It seems that the Summers’ incident has an indeterminate half-life for it is still riling up those opposed to the argument Summers put forth. The latest blog post that I’ve read is enthusiastically endorsed by the usual suspects and is written by Echidne of the Snakes (which surprised me because I enjoyed her comments when she visited here to debate with Razib during the hottest moments of the Summers’ Tempest) guest-blogging at Alas, A Blog.
She takes aim at Steven Pinker for his defense of Larry Summers’ remarks. Here is our summary, and linkfest, to that debate. The point of her essay is that Pinker is hopelessly biased and his arguments are illogical and thus his views on the matter are immaterial. So, because her essay is so centered on the issues of bias and logic I thought I’d return the favor and apply the same standards to her essay and see if her analysis of Pinker is correct.
She opens her essay by linking to Pinker’s article in The New Republic:
Summers did not, of course, say that women are “natively inferior,” that “they just can’t cut it,” that they suffer “an inherent cognitive deficit in the sciences,” or that men have “a monopoly on basic math ability,” as many academics and journalists assumed. Only a madman could believe such things.
She starts her rebuttal by focusing on the above quote from Pinker. She writes:
I remember hearing a radio interview with Pinker when his book The Blank Slate came out, and he used the same madman-argument to clear the deck of any accusation that he might be an essentialist. As few researchers would call themselves madmen, this clever trick means that we can now dispense with any exploration of Professor Pinker’s own possible biases, and can go on to study the biases of his opponents.
However, there is much context lost by clipping the quote. Let me do my part to remedy the oversight. Here’s the remainder of the quote that Echidne omitted:
Summers’s analysis of why there might be fewer women in mathematics and science is commonplace among economists who study gender disparities in employment, though it is rarely mentioned in the press or in academia when it comes to discussions of the gender gap in science and engineering. The fact that women make up only 20 percent of the workforce in science, engineering, and technology development has at least three possible (and not mutually exclusive) explanations. One is the persistence of discrimination, discouragement, and other barriers. In popular discussions of gender imbalances in the workforce, this is the explanation most mentioned. Although no one can deny that women in science still face these injustices, there are reasons to doubt they are the only explanation. A second possibility is that gender disparities can arise in the absence of discrimination as long as men and women differ, on average, in their mixture of talents, temperaments, and interests–whether this difference is the result of biology, socialization, or an interaction of the two. A third explanation is that child-rearing, still disproportionately shouldered by women, does not easily co-exist with professions that demand Herculean commitments of time. These considerations speak against the reflex of attributing every gender disparity to gender discrimination and call for research aimed at evaluating the explanations.
Let’s review her logic here. Her premise is that Dr. Pinker’s defense of Dr. Summers is not logically coherent because it is ridden with poor reasoning and that Pinker’s bias infuses, and thus invalidates, the thrust and conclusion of his article. Rather than tediously repeating the details of remarks that are already quoted, I’ll restate them in LogLish (Logical English.)
The {Condition} C of women in {field} S has at least 3 explanations. Reason X, Reason Y and Reason Z. It is posited that Reason X is not sufficient to explain the {Condition} C. If Reason X cannot explain {Condition} C then we should investigate Reason Y and Reason Z.
Echidne dispenses with Pinker’s defense of Summers by attacking Pinker’s attempt to correct misreporting and not by addressing the logical problem set laid out by Summers or Pinker’s treatment of that problem set. Let’s look at the LogLish version of Pinker’s correction of the prevalent misstatement:
The {Condition} C of women in {field} S is caused by women’s lack of Attribute A, Attribute B, and Atribute C. Only a madman would take such a position.
This is an entirely appropriate position to take if you are being mischaracterized by your opponents. State at the outset that you do not hold these beliefs. It does not follow, as Echidne argues, that the remainder of your argument cannot be criticized. The only conclusion that can be reached is that your argument cannot be criticized for the reasons you’ve pre-emptively disavowed. This tactic removes lying from the arsenal your opponents can use against you. This in no way removes from Pinker’s opponents their ability to further explore Pinker’s text for bias. Go to town.
Let’s continue to explore the essay’s poor reasoning.
But Pinker’s summary of the issues is partial: he fails to address all the reasoned responses from feminists and progressives, and he fails to mention the truly outrageous statements on many of the anti-feminist and conservative websites and blogs. This makes the unreasonableness appear solely something that takes place among the liberals and feminists, not something that might even infect careful researchers such as Professor Pinker.
Let’s go back to Summers’ original problem set:
The {Condition} C of women in {field} S has at least 3 explanations. Reason X, Reason Y and Reason Z. It is posited that Reason X is not sufficient to explain the {Condition} C. If Reason X cannot explain {Condition} C then we should investigate Reason Y and Reason Z.
Let’s reframe this into propositional logic:
Premise #1 Reason X partially contributes to Condition C.
Premise #2 Reason Y partially contributes to Condition C.
Premise #3 Reason Z partially contributes to Condition C.
Conclusion: Reasons X, Y, & Z explain Condition C.
Those attacking Summers are putting forth a self identity, something that is a valid statement, true in every case:
Reason X wholly explains Condition C.
They have dismissed Summers’argument and substitued what they think is a true and valid statement in its stead. By doing so, they open themselves up to a test of statement falsification, which Summers and many others feel has already been achieved. So, with the statement being shown to be untrue the issue focuses on which of the 3 premises offered by Summers is incorrect. Those opposed to this argument have completely removed a premise from consideration without a test of falsification. If at this point you think that this essay is begging the question because I’m not offering supportiing links in support of the invalidation of the self-identity you can stop reading and Google the issue for yourself. This is a blog entry, not a journal submission.
So, Pinker need not address all the proffered reasoned responses from feminists and progressives in support of the self-identity that has been argued is insufficient to explain the phenomena. He could give them full hearing and that would still leave the self-identity being invalid. Nor is there any need for Pinker to pay attention to the excesses of the feminist opponents who accept the invalidity of the self-identity and the validity of the 3 premises. The exuberance of these parties has no bearing on the validity of the argument. Similarly, the reasoned support of an invalid self-identity is also immaterial. So, Echidne’s argument, as she has offered in this point, for bias on the part of Pinker is i
nvalid. She continues:
In any case, our careful researcher then goes on to summarize various studies which demonstrate gender differences on the average. He doesn’t summarize the studies which don’t support these findings or the studies which address the whole question of what we are actually trying to measure with the various tests. All this reads “biased” in my book.
Again we come back to the test of invalid self-identity versus a proposed argument. Pinker is writing for an audience that holds to the validity of the self-identity and thus he is attempting, for their benefit, to falsify the self-identity. He only cites evidence that demonstrates gender differences in order to remove the self-identity from consideration and that is at the heart of this argument, and Echidne makes a grave error in her analysis by not recognizing this. Pinker, and Summers, are not out to replace one instance of self-identity with another instance, namely that all variance is explained by gender difference and they have no need to bring in all sorts of irrelevencies that only confuse the issue. The countervailing evidence on gender differentiation is not required, as it would be needed to invalidate an alternative self-identity, and does not show bias, for both Pinker and Summers argue a multifactorial causation scheme. If one premise is falsified that doesn’t reinstate the already invalidated self-identity favored by Summers’ opponents. That dog is already dead.
There are no feminists who believe that women and men are biologically exactly the same, though there seem to be a very large number of anti-feminists who never see the most obvious difference between the two sexes which is the fact that women give birth. Anti-feminists want to have more science to find out what really distinguishes the sexes, all the while letting their eyes glide over the pregnant bellies of their coworkers or the countless young women pushing prams outside.
It’s hard to decide what to make of this statement. It’s either a disingenuous tactic that by pointing at obvious gender differences seeks to save from criticism the central issue that is at the heart of the Summers’ dispute or it’s a testament to the fact that Pinker and other researchers have been effective in dispelling the notion that we are Blank Slates but in no way does it invalidate the position that this was a widely held view, and still is when we consider non-obvious biological differences. It seems that Eschidne is out to punish Pinker for being effective in shedding light on the issue. The subtlety of her statement that many anti-feminists are oblivious to the fact that women give birth is lost on me. Perhaps a little more exposition on her part would have been helpful in clearing up her point.
The differences that gender science may find are going to be put to political uses pretty fast. Even if the results are based on faulty methods and data, the harm the political applications will do is real. This is the reason why it is so important to insist on transparency and high methodical competency from all practitioners of gender science, and why it is very important not to have a value bias among this group towards one sex or the other. Currently there is such a general bias, as even a cursory reading of the studies reveals, and that is one of slight misogyny. In other words, not all science is somehow above politics or even above cheating, and all science should be approached with a very critical mind.
This isn’t a direct criticism of Pinker, but more of a general smear. The hypocrisy of the statement though, is quite rank, for she in no way acknowledges the extreme bias that is present with existing, wholly environmental, studies. She seems to posit that they are value neutral and bias can only be introduced in the study of genetics. Need I go on?
Actually, I agree with Pinker on one of his arguments: that we should encourage good science on innate gender differences. The real question is how to do this. How would Pinker create a study which would tell us, for once and for all, what the real cognitive differences between men and women are? We actually don’t have the tools to do this today, and this is the main reason why I find Pinker’s elegant impartiality so insulting.
She may find Pinker’s impartiality insulting but she surely hasn’t demonstrated to her readers that he is biased or that even if bias is present, that the bias invalidates his arguments. In the end, I find her argument to be invalid.
Posted by TangoMan at 09:48 PM