The comments in this post have inspired me to flesh out my thoughts on diversity more fully. In that post and elsewhere, I've been called out by Mark Olson for only supporting particular kinds of "diversity", and choosing those kinds due particular political commitments I hold. This post concedes both points. However, it argues why that is not only not a bad thing, but likely unavoidable, and in doing so expands on the concept of "salient diversity" I alluded to in the comments.
I. Pure Diversity
Even as an abstract matter, diversity is beneficial, if only from the generic observation of a liberal education that it is good to be exposed to a wide variety of things. It is good that I have at least some background in music, biology, and economics (among other disciplines), even though I don't plan on directly utilizing any of them in my career, because they broaden my mental horizons, enable me to more fully communicate with people for whom these areas are important, and expose me to new manners of manipulating and expressing myself in the world that I may (unexpectedly, perhaps) find useful or fruitful. The same thing can be said about meeting different types of people, who also can broaden my horizons, improve my ability to relate with other people unlike myself, and give me new and useful perspectives on important issues.
I associate these benefits with "pure" diversity because they accrue regardless of how, exactly, the two parties are different from each other or from their normal experiences. That is to say, interacting with someone who seems genuinely and materially different from myself helps my ability to relate and communicate with all of those who seem significantly unlike myself, regardless of whether that someone's difference is that of race, nationality, occupation, class, or anything else. The only operating factor is the fact of difference, not the specificity of what that difference is.
This matters because it establishes that at least some of the benefits of diversity still occur even though it is obviously impossible for anyone to encounter every sort of person or all types of difference. Even though I can't take every class here and there are whole departments which I have never set foot in, I've benefited from Carleton's liberal education in the aforementioned ways, because I've taken some different courses and some areas of study beyond that which I'd normally be inclined. So long as diversity is pursued in some form, the people will reap the benefits, even though necessarily only some types of diversity will be represented in anyone's experience. However, this does not mean that particular kinds of diversity don't bring their own, unique benefits. They do, and this is something I will explore in the next section.
II. Two Models of Diversity
The act of pursuing diversity necessarily requires the pursuing actor to make decisions over what type of diversity they want. Since it is impossible to represent every sort of difference, these decisions always contain a normative component -- some differences will be privileged, others, subordinated. There are, I think, two broad models one could use as guidelines for diversity pursuit: The "distance" model, and the "relational" model.
A. The Distance Model
The distance model of diversity takes as its starting point the idea that our differences are measurable, and hence some people are fundamentally more different (more diverse) than others. Since we're viewing diversity as a positive good, then, the goal should be to incorporate into the system or institution the most different people possible, and the model looks for those whose experiences or life position are most "distant" from our own.
The distance model first requires is to construct a measuring point -- a norm from which we measure difference and distance. So, a college might start with who it imagines to be its paradigmatic student, Jane: White (Caucasian), American, female, upper-middle class, Christian, able-bodied, moderately-liberal heterosexual. It would then actively pursue students who were as distant from this image as possible. A White American upper-middle class heterosexual male moderate conservative would be a little different. A poor lesbian Latina Marxist would be far more different. A poor disabled gay nationalist Rwandan who has traditional "pagan" beliefs, of course, trumps all.
There is obviously some appeal to this. In terms of the benefits of pure diversity, it is likely that Jane would benefit more from the Rwandan prospective student than either the Latina Marxist or the male moderate conservative. Moreover, the distance model seems to carry with it the benefit of objectivity. We're looking for diversity. The Rwandan is clearly the most diverse. Hence, he should be admitted.
But as it turns out, the model is deeply flawed. For starters, there is a big problem is with creating the "paradigmatic student" at all. The establishment of certain students as "normal" (with the concurrent implication that others are, to varying degrees, "deviant") is inherently problematic, with unavoidable messages of privilege and subordination. If Jane is the norm, does that mean the college should structure its educational mission around her? The distance model's answer is clearly yes: we're looking for the "most distant" person as it relates to Jane, a yardstick which might not hold true for other students. If the college admitted the Latina Marxist, neither Jane nor the Rwandan would be "most different" for her -- that position would seem more likely to be occupied by, perhaps, a wealthy libertarian straight Asian male. By establishing one type of student as the norm, the college ends up objectifying all the others, who are perceived solely as tools for the educational benefit of a few.
Furthermore, upon closer examination, the supposed objectivity of the model is questionable as well. Obviously, the college has to choose certain attributes to be calculated as part of what makes one "diverse." But not every attribute can be chosen -- favorite color, height, body type, and non-academic interests, for example, were not directly included on my original list. The choice to include certain characteristics but exclude others is both subjective and political. And even if one could settle upon an agreed upon list of salient characteristics, that would represent only marginal progress, for two reasons:
(1) It doesn't account for what Stentor aptly calls non-Pareto cases. In the above example, each person listed was indisputably more "different" than the one before -- I held the other variables constant. But outside perhaps the absolute poles, most of the time people mix-and-match along categories: we're White and liberal, but Jewish, French and disabled, versus Black and poor, but American, able-bodied, and straight. These differences aren't in themselves fungible, and so they provide no basis for comparison in all but the most extreme cases.
(2) It doesn't account for the fact that people weigh these differences differently. This came into play already in the original decision of which categories to include (excluding, say, favorite color is implicitly saying one doesn't find that it has any meaningful weight at all). But even if we have a set of characteristics we agree are all somewhat important, we might not agree that they are all worth the same amount or which ones are worth more than the others. Jane might relate to her own nationality, sexual orientation, and class more than her religion, political ideology, or able-bodied status. Hence, she'd view a poor gay Frenchman (who is also an able-bodied, moderately liberal Christian) as significantly more different than a disabled atheist conservative (who is also an American, heterosexual, and middle-class). If she weighs any of them highly enough, it might even cause her to see someone with just one salient difference (nationality, say) as more distant than someone with many differences she is less concerned about (religion and political ideology and able-bodied status). And of course, someone who has the exact same characteristics as Jane could have a precisely inverted weighing scheme. So unless we are to expand the "paradigmatic student" so that it also includes a "paradigmatic weighing mechanism" (which would be impossible to do objectively), its objectivity becomes mythological.
B. The Relational Model
The relational model, by contrast, does not take as its position that there are objective distances in the differences between people. Nor, for that matter, does it assume that certain differences are objectively more salient than others. The relational model takes as its starting point that different differences can matter in various circumstances, and pursues specific types of difference insofar as those differences can contribute to the ascertainment of particular political or social goals.
A good example for this model is a basketball team, which wants on its roster a diverse number of positions represented (i.e., it wants to have some guards, some forwards, and some centers). It is not creating a "paradigmatic player" (Shaquille O'Neal, say) and then trying to find qualified players who are as different from him as possible. Rather, the quest for diversity here stems from the observation that, as good as Shaq is, a starting lineup of five Shaq clones likely would not create a good basketball team. We need players who can shoot the three, who can create off the dribble, who can shoot free throws(!), who can make fast breaks, and who can defend the perimeter. What we're looking for is a group of players who as a team (student body, corporation, legislature, etc.) are best able to fulfill the goals we want them to fulfill. I call it the relational model because it looks at diversity not in terms of objective distance between people, but rather based on the idea of a network of individuals, none of whom are considered the center, whose relationships to each other are defined by the particular roles they can play specific systems.
In a relational model, selecting "players" (admitees, hires) is what the late Harvard Law Professor Lon Fuller called a "polycentric problem." Polycentrism occurs when the selection of one person changes the "ranking" of all the other potential candidates. The Los Angeles Lakers have Shaquille O'Neal. Their next priority in player choice is likely not to be a center, even a really good one, because they already have the center role filled (back-ups excluded) and what they need now is a guard. If their first two players were Kobe Bryant and Dwayne Wade, by contrast, then their next choice might well be a center.
But it's even more complicated. Some people can fill a given "role" in certain networks but not others. For example, Kobe Bryant can normally fill the role of guard rather well. But, as the Lakers can tell you, if your center is Shaquille O'Neal, then Kobe Bryant maybe isn't your best choice at guard. The players, in other words, can't just fill their role abstractly. They have to be good in relation to one another. This lends the relational model considerable dynamism compared to the more static distance model.
So the relational model asks "what set of people would most effectively allow us to do the things we want to do", noting that in any complex system people will likely have to fill different roles, that different people are best suited for different roles, and that different people are best suited for different rules depending on who else is filling the other roles. The basketball team is a simplified example, however, because by and large "things we want to do" is simple and agreed-upon: we want to win our games. That isn't always the only consideration -- a team might want to have high attendance, and thus plays a home-town hero to build fan loyalty -- but it generally has primacy.
In other environments, however, "what we want to do" is contested and variable. An institution has to decide what its goals are before it can pursue the type of personnel set necessary to do it. Since these goals are contested, it is a normative decision to adopt one set over another. What goals you adopt is premised on what you think is important. So in that sense, it's perfectly fair to say that institutions constructing diversity programs are pursuing an agenda -- and a specific one at that. But it's clearly necessary to adopt such an agenda, and doing so channels decision-making procedure towards particularly types of diversity.
For example, I think inter-racial dialogue is important, and I think colleges are a key institution that constructs people's ability and inclination to engage in public dialogue once they enter society. Hence, I will specifically support the types of collegiate diversity that foster such dialogue and create people who can communicate with persons of different racial backgrounds in the public sphere (something which requires a critical mass of people filling the "role" of each major racial cleavage in American society). If you do not think inter-racial dialogue is important (or is less important than other, competing goals), then you will not support this type of diversification (or won't support it for that reason). And on the other hand, in institutions that I do not believe play primary roles in constructing society's discursive horizons (say, a graduate program in eastern European studies, or for that matter a basketball team), I won't support that same type of diversification (or at least, not for that reason).
The last example indicates that different differences matter in different places. As I noted in my discussion of the difference model, we can legitimately put different weight on various differences depending on whether we feel they are more or less salient. This is not an objective procedure in either the distance or relational model, but it is at least mitigated in the relational model because its localized -- we're only asking for relative weights in a particular situation, not as some sort of metaphysical truth that exists across all cases. Just as it is likely that all major differences will be valued by some people, some of the time, it is also likely that these differences will be seen as salient by at least some institutions, some of the time. This has the effect of expanding diversity conceptually beyond the strictures of the American affirmative action debate and places it instead as a consideration for all types of decision-making.
The first section establishes that diversity confers at least some benefits no matter how the diversity is constructed. The mere fact of being exposed to different people, different disciplines, different ideas, and different experiences is inherently salutary. The second section, noting that we have to pursue diversity in some fashion, identifies and analyzes two competing frameworks of diversification. If it isn't obvious already, I find the relational model greatly superior to the distance model as a framework for how we ought to construct social relationships. The latter not only avoids the creation of a particular norm and the exclusionary implications that would follow from it, but it also displays more flexibility to adopt itself to differing situations. A corporation might not care how many tall guys it employs, a basketball team definitely does. A chemistry department might not care how many Ukrainian immigrants it has among its affiliates, an eastern European studies department could find that very relevant. What matters about diversity is not any objective measurement of difference, but the manner in which difference is operationalized to enable different functions and achieve social goals.