A Thousand Faces
Stepping back from my tendency to relabel the results of research into human neuropsychology as “magic,” I wanted to talk a little bit today about the methodology behind some of the weird shit that interests me.
In the interest of full disclosure, I’m a neuropsych nerd, but by the standards of academic neuroscientists, I’m a dilettante. My approach to cognitive neuroscience is, like my approach to just about everything in which I’m interested, idiosyncratic and based around a set of premises that I don’t think most people start from, and goals not everyone shares. I’m interested in things like philosophy, psychology, and linguistics because they’re like lockpicks for the human skull, and while historically I’ve focused on opening up other people’s heads to see what’s inside (and occasionally to rewire their brains, mostly for their benefit), I’ve been popping the hood of the Rollins machine and tinkering with my own internal workings much more frequently of late.
I got diagnosed with autism spectrum disorder midway through the process of deprescribing from psych drugs. On reflection, I think this was God finally acknowledging my revealed preference for doing literally everything the hard way, shrugging his shoulders, and saying “Okay, Rollins, if you insist.”
My therapist had gently explained that autistic people always have trouble with empathy, because we don’t read faces well. My response to that, in keeping with my general attitude toward people who say I can’t do something, was to nod politely and chuckle internally. I find it’s much more satisfying to outperform people who say shit like that and then be really polite to them afterward than it is to laugh in their faces.
I took a month to agonize about the “everything you know is a lie” aspect of the diagnosis, got over myself, and took a random walk down every portal to neuropsychology research that looked promising.
I had first encountered the Facial Action Coding System years ago, while reading about the history of cel animation. It was invented by Paul Ekman, a giant of twentieth century psychological research. I glanced at a photo of him and noted that he appears to have had botox injections (which selectively paralyze facial muscles, making face-reading impossible) in recent years. I took that as a sign that he had some confidence in his system (and that he may be a little squirrely on a personal level).
The top-of-the-line Ekman International course on face-reading costs $300. It’s not subtle, but it works. It consists of photo animations which display first a neutral face, and then, for a split second, the same actor making one of Ekman’s seven basic expressions (anger, disgust, happiness, contempt, sadness, surprise, and fear) before returning to neutral. There are practice tests and timed tests, and the whole program is supposed to take about two weeks. I aced the tests after about three days.
Then, as I am wont to do the moment someone teaches me something useful, I started trying to figure out how to creatively misapply it.
I am not up on the latest research in the field, so this may not be a new insight, but I think you can fairly break down people’s faces into a slightly more refined typology than Ekman’s baseline. I think you can go beyond just the expressions, down to what I think of as “gesture sets.” What I mean by a gesture set is that expressions aren’t generic. Rather, they’re typical of particular groups, and those groups are themselves incredibly modular. The “contempt” face of a patrician Republican and a patrician Democrat are slightly different, for example. The “happy” face of a woman who has just gotten out of a domestic violence situation is characteristically different from the “happy” face of the same woman two years down the line, when she’s been doing therapy and is in a stable relationship in a new city, but it’s noticeably similar to other women in the same situation.
There are a couple of other insights that flow from this. One is that communities have a particular gesture set. At this point, I can tell with a fair degree of accuracy whether a person lives in a city or a rural area, and if a city, whether it’s a major one or a small city. The faces of people in cities are harder, and the bigger the city, the harder the face. Population density is only one factor; Europeans have different gesture sets than Americans, and I can reliably distinguish between a similarly dressed native of an Asian country and an Asian-American on sight by their gesture set.
The practical implications of this observation are staggering. For example, here’s a technique I discovered based on another modularization of gesture sets and a seemingly unrelated concept, assortive mating (the observation that people find romantic partners who are like themselves). If I were using the magic metaphor, I’d call this technique divination using Tinder as the oracle: Figure out who you match with on dating apps, which have an assortment of photographs of the same person, and reverse-engineer the face-reading technique. Once you know who likes you, analyze the gesture sets in their pictures, and assuming the gesture sets of your matches correspond to your own, you can track how you’re progressing as a human being over time by watching how the kind of women who swipe right on you change. It’s not a perfect system; there are catfishes and bots, and even sincere people often manipulate their photographs in order to seem more appealing. But it works well enough for use “in the field.”
How about social languages? Once you realize that the geography of a human settlement has a direct correlation with the characteristic facial expressions of the inhabitants, you can figure out what the social language characteristic of that settlement is, and effectively communicate with people from there. Different professions have different gesture sets, too; tradesmen don’t have the same gesture set as UPS drivers, but the two gesture sets have more in common than the gesture set of a college professor. That is the way to determine which social language a person speaks, even when they aren’t wearing their work clothes. And understanding gesture sets is the practical component that makes my theory of social dynamics work.
None of this is rocket science, and all of it is available to everyone. It’s ways of thinking that aren’t necessarily familiar to everyone, but we can all do them. Maybe if more of us did, it’d be a world in which we communicated with one another more effectively.
A little housekeeping: Henceforth, posts in The God of Death and Second Chances, my serial Modern Noir novel, will go up on Wednesdays. I’d like to spend a little more time editing those rather than posting them and tweaking them once they’re already on the site, because I don’t want people who have opted in to have to read what amounts to a first draft (remember, you have to select it in your settings if you want TGoDaSC posts sent to your inbox).
(I will, of course, continue putting up half-baked brain droppings and altering them on the fly in The Wonderland Rules.)
-Jay
> My therapist had gently explained that autistic people always have trouble with empathy, because we don’t read faces well.
My therapist phrase it like this: it is definitely possible to "read" facial emotions in high fidelity, but it is hard for ASD, compared to the general population, to infer the cause of such a gesture without previous direct experience. Alternatively, it can be seen as inference (horizontal generalized knowledge), rather than abstraction (vertical pattern overfitting), that these therapists are evaluating. They want a person to have an "emotional cloud" that has a sliding gradient between emotions, rather than "dead" specific classification as certain isolated archetypes.
https://s3-us-west-1.amazonaws.com/vocs/map.html#modal https://www.ocf.berkeley.edu/~acowen/music.html#modal
That's an interesting insight. I have noticed that I tend to get along with people more easily when they are from places similar to me (sparsely populated rural areas and smallish towns.) I had been chalking that up largely to just culture, but hadn't considered that being able to more easily read each other, or send fewer misleading signals, through facial expressions was part of that. Some people seem to have a really hard time picking up on whether or not I am joking, but I couldn't figure why other than some people don't have a close enough sense of humor while others do. The idea that they can't quite read the facial expression of "deadpan humor" I put out while others recognize it easily explains a bit there.
It makes sense, too, as one often reads of how people in various areas tend to be e.g. extremely grumpy seeming to outsiders but insiders don't see it, and this seems to work a little differently than the usual "city people mean, country people nice" dichotomy that seems rooted in other issues.