What (Epistemic) Benefit Inclusion?

Kenneth R. Howe
University of Colorado at Boulder

In his 1995 presidential address to the Philosophy of Education Society,^[1] Harvey Siegel advanced several theses regarding inclusion. In this essay I will challenge what I take to be the central one: Siegel's claim that there are good reasons for embracing inclusion but they are moral, not epistemic. In general, we are required to include people in our various conversations to show respect for them, but including them in no way guarantees that what they have to say will get us any closer to the truth. As Siegel puts his thesis:

Inclusion is not an epistemic virtue...there is no necessary connection between inclusion and epistemic worthiness or exclusion and epistemic defectiveness...Rather, inclusion is a moral virtue, and should be valued as such. People and groups deserve inclusion not because of any special epistemic privilege they enjoy, nor because including them necessarily increases the probability of obtaining true or justified theories.^[2]

In the arguments to follow, I begin by suggesting that Siegel's moral/epistemic distinction is a shaky foundation on which to erect his thesis, and will return to this theme throughout. (This distinction recurrently gets him into trouble.) I then offer several counter examples to Siegel's thesis from social science. For the most part, I focus on the thesis Siegel puts forward, and am not much concerned to defend the thinkers he is reacting against. (Lorraine Code is the exception to this.)

THE MORAL VERSUS THE EPISTEMIC: A SHAKY FOUNDATION

Siegel apparently assumes that we will find his distinction between the moral and the epistemic unproblematic,^[3] for he presupposes rather than establishes it. I do not find it unproblematic. Although I can make sense of a distinction between moral claims and scientific claims - in terms of the principles and evidence that would brought to bear - I do not see how to strictly separate the two. (I can also distinguish a dragonfly from an orangutan, but it doesn't follow that only one can be an animal.) I do not see how to strictly separate the two, that is, without presupposing a noncognitivist conception of ethics, such that moral claims are to be excluded from the (epistemic) realm of truth, knowledge, and reason.^[4]

I am not sure whether Siegel embraces noncognitivism in ethics (I think he doesn't), but he faces difficulties either way. If he does embrace noncognitivism, then he may be charged with inconsistency. For it is difficult to see how noncognitivism can be squared with various claims he makes that attribute a genuine cognitive status to morality. For example, he suggests in one place that rules against torture might (should) be universal and transcendent; in another place he suggests that "truth" might attach to matters of social justice. But if Siegel rejects noncognitivism in ethics such that moral claims may be universal, transcendent, and true, then the epistemic and the moral are intermingled. In this case, the question of whether the practice of inclusion is either a moral or an epistemic virtue is the wrong one to be asking.

HOW INCLUSION YIELDS EPISTEMIC BENEFITS: THREE CASES

The possibility that escapes Siegel is that inclusion might be both a moral and an epistemic virtue.^[5] That it is a moral requirement Siegel concedes, of course. So, except insofar as the moral and epistemic dovetail, I won't devote much attention to that. I turn instead to establishing how inclusion yields epistemic benefits.

At least three cases may be distinguished: (1) including persons in a community of scholars, (2) including persons as research participants so as to avoid biased samples, and (3) including persons as a general methodological requirement in social science. (Although Siegel does not himself explicitly identify these cases as such, he nonetheless broaches each.)

1. Inclusion in a Scholarly Community. Siegel employs the well-worn flat earth example, among others, to dismiss the view that there is any "necessary connection between inclusion and epistemic goodness."^[6] His general point is that widely held and talked about beliefs like "the Earth is flat" can and have proven utterly false, inclusion notwithstanding. In a general vein, Siegel says, "Inclusion is not necessary for good science; exclusion does not guarantee bad science."^[7]

If Siegel has logical entailment in mind, then he is surely right that there is no "necessary connection" between inclusion and the truth, and that inclusion cannot provide any "guarantee." But who would claim otherwise? In this regard, it would be helpful if Siegel provided an example of what such a "necessary connection" or "guarantee" looks like. Interestingly, Siegel does concede that "inclusive research communities, at least under certain sorts of circumstances, have a better chance of obtaining worthwhile results than exclusive communities."^[8] Although Siegel wants "necessary connections" and "guarantees," a "better chance" seems to be all that inclusionists need to win their point, particularly if "certain sorts of circumstances" widely hold (an issue I return to later).

In this connection, Siegel does not tell us just how strictly we should interpret his claim that "inclusion is not necessary for good science." Suppose we interpret this in the strictest sense, such that we can have scientific communities composed of single individuals. Now, there is no "guarantee," I take it, that scientific communities of one would fail to produce credible results. Does Siegel mean to suggest that such communities possess just as much "epistemic goodness" as the large and complex cooperatives we are familiar with? If not, is there then a necessary connection between some level of inclusion and "epistemic goodness?"

What Siegel has to say about standards of scholarship and expertise in a given discipline is germane here, for his rejoinder would likely be that including individuals qualified to participate in the conversation of a given discipline does have epistemic pay offs. Siegel's position, then, should be modified to say something like the following: inclusion does not insure epistemic pay offs all by itself, that is, inclusion is not a sufficient condition of epistemic goodness.

Now, this is a version of Siegel's thesis with which it is hard to disagree. I certainly don't, and I would join Siegel in criticizing those inclusionists who embrace it (if there are any, that is). But this makes Siegel's general motivation and point difficult to fathom. For, on this version of the thesis, inclusion is indistinguishable from others things researchers do in the name of enhancing epistemic goodness. For example, representative sampling is not a sufficient condition of epistemic goodness, and neither is precise measurement. Observing such methodological canons simply provides a "better chance" (but no "guarantee") of producing the truth. And if these were all a researcher did to insure epistemic goodness, their effects would be negligible or non-existent. Likewise with inclusion.

Siegel might well respond here by invoking the distinction between "rules governing the conduct of inquiry" and "criteria for evaluating the products of inquiry."^[9] Employing this distinction, ordinary methodological canons can (presumably) be identified with the latter, whereas inclusion can be identified with the former. In "this sense," Siegel concedes, "inclusion can be, and often is, an epistemic or methodological virtue."^[10]

In this sense? I must say, I don't find the distinction between rules governing the conduct of inquiry and criteria for evaluating the products of inquiry any less problematic than the moral/epistemic distinction. (And isn't just the old "context of discovery" versus "context of justification" distinction?) But that this is an untenable distinction is more than I need establish for present purposes. In general, observing methodological canons adds credence - "epistemic goodness" - to scientific conclusions, but does not thereby guarantee their truth. So far as I can tell, Siegel has done nothing to distinguish inclusion from other methodological canons in this sense. (And here it is worth observing that scientists are in the habit of forming new communities and adding members to existing ones when they have pressing problems to solve. The Manhattan Project is one example, and research on AIDS is another.)

2. Unbiased Sampling. The example of unbiased sampling is particularly worth pursuing because it is one that Siegel specifically addresses. The objection is that Siegel's claim that inclusion has no epistemic force is surely wrong, for sampling so as to include the diversity over which one wishes to generalize is a firmly established methodological canon.

Siegel's response is interesting for its utter failure to grasp the point of the objection.^[11] Siegel seems to think it amounts to something like this: Say a group of geologists is gathering a sample of volcanic rock in order to generalize about its density. They obviously had better make sure that they collect their sample from different volcanoes, located in different parts of the world, that erupted at different times, and so forth. Collecting the sample in this way will necessarily (?) add to epistemic goodness relative to the situation as described (one involving a community of competent scientists observing the rules of good scientific practice). As I understand Siegel, what will not necessarily add to epistemic goodness relative to this situation is merely including the perspectives of more people (like me, for example).

This misses the point because the objection has to do with social science, the Kohlberg-Gilligan controversy specifically, and with sampling human beings. In particular, it is inclusion in the sample, not (directly) in the community of scholars, that is at issue. Viewed as a claim specifically about social research, the objection goes like this: Excluding the voices of women through biased sampling is an epistemic flaw because generalizing about what and how human beings think, feel, and reason in general requires that the voices of women be represented. Thus, inclusion in this context is an epistemic good, and this case can be easily extended so as to apply to other hitherto excluded groups. Accordingly, Siegel's blanket claim that inclusion is not an epistemic good is mistaken.

We might well get another concession from Siegel to the effect that, yes, inclusion is indeed an epistemic good in this sense. But he might (likely would) contend that, epistemologically (versus morally) speaking, the requirement for social scientists to include a diversity of human beings in their samples is really no different from the requirement for geologists to include a diversity of rocks in theirs. This brings Siegel's moral/epistemic distinction to the fore once again, particularly regarding its implications for the special case of social research.

3. Inclusion as a General Requirement in Social Science. Siegel nowhere explicitly distinguishes social science from natural science, but it is with respect to social science where he seems to concede most. For example, social science is one instance of the "sort of circumstances" in which inclusion may be an epistemic virtue (but only vis a vis the conduct of inquiry). In another place Siegel remarks, "For many discourses - perhaps most - everyone is qualified. This is especially true of discourses concerning social and political values and practices, the outcomes of which affect everyone."^[12] Perhaps social research in general is (or ought to be) one such discourse. Perhaps it in general exemplifies the "sort of circumstances" in which inclusion is an epistemic virtue. I take up these suggestions below. To set the stage, I first briefly revisit the problems engendered by Siegel's moral/epistemic distinction, and point to the ironic conclusion to which it leads.

Consider the following situation: A group of theoretical physicists are engaged in a conversation concerning some esoteric topic about which I have no knowledge beyond merely having heard of the issue, say, string theory. Now consider the question of the obligation they have to include me. In my view, these physicists have no moral obligation (beyond mere politeness) to include me precisely because I have nothing to contribute epistemically.

Interestingly, Siegel seems to agree that it is legitimate for me to be excluded. He holds that, in general, "participants in...conversations can surely distinguish between those who clearly can, and those who clearly cannot contribute meaningfully to their conversations, and are surely within their rights to prefer the former."^[13] There is a related point on which Siegel and I (interestingly) agree: The moral obligation of inclusion is a result of a person's or group of persons' interests being at stake. As he puts it, "inclusion should be embraced as a conversational ideal because it is morally wrong to exclude people from, or silence them in, conversations in which they have an interest or stake."^[14] These grounds for inclusion are not satisfied in the physicist example. I not only don't know (or care) much about string theory. The nature and results of the controversy are far removed from, and are not likely to impinge on, my day-to-day concerns.

I say Siegel and I "interestingly" agree on these points because his moral/epistemic distinction would seem to lead in a different direction. In particular, shouldn't epistemic goodness and moral goodness be kept distinct? Where epistemic goodness and moral goodness are kept distinct, shouldn't the moral virtue of inclusion always come in to play, independent of epistemic concerns? Don't the physicists, then, have some free floating moral obligation to include me?

Regarding the role of interests, Siegel would apparently put them on the moral side of the moral/epistemic distinction, such that an individual or group can have a stake in some conversation and yet not be able to contribute anything to it, epistemically speaking. Although he doesn't explicitly say this, he provides no alternative account of the role of interests. Furthermore, if there were (epistemic) grounds for deciding who really has a stake (whether they claim to or not), interests would be epistemically laden and would serve as a bridge, as it were, between the moral and the epistemic.

Placing interests on the moral side of the moral/epistemic distinction implies that anyone who claims to have an interest does have an interest. It also buttresses the notion of a free floating obligation to include persons, whatever their epistemic credentials. It is more than a little ironic that this is where Siegel's argument leads.

AN ALTERNATIVE VIEW

On the heels of these polemics, I owe Harvey Siegel some alternative account, in which I, too, take some risks. The general question of the epistemic status of moral/political claims, clearly at issue, is far too huge to take on here. So, I confine myself to how social science figures into this. I also confine myself to showing that inclusion is an epistemic good, since whether it is a moral good is not in dispute. I will reiterate several points I have already made, but will reinforce them from within a more clearly articulated framework.

Lorraine Code^[15] captures the crux of the issue as well as anyone, and her view is an especially good one to consider because Siegel picks it out as exemplary of the kind of view he wishes to reject. The crux is this: How people know and know about other people is a peculiar but genuine form of knowledge. In this way, Code is an advocate of the kind of social science that Rabinow and Sullivan associate with the "interpretive turn."^[16] Code adopts a hermeneutical form of knowledge for social science and, along with it, advocates "overcoming epistemology" in its traditional form^[17] - a form that she identifies with working out necessary and sufficient conditions for the schema "S knows that p."

Code identifies this schema with the paradigm of knowing in physics, and contends that it wrongly presupposes that S can be wholly neutral, can be without any peculiar "subjectivity." But the problem is much more acute in social science, for the objects of inquiry, the p's, also possess subjectivities, and this helps account for why knowing other people is a very different thing from knowing, say, "it is a hot sunny day." Code suggests that given the way human beings develop and conduct their day to day lives, this may be the most fundamental kind of knowing. At any rate, she observes, "[I]t is certainly no more preposterous to argue that people should try to know physical objects in the nuanced way that they know their friends than it is to argue that they should try to know people in the unsubtle way they often claim to know physical objects."^[18]

Now, it is a short step from this interpretivist conception of knowing other people to seeing inclusion as an epistemic virtue, both with respect to inclusion in the community of scholars and with respect to inclusion in social science samples. All that needs to be granted is that members of a scholarly community wouldn't (couldn't) know and know about identifiable individuals and groups outside their community without engaging them in dialogue. (This, it should be noted, does not require invoking different "conceptual schemes." I just take it to be a fact, pervasive in human experience. It is, of course, possible to come to know about others second-hand, on the basis of accounts that employ an interpretivist methodology.)

Take the case of the historic exclusion of women from social scientific communities. Code suggests that as a result of white men dominating these communities and not engaging in dialogue with anyone but themselves, a certain set of assumptions and practices has evolved such that women should be included so as to serve as an "experimental control." Recall that Siegel concedes inclusion may be an epistemic virtue in this sense, which he then identifies with "rules governing the conduct of inquiry." As I observed before, this is an ad hoc way for Siegel to save his thesis, for it is by no means clear how inclusion may be distinguished from other methodological principles on these grounds. And Code adds a particularly insightful rejoinder in this vein when she observes that there is a certain selectivity in labeling and rejecting accounts of knowledge as "genetic" and therefore epistemologically irrelevant. She writes:

"Science has proved..." carries a presumption in favor of it reliability because of its objectivity and value-neutrality - a presumption that these facts can stand up to scrutiny because they are products of an objective, disinterested process of inquiry. (It is ironic that this patently "genetic" appeal - that is, to the genesis of cognitive products in a certain kind of process - is normally cited to discredit other genetic accounts!)^[19]

The idea that a community of scholars wouldn't (couldn't) know and know about identifiable individuals and groups of individuals without the benefit of an overarching interpretivist methodology^[20] requires that they be included in social research samples for reasons different from those that require our geologists to include a representative sample of volcanic rock. Interestingly, if the social science case were faithfully modeled on the geology case, then, contra Siegel, there would be no moral requirement whatsoever to include a diversity of persons in the conduct of social science; the reason for including them, like the reason for including a diversity of rocks, would be exclusively epistemic.

Again, the moral requirement for including a diversity of persons cannot be detached from the fact that they know things, in particular, what their lives are like, what their interests are, and how both might be at stake in the practice of social science. I don't mean to suggest by this that individuals or groups are infallible authorities on their lives or even their interests. (There is something, after all, to the concept of "false consciousness.") But it is surely unwarranted to conclude from this that how people see themselves and define their interests has no epistemic value.

By comparison to physics, social science practiced in line with the "interpretive turn" is messy and controversial, and its findings are highly tentative. But this is a feature of our epistemic condition with which we can be content, provided, that is, we can settle for a "better chance" of obtaining the truth rather than requiring a "guarantee." In my view, it is not as if there is a choice.

CONCLUSION

To briefly recap, the notion that inclusion is exclusively a moral virtue implies that it should operate independent of epistemic virtue. If not, then moral and epistemic goodness are not exclusive after all. But Siegel seems to want to have it both ways: to separate strictly the moral and the epistemic and yet to hold that individuals may be legitimately excluded from conversations for lack of epistemic credentials.

When confronted with the methodology of social science, which provides an especially strong challenge to his thesis, Siegel takes refuge in the distinction between "rules governing the conduct of inquiry" and "criteria for judging its products." He then allows that in "certain sorts of circumstances" inclusion may be an epistemic good after all, at least vis à vis the conduct of inquiry. As it turns out, the circumstances in which inclusion provides epistemic benefits in this sense turn out to be pervasive indeed. (In addition to social science, history would qualify, and there are no doubt other examples as well.) Moreover, the distinction between rules governing the conduct of inquiry and criteria for judging its products looks to be little more than an ad hoc maneuver, employed by Siegel to save his teetering thesis.

The alternative to Siegel's thesis goes like this: There is a moral obligation to include people because they know things and have interests associated with these things - things that the traditional powers that be don't (and can't) know so long as they conduct inquiry in a way that keeps them insulated. Were inclusion exclusively a moral obligation, detached from any claims to knowledge, it is difficult to see why it should not apply universally. On the other hand, it is also difficult to see how it could amount to anything more morally substantive than polite regard - or disregard - for what people have to say.

True, inclusion does not "guarantee" epistemic goodness all by itself. Here, it seems, Siegel and I are agreed.

***For a response to this essay, see Siegel.

---

[1] Harvey Siegel, "What Price Inclusion?" in Philosophy of Education 1995, ed. Alven Neiman (Urbana, Ill.: Philosophy of Education Society, 1996).

[2] Ibid., 4.

[3] Both respondents to Siegel's paper; Kathryn Pauly Morgan, "We've Come to See the Wizard! Revelations of the Enlightenment Epistemologist" and Sharon Bailin, "Inclusion and Epistemology: The Price is Right," allude to this problem in Siegel's view, especially Bailin. I develop it in a more sustained fashion.

[4] The positivists' "emotive theory" is, of course, the most notorious example of this kind of view.

[5] This point is explicitly noted by Bailin, "Inclusion and Epistemology."

[6] Siegel, "What Price Inclusion?" 4.

[7] Ibid., 5.

[8] Ibid., 4.

[9] Ibid., footnote 17, 17.

[10] Ibid., 5.

[11] Speaking for myself, at least Siegel missed the point I was making when I raised it at the session in which he delivered his paper.

[12] Siegel, "What Price Inclusion?" 13.

[13] Ibid., 12.

[14] Ibid., 3.

[15] Lorraine Code, "Taking Subjectivity into Account," in Feminist Epistemologies, ed. Linda Alcoff and Elizabeth Potter (New York: Routledge, 1993).

[16] Paul Rabinow and William M. Sullivan, "The Interpretive Turn: Emergence of an Approach," in Interpretive Social Science: A Reader, ed. Paul Rabinow and William M. Sullivan (Berkeley: University of California Press, 1979).

[17] "Overcoming epistemology" is the title and theme of the introduction to Charles Taylor's latest book, Philosophical Arguments (Cambridge: Harvard University Press, 1995). For his earlier, now classic, defense of interpretivism in the social sciences, see "Interpretation and the Sciences of Man," in Interpretive Social Science: A Reader, ed. Paul Rabinow and William M. Sullivan (Berkeley: University of California Press, 1979).

[18] Code, "Taking Subjectivity into Account," 37.

[19] Ibid., 27

[20] I use the qualifier "overarching" because I do not mean to suggest that any and every thing that social scientists do must be explicitly interpretive. Rather, it must be ultimately grounded in an interpretive (or "qualitative," to use the education jargon) framework. See Kenneth R. Howe, "Against the Quantitative-Qualitative Incompatibility Thesis (or Dogmas Die Hard)," Educational Researcher 17, no. 8 (1985): 10-16 and Kenneth R. Howe, "Two Dogmas of Educational Research," Educational Researcher 14, no. 8 (1988): 10-18.