Thy Bloody Awful Symmetry
As well as the whole Michael Hardt/ David Hasselhoff thing below, my mind was turned to thoughts of evolutionary psychology by an article in yesterday’s New York Times. Fundamentally, it’s exactly the sort of work I was planning on doing; somebody’s taking a look at the actual experimental methodology that supports such convenient factoids as “men are more concerned about sexual jealousy, while women worry more about emotional infidelity”. It turns out that this “result” is incredibly fragile as to the situation of the experiment; if you sit people down, ask them the question straight out, and give them time to think, then men and women assign themselves correctly to their gender roles, whereas if you catch them off guard in order to get a more “instinctive” response, the differentiation “predicted”by an amazingly tendentious just-so story about cavemen in Africa just doesn’t show up. (I’d note in passing that the EP crowd are often in the forefront of moaning about “double-blind trials” when they’re on the attack on some other point; the methodology of having an experimenter with an agenda ask a question face to face and then write the answer down himself is about as far from double blind as it gets).
In any case, the main point of the article linked above is to show what total and utter patronising knobheads evolutionary psychologists can be when pulled up on a point of science (read it, honestly, the guy starts comparing himself to Galileo!). But it dovetails quite nicely with a couple of points I’d like to make about some other sacred cows of evolutionary psychology; specifically, some of those claims which the pop science gang like to make about the “genetic” foundations of human beauty.
It’s a shame that I’m too mean to cough up for the version of this weblog which would allow me to put up pictures, but there you go … but you don’t have to search far on the web to find someone claiming it to be an established “fact” that facial attractiveness is a function of facial symmetry. Coincidentally, you also don’t have to go far on the web to find a picture of Elvis Presley (bloody great asymmetric sneer) or Cindy Crawford (bloody great asymmetric mole on face). So what gives?
Apparently people with symmetric bodies have “good genes”. Don’t ask me, I’m a stranger here myself. But let’s assume for the meantime that in some way, a little glitch in the building of the face of a foetus is evidence of a deep-seated horrible lurgey in the genes which is just waiting to show up as sickle-cell anaemia or low resistance to malaria or something. The question I’m interested in is, how did anyone find out that people with symmetrical faces are the most beautiful people of all?
Note at this stage, that I’m not interested in studies which claim to have shown that symmetrical people have more sex than anyone else. Randy Thornhill claims that this is the case, and it might be the case even though the experiments which claim to demonstrate it come from the same guy who brought you a theory of rape which doesn’t work at all as a theory of sexual assault not involving penetration. Personally, I think that Thornhill is all over the place, and I’ll explain why in future (there’s a clue in this sentence for the impatient), but I want to establish that it doesn’t effect my current argument if the symmetrical are shagging wild all over the place. The claim that “beauty” is “whatever gets you laid” is one that the EP crowd is committed to, not me. But this is by the by.
Absent the sex life studies, the evidence for “beauty” being this, that, or the other, has to come from what actual people judge to be beautiful. So, the best method for carrying out this experiment would have to be to get a bunch of people, show them a bunch of photographs of people, and get them to pick out the beautiful ones. Then you count the number of points each photograph gets and have a look at which ones are picked the most often, right?
If you ask people to pick out the photographs from a set which strike them as the most beautiful, you’re actually asking them to perform cognitive acts, not one. You’re asking your experimental subjects to:
a) notice a picture of a face
b) judge whether it’s beautiful or not.
The first of these is not a trivial act, as anyone who’s observed a baby younger than about two months will testify. The extent to which you’re going to carry out the act of picking a picture for the beautiful pile depends on the extent to which it catches your attention as well as what you actually think of the face. There will be an error in your results from people “misclassifying” faces because they weren’t really paying attention to them. There are all sorts of misjudgements that it’s possible to make when looking at a two-dimensional image of a three-dimensional object; as the post below demonstrates, I quite seriously misclassified a picture of Michael Hardt’s hairstyle as “bouffant” when it wasn’t.
So far so good. Now, readers with extremely advanced degrees in econometrics won’t be asking … what do we know about this error? Importantly, is it unbiased — can we assume for modelling purposes that it can be ignored as something that will in a large enough sample?
I’m arguing, no. One of the things that, broadly construed, evolutionary psychology has usefully done for us is to dig up some important insights into the neuropsychology of visual perception. Particularly, it’s been noted (as in, anatomically observed) that there is a mechanism in the brain which is specifically adapted for distinguishing between symmetrical things and non-symmetrical things. I find the “evolutionary psychology” (in actual fact, ethology, the rather more serious parent discipline which looks at behaviour without making tendentious and unsupported claims) argument quite convincing in this regard. The reason we have a symmetry-detector is that very few things in nature are symmetrical except animals, and animals are only symmetric when they’re looking straight at you. Since the fact that something is looking at you is almost always a useful thing to know, we have been provided with a very acute sense of whether a thing is exactly symmetrical or not. Symmetry is a property which “jumps out of the page”.
So, given that photographs of symmetrical faces are more likely to be noticed, the errors are not going to be evenly distributed. In any study which is asking you to pick out a “noticeable” characteristic the symmetrical pictures are always going to be over-represented, because symmetry is a noticeable property. Furthermore, this property is highly likely to account for the fact that babies tend to look longer at the same photos which adults pick out of a pile as being most attractive, another factoid often advanced as evidence for the beauty=symmetry hypothesis.
I have no particular investment in believing that there is nothing aesthetically attractive about symmetry; I spend a lot of time with a sneer on my face, but that’s mainly because I read a lot of right-wing weblogs. But the fact that nobody saw fit to inquire into this possible source of experimental failure tends to suggest to me that people want to believe in the “evolutionary” arguments for reasons other than those of pure science. And when you get people like Todd Shackelford responding to the Northeastern study by just saying “”I guess, to state it plainly, I think the paper is in large part ludicrous .. It’s clear to me that they have an agenda they’re pushing.”, I think I’m on to something.