The Use and Misuse of Social Science (herein of “verbal violence”)

In his most recent post, Frank links to his approving reference to Ben Barres’s claim that “When faculty tell their students that they are innately inferior based on . . . gender they are crossing a line that should not be crossed –the line that divides responsible free speech from verbal violence.” So does my last post, suggesting greater variability in results on math tests based on gender amount to such a claim? Are those who say such things asserting that women are inherently inferior at mathematics?

No. I don’t think so.


From what I gather there is a fair amount of evidence that if you take a sufficiently large group of men and a sufficiently large group of women, their mean mathematical performance will be about equal. This does not mean, of course, that every individual’s ability will be equal. Some will be good at math and some will be bad at math. Indeed, there is no particular reason that any member of either group actually performs at the mean that we ascribe to “men” or “women.” An average is by definition a mathematical property of a set not of an individual.

My understanding – based on the Science study touted by the NYT – is that at the extremes there are more males than females. This does not mean that all the geniuses and all the dunces are males. It does not mean that the smartest male is smarter than the smartest female or vice versa. Ditto for the bottom of the distribution.

“Ah,” one might object, “you are still claiming that men are more likely to be math geniuses than women! This is precisely the kind of ‘verbal violence’ that Barres was warning against!”

Um, not really. It is true that if faced with an individual about whom I knew absolutely nothing except his or her gender, I might hazard a guess that there is a slightly higher probability of their being a mathematical genius or a dunce if they are male. A very, very slight probability. On the other hand, I would hasten to add that gender is unlikely to be the best predictor of math ability, and if I really wanted to make some sort of claim about ability I would want to look at a bunch of factors other than gender that are likely to give me more information.

Indeed, if I was faced with an individual the best way of determining their mathematical aptitude would not be to observe their gender. It would be to observe their mathematical performance. Any person – male or female – might be a genius or might be a dunce. Chances are they are someplace in the middle. Regardless, the best thing to do is to observe their actual ability. In short, when making individual assessments, the claims that we make about large populations are largely beside the point.

Then why even pursue such inquiries? Aren’t we just playing with fire, giving ammunition to reactionary forces that lack my (admittedly pathetic) mathematical nuance?

The short answer is that when we are thinking about the best intellectual, policy, emotional, and moral response to social facts – like the male dominance of engineering faculties, for example – social data is useful. Notice, this is not the same thing as claiming that men are inherently better at math than women, nor does it amount to the claim that there is some biological determinism at work in the variability. I am frankly agnostic on the question of etiology. Nature v. nurture; who knows? Not me. I suspect the answer is both, but that is a gut instinct not a reasoned position. It is also not the same thing as claiming that the variation in performance among men entirely explains social facts like the engineering faculty. I doubt it. I certainly haven’t seen evidence that it works as a solo cause. On the other hand, I doubt that it is irrelevant.

All of this means that as a functional matter, I tend toward the libertarian end of the spectrum when it comes to academic freedom. As a philosophical matter, I agree with Frank that ultimately what counts as a worthwhile inquiry and what is a pernicious waste of time (and I do think that there are things that are a pernicious waste of time) has reference to our current social condition and our moral judgments. That said, however, in practice I think that we ought to be suspicious about attacks on this or that research agenda as “verbal violence.”

You may also like...

7 Responses

  1. A.J. Sutter says:

    It looks like there’s a shift between the premise of your article and your conclusion. The “verbal violence” (I’m agnostic about the meaning of this phrase) referred to at the beginning arises “[w]hen faculty tell their students that they are innately inferior” — i.e., a statement of conclusions. Your last sentence, though, counsels caution before attacking “this or that research agenda as ‘verbal violence.’”

    I presume you mean a research agenda that doesn’t have a priori conclusions? So you don’t appear to be disputing Barres. Is that your intention? (BTW do you distinguish research agendas that do have a priori conclusions? I think Frank, or someone on this blog, has written about those before.)

  2. I’d say that your nuance is actually pretty good. And I’m a professional mathematician!

    There’s one overtone I’d add to your description of the “very, very slight probability” difference, and that other factors are unlikely to be the best predictor. It’s the way this sort of thing can show up in the aggregates.

    Scientifically-versed types know that gravity is incredibly weak — much weaker than electromagnetism. Two protons near each other push apart with an electric force many orders of magnitude stronger than the gravitational force that pulls them together. But when you get a lot of protons and electrons (and neutrons) together into macroscopic objects, they usually come in pretty equal numbers of positive and negative charges. All those powerful electromagnetic forces cancel each other off and the feeble gravitational force shines through to keep the Earth from flying away from the Sun.

    It’s the same sort of thing here: most variables have a much bigger effect on mathematical performance than sex. But when you take a very large number of men and women, those other factors will be distributed all over the place. Some will pull performance (at the tails) up, some will pull it down, and in the end it’s a wash. What you’re left with is the aggregate of many many instances of the tiny effect sex has.

    I’m sure that a lawyer could come up with a better analogy for a blawg than my physicsy one above, but I’ll have to leave that for someone else.

    As for “verbal violence”, it doesn’t stop there. When I said in the wake of Summers’ original gaffe that the question is open to research, I was told by LitCritters that I was “exactly the same as the Nazis”. These are graduate students and young professors in the humanities departments (including the law school) at Yale. In their mind there actually are questions one is not allowed to ask. Where’s Godwin when you need him?

  3. jerry says:

    Then why even pursue such inquiries? Aren’t we just playing with fire, giving ammunition to reactionary forces that lack my (admittedly pathetic) mathematical nuance?

    Well, it does seem to be a well understood and agreed upon fact that there is a lot of sex-linked stuff on those X and Y chromosomes that play enormous roles in our lives. And all understood and agreed upon that certain elements that seem to be genetic are in fact culturally based.

    It seems like a reasonable, ethical, moral, and productive area for research to investigate those differences, both to better understand our species and the rest of Animalia, and to understand the differences in nature and nurture.

    What does “giving ammunition to reactionary forces” have to do with anything?

    Isn’t that the definition of self-censoring, and “political correctness?” Do you have any evidence that with the absence of your research, reactionary forces will do less well, or find nothing else to be reactionary about?

    Assuming you’re a tenured researcher, isn’t pursuing such productive, moral, ethical research the definition of academic freedom and more importantly a duty you owe to the society that gave you tenure?

  4. Anon says:

    “Um, not really. It is true that if faced with an individual about whom I knew absolutely nothing except his or her gender, I might hazard a guess that there is a slightly higher probability of their being a mathematical genius or a dunce if they are male. A very, very slight probability.”

    On what basis do you conclude that there is only a “very, very slight” higher probability? If the tails of the distributions are really different, the probability could be quite a bit higher, given the bell curve nature of the distribution. Of course, the probability of being a genius would be quite small for either men or women, but the relative probabilities is an entirely different matter. And if you are saying that the probability would be only slightly different for men, then you are likely saying that the distributions are not different, or at least not different in the manner you discussed.

  5. jerry: the answer to your question is in the last paragraph of my previous comment. The fear (unfounded, in my view) is that the notion of biologically-based differences will lead to justifications for atrocities.

    The Nazis claimed to have identified scientific differences between Jews and Aryans which made the former inferior to the latter, and thus worth of segregation and extermination. In this case, some people see a straight line between “the right tail is fatter for men than women” and “we shouldn’t waste money on those girls’ educations because they’re not as likely to be as good anyway”. I think those people are crazy, but that’s the qualm Prof. Oman was referring to.

  6. John,

    The Nazis were hardly the only people to have identified those differences. In fact, the historical evidence is robust that eugenics was far more advanced in the U.S. than it was in Germany during the first several decades of the 20th century. There are a number of excellent sources for this, particularly Martin Pernick and Alexandra Minna Stern’s work.

  7. Frank says:

    here’s an interesting discussion on the topic:

    http://www.slate.com/id/2194486/entry/2194487/

    bottom line of it:

    “Even on the most hotly contested questions—like whether women have better verbal skills, or are hard-wired for empathy, or have cognitive differences that limit their advancement in math and science—the case for large, innate disparities is messy and, for the most part, underwhelming.”