the dominant race, viz:
"despite their higher average IQ, Asians have probably failed to become the culturally dominant race because their innate personality traits work against them. They are too passive, too unquestioning, too lacking in initiative. The shape of their IQ with higher non-verbal scores and lower verbal scores may be wholly or partially the cause of these personality traits or, conversely, the shape of the IQ is simply an expression of the personality traits. Other biological traits such as low testosterone levels may also promote such behaviour."
If East Asians are truly less able to interpret facial expressions than whites, this could provide an explanation of why, despite their superior IQ distribution, they have failed to become the dominant racial type in terms of social development and intellectual and technological advancement. The difference in ability to interpret facial expressions may mean that East Asians are less socially adept than whites with a consequent need for different social structures to Whites.

How could such a perceptual difference arise?

Some behavioural signals are almost certainly entirely customary rather than innate. For example, Britons and Americans nod their heads to signify agreement and see black as the colour of death, the Chinese shake the head to signify agreement and see white as the colour of death. It is conceivable that there are differences in brain function which determine such differences but that is improbable going on impossible.

Conversely, a trait such as the interpretation of emotions from facial expressions is most unlikely to be culturally determined. We recognise emotions from facial expressions for the same reasons that our nearest primate relatives, the apes, recognise behaviours to indicate calm, threat and so on. It is simply part of the species’ template. Unless human beings have some form of mental abnormality such as autism, they recognise the meaning of facial expressions without consciously thinking about them. Nor do people have to be actively taught how to recognise facial expressions, although it may be that the facial expressions become associated with certain types of behaviour as the child develops, and from that information the child extrapolates from particular instances where an expression occurs to using the expression as a general signifier of an emotion rather than the response to an event. (The behaviour of children supposedly brought up without human contact – raised by animals of one sort or another for example – suggests that this may be the case).

Now you understand.
But even if the identification of expressions did occur in that way, it would not explain the differences in mode of scanning which is the most plausible cause of the difference in identification success. There is absolutely no evidence of cultural practices which would lead people of one racial type to behave in one way when they scanned a face and people of another racial type in another way. Indeed, it is difficult to even envisage such a cultural practice because the behaviour of scanning the faces of others is such an innate and natural thing, something which can be seen in very young babies.

But if the difference in scanning is genuine how did it arise? If it is not cultural it must be genetic. A trait which was advantageous would be preferentially selected and spread. Why would it be advantageous? Perhaps the range of possible physical expression in East Asian faces is less than it is in Whites. Suppose further that the range of expression in East Asian faces is reduced around the mouth. The most naturally efficient thing for East Asians to do would be to concentrate on the eyes. Natural selection would work on that trait, favouring those best able to interpret from the eyes.

That leaves the question of why East Asian faces might be less expressive. If, as has been suggested, the environment in which East Asians evolved was abnormally cold, and, as a response to this environment, the East Asian face came to contain more fat and internal physical facial features to guard against the cold, this may have so changed the morphology of the face that it restricted the ability of East Asians to communicate through facial expressions. It is possible that the old white jibe that “they all look alike” may have some truth in it.

An alternative explanation could be some general difference between the European and East Asian languages. Perhaps East Asian languages cause their users to move the mouth less energetically than do European languages, and this gives less non-verbal information from around the mouth area, causing East Asians to concentrate on the part of the face which does give more accessible information, namely the eyes.

"Listen to my eyes!"
Different languages use facial muscles in different ways. This affects the shape and mobility of the face, which in turn will cause facial expressions to differ. These may be very subtle differences in terms of physical difference, but very significant differences in terms of perception by others.

It is also possible that differences in the perception of facial expressions vary not merely amongst racial groups but also amongst different cultures or even different groups within a population, such as classes or those with distinct accents or dialects.

Yet another explanation may rest on the East Asian’s visual bias, as shown in their disproportionately high strength when dealing with non-verbal questions in IQ tests. It could be that the East Asian concentrates on the eyes because that is the sense most important to them.

Finally, there is the possibility of functional redundancy.

As any dog or cat owner will vouch for, animals can be incredibly sensitive to identifying human emotional states. They do this entirely by picking up non-language signals. That ability they extend to other animals, both of their own and other species. The ancestors of homo sapiens, earlier forms of homo such as homo habilis and homo erectus, must have been in much the same boat as animals. Their language skills would be much less than that of modern man and, like animals, interpretation of non-verbal signals, such as facial expressions, would be much more important to them than it would be to homo sapiens in a primitive state and vastly more important than such abilities are to men living in sophisticated societies.

As human beings evolve, perhaps there is less need for accurate interpretation of emotions because the reliance of human beings on one another for survival lessened as societies became ever more sophisticated – there is a big difference between living in a tribe or band of 50-200 people where every individual is important to the survival of the tribe and living in a large city where the loss of an individual will not harm the community.

Your emotions are not important. The train leaving on time is.
It could be that East Asians – with their superior average IQ – simply became less efficient at such social skills because they became less necessary to the type of society they naturally created. By that I do not mean that the society they created was the most advanced possible – indeed, the reality of East Asian societies suggests that they put a block on technological and intellectual advance beyond a certain point. Rather, I am suggesting that the society their innate behaviour created was an efficient means of managing East Asian populations, and that was all it needed to maintain the society.

Possible variations in the perception of other non-verbal behaviour

Apart from interpreting facial expressions and using the overt meaning of language, individuals have many other ways of assessing emotion in others. Human beings definitely use body language and the nuances of language structure (syntax, grammar) and responses to the quality of voice (pitch, timbre, speed, and so on). They may also use less obvious clues, such as pheromones.

This raises a problem for the Glasgow research. They have measured only one means of interpreting emotions, in which one racial group is apparently less competent than another. That is significant as far as it goes. Where it does not go is into real life situations, where the whole range of verbal and non-verbal clues are available to allow the individual to make a judgement on the emotional state of others. In addition, in real-life, human beings do not have to rely only on their own judgement to make such decisions. They can ask others. It could be that East Asians, while deficient compared to Whites when it comes to facial recognition, are as effective as Whites at identifying emotions, more effective than Whites, or even less effective than whites when more than facial scrutiny is employed, with variations in ability arising from different combinations of the various clues humans give to their emotional state, for example, facial expressions plus body language might trump facial expressions and quality of voice in one racial group but not the other.

Language is a signifier but what does it signify?

It is one thing to call things by the same name, quite another for the things called by the same name to be the same thing to each individual. To begin with there are the difficulties of exactly translating ideas from one language to another. For example, the word for Disgust in Chinese may have different connotations to the English word Disgust, or the English word Disgust may have different shades of meaning for those who have English as a first language but who come from significantly different cultures, for example, a white Englishman and a white Barbadian.

It is certainly true that languages are not equal in their functionality. Consider the case of the Piraha, an Amazonian tribe with several hundred members. They have been in contact with Brazilian culture for two centuries or more, yet they display some very odd traits, one of which is to have no sense of number! An American linguistic anthropologist Daniel Everett has studied them for 27 years. Apart from their innumeracy, Everett reports that...
"...the Piraha is the only people known to have no distinct words for colours. They have no written language, and no collective memory going back more than two generations. They don’t sleep for more than two hours at a time during the night or day. Even when food is available, they frequently starve themselves and their children… They communicate almost as much by singing, whistling and humming as by normal speech. They frequently change their names, because they believe spirits regularly take them over and intrinsically change who they are. They do not believe that outsiders understand their language even after they have just carried on conversations with them. They have no creation myths tell no fictional stories and have no art. All of their pronouns appear to be borrowed from a neighbouring language."
Life Without Numbers, The Globe and Mail – Friday, August 20, 2004)
The Piraha’s innumeracy is particularly interesting.
"Their lack of numbering terms and skills is highlighted in a report by Columbia University cognitive psychologist Peter Gordon that appears today in Science. Intrigued by anecdotal reports that Prof. Everett and his wife Karen had presented about the matchlessness of Piraha life, Prof. Gordon conducted a number of experiments over a three-year period. He found that a group of male tribe members — women and children were not involved because of certain cultural taboos — could not perform the most elementary mathematical operations. When faced with a line of batteries and asked to duplicate the number they saw, the men could not get beyond two or three before starting to make mistakes. They had difficulty drawing straight lines to copy a number of lines they were presented with. They couldn't remember which of two boxes had more or less fish symbols on it, even when they were about to be rewarded for their knowledge. A significant part of the difficulty related to their number-impoverished vocabulary. Although they would say one word to indicate a single thing and another for two things, those words didn't necessarily mean one or two in any usual sense. 'It is more like ones and twos,' according to Gordon."
If the Whorfian theory is correct, or at least describes a quality which profoundly affects the way the world is perceived, other behavioural divisions between the various populations of Man must exist. (The ideas of a universal grammar and a universal language instinct are not necessarily incompatible with the idea that a particular language determines thought, for there could be a basic language template that is then altered by experience. Moreover, it is conceivable that natural selection creates subtle brain differences between populations to accommodate differences in language). To any Whorfian differences in populations may be added the vast differences in cultural expression, some of which could be laid at the door of linguistic determinism of thought.

Daniel Everett engaged in his research.

The political and social implications of the Glasgow research

Assuming the research is sound, the implications are profound. The Glasgow researchers conclude:
"Our results question the universality of human facial expressions of emotion, highlighting their true complexity, with critical consequences for crosscultural communication and globalization."
Just so. If human beings do not share a common understanding on such a basic level as the recognition of emotions, the scope for inter-racial friction is vast. It would mean that multi-racial populations must be forever conglomerations of racial groups estranged from one another to varying degrees. It would mean that racial wars will always remain a possibility and that the possibilities of such wars will be enhanced by the settlement of different races on the same territory. It could cause warfare between states dominated by different racial types if one or both see those of their own racial type being, in their view, mistreated by the other state.

It could be objected that the Glasgow research does not show that there is no shared facial recognition between Whites and East Asians. East Asians recognised Happiness and Sadness as efficiently as Whites, and even in the case of Fear, Surprise, Disgust, and Anger, they were correct two thirds of the time, (although it is telling that East Asians made fewer mistakes when presented with East Asian faces). The liberal searching for a light at the end of the racial difference tunnel would undoubtedly point to the fact that East Asians identified emotions in the same way as Whites most of the time and that this agreement between the races proved a common biological emotional template.

The problem with that argument is that identifying emotions wrongly a third of the time is not a small margin of error. It would be a severe handicap to any understanding between people of different races.

Liberals have a tendency to see things upside down.
It is not that the research shows that different races have nothing in common when it comes to recognising emotions from facial expressions, it is the degree of difference which is important. An analogy could be made with IQ. Every race has some of whatever it is that IQ tests measure, but the distribution of IQ varies according to race with the descending hierarchy being East Asians-Whites-Blacks. (In addition, the shape of IQ varies between races with, for example, on average, Whites scoring higher on verbal tests and East Asians on visual tests.)

These racial differences in IQ are extremely important at both the individual and group level because they affect the way individuals and nations perform. Low IQ equals poor life outcomes for individuals in any society, and societies where the average IQ is low are invariably poor.

Similarly, if substantial differences in the ability to recognise emotions in others exist, that may have substantial effects on how different races perform in both the organisation of societies in which they dominate and societies in which they are in a minority. The societies in which they dominate may need a structure which is inimical to intellectual and technological development beyond a certain point. Living as part of a minority, being unable to connect on an emotional level with the majority of the population, could be as much a life definer as a low IQ.

If similar racial differences exist in the ability to interpret language, body language, tones of voice, and so on the opportunities for racial misunderstanding will be multiplied and amplified.

The idea that people of different races have considerable difficulty from misunderstanding the intentions of other races is given credence by the strong propensity of human beings of the same race to live together when they have the choice and the universal racial suspicion found in racially mixed societies. In short, in the real world human beings behave just as one would expect them to behave if the findings of the Glasgow study are correct.

If the Glasgow study is replicated and more work is done demonstrating other innate behavioural differences between races, it would leave the present elite ideology of globalism in an intellectual mess. It would undermine utterly the liberal internationalist idea that human beings are all of a piece and may be readily placed in any society. That would not of course immediately cause the elites to throw up their hands and say we have been wrong, most grievously wrong, but over a generation or so the elite position could be changed by such academic research.

Innate racial behavioural differences are of course not the sole explanation for racial conflict – my other three favourite candidates are (i) the simple brute need to occupy a territory to gain physical security and enjoy its resources, (ii) the aesthetic sense which favours those who resemble the individual, and (iii) the sense and the sociological pressures which arise from the need of any social animal to maintain a viable group. Nonetheless, innate differences in behaviour must rank as a powerful driver of racial conflict.

Be cautious

The research needs to be treated with caution. As yet it has not been replicated and it is based on a very small sample. However, much research in the social and biological sciences uses similarly small samples which are treated as legitimate. Moreover, the nature of what was being tested in this research – the recognition of facial expressions and the controlled physical measurement of the mode of scanning faces – plausibly allowed for objective data to be extracted, while the judgements required of the subjects involved nothing that is obviously contentious, for they were simply being asked to interpret facial expressions and, consequently, questions of moral or political bias did not arise, as they often do in socio-biological research. But even if a participant had wanted to produce a desired outcome in their particular case, he or she could not have done so without the collusion of at least of one of the two participating racial groups.

Nonetheless, the small sample size is a problem because the racial groups are from a few societies, most notably in the case of the East Asians, with twelve coming from China and one from Japan. The research needs to be replicated, ideally with substantially larger numbers of subjects, and subjects should be drawn from a wide range of societies to test whether the differences are stable across cultures, for example, compare Japanese-Americans with Japanese natives or White Englishmen with White Italians.

There is also the objection that viewing still images in an artificial environment is entirely different from interpreting facial expressions when interacting with others in ordinary life. This is not strictly relevant to the question of whether different races adopt different scanning behaviour or have significant differences in their success in identifying emotions. The mental processes which allow identification of emotions will operate in the same fashion in any situation. Of course, in real-life situations there will be distractions not found under laboratory circumstances which may cause facial expressions to be missed completely or not properly heeded because of a lack of concentration. But that would say nothing directly about either the efficiency of recognition or the method of scanning faces.

All that real-life situations might show is that the White and East Asian methods of scanning faces and interpreting emotions is differently affected by the distractions of real life situations. For example, it could be that concentrating on the whole face requires more concentration than simply taking information from the eyes. But there still remains the problem I have already mentioned, namely, that in real-life situations human beings use multiple clues to judge the emotional state of someone else. The ability of different racial groups to perform using multiple behavioural clues could perhaps be tested by using film of people using the full range of behavioural clues, and asking research subjects to evaluate the emotional state of the person in the film.

Further research

It would be interesting to see the same tests applied to other racial groups. Because many racial genetic differences, such as IQ distribution and testosterone levels, place the three main human races in the order of Black-White-East Asian, I think it probable that Blacks would be more adept at facial expression recognition than Whites. This would plausibly fit in with their higher extroversion scores if it could be shown that ability at facial recognition is a potent trigger for emotional displays. Blacks are also probably better at interpreting other non-verbal behavioural cues.

In addition to replicating and expanding the Glasgow team’s research, there is ample room for related work such as studies of the interpretation of other non-verbal clues to emotional states, such as body language and voice elements, to see whether they also vary between racial groups.

Despite the slender nature of the evidence presently available, the Glasgow research has what might be called "the ring of plausibility." There is clear evidence that there are behavioural differences between races which appear to be innate – the variation in IQs between racial groups being the most famous – and many instances of objective physical biological differences, from the considerable external racial differences which anyone can see to the covert physiological differences such as sickle cell anemia in West Coast Africans, such as Nigerians. That beings who have evolved such differences might well have followed different evolutionary paths in the matter of perceiving emotions does not seem inherently far-fetched, because, provided a behaviour has a genetic base, it will be subject to natural selection.


1. "Cultural Confusions Show Facial Expressions are Not Universal," published in Current Biology which charges for its articles. A free copy of the draft report can be found here.
2. The research paper uses Western Caucasian for White.
3. Thirteen Western Caucasian (13 Europeans, 7 females) and thirteen East Asian (12 Chinese, 1 Japanese, 8 females) observers participated (mean age = 24 years 5 months; 23 years 2 months, respectively).
4. To conclude that the genes have no part to play in determining behaviour would imply that all behaviour is the product of mind and that mind is somehow divorced from the physical body and consequently not subject to natural selection mediated through the genes.

Originally published at Living In A Madhouse

No comments:

Post a Comment


by Hewitt E. Moore @hewittemoore Sen. Elizabeth Warren (D-Mass.), a potential 2020 presidential candidate, addressed her claims ...