Epistemic trustworthiness is defined as a complex character state that supervenes on a relation between first- and second-order beliefs, including beliefs about others as epistemic agents. In contexts shaped by unjust power relations, its second-order components create a mutually supporting link between a deficiency in epistemic character and unjust epistemic exclusion on the basis of group membership. In this way, a deficiency in the virtue of epistemic trustworthiness plays into social/epistemic interactions that perpetuate social injustice. Overcoming that deficiency and, along with it, normalized practices of epistemic exclusion, requires developing a self-critical perspective on the partial, socially-located character of one’s perspective and the consequent epistemic value of inclusiveness.
In all aspects of our lives, we function, in part, as epistemic agents. Epistemic functioning often, and perhaps always, involves the exchange of epistemic goods, and a shared acceptance of the epistemic norms that guide the practices that yield those goods. Social functioning, then, requires epistemic cooperation; and epistemic cooperation requires trust. My interest here concerns what is required for an epistemic agent to be worthy of that trust. That question is, in part, a question about moral character and the relation between moral character and social practices. But it is also about epistemic character, and the relation between epistemic character and social practices; my focus in this paper is this epistemic dimension of trustworthiness.
I will argue that epistemic trustworthiness is socially inculcated, complex character-based, veritistic virtue that supervenes on a relation among first-order beliefs, and (often tacit) second-order beliefs about one's own and others' epistemic competencies. Since attitudes about individuals' epistemic competencies, in part, determine who is granted full membership in an epistemic community, and who is marginalized or excluded, second-order epistemic attitudes link individual epistemic agency to broader social practices, and to social problems such as sexism, racism, ageism, and so on. That link is evident in what I will call the problem of 'epistemic exclusion on the basis of social location' (or 'epistemic exclusion' for short). This problem is expressed through a frequently normalized pattern in epistemic interaction in virtue of which epistemic authority is granted to, or withheld from, individuals on the basis of their perceived membership in a socially constructed group. Understanding how epistemic agency and epistemic character play into the [End Page 109] social/epistemic interaction that perpetuates social injustice should, in turn, yield some understanding of how epistemic character can be a vehicle for social change.
2. Epistemic Trustworthiness and the Epistemic Principle of Charity
Let us assume, for the sake of argument, that all participants in an epistemic community are morally trustworthy in the sense that they are benevolent, and they sincerely believe what they say, say what they sincerely believe, and behave consistently with those beliefs (when truth and understanding alone are the aim). Imagine that community member S asserts that P. It is in virtue of our assumption that S is morally trustworthy, as defined above, that we experience her assertion as not merely providing the information that 'S says that P'; we experience it as showing that 'S believes that P'. But if we are to engage epistemically with S's asserting that P, more than that assumption of moral trustworthiness is required. We must experience her assertion as a reason for us to believe that P, or at least as a reason for us to consider the possibility that P to be worth taking seriously. That is, normal practices of epistemic interaction and cooperation require that members of an epistemic community typically extend to one another the presumption that they meet some threshold level of epistemic credibility.1 In extending that presumption, we extend what I consider to be an epistemic principle of charity; extending that epistemic principle of charity, therefore, is a 'ground-level' requirement of standard epistemic practices.
The idea that a principle of charity is required for social cooperation is by now quite familiar, through Donald Davidson's work in radical interpretation.2 Davidson's view is that it is possible for me to interpret your speech behavior as meaningful linguistic behavior only if I presume that your utterances express mainly true beliefs, and therefore, only if I presume that we share a body of belief. In his work on testimony, C.A.J. Coady (1992) uses a version of that principle to argue that communication requires that we presume testimony to be generally reliable.3 Although similar in spirit, the use I wish to make of a principle of charity here differs in the following respect: I am not addressing necessary conditions for meaningful linguistic exchange, but rather, I want to consider what makes productive epistemic exchange possible in a linguistic community. My suggestion is that, in order to see one another as potential epistemic partners, or as co-members of an epistemic community, we must presume one another's behavioral expressions of epistemic authority to be more or less reliable. The epistemic principle of charity at issue here, then, involves presuming that others' second-order epistemic evaluative attitudes are, by and large, accurate.
Consider a situation in which we self-consciously decide whether or not to invite an individual to participate with us in a given inquiry. In such a situation, what we would like to know is not only whether she is open, honest, and benevolent. Also, and more importantly for present purposes, we would like to know whether she would behave as though she were epistemically competent in an area in which she was not epistemically competent, not because of a moral failing, but because of an inaccurate conception of her own epistemic abilities. If someone frequently and sincerely claims epistemic authority that she does not possess, she is not worthy of trust, in an epistemic sense; and we should [End Page 110] not believe that P, or even seriously consider the possibility that P, on the strength of her testimony alone. That is, we should not extend to her the epistemic principle of charity. If she usually asserts epistemic authority only when she does possess it, she is worthy of trust, and we should believe that P, or at least seriously consider the possibility that P, on the strength of her testimony alone; that is, we should extend to her the epistemic principle of charity.
In extending that principle of charity to individual A, then, we presume that A generally (or: to a reasonable degree as defined by the above-mentioned threshold) expresses epistemic authority when and only when in fact she possesses that authority. It follows that it is a necessary condition of effective and productive epistemic practice that participants be epistemically trustworthy (in a sense to be more fully defined below), and that engaging in cooperative epistemic activities requires presuming that we, and those with whom we interact, are epistemically trustworthy.
This line of thinking suggests the following as an initial characterization of ideal epistemic trustworthiness. Assume that agent A is open, honest, benevolent, and rational (in the sense that she is not self-deceived, and she successfully, consistently integrates her epistemic self-conception into her behavior). Then,
ET (ideal): A is epistemically trustworthy if and only if A is disposed to behave (when contextually appropriate) as though her epistemic status is S if and only if her epistemic status is S.
ET implies that an (ideally) epistemically trustworthy agent (sincerely) confidently asserts that P only if she knows that P, expresses doubt about P only if she has reason to doubt that P, asserts that P is possible only if P is consistent with her standing beliefs, and so on. But of course none of us is perfect; any given individual will be more or less epistemically trustworthy, at a time, relative to a particular domain or subject matter. A may have a very clear sense of her epistemic competencies in botany, but a very poor sense of her epistemic competencies in reading social cues. Her competence in botany may improve, and her sense of that competence may or may not track that improvement.
Further, different degrees of trustworthiness regarding a given domain or subject matter are required by or appropriate to different contexts. When A asserts that P, the more implausible that P is, or the more that is at stake in the question whether P, the higher degree of trustworthiness we require of A, and the more confident we must be about that attribution before we are willing to accept her claims. So, for example, when you ask, out of curiosity about the food industry in the United States, whether the meal I am serving contains any peanut products, you will be more likely to take my negative reply on trust than you would if you were seriously allergic to peanuts. And if I am trustworthy, in the latter situation I will offer a negative reply with confidence only if I am absolutely certain of my answer. If you know me well enough to know that I am trustworthy, then you know that I would not answer with confidence unless I were certain that P (that is, unless I knew that I knew that P). On the other hand, if you do not know me well, or you have reservations about my epistemic character, you may decline the meal unless you know, independently, that it contains no peanut products. Lacking experience of me as an epistemic agent, you may not accept the meal even after I offer evidence to support [End Page 111] my claim that it is peanut-free.
This suggests that the above characterization of epistemic trustworthiness, ET (ideal), should be qualified as follows:
ET (qualified): A is epistemically trustworthy in circumstances C with respect to domain D, if and only if A is disposed to behave as though A's epistemic status in C with respect to D is S if and only if A's epistemic status in C with respect to D is S.
The role of second-order beliefs here should not be over-intellectualized. I am not claiming that as we form first-order beliefs, accept or doubt others' claims, and so on, we are always deliberately attending to what we are doing, self-consciously assessing whether we are accurately representing our epistemic competencies, or explicitly deliberating about whether or not to extend the epistemic principle of charity to others.4 Usually, we are simply attending to the first-order subject matter. Our epistemic character determines how we do that, usually without attracting attention to itself. So our epistemic activities and attitudes are 'framed' by a usually unarticulated, continually evolving 'sense' of the status of our (also continually evolving) epistemic character with respect to various kinds of issues, in various contexts of inquiry; and inflected by our continual, often tacit, 'sense' of the epistemic status of others (again duly qualified by domain, context, and etc).
In general, then, A's epistemic behavior expresses the degree to which she does or does not tacitly grasp her epistemic strengths and weaknesses at a time, in a context, regarding a subject matter or domain. The extent to which her second-order, epistemically self-critical attitudes are accurate, on average, determines the degree of her overall epistemic trustworthiness. When we extend the epistemic principle of charity to A, or withhold it from A, we thereby express a (usually but not always tacit) judgment regarding the degree to which A is or is not epistemically trustworthy.
3. Epistemic Trustworthiness as a Social Epistemic Virtue
The characterization of epistemic trustworthiness thus far implies that it is both veritistic (as on an externalist account of epistemic virtue) and character-based (as on an internalist account).5 It is veritistic, or truth-connected, on two levels: first, if an agent is (completely, ideally) epistemically trustworthy, then her epistemic behavior6 will accurately reflect her epistemic status. The explicit and tacit beliefs that the behavior expresses must therefore be true. If those second-order epistemic beliefs are true, then when our agent claims to know that P, P must be true (and when she claims that there is reason to doubt that P, there is reason to doubt that P; and so on). ET therefore implies first-and second-order epistemic success, and therefore, a truth-connection between first-order beliefs and their objects and between second-order self-critical referring epistemic attitudes and their objects (viz., first-order beliefs).
That epistemic trustworthiness is character-based is clear from its relation to simpler, clearly character-based dispositions and traits. Consider, for example, excessive diffidence, or excessive self-confidence. Such traits will affect whether or not A remains unsure about P even after significant consideration; or very quickly comes to feel certain that P, even without good reason. In either case, her epistemic trustworthiness is compromised. [End Page 112] But if A is epistemically trustworthy, to some reasonably expected degree, she will have developed a stable character state in virtue of which she is disposed to be confident to the degree appropriate to the circumstances, given her relevant epistemic abilities, regarding what and whom she should believe (doubt, etc.) in what contexts. As a result, she will express herself in a way that reasonably accurately reflects her epistemic status with respect to most domains under most normal circumstances.
Notice that the character states that underlie the dispositions responsible for A's first-order epistemic attitudes also underlie the dispositions responsible for her tacit epistemic assessment of those first-order attitudes. If A is overly confident, she will tacitly believe her unsupported first-order belief to be well-supported, and she will behave as though she possesses epistemic authority which she lacks, with respect to both her first-order beliefs and her judgments regarding other epistemic agents. If she is overly diffident, she will feel that she has reason to withhold judgment regarding P when in fact she possesses good reason to believe that P. She will behave as though she lacks epistemic abilities that she does in fact possess. However, if she is (to a reasonable degree) epistemically trustworthy, she will have an accurate (enough) sense of her epistemic status with respect to the given inquiry, under the circumstances, and her behavior will express it unambiguously. Along with 'the facts' and one's 'faculties', then, one's epistemic character 'simultaneously' determines first-order beliefs, second-order epistemic attitudes about them and the 'fit', or lack thereof, between the two 'levels' of attitude.
When we extend or withhold the epistemic principle of charity, we express (some level of) confidence regarding the epistemic status of our beliefs concerning others' epistemic character and abilities. If we are to be epistemically trustworthy, that confidence must be warranted, just as our confidence regarding our own epistemic status must be warranted. This implies that the above, qualified, definition of ET, is still oversimplified. The degree of our own epistemic trustworthiness regarding propositions in domain D in a given situation depends not only on the accuracy of (our behavioral expression of) our assessment of our epistemic status with respect to D; it further depends on our assessment of the epistemic status of others in relation to D, and our level of confidence regarding that assessment. If I confidently dismiss the views of a D-expert without serious consideration, I reveal my close-mindedness, dogmatism, or, at best, epistemic impatience. I thereby reveal that I cannot be trusted to make sound judgments regarding when to listen seriously to whom about D, or whose input and criticism to seek out as I form my own D-related views.
Hence I can be epistemically trustworthy regarding D in C only if I am disposed (to a reasonable degree) to appropriately extend the principle of epistemic charity to others regarding D in C, and therefore only if my beliefs about others' epistemic character and competence (regarding D in C) are reasonably accurate. And whether or not I am so disposed depends, again, on my epistemic character traits. If tend to be overly generous, deferential or gullible, I may uncritically accept whatever A says, even when she is untrustworthy; if I am stingy, overly suspicious or overly skeptical, I may persist in doubting what a trustworthy A confidently asserts and supports with strong evidence or argument.
Epistemic trustworthiness is a social epistemic virtue, then, insofar as it depends [End Page 113] on appropriate attitudes towards others, as well as toward oneself, as epistemic agents. Another way in which it is a social epistemic virtue is that the character traits and skills required for epistemic trustworthiness are developed over time, through interaction with others in the context of normative social practices. One develops one's epistemic character through practice and habituation, by internalizing norms that determine what degree of confidence is appropriate when, and what degree of epistemic deference or skepticism should be extended toward which others under what circumstances, regarding what kinds of subject matter. This process depends, in part, on learning to 'read' and respond to the evaluative responses one receives from those others to whom one has come, through the same complex process, to extend the principle of epistemic charity. We form, test, and revise our epistemic conception of ourselves and others in part by checking our first-order beliefs against theirs, in light of our conception of their epistemic characters, and what we sense to be their conceptions of our epistemic character. We strive to bring our beliefs in line with the beliefs of those whom we trust and admire, and therefore those to whom we extend the epistemic principle of charity; we expect our beliefs to differ, and possibly deliberately distance them, from beliefs of those from whom we withhold the principle of epistemic charity. And of course the same is true of others as they interact with us.
4. Epistemic Exclusion
In a culture plagued by racism, sexism, and other forms of oppression, most community members have internalized, to varying degrees, pernicious stereotypes regarding the cognitive and moral traits of socially constructed 'kinds' of people. Typically, those stereotypes form an interlocking system in which dominant groups are defined to represent the human 'standard' or ideal, and others are defined in relation to, and, to varying degrees, as inferior to, that standard. For example, traditional gender stereotypes represent men as rational, and women as emotional, where 'emotional' is interpreted as implying 'irrational'; an ongoing legacy of the attempt to justify slavery in the US, is that African-Americans are traditionally represented as less intelligent than European Americans; and the poor are considered to be 'unsophisticated' in relation to those who are upper-middle-class and wealthy. As a result of internalizing that interlocking system of stereotypes, we are trained, to varying degrees, to perceive individuals as members of groups, and in virtue of that membership, as prima facie more or less appropriate recipients of the epistemic principle of charity.
In an unjustly hierarchical society, where individuals are perceived as members of different 'human kinds', then, the epistemic principle of charity will often be extended when it should not, and withheld when it should be extended. Where the pressures of socialization are strong and subtle, even morally trustworthy agents (as defined above) can be expected to fall into this pattern of epistemic behavior to varying degrees, despite good intentions. These pressures will therefore encourage members of dominant groups to develop undue epistemic confidence regarding their own abilities in relation to those of members of other groups, and members of oppressed groups to develop undue diffidence. If you are young and impressionable, and people whom you perceive as authoritative [End Page 114] systematically and confidently treat you as having little cognitive/epistemic potential, it is not unlikely that you will internalize that expectation about yourself. Depending on other aspects of your history and character, your interactions with those people over time may reinforce the resulting deficit in epistemic trustworthiness for both of you: each of you is 'off' in your epistemic assessment of self-in-relation-to-other, and the consequent behavior of each reinforces those assessments. On the other hand, if you are identified with a dominant group, and you witness your role models treating members of other groups, but not of yours, as epistemically inferior, you may well develop unwarranted confidence in your epistemic abilities in relation to theirs.
In a way this skewed use of the epistemic principle of charity is quite natural, because the norms that guide and regulate epistemic practices must rely on presuppositions about the nature of knowers and of knowledge anyway. For example, in many epistemic domains, observation is a preferred source of knowledge of our surroundings. Standard epistemic practices in those domains therefore presuppose that human perception is a more or less reliable source of information about our surroundings. Together with our understanding of perception, those, and other, presuppositions imply that under various circumstances, some perceivers are in a better position to observe objective features of the world than others, and therefore to provide testimony that is more epistemically valuable than others situated differently, in relevant ways. So, for example, if A perceives S's face nearby and in full daylight, and B glimpses S from the side, at a distance, then A's identification of S is (correctly) considered more reliable than B's.
The same example illustrates that, as a matter of standard practice, we take the testimony of others to be a source of knowledge. Exactly why, when, and to what extent testimony is normally treated as a reliable source of knowledge is, of course, a subject of ongoing debate;7 nonetheless it is clear that, in our everyday epistemic practices, we do rely on testimony. As in the perceptual case above, so in other cases of testimony: we make discriminating judgments regarding the different epistemic value of different (actual or potential) testifiers in particular situations, regarding particular domains. In many contexts, for example, we privilege the testimony of adults over that of children, natives over tourists, experts over laypersons, and so on.
Because epistemic norms embody background presuppositions that allow us to discriminate among more and less reliable sources of testimony in a given situation or domain, it is to be expected that if we have internalized pernicious stereotypes about 'human kinds', those stereotypes are likely to work their way into those background presuppositions, and thereby lay the groundwork for the problem of epistemic exclusion. We will be disposed to expect members of some groups to be more, or less, epistemically valuable than members of other groups in particular circumstances or with respect to various domains, and those expectations will inflect our interpretation of their epistemic behavior, and our dispositions to extend, and withhold, the epistemic principle of charity to particular others.
Consider, for example, stereotypes that represent women, and men of color, to be not-fully-rational, in relation to middle-and-upper-class white men. Against the background of those stereotypes, it is as epistemically appropriate to favor white men over women, and men of color, for roles for which rationality is a primary qualification, as it is to [End Page 115] favor the testimony of an eye-witness with excellent vision over the testimony of a myopic witness without corrective lenses. Just as children are not of the cognitive (or moral or psychological) 'kind' to reasonably be considered epistemically competent to serve as political leaders, so, from traditional perspectives, women, and men of color, are not be considered to be of the appropriate cognitive 'kind' for such roles. Usually (but not always), such unjustly differential epistemic receptivity is not deliberate, or even conscious: some people are simply 'seen' as more likely to be epistemically competent in a given domain than others.
Further, if women, and men of color, are tacitly assumed to be not-fully-rational, and are steered away from various fields and positions of authority, then their consequent under-representation in those fields and positions is easily interpreted as evidence that they are (biologically determined to be) cognitively, psychologically, or emotionally unsuited to those fields and roles. If women passionately object to their exclusion, their dissent will seem, to some, to be further evidence of their irrational tendencies. If African American males are assumed to have superior native athletic abilities but inferior cognitive abilities, those African American males who in fact are athletically talented will be encouraged to develop their skills, while their cognitive potential may remain unnoticed, because unexpected, and often, as a consequence, undeveloped. African American males will then be highly visible in athletics and under-represented in the academy and in 'high-profile' professions, thereby reinforcing the pernicious stereotype that normalizes the epistemic exclusion in the first place. Unjust epistemic exclusion on the basis of social location is therefore self-perpetuating: its consequences perpetuate the inequalities that fulfill, and therefore seem to justify, the discriminatory expectations that, in turn, perpetuate unjust epistemic exclusion.
To sum up, where unjust power relations are in play, the link between individual epistemic agency and social epistemic practices forged by attitudes about the epistemic capacities of self and diverse others, creates a mutually supporting 'feedback loop' between a widespread, socially inculcated habitual failure of epistemic trustworthiness, on the one hand, and patterns of epistemic interactions, on the other, which perpetuate those power relations. Since these retrograde patterns of epistemic interaction are structurally in line with normal modes of social/epistemic interaction, they easily become normalized. As a result, their hold on an epistemic community is easily concealed, and difficult to break.
An obvious epistemic consequence of unjust epistemic exclusion is that the community loses the potentially valuable contribution of each excluded or marginalized individual in its epistemic life, which may eventually cause incalculable overall epistemic (and therefore other kinds of) losses to the community. Those marginalized individuals will rarely develop their intellectual potentials to the extent that individuals privileged by their social location will; since educational and other resources will not be as readily available to them, they may not have developed the 'know-how' required to use resources that do become available, and in some cases their own and others' expectations will dissuade them from trying to use those resources. To the extent that the quality of education available to community members depends on their socio-economic status, and socio-economic status depends in part on education, this problem of exclusion will be reproduced over time and over generations. [End Page 116]
Another consequence, already indicated, is that insofar as an individual colludes in perpetuating the problem of unjust epistemic exclusion, to that extent her own epistemic character suffers. She does not epistemically engage appropriately with the testimony of others, her second-order epistemic attitudes toward herself-in-relation-to-others become skewed, and her epistemic self-assessment does not 'fit' her first-order beliefs. Since her treatment of others conforms to the retrograde pattern described above, it reinforces similar deficiencies in others, with the result that the overall quality of inquiry in the epistemic community, again, is impoverished.
5. Epistemic Implications: The Importance of Multiple Perspectives.
Still another way that epistemic exclusion degrades the epistemic life of a community may not be as obvious initially as those just mentioned. Participating in epistemic activity often requires that we recognize when a particular inquiry requires cooperation among diverse inquirers. For example, if the goal is to analyze well water in a particular location, one need only send off a sample to the lab and wait for the computer analysis. If the goal is to understand the degradation of a particular ecosystem, the investigative team that should be assembled should include, at least, a multi-disciplinary group of natural and social scientists, none of whom has financial ties to invested industries, or obligations to invested political parties or politicians.
If the goal is to understand the social problems of a given community, the investigating group should be socially and socio-economically diverse. The reason for this is not that all members of a given racial group, or gender, or economic class, share beliefs or have the same innate 'way of knowing'. Rather, this diversity is needed because perspectives are necessarily partial, and are shaped through individual epistemic histories. Despite their many differences, those who are situated similarly socio-economically will probably have some kinds of shared patterns of experience in their epistemic histories, and particularly, those due to, or in response to, their identifying with, or being identified with, an unjustly stereotyped group (whether the stereotypes typically benefit, or disadvantage, members of the group). Hence we should expect that, oftentimes, and with respect to social domains in particular, those differently situated will be disposed to ask different kinds of questions, experience the 'same' interactions differently, notice different kinds of patterns, describe what they notice differently, and so on.
For example, many have argued that as science becomes more diverse, different kinds of questions are asked, data once seen as irrelevant become interesting, 'old' data are described differently, their evidential significance is interpreted differently, with ultimate positive results for the growth and development of scientific knowledge and understanding. Relatively recent work in biology suggests that, until the 1980's, reproductive theory tended to reproduce, and naturalize, traditional gender stereotypes by describing ova as "passive" and sperm as "active", thereby ignoring the ("active") role of microvilli on the surface of the ova. This occurred in spite of the fact that the presence of microvilli was first detected in 1895; those data suggesting an "active" role for ova in reproduction had been dismissed as irrelevant, or simply ignored.8 A number of studies [End Page 117] reach parallel conclusions about the influence of traditional gender assumptions in evolutionary theory, microbiology, primatology, and archeology, to name only a few.9 The claim is that, as science has become more diverse, feminist theorists (who are often but not always women) recognize the influence of gender biases shared by all members of earlier, more traditionally-minded research teams, which, because ubiquitous, had gone unnoticed (or, if noticed, were considered to be 'facts', and not biases). Since feminist scientists self-consciously reject traditional gender stereotypes, they are in a position to notice and identify, at least to some extent, the pervasive, distorting effects of those stereotypes on the affected work, and to intentionally re-view the material from a fresh, less restricted perspective.
When the epistemic goal is to understand and create policy in response to a social problem (whether in a scientific community or a community more broadly construed), a breakdown of epistemic trustworthiness of the sort that perpetuates unjust epistemic exclusion should be expected to create a climate in which participants are likely to ignore, or simply 'not see', the need for including the perspectives of members of unjustly stereotyped groups; or to misunderstand and devalue such perspectives when they are included. As a result, often those whose perspectives are most needed in order to arrive at an appropriately rich, comprehensive, insightful understanding of the social group or phenomenon at issue, are the very perspectives those problems themselves marginalize.
Consider the following scenario. A racist incident has occurred on the campus of a small college, and an investigative team of faculty and administrators is assembled to determine whether it is an isolated incident or a symptom of deeper, systemic racism in campus culture. In one discussion, Alice, the only faculty person of color in the group, describes her sense that students tend to challenge her authority more than they challenge the authority of white faculty, and resent her when she corrects them in class. She knows that other faculty of color share those experiences. But when she describes particular classroom situations that she experienced as subtly racist, she feels she is understood only by some of those colleagues who themselves have experienced subtle expressions of other forms of bias (such as sexism, homophobia, etc.) in the classroom. Another, white, faculty person gently reminds Alice that all faculty sometimes experience student disrespect and resentment, and since Alice really cannot know what the students are thinking, it could be that her sense that they disrespect and resent faculty of color more than others may be due more to her own insecurity than to any attitudes held by the students. In the end the group finds that, while there are occasional scattered racist incidents on campus, there is no evidence that racism is endemic to the culture of the institution.
When the group minimizes the significance of Alice's concerns, they assume that they can 'read' students' attitudes and characters better than she can, and thereby assume that they are in a better position to judge what really has gone on in the interactions between her and her students than she is. So they withhold, to a degree, the epistemic principle of charity from Alice. That is, they assume they know better than she does what she could and could not know in that situation. They assume that they are better positioned to make such judgments than she is, since they consider themselves to be uninvolved and so to occupy a more 'objective' perspective on the situation than does Alice. But if Alice is right, of course, students' attitudes toward the white faculty are as biased in favor of [End Page 118] the white faculty, on the basis of their membership in a group, as they are biased against Alice and other faculty of color, on the basis of their membership in a group. From her perspective, the fact that the white faculty experience students as respectful, for the most part, while she and other faculty of color do not, confirms her perception of pervasive, subtle racism in the student body.
Note that the difference between Alice's position and that of her more privileged colleagues is not only that she experiences racism, while they do not experience racism. She sees that she experiences racism that they do not experience, and she expresses concern about this contrast. They fail to see that the difference between their experiences of the students is relevant to understanding the social situation on campus, since they assume that Alice's interpretation of the students is incorrect. The fact that they make that assumption indicates that they further fail to see that the contrast between their perception of student attitudes and her perception of student attitudes is precisely what one would expect if her interpretation of the students' attitudes were correct. This illustrates a specifically epistemic aspect of their privilege: not only are they spared the students' contempt, they are spared having to see that it is real, and Alice is not so spared. As a result, they do not experience the kind of epistemic pressure that brings Alice to acquire a self-critical stance from which to see the contrast between her perspective and theirs. That is, they have the privilege of remaining ignorant of their privilege, its effects on their own perspectives, and its effects on the perspectives of those less privileged. Therefore, just as the participation of feminist scientists was necessary to reveal the bias inherent in standard approaches to reproductive theory, so here, Alice's participation is necessary to reveal the bias in the 'normal' student and faculty attitudes. She occupies a standpoint from which it is possible to see the distortion caused by taking the 'dominant' partial perspective to represent 'the objective point of view'.
Let us look more closely at Alice's situation. She has experienced racism all her life; is well-educated; has been teaching for several years in predominantly white institutions; is well-read in literature regarding race and racism; and may well have worked with a therapist to learn strategies for coping with being a person of color in a white-supremacist society. She has become self-critical regarding the possibility of over-interpreting the racial content of interactions, and self-aware regarding the ways in which the effects of racism sometimes distort her own perspective. She realizes that there have been times when she has experienced interactions to be racist when they were not. She also knows that it is likely that what she perceives to be subtle expressions of racism are just that, and that, for the most part, they will not be recognized by her white colleagues.
Alice has achieved this critical second-order perspective, in part, because it has been necessary for her to learn to 'see how things look' from the 'dominant white perspective', precisely because that perspective is typically taken to be definitive and therefore, in some sense, binding. That is, she has developed "double consciousness".10 She is aware that she constantly confronts and negotiates differences between her ways of perceiving situations, including her role in them, on the one hand; and the way that people are expected to perceive them, on the other. Through this awareness, in part, she is able to develop (some admirable degree of) epistemic trustworthiness. Because her most privileged colleagues do not have to negotiate these perspectival differences, they need [End Page 119] not develop the kind of self-critical stance regarding perspectival differences that Alice develops. They can remain, to a degree, comfortably untrustworthy.
Note that there is no need to attribute to the white members of the group overtly racist attitudes or deliberate attempts to sway the outcome of the study, any more than there was a need to attribute explicitly sexist intentions to the cell biologists who assumed that ova are "passive". Well-intentioned, intelligent, epistemically skillful people, including many of us who are white, have unintentionally created or participated in situations not terribly unlike the one just described, oversimplified as it may be. We may be complicit in this kind of self-perpetuating pattern of harmful behavior even if we have learned, intellectually, that 'race' is not a biologically significant category, and that our culture has a racist history that continues to influence social structures and attitudes today.
The primary difference between Alice and her white colleagues, then, is that she is attuned to the necessarily partial character of perspectives, where that partial character is defined in part by social location; many of them are not so attuned. She is aware of this difference between them; many of them are not so aware. For all of these reasons, she has developed a higher degree of epistemic trustworthiness regarding a range of social phenomena than they have developed; she is a better partner in social inquiry than they are. Yet the very factors responsible for her epistemic privilege in social inquiry are likely to prevent her from full participation in social inquiry.
We have seen that inclusive epistemic interaction among a number of diversely located, intentionally critical epistemic agents has the potential to generate a richer, more comprehensive and more nuanced understanding than does the epistemic work of a single individual or a homogeneous group.11 The question remains: how do Alice's colleagues learn to achieve that kind of critical perspective? How do they overcome the resulting deficit in epistemic trustworthiness in a context implicitly designed to conceal it?
A word of caution is in order at this point. Remember that we are discussing agents assumed to be morally trustworthy and hence well intentioned. The question at hand, then, concerns patterns of specifically epistemic behavior and the underlying internalized interpretive framework into which pernicious stereotypes are logically integrated. Hence we are asking how people can change their 'deepest' social belief system. With this question, we come face to face with the problem of voluntarism: to what extent does it make sense to assume that we can decide what to believe, and effectively will to change our beliefs? Although this problem is beyond the scope of this paper, any adequate response to our current question surely must be sensitive to its presence.12
Certainly it is necessary for Alice's colleagues to engage with empirical evidence that falsifies the stereotypes that conceal their privilege and its epistemic consequences. But providing what some consider to be such evidence is not sufficient to ensure that others will see it in that way, particularly if the stereotypes to be falsified are embedded in the conceptual framework against the background of which evidence will be interpreted.13 'Theoretical information' that undermines pernicious stereotypes also should be provided. But, again, if these stereotypes are deeply internalized, opposing theory can be dismissed [End Page 120] as faulty, disarmed through reinterpretation, or conceptually compartmentalized (as perhaps has been the case with biological theory that denies biological relevance to socially constructed racial categories). Nor will it be sufficient, as in the example above, simply to include a person of color in the investigating team. Inclusivity must be epistemically motivated, that is, motivated by the epistemic value of including multiple perspectives.
In Alice's case, it is likely that her colleagues will try to understand her perspective and will be open to seeing the partiality of their own. Perhaps what is needed is that we make a point of extending the epistemic principle of charity fairly, by formulating rules such as 'do not assume that people in a position of power always are right'; or 'do not assume that people in traditionally marginalized groups are wrong'; or even, 'when someone has something to gain by your believing them (such as maintaining the 'status quo'), think twice about extending the epistemic principle of charity' (as we do, for example, with respect to product safety studies sponsored by the products' manufacturers).14 But these rules need not have the desired effect. One reason is that frequently, people in positions of power are right, and people in subordinate positions may be wrong; social location is not always relevant to epistemic status. Further, because of their social locations, both those in positions of power and those in marginalized positions stand to gain by being taken epistemically seriously. 'Standing to gain by being believed' as opposed to 'not standing to gain by being believed', is not an appropriate distinction on which to formulate a criterion on the basis of which to extend the principle of charity.
What about the rule 'be wary of extending the principle of charity to those who stand to gain by remaining ignorant about the issues under discussion'? This seems more promising. But those whose perpetuation of unjust exclusion is most formidable, of course, are those members of dominant groups who are unaware of their ignorance. They are untrustworthy in virtue of their confidence in their epistemic assessments of others and themselves. This suggests that we have been going about the project of formulating rules in the wrong way: rather than trying to formulate rules that guide us in identifying situations in which we should or should not extend the principle of charity, we may need to take a more Aristotelian, virtue-theoretic approach and formulate rules that guide us in altering the very epistemic character traits that are introduced, or at least reinforced, by unjust epistemic exclusion in the first place.
This change in approach also avoids potential objections that rise from concerns about doxastic voluntarism: we may not have control over what or whom we come to believe, under what circumstances, but we can decide to take control, gradually and to some degree, of the process by which we change our 'habits of mind'. One whose privilege may engender false confidence should, in the words of Uma Narayan (1988), adopt a principle of "methodological humility", that is, such a person should "always sincerely conduct herself under the assumption that [in virtue of her privilege] … she may be missing something" (1988, 38), especially in discussions of social oppression.15 This procedural rule would flag situations in which she is disposed to see someone who doesn't share her privilege as making a 'mistake' (as in the case of Alice's perceived misinterpretation of her students' attitudes), as situations in which she should deliberately resist that disposition. In contrast to the act-based rules mentioned above, accepting the character-based rule of abiding by the principle of methodological humility does not require that we have [End Page 121] already made the conceptual shift that this methodological principle is supposed to bring about. Rather, it requires good will, a recognition that one belongs to a privileged group, and a recognition that our 'automatic' responses to others are in part conditioned by internalization of norms through socialization.
I am not as confident of my ability to discern what is methodologically advisable for breaking the cycle of epistemic exclusion on the part of those in less privileged groups. A principle of methodological courage may seem appropriate, that is, a principle such as "resist the tendencies of those in positions more privileged than yours to dismiss your point of view as less epistemically valuable than yours." However, while that principle may seem warranted from an epistemological point of view, it may not necessarily be advisable from a psychological or emotional point of view. This is because following it could create psychological and emotional pressures in addition to those already experienced by those likely to be subject to unjust epistemic exclusion.16
I have argued that where social problems exist, the link between epistemic character and epistemic/social practices creates a mutually supporting feedback loop between a widespread, socially inculcated habitual failure of epistemic trustworthiness, on the one hand, and the pattern of epistemic interaction that results in unjust epistemic exclusion, on the other. The discussion above shows how that same link instead can become a vehicle for a mutually supporting, epistemically constructive feedback loop between the inculcation of epistemic trustworthiness, on the one hand, and inclusive epistemic behaviors that would displace unjust epistemic exclusion. Taking advantage of this possibility requires self-conscious cultivation of the epistemic dispositions most directly operative in epistemic trustworthiness, initially by means of methodological guidance from principles such as the principle of humility. Intentional adherence to such principles would make possible the openness necessary for genuine epistemic engagement with the various forms of empirical evidence and theoretical understanding that reveal and undermine the pernicious 'human-kinds' stereotypes driving the practice of unjust epistemic exclusion.
Nancy Daukas is Associate Professor of Philosophy at Guilford College in Greensboro, NC. Her research interests span various areas of epistemology and feminist philosophy, and her publications include “Classroom Relativism as a Pedagogical Opportunity” and “Skepticism, Contextualism, and the Epistemic ‘Ordinary’.”
I am deeply indebted to Alison Wylie for her extensive and insightful suggestions and criticisms of earlier drafts. Many thanks also to Alvin Goldman and Jim Brown for helpful comments on an earlier version of this paper. Thanks also to the participants of the June 2006 Episteme Conference at the University of Toronto for stimulating and insightful comments and criticisms. Finally, I wish to express my appreciation to Lisa McLeod, Vance Ricks, and Jonathan Malino for continuing discussion of some of the ideas behind this paper.
1. Thanks to Lorraine Code for suggesting this way of articulating this point.
2. See Davidson (1983), and essays in Davidson (1984), esp. “Belief and the Basis of Meaning,” “Radical Interpretation,” “The Method of Truth in Metaphysics,” “Thought and Talk.”
3. See Coady (1992), esp. chapter 9.
4. This is one reason that Sosa’s attempt to integrate second-order epistemic attitudes into an account of epistemic agency in (1991), Part IV, is unsatisfying.
5. In this way my view shares company with Zagzebski’s, in (1996). My most important disagreement with Zagzebski concerns the status of justification, for reasons not relevant here.
6. ‘Epistemic behavior’ should be understood broadly: it includes not only speech acts asserting knowledge claims, but ‘acting as though one knows (doubts, wonders, suspects, etc.)’. It includes behavior such as asking, or not asking, a question; answering, or not, a question; going to look for something, or not; attending a class or seminar, or not; etc.
7. For a review of the terrain, see Goldman (2002) chapter 7.
8. The Biology and Gender Study Group (1988). See also Martin (1991).
9. The relevant literature is vast; see, for example, Bleier (1984), Fausto-Sterling (1992), Hubbard(1990), Longino (1990), Longino & Doell (1983), Spanier (1995), The Gender and Biology Study Group (1988), Wylie (2002), chapter 14.
10. DuBois, W.E.B. (1994), p. 2. Thanks to Lisa McLeod for this references, and for helpful discussion of DuBois’ remarks on ‘double consciousness’.
11. Compare to the idea of ‘strong objectivity, and, more generally, ‘feminist standpoint epistemology’. See Harding (1991), chapter 6, and (1993). For representative and recent work in Feminist Standpoint Theory, see the collection of essays in Harding (2004), including Haraway (1991), Collins (1986), Harding (1991), Hartsock (1983), Wylie (2003).
12. See Zagzebski (1996) for a clear discussion of the problem of voluntarism.
13. Here I take issue with Fricker’s suggestion in (2003) that pernicious stereotypes can be isolated and rejected as empirically unjustified. In that article, Fricker argues in favor of a virtue-theoretic approach to what I am calling the problem of unjust epistemic exclusion, and emphasizes the importance of becoming a virtuous hearer. With respect to those points, I am in complete agreement with Fricker.
14. Here I discuss a suggestion made in conversation by Elizabeth Anderson.
15. Many thanks to Alison Wylie for introducing me to this piece of Narayan’s, and particularly for pointing out the relevance of Narayan’s principle of ‘methodological humility’ to my project.
16. See Narayan (1988) for a discussion of the emotional price of epistemic injustice for those who experience it, as well as the potential epistemic value of one’s emotional response to experiencing it.