For the past 10 years, feminist philosophers have tried to place feminism on a scientific footing. In pursuit of that goal, they have turned to social psychology, and in particular, to research on implicit bias and stereotype threat. Among the two, the hypothesis of implicit bias seems to enjoy slightly greater credibility. Even a philosopher as capable as David Papineau wrote, in a review for the Times Literary Supplement, that
…as Jennifer Saul stresses in her measured contribution to Women in Philosophy, there is a wealth of evidence to show that many well-meaning people, including academics, display “implicit bias” against historically unprivileged groups even when they are trying to treat them fairly.
However, as an explanation of discriminatory behavior, implicit bias has recently received a fair amount of critical scrutiny. Philosophers have performed such scrutiny here and here (for an ungated version of the latter, look here). Psychologists were, of course, first (see the many references in the aforementioned articles, and also this and this).
Some of this skepticism has made it to popular philosophy blogs such as Brian Leiter’s (here and here), and to the mainstream right-wing press (here and here). Just recently, Daily Nous, too, has started to report on it. Hence, perhaps the day is coming when The Stanford Encyclopedia of Philosophy (February 2015) entry on implicit bias will be updated to include in its main text some of the recent, and not-so-recent, negative news regarding the Implicit Association Test (IAT), which is the most well-known test for measuring implicit bias (somewhat reminiscent of Carl Gustav Jung’s word association test, which you may have seen at work in the David Cronenberg film A Dangerous Method). For the time being, the Encyclopedia relegates the skeptical perspective entirely to a very brief footnote, while the main text states that:
Several large reviews have found the IAT in particular to be reliable, relatively unsusceptible to intentional distortion, and most ominously, predictive of a range of discriminatory behavior, in some cases better than self-report… Overall, the IAT appears to predict many distinct kinds of behavior.
The truth is that “several large reviews” (cited in the aforementioned critical articles) cast serious doubt on the predictive power of the IAT.
In a similar vein, criticism of implicit bias research is scarcely mentioned in the two-volume work recently published by Oxford University Press (OUP). By the way, when did OUP’s philosophy section last publish a two-volume work on some controversial idea propounded by conservative philosophers?
Some feminist philosophers (for example, Edouard Machery , Sally Haslanger, and Anne Jacobson) seem to be hopeful that a “trait” view of attitudes can provide a more solid notion of implicit bias. On the trait view, attitudes are dispositions to think, feel, and act in certain ways (positively or negatively, depending on the attitude). But, as both Machery and Haslanger acknowledge, the trait view does not allow for a distinction between implicit and explicit attitudes. On the trait view, either one has the attitude, or one does not have it.
Moreover, attitudes understood as dispositions are unlikely to be of much explanatory value. For suppose that we find people behaving in a biased way toward a certain group; say, they strongly prefer to hire the female candidate for a job, even though the male candidate is equally qualified (as in this experiment). Certainly, a poor explanation would be that the people in question have a disposition to act in such a biased way. A better explanation would invoke the psychological basis of the disposition. But notice that the explanatory work is then no longer performed by the attitudes themselves.
In other words, what feminist philosophers are doing when they adopt the trait view of attitudes is changing the subject, from a mental state that could in principle explain an objectionable behavioral disposition to the behavioral disposition itself. But that disposition does not explain itself. Hence, while implicit bias may be resurrected as a trait or disposition, it can only do so by becoming idle in the explanation of biased behavior.