Image from: Baron-Cohen, The Essential Difference, 2004, p.198

Reading the Mind in Claudia Schiffer’s Eyes — The Greatest Rubbish in the History of Psychology

Danton: What do I know! We know little about each other. We are pachyderms, we reach out to each other, but it is a futile effort, we only rub the coarse leather against each other – we are very lonely.
Julie: You know me, Danton.
Danton: Yes, what one calls to know. You have dark eyes and curly hair and a fine complexion and you always say to me: dear George! But (he points to her forehead and eyes) there, there, what lies behind that? Go, we have rough senses. Know each other? We would have to crack open the tops of our skulls and pull the thoughts out of each other’s brain fibers.
“There are and can be only two ways of searching into and discovering truth. The one flies from the senses and particulars to the most general axioms, and from these principles, the truth of which it takes for settled and immovable, proceeds to judgment and to the discovery of middle axioms. And this way is now in fashion. The other drives axioms from the senses and particulars, rising by a gradual and unbroken ascent, so that it arrives at the most general axioms last of all. This is the true way, but as yet untried.”
— Francis Bacon, Novum Organum

It is one of the specific properties of rubbish that it often, maybe even more often than non-rubbish, has the ability to become common knowledge. It is very usual that people who arrive in my acting classes come with the “common knowledge” that acting is for sure about making specific facial expressions. They then make all sorts of grimaces that are supposed to “mean” something as if they were words, “mean” joy, “mean” sadness and “mean” anger. That’s interesting, because if you look at movies, no halfways good actor actually does that. So they would have had the chance to know that this is not what acting is about. The fascinating thing is that they didn’t manage to gain that insight from the many movies they for sure have watched, instead they persisted with strange rubbish. It seems to be a common human feature that the cliché often outbeats the observation. It blinds them and part of their acting education must be to unlearn the blinding rubbish in order to learn to see. 

The skill of seeing humans will be a topic of this Organum again and again. To give it a proper launch I will leave my little field of acting and start with an excursion to psychological diagnostics. Of course rubbish becomes visible whatever field you look at. But there are differences in the potential of that rubbish to do harm depending on the influence the field has. Art academies are stuck with rubbish, but their impact on society is close to zero, so who actually cares. Psychological diagnostics however have a degree of  influence on the lives of individuals and, on a larger scale, on what our society considers as normal and as healthy, that its potential to do harm, were it based on rubbish, would be immense. 

Have a look at this pair of eyes. Please choose which word describes best what the person in the picture is thinking or feeling. Is the person a) aghast b) baffled c) distrustful  or d) terrified? Write down your answer before you continue reading. 

The  image is image 34 of 36 images in the “Reading the Mind in the Eyes Test”, a test developed by a team around clinical psychologist Simon Baron-Cohen. Sir Simon Philip Baron-Cohen FBA FBPsS FMedSci is a British professor of developmental psychopathology at the University of Cambridge. He is the director of the university’s Autism Research Centre and a Fellow of Trinity College. The “Reading the Mind in the Eyes Test” is used as a measure for adult “mentalizing” which is commonly described as the capacity of humans to assume mental states to other people and which is also sometimes called to have a “theory of mind”. The test inversely correlates with the Autism Spectrum Quotient (AQ), a measure of autistic traits in adults of normal intelligence. The “Reading the Mind in the Eyes Test” is described as having the “power to detect subtle individual differences in social sensitivity” and it is sometimes also described as measuring empathy. It is widely used, in research and in diagnostics, for example as part of the procedure to diagnose autism for adults, and in an adapted version also for children. 

So, what is the correct answer and how good are your social sensitivity and empathy? The correct answer for this pair of eyes is “distrustful”. So, if you chose that answer, you can take it as a first hint to have good social sensitivity and empathy. You can. If you are willing to trust psychological tests and to hand over the estimation of who you are and what you can and can’t do to psychologists. Here is the full image of the pair of eyes. It is Claudia Schiffer on the cover of Penthouse Magazine December 1993.

I don’t think that there is much of a necessity to prove that it is the same image. A simple cut out shows enough similarity to make the question inevitable, how two pairs of eyes that similar can be “distrustful” and part of such a smiling face at the same time. 

When I recognized Claudia Schiffer on image 34, I wondered how these psychologists would have had a photo of her on which she would look distrustful. I considered that as not too probable, maybe some Paparazzo photo, but not a photo that obviously had studio quality. I did an image search, I found the image. Actually Claudia Schiffer used to look more or less like that on every third image, maybe because she looks better from half-side than from in front (concavity of her nose etc). Although I insist that it is of no relevance if the image is actually the same or simply super similar, I rebuilt the test image from the image on the cover of the magazine. Please replicate yourself, here is what to do:

  • turn the image counterclockwise a few degrees (in the original the axis between the pupils is horizontal, in the test image it is slightly rising)
  • make the equivalent cut out of the eyes region
  • turn the image black and white
  • increase contrast
  • make the pupils black
  • make the radius of the left iris bigger (if you look closely at the image you see that the proximal side of the left iris is not round), make everything right from the right pupil to the outer corner of the right eye black)
  • add light reflection in the left lower quadrants of the irises
  • find a solution for the hair on the right and for a sharper edge between the face and the background on the left 

Although this might at first glance seem quite a lot of change in the image, it is, concerning the face, actually only 

  • cosmetics (change of colour without changing spatial relations, but with quite a loss of information in the distal corner of the right eye)
  • a slight rotation of the whole image (so also no change of within-image relations)
  • one (1!) clear change of a spatial relation within the image (the enlarging of the left iris and therefore a smaller distance of the left iris to the proximal corner of the left eye). All other spatial relations remain the same.

Even if it were two different images, we would have to understand  how these of all differences would carry out the quite extraordinary shift from “friendly” to “distrustful”. Are the measures of the “from friendly to distrustful scale” in cosmetics? Or in the deviation of the line between the pupils from the horizontal line? Or in the degree of how much the eyes appear to be turned to the side in relation to the skull? And what role does the location of the light reflection play and how is this supposed to be shaped by an inner mental state? 

I then had a look at how the test was designed. Surprise, surprise, they took images from magazines! And here it comes:

They assumed 4 different mental states for each image, one of them assumed as right, the others assumed as wrong. Then these images with the four attributions were shown to 8 people. If 5 of the 8 people assumed as well that the one mental state that was chosen rightly described the mental state of the image, the image became part of the test, if less people assumed so, the image didn’t make it into the test. So, to make that clear: The whole test is built on nothing but assumptions. They determined  the right answer for a test that would determine self-images and therefore lives by what a small majority of self-entitled people felt to be right. No one ever tested if the “mental states” were actually there.

This looks like rubbish on every level I can think of. Here’s a quick list, it’s not sorted by relevance.

  1. It is against everything we know about how our mind grasps for environmental data. The untrained common brain does not use tweezers. No mind except maybe a super trained one can isolate a less than a second long expression of only the eyes from the overall multiparameter all-time-moving image of a person embedded in a situation and an environment one also can’t manage to have no knowledge of, including the layers of knowledge one has of oneself in that situation. You might be able to see it, but you cannot isolate it. And also: Why would you?
  2. Such a pair of eyes can be part of many, very very different images which all will lead to different assumptions of what is going on in the person.  This is basic knowledge of the director, is called the Kuleshov Effect, named after Lev Kuleshov, a Russian filmmaker in the early twentieth century. And here we have Alfred Hitchcock talking about it. Maybe add Kuleshov and Hitchcock to the psychology study books?
  3. It might also be worth asking if it isn’t observable with the plain eye that no one actually does read people like that. The advice for the test is to answer superquickly. But even the quickest verbal categorization is, in the temporal categories of human reactions, actually among the slow ones. If it were necessary part of everyday human-to-human behaviour, wouldn’t it produce a different image of the behaving human in time than what we actually see, some sort of stop-and-go behaviour? Don’t humans, if they react to others, do something much less specific and on a far too low cognitive level to differentiate between terms as specific as “aghast”, “baffled”, “distrustful” and “terrified”?  Dogs also get along socially quite well, they are even able to play which is really complex social behaviour. That is why I would be strongly hesitant to even take into consideration that for human-to human social sensitivity you would have to know in words what is going on in the other. I as somebody whose work consists of observing humans moving around all the time, like they do in real life, can’t resist to think that it might be no coincidence that it was psychologists who thought that up: It might be possible to think that this is what happens, when you are in the super artificial severely expression-reduced situation of a one-to-one encounter in sitting. But that’s not at all what the majority of our social situations, and especially those in which right reaction is crucial, are made of.
  4. What is a “capability to assume” supposed to mean on the very basic level of language? Everybody can assume everything. There is no measurable capability in that, a verification of the assumption has to be part of the verification of the capability, otherwise the assumption contains no capability.
  5. Is there any evidence for a universal expression of “distrustfulness” in human eyes? I would really like to see that data, even the most basic emotional expressions stand on weak grounds concerning their universality. Paul Ekman made some honourable observations, but you can’t say it’s universal because you see it everywhere. You might simply not see all those who don’t do it.
  6. And does “distrustfulness” contain all degrees of distrustfulness from distrusting the partner to buy the right sort of spaghetti for dinner to distrusting the man with a gun at night in a lonely street who tells you he only wants your money while all the time staring at your boobs? Is that the same expression? So no matter what situation might come to the test participant’s mind, it would be the same eye expression?
  7. Variability of eye expression is far higher from individual to individual than for one individual in different situations. What we mainly see if we look at the different images from the test is the high variability of individual phenotypes. Where within that high variability would be a place for 4 x 36 = 144 distinct eye expressions that would remain universal across individuals as much as that they could be used as a target of measurement in social evaluation of the other? (This is basic logic: The test has 4 answering options for each of the 36 images. It is necessary for the test that the three that are not “the right one” also have specific expressions, because if they had no specific expression they could also express in the same way as “the right one”, would therefore not be distinguishable from it, so there would be no “right one”. Each one of them has to be at least “not anyone of the other ones”, which, in the case of 143 other ones equals specificity.)  And which aspects of the image would carry that info? 144 can, if broken down in the most possible effective way be split up into  2⁴ ∗ 3², so it would have to be at least 4 items with an “either-or” and 2 items with three options. This is the mathematically necessary minimum for 144 distinct expressions. 4 items with an “either-or” and 2 items with three options doesn’t sound like too much, but now try to find even one thing that remains specific when everything that can be attributed to individual phenotype has been excluded.  (For these considerations it doesn’t make any difference if the information is read consciously or unconsciously. Even if read unconsciously it has to be somewhere in the image. The unconscious can’t for sure not read non-info.)
  8. How would you know that a certain expression of the eyes is a specific expression if you don’t know how the person looks when neutral? Humans have all sorts of eye expression and it is simply their mental default mode. Ugly people can still feel happy and an attractive one could look seductive to you even if her mental state is currently dominated by her desire to fart. If psychologists are not aware of that, they should maybe seat themselves in subways instead of universities for a year or two.
  9. A situation where a human can only see the eyes of a person, without movement, without any other (ongoingly fed in) knowledge about the situation and the environment, including the knowledge of the observer about himself, does not exist and has never existed. So why should this ability have evolved after all?
  10. We humans are all behaviour and movement. We have ongoing eye expression all of the time, because we use our eyes all of the time, mainly for seeing things, not for “expressing”. So, how long does such a behavioural unit have to last in order to be selected as worthy to correspond to a mental “state” while so many others are not? By what criteria is this one to be selected and not the many others in the seconds before and after it? Some mean something and the others mean nothing? What is the theory here?
  11. Is it a 21st century-worthy way of thinking about the human brain as that it contains “distrustfulness”? How would that look in the brain and where would that be located?
  12. It doesn’t even make sense on the level of Baron-Cohen’s own theories. Why should the ASD people, the ones who are supposed to be better in seeing patterns and details (so actually the ones who should be better able to use the perceptual tweezers), do worse in a test where the only perceivable differences lie in spatial relations, in distances and angles? There is no social talent required in that, it’s just learning to distinguish 144 different things, who would not be able to do that?

I have, during my lifetime, developed a certain respect for brashness. Brashness puts the world to the test and delivers us important data about the state of humankind. Baron-Cohen and his team have actually put their whole field at a test. They have carried out a “Reading the Mind in the Eyes Test”-test, a long term and large scale test within the field of psychology. This test measured the quality of procedures of the field, both in research and in diagnostics. So, let’s have a look at some numbers: Google scholar shows 3 million search results for “Reading the Mind in the Eyes Test”. There is quite some research built on it, many, many, many people who ought to be expected to be able to recognize at least one of the flaws of the test, had the test in their hands. How could it happen that they all didn’t notice?

I have found no data on diagnostics, so I remain with questions: How many psychologists have administered this test how often? By what curricula were they educated? Is there no such thing as a code of ethics that obliges psychologists to look at a test design before they use the test? How many people, in passing this test, were told that they have a deficit? Of how many autism diagnoses was this test part? How many people were, in passing this test, told that they are good at something that doesn’t exist? How many children are affected? How has this influenced people’s mental health and life outcome? I would consider it as an action of non-integrity if the field of psychology would now claim that no harm was done. It is their own claim that diagnoses have impact on lives, not mine.

The “Reading the Mind in the Eyes Test” and all research that has been built on it must be withdrawn, using it in diagnostic procedures must be stopped, diagnoses based on it must be repealed, an investigation of the damage it has caused must be made. Apologies would be at place. One might also allow oneself to think that where this can happen, there could be more that goes wrong. There was a replication crisis, maybe we should add a diagnostic crisis? Maybe good to have it now, thick books are heavy, and it might be good to go through that before DSM will contain a special diagnosis for every single individual on the planet. 

Meanwhile, if somebody comes up with the idea that you need a psychological diagnosis, you’d maybe better run. If you don’t know where to run, maybe run to your TV and watch a movie. Theatre people have known-in-doing for two thousand years that we can only know of other humans by grasping for multiple parameters unfolding in ongoing action along the axis of time. That is the basic implicit wisdom of the dramatic arts. There have been great psychologists long before the term was invented and our great playwrights have understood more about human behaviour and put it into much better functioning vehicles than some of those who carry  a “psychologist” academic title today. 

In order to keep it light I close with a fun fact: Did you know that the other famous Baron-Cohen, the comedian Sacha also occupied himself with facial expressions? He tried to specify the pre 9/11 and post 9/11 facial expressions. I personally find the approaches of both Baron-Cohens equally scientific.