What Summers did wrong

Having written all that I did in defence of Summers' argument, I still don't have a great deal of sympathy for him. He did make several very important mistakes, which are things he should know better than to do, and go some way towards explaining the depth of hot water he's ended up in.

Forgetting who he is
Though he wasn't speaking 'in an official capacity', and he said so several times, he can't escape from his job. When one is the president of a University, it's impossible to completely divorce any public speech from that status, and naturally anything said in a public forum runs the risk of being taken as the view of the University. For this reason, I think the most justified of his critics are the Harvard faculty, who last week officially censured him. I think this was a reasonable thing to do, to make clear that the views expressed were not the views of the faculty at large, and I do wish his defenders wouldn't go calling these people things like academic Stalinists.
Forgetting who he isn't
He's an economist by training, not a psychologist. This doesn't mean he's not entitled to opinions about psychology, but when speaking to a conference, one should stick to one's expertise. Pinker, on the other hand, is a psychologist, and therefore much better placed to talk about such issues. I note that when Pinker presented essentially the same logical point to two audiences (a small group first and then a couple of thousand people in Severance Hall) last week he was able to explain it clearly and persuasively, no-one walked out on him, and the questions he faced were much less angry.
Forgetting where he was
I would have very much enjoyed spending an evening down the pub discussing what he had to say. The trouble is, he didn't present these ideas in a private, informal conversation. He got up in front of a conference and presented a lot of half-baked speculations on things outside his area of expertise. At best, this was pretty insulting to the audience, whose time he wasted. This was only made worse by the fact that he was taking on a controversial subject. I don't understand how he could not have expected this to end up in the papers, and upsetting people.
Not giving enough time to the serious systemic reasons why more women don't want to go into academic science careers
In Summers' defence, I imagine this had all been gone through by other speakers at the conference, but once again he must have realised that because of his status his remarks would end up being discussed in isolation. In that vein, and for the sake of giving his own staff confidence, he could have helped himself a great deal by devoting a few minutes to lay out the many things we're doing at Harvard to promote the crucial objective of diversity, instead of just glossing over it. He could also have given a little space to discussion of some of the less controversial partial explanations of the gender disparity, particularly ones which are symptomatic of features of these workplaces which are actually bad for everyone. I won't go into this here, but just point to two pieces I've read on the subject recently: Meg Urry on the impact of subtle discrimination and too visibly being the only representative a particular group, and Brad DeLong on the insane level of job competition at the top of academia.
Suppressing the transcript
I think this was actually his biggest single error. Having argued that his actual words were less offensive than the secondhand account I first heard of his words, it's important to recognise one of the things that gave the secondhand accounts power. For over a month there was no transcript to refer to, because it took him that long to release it. Not only did that make it impossible for many people to see what had actually been said, but actions like these are generally consistent with someone who has something to hide, so it lent credence to those arguing that his words were disgustingly offensive.
The use of stupid anecdotes
I'll quote what I think was the single worst part of the whole speech: ...I guess my experience with my two and a half year old twin daughters who were not given dolls and who were given trucks, and found themselves saying to each other, look, daddy truck is carrying the baby truck, tells me something. Again this is a matter of context, because down the pub this would be a perfectly acceptable point to raise, and it would just have to be countered with an example of the opposite. On a podium in front of an academic audience, this comes across as, well, not worthy of even mentioning. The use of such an anecdote in that context immediately undermines what the speaker is saying, and I can certainly understand why it would have made plenty of the audience switch off and disregard the rest of his remarks. I think I might have done the same.

So in summary, while I think a lot of the outside reaction to Summers' remarks are completely out of proportion to the relative inoffensiveness of his overall message, I do think he got himself into this situation. I also have a great deal of sympathy for the faculty who have felt the need to distance themselves from their president, and I think outsiders chiming in to accuse them of censorship are missing the point. If 'some professor' had said these things I would have to ask why anyone felt the need to argue, but when it's the president of a University the context is immediately changed.

Update: I've just seen an editorial in the Monitor that takes pretty much this line. I think it's the first thing I've seen in the mainstream press that actually tries to paint it this way rather than being outraged by either the supposed censorship of Summers or the supposed disgustingness of his words, even if it does caricature his words a little.


Trackback URL for this entry is:


Thanks for this; a good pair of posts! (And not just because you seem to agree with me ;) )

Posted: March 21, 2005 08:08 PM

I definitely agree. I don't really care what Summers thinks about gender issues -- he may be right or wrong. He shouldn't be denounced for stating it, the ideas should be analyzed and debated.

Academia is all about bringing up new ideas and examining them. But if I were him, I don't think I would have chosen that point to bring them up. :)

Posted: March 22, 2005 11:02 AM


That actually wasn't quite what I was trying to say. Because Summers is in effect the boss of a number of the people most directly affected by this issue, his views do matter. What I am arguing is that if Summers were just 'some professor' this whole flap would be ridiculous, but the position he holds makes it an important issue.

Posted: March 22, 2005 11:28 AM

Eldan, thank you for this pair of posts, I found it to be an interesting and well-reasoned take on the controversy.

This comment is rather off-topic and closer to some of the implications of the gender disparity issue. I found myself reading a Time Magazine article (registration required) that sought to show how the aptitudes of women and men are partially determined by forming aversions to material the person was introduced to at an age before a particular brain capacity had matured (i.e. spatial reasoning maturing later in the typical female brain). Regardless of many of the points and studies shown in the article, the main irritant was the presentation of all female brains/attitudes developing in the same way and all male brains/attitudes developing in the same way. When you referenced statistics in your post, you were careful to point out the deviations rather than taking the normal of whatever set of data was collected and attributing it to the entire female/male population. One of the consequences of this absolute classification was a portion of the article advocating segregating girls from boys in education so that the individual learning styles could be catered to as a function of where the child's brain maturity was. Aside from the usual questions that would arise at this point ("Are the children receiving an equal education?" "Will the students have difficulty operating in a mixed-sex environment as a result of this segregation?"), the issue of deviations was never addressed. What about the portion of girls who develop spatial reasoning at an earlier age? What about the boys that develop better hearing in the vocal range at an earlier age? Over the years, most learning environments have sought to cater to a wider range of learning styles. The advocates of single-sex education in this article did not appear to consider the idea of testing the individual students to determine what learning environment to place them in. Our society enjoys the neatness of absolute classifications, in this case resulting in a significant minority being placed in a worse learning environment in order to cater to the current statistical norm.

Posted: March 22, 2005 05:05 PM

I think that is a very important point, and it's another thing that Pinker took care to stress and Summers would have done well to discuss. On any measure, individual differences are bigger than male-female differences, even when there is a significant difference between the means of the two groups. When we're looking at a difference between variances instead, that tells us absolutely nothing about the relationship between any two individuals. So all of the research I've ever seen supports the same conclusion with respect to the treatment of individuals: that it's always wrong to use an individual's group membership in deciding what will be right for them.

Putting people into single sex schools on the assumption that they will have a particular learning style because of their gender risks reinforcing both real gender differences and gender prejudice.

Posted: March 23, 2005 10:26 AM

Just got round to reading the full transcript of the speech...

Eldan: I pretty much agree with both your posts

Melinda: like you, I was really wound up by the arrogance of his tone

I thought the argument about the difference in variance between men and women was pretty weak, at least in the manner in which it was presented. First, it seems like an enormous leap to go from the observation that there are tests where the standard deviations of male scores are higher than female scores (I'd like to see the evidence for this) to the suggestion that aptitude for "high-powered jobs" has a wider distribution in men than women. But he did present the argument speculatively, so this doesn't seem a particularly big deal. More importantly, I object to the way in which he presents the "differing variance" hypothesis as being in some way causally explanatory, and necessarily independent of socialisation and discrimination. This is obviously nonsense. If it's true that the distributions between men and women differ in variance, if not in mean, then the question remains why this is so. And one likely explanation is factors such as socialisation and discrimination. (E.g., as a result of socialisation you may not find so many women at either extreme of the distribution, or there may be two or more forms of discrimination independently contributing to the over-representation of men at each extreme, etc. etc.)

Posted: March 28, 2005 05:27 AM

Post a comment