The argument against selecting for diversity is an example of a more general principle:
Even if ability is equally distributed, if you limit your search you will often fail to hire the best person available.
At the risk of belaboring the obvious, an example may help. Let's first move away from the controversial topics of gender and race diversity C the thrust of current efforts C and look at something more neutral: regional diversity. Let us assume that Duke is as rich as Croesus C or, at least, Harvard C and can attract any potential faculty member it wants. And, to be specific, consider just two regions, East and West of the Mississippi, and a single discipline, physics.
Now suppose it turns out that there are equal numbers of able physicists on both sides of the country but only 20% of current Duke physicists come from the West. We can deal with this disparity in two ways: (1) Ignore it, and just hire the best possible physicist, or (2) address the problem directly by hiring only the best physicist from the West.
Look first at strategy 2, hiring for diversity. Since the best physicist available will be from the West only half the time, it follows that 50% of the time Duke will hire less than the best physicist available. Even if there are more good physicists in the West C say 70% C hiring with diversity as the goal will cause Duke to miss the best 30% of the time. Hiring for diversity will eventually achieve the 50:50 goal, but at some inevitable cost in terms of quality. In short hiring for diversity is necessarily in opposition to hiring for excellence.
Now consider strategy 1: Because ability is assumed to be equally distributed across regions, hiring for excellence will eventually result in a 50:50 mix of Duke physicists. Diversity will occur as an outcome, not as a goal, at no cost in terms of quality. This seems to many of us to be the proper course.
There is of course another way to resolve this dilemma and that is to redefine excellence. Some might argue that an excellent physicist is not just someone good at physics but someone whose other attributes C region, gender, race C satisfy some non-physics criterion. "Excellence," in this new definition, represents a balance between these two sets of criteria.
This approach is not without
precedent. After all, physics departments don't just hire any old
great physicist, they look for someone whose area of interest
and professional links to other department members will make him or her a valuable contributor to the work of the department as a whole. But these other criteria usually have some relation to physics and they are defended explicitly and specifically. Unfortunately our current preferential hiring policies are rarely defended in this way. The usual tactic is to hold up "diversity" as a self-evident" ood. What we need is an explicit defense of the specific preference that is being advocated. What are the goals social, political, cultural or whatever C to be served by restricting hiring to a particular group? How should these goals be weighed against excellence in the discipline? Let's hear a real defense of the diversity policy, not just denials of the ineluctable conflict between diversity and disciplinary excellence.