Cordelia Fine, a professor of the history and philosophy of science at the University of Melbourne, has carved out a niche for herself by attacking the notion that there are any evolved and genetically-based differences between males and females. Her books have been best-sellers (Testosterone Rex won the Royal Society book prize), probably because her conclusions appeal to those of a certain ideology. But those conclusions are flawed (see here, for instance). Fine’s critiques of some studies purporting to show sex differences are often good, but they’re combined with misguided characterizations of other work as well as the ignoring of results that go against her men-and-women-are-pretty-much-the-same thesis. In other words, Fine is tendentious, not objective, and her claims must always be taken with a grain of salt.

This is all on view in her op-ed with Daphna Joel (a professor of psychology and neuroscience at Tel Aviv University) in Monday’s New York Times (click on screenshot below).

If you accept that, on average, males and females differ in how they behave, what they think, and how they feel, then there are only two reasons for this: there are evolved differences, or those differences come about when “no difference” brains are differentially conditioned by society. It is, however, ridiculous to deny that there are evolutionary differences between the sexes to at least some of human behaviors, including sexual behavior. And any differences, be they cultural, genetic, or a combination of both factors, must be instantiated by a difference in brain structure, even if we don’t yet have the tools to see those differences. After all, behaviors and preferences come from the brain. So what are Joel and Fine talking about?

It turns out that their article is slippery in two ways. First, it conflates average differences between the sexes in behavior, emotions, and mentation with whether an individual can be diagnosed as male or female. So while Joel and Fine admit (grudgingly) that there are differences between men and women in both brain structure and behavior, they harp relentlessly on whether a single person can, from inspecting that individual’s behaviors and brain, be put neatly into the “male” or “female” class. Actually (see below), we’re already close to that.

But that’s a bogus problem, for the general claim about male/female differences rests on averages, not whether an individual can be diagnosed with 100% accuracy. Read this passage of their op-ed for some weasel-words, for instance:

The key point here is that although there are sex differences in brain and behavior [JAC: note the admission], when you move away from group-level differences in single features and focus at the level of the individual brain or person, you find that the differences, regardless of their origins, usually “mix up” rather than “add up.” (The reason for this mixing-up of characteristics is that the genetic and hormonal effects of sex on brain and behavior depend on, and interact with, many other factors.) This yields many types of brain and behavior, which neither fall into a “male” and a “female” type, nor line up tidily along a male-female continuum. Even when you home in on only two psychological characteristics, people don’t fall in line on a continuum from, say, extreme systemizer or “things-oriented” — supposedly the “male” pole — to extreme empathizer or “people-oriented”— the “female” pole. Rather, as recent studies have shown, people’s self-reported tendency to empathize tells you almost nothing about their self-reported tendency to systemize, and people may be highly oriented toward both things and people, to mainly one of these, or to neither.

The notion of fundamentally female and male brains or natures is a misconception. Brains and behavior are the product of the combined, continuous interactions of innumerable causal influences, that include, but go well beyond, sex-linked factors.

Note the word “fundamentally” here, which is a strawman claim that few people would embrace: the claim that every person’s brains and behaviors slot them neatly and diagnostically into either a “male versus female” binary. The questions that most researchers ask, however, is whether there are average differences in behavior and brains, and, if so, how large are those differences.

We all know that there are average height differences between men and women, with men being about 7-9% taller than women in nearly every country in the world, yet you can’t tell from the height of a single individual whether it was male or female. I’m a short male (5 feet 8 inches), and there are plenty of women taller than I.  I could claim, as do Joel and Fine, that “the notion of fundamentally female and male heights is a misconception,” and I’d be right. But that would be missing the real difference, which is hugely significant and, of course, raises scientific questions: why is there that difference? Is it the result of natural selection, and, if so, what kind? And of course if you combine height with other traits, like genitals and chromosomes, you get close to 100% diagnosability.

Likewise with Joel and Fine. By conflating average differences—which could be substantial, and important in explaining, say, male versus female preferences and differences in sexual behavior—with diagnosability of single individuals, they are somehow conveying the message that there aren’t differences between men and women’s brains and behavior. They are blank slate-ists, and they know what they’re doing. But they’re doing it for the wrong reasons: their motivation seems to be that the admission of some differences between men and women’s brains and behaviors will somehow justify sexism. This becomes clear at the end when they describe their social program (my emphasis in the passage below):

The claim that science tells us that the possibility of greater merging of gender roles is unlikely because of “natural” differences between the sexes, focuses on average sex differences in the population — often in combination with the implicit assumption that whatever we think men are “more” of, is what is most valuable for male-dominated roles. (Why else would organizations offer confidence workshops for women, rather than modesty training for men?) But the world is inhabited by individuals whose unique mosaics of characteristics can’t be predicted on the basis of their sex. So let’s keep working on overcoming gender stereotypes, bias, discrimination, and structural barriers before concluding that sex, despite being a poor guide to our brains and psychological characteristics, is a strong determinant of social structure.

But some day we may be able to tell one’s sex with substantial accuracy by looking at one’s brain, either in vivo or in vitro. And who could argue that of course we should try to overcome gender stereotypes, bias, discrimination, and so on? Who doesn’t want equal opportunity for people of different genders and ethnicities? But Fine and Joel seem to be telling us as well to simply stop looking for average differences.  

Joel and Fine’s tendentious piece reminds me of those people who deny genetic differences between ethnic groups because there are not single diagnostic differences that can tell you your ancestry. But their are small differences among many genes, and taking them all together you can discern someone’s genetic background with remarkable accuracy. Yes, you can’t diagnose someone’s ethnicity from one or two traits or genes, but you can do so with groups of genes. And, I think, once we have a better handle on brain structure, and can combine different aspects of brain function and morphology, we’ll be able to do that with brains. It’s interesting that the motivation for the genetic blank slaters is the same as that of people like Joel and Fine: they think that if we see differences, especially genetic ones, it will somehow justify racism and sexism. As I’ve said before, it needn’t do that, for we should not base moral equality on biology.

My second plaint is about the science they cite. Now I haven’t checked all their scientific claims in this article, but I did check one. It’s this one:

In 2015, one of us, Daphna Joel, led an analysis of four large data sets of brain scans, and found that the sex differences you see overall between men’s and women’s brains aren’t neatly and consistently seen in individual brains. In other words, humans generally don’t have brains with mostly or exclusively “female-typical” features or “male-typical” features. Instead, what’s most common in both females and males are brains with “mosaics” of features, some of them more common in males and some more common in females.

Daphna Joel and colleagues then applied the same kind of analysis to large data sets of psychological variables, to ask: Do sex differences in personality characteristics, attitudes, preferences, and behaviors add up in a consistent way to create two types of humans, each with its own set of psychological features? The answer, again, was no: As for brain structure, the differences created mosaics of feminine and masculine personality traits, attitudes, interests, and behaviors. For example, in the data set on 4,860 adolescents from the National Longitudinal Study of Adolescent Health, the variables on which young women and men differed the most included worry about weight, depression, delinquency, impulsivity, gambling, involvement in housework, engagement in sports, and a femininity score. So far, so gender normative. But: Not a single person had only feminine or only masculine scores on these variables. Rather, what was typical of both men and women (70 percent of them, to be exact) was a mosaic of feminine and masculine characteristics.

Here’s the paper at issue (click on screenshot to see it):

Yet on the same page of Proc. Nat. Acad. Sci. where their article appears, there is a note that there are four rebuttals to the paper of Joel et al.:

This article has a reply. Please see:

Multivariate revisit to “sex beyond the genitalia”

Yes, there is a female and a male brain: Morphology versus functionality

Patterns in the human brain mosaic discriminate males from females

Joel et al.’s method systematically fails to detect large, consistent sex differences

The titles more or less tell you what’s going on: multivariate analyses are actually quite good at discriminating male and female brains into two groups. (I can’t find a reply by Joel et al. to these critiques, but one may exist.) Joel and Fine do not mention these credible criticisms of their paper; they just pretend that their statement stands clear and unrefuted. I find that a sleazy way to behave, and had I vetted the editorial for the NYT, I would have insisted that Joel and Fine at least point it out.

So be it. We needn’t base our morality on our biology, but we needn’t base our facts on our ideology, either.

UPDATE:  Bill Boecklen (comment #13) wanted to post this figure to illustrate his point, but neither of us know how to post figures. Below is what he said and his figure:

I suspect the Fine-Joel argument may result, in part, from a mathematical artifact – the projection of a multidimensional space onto a line. The line, of course, is the male-female continuum. The multidimensional space represents all the characters that distinguish males from females. The observation that an individual cannot be assigned into male or female categories with probability =1.0 does not in any way suggest that there are not non-overlapping categories in n-space. Consider the following graph of two non-overlapping groups in 2-space. A projection onto either axis will result in overlapping categories in 1-space:

h/t: Cesar