« Reading List: Europa's Lost Expedition | Main | Reading List: Three Laws Lethal »

Tuesday, December 17, 2019

Reading List: How to Judge People by What they Look Like

Dutton, Edward. How to Judge People by What they Look Like. Oulu, Finland: Thomas Edward Press, 2018. ISBN 978-1-9770-6797-5.
In The Picture of Dorian Gray, Oscar Wilde wrote,

People say sometimes that Beauty is only superficial. That may be so. But at least it is not as superficial as Thought. To me, Beauty is the wonder of wonders. It is only shallow people who do not judge by appearances.

From childhood, however, we have been exhorted not to judge people by their appearances. In Skin in the Game (August 2019), Nassim Nicholas Taleb advises choosing the surgeon who “doesn't look like a surgeon” because their success is more likely due to competence than first impressions.

Despite this, physiognomy, assessing a person's characteristics from their appearance, is as natural to humans as breathing, and has been an instinctual part of human behaviour as old as our species. Thinkers and writers from Aristotle through the great novelists of the 19th century believed that an individual's character was reflected in, and could be inferred from their appearance, and crafted and described their characters accordingly. Jules Verne would often spend a paragraph describing the appearance of his characters and what that implied for their behaviour.

Is physiognomy all nonsense, a pseudoscience like phrenology, which purported to predict mental characteristics by measuring bumps on the skull which were claimed indicate the development of “cerebral organs” with specific functions? Or, is there something to it, after all? Humans are a social species and, as such, have evolved to be exquisitely sensitive to signals sent by others of their kind, conveyed through subtle means such as a tone of voice, facial expression, or posture. Might we also be able to perceive and interpret messages which indicate properties such as honesty, intelligence, courage, impulsiveness, criminality, diligence, and more? Such an ability, if possible, would be advantageous to individuals in interacting with others and, contributing to success in reproducing and raising offspring, would be selected for by evolution.

In this short book (or long essay—the text is just 85 pages), the author examines the evidence and concludes that there are legitimate correlations between appearance and behaviour, and that human instincts are picking up genuine signals which are useful in interacting with others. This seems perfectly plausible: the development of the human body and face are controlled by the genetic inheritance of the individual and modulated through the effects of hormones, and it is well-established that both genetics and hormones are correlated with a variety of behavioural traits.

Let's consider a reasonably straightforward example. A study published in 2008 found a statistically significant correlation between the width of the face (cheekbone to cheekbone distance compared to brow to upper lip) and aggressiveness (measured by the number of penalty minutes received) among a sample of 90 ice hockey players. Now, a wide face is also known to correlate with a high testosterone level in males, and testosterone correlates with aggressiveness and selfishness. So, it shouldn't be surprising to find the wide face morphology correlated with the consequences of high-testosterone behaviour.

In fact, testosterone and other hormone levels play a substantial part in many of the correlations between appearance and behaviour discussed by the author. Many people believe they can identify, with reasonable reliability, homosexuals just from their appearance: the term “gaydar” has come into use for this ability. In 2017, researchers trained an artificial intelligence program with a set of photographs of individuals with known sexual orientations and then tested the program on a set of more than 35,000 images. The program correctly identified the sexual orientation of men 81% of the time and women with 74% accuracy.

Of course, appearance goes well beyond factors which are inherited or determined by hormones. Tattoos, body piercings, and other irreversible modifications of appearance correlate with low time preference, which correlates with low intelligence and the other characteristics of r-selected lifestyle. Choices of clothing indicate an individual's self-identification, although fashion trends change rapidly and differ from region to region, so misinterpretation is a risk.

The author surveys a wide variety of characteristics including fat/thin body type, musculature, skin and hair, height, face shape, breast size in women, baldness and beards in men, eye spacing, tattoos, hair colour, facial symmetry, handedness, and finger length ratio, and presents citations to research, most published recently, supporting correlations between these aspects of appearance and behaviour. He cautions that while people may be good at sensing and interpreting these subtle signals among members of their own race, there are substantial and consistent differences between the races, and no inferences can be drawn from them, nor are members of one race generally able to read the signals from members of another.

One gets the sense (although less strongly) that this is another field where advances in genetics and data science are piling up a mass of evidence which will roll over the stubborn defenders of the “blank slate” like a truth tsunami. And again, this is an area where people's instincts, honed by millennia of evolution, are still relied upon despite the scorn of “experts”. (So afraid were the authors of the Wikipedia page on physiognomy [retrieved 2019-12-16] of the “computer gaydar” paper mentioned above that they declined to cite the peer reviewed paper in the Journal of Personality and Social Psychology but instead linked to a BBC News piece which dismissed it as “dangerous” and “junk science”. Go on whistling, folks, as the wave draws near and begins to crest….)

Is the case for physiognomy definitively made? I think not, and as I suspect the author would agree, there are many aspects of appearance and a multitude of personality traits, some of which may be significantly correlated and others not at all. Still, there is evidence for some linkage, and it appears to be growing as more work in the area (which is perilous to the careers of those who dare investigate it) accumulates. The scientific evidence, summarised here, seems to be, as so often happens, confirming the instincts honed over hundreds of generations by the inexorable process of evolution: you can form some conclusions just by observing people, and this information is useful in the competition which is life on Earth. Meanwhile, when choosing programmers for a project team, the one who shows up whose eyebrows almost meet their hairline, sporting a plastic baseball cap worn backward with the adjustment strap on the smallest peg, with a scraggly soybeard, pierced nose, and visible tattoos isn't likely to be my pick. She's probably a WordPress developer.

Posted at December 17, 2019 00:22