Objectivity, Diversity, and Uptake: On the Status of Women in Philosophy

This paper argues that diversity and uptake are required for objectivity. In philosophy, women are underrepresented with respect to teaching, publishing, and citations. This undermines the objectivity of our research output. To improve women’s representation and objectivity in philosophy, we should take steps to increase women’s numbers and institute uptake-conducive conditions. In concrete terms, this means fostering an appreciation for diversity, diversifying evaluators, integrating women’s contributions into mainstream discourse, and reducing implicit bias.


Introduction
It's well known-and empirically verified-that women are underrepresented in philosophy, with respect to teaching, publishing, and citations. This obviously isn't an ideal state of affairs. There's a good case to be made that the gender gap is unfair and harmful to women. These are moral concerns. But there's another reason to disapprove of the situation, though it might be less obvious. Namely, the underrepresentation of women in philosophy undermines the objectivity of research in the field, because diversity facilitates objectivity. This is an epistemic concern. If this concern is legitimate, then as researchers we have a shared interest in reducing the gender gap, since we're invested in objectivity as a research goal.
In this paper, I'm going to expand on why we have an epistemic motive to promote women's representation in philosophy. In section 2, I'll present some empirical research that quantifies the gender gap in different areas of philosophy. In the remaining sections, I'll offer suggestions as to how to rectify this problem, beginning with more theoretical considerations, and moving toward concrete solutions.
In section 3, I argue, following standpoint feminist epistemologists, that the solution to gender inequality isn't to strive for value-neutrality, since valueneutrality isn't possible. All knowledge is value-laden. But not all values all equal; we do need a means of discriminating between constructive values and destructive ones. Hence, in section 4, I defend Sandra Harding's (2015) proposal that diversity facilitates objectivity, that is, it produces 'stronger objectivity' (versus 'weaker objectivity'). I call this the 'Strong Objectivity Thesis' (henceforth, SOT). After illuminating this principle from Harding's perspective, I suggest refinements to increase its plausibility. Specifically, I draw on empirical research on workgroup task performance to show that diversity does in fact enhance objectivity (in terms of optimal task performance), but this outcome is conditional upon 'uptake-conducive' background conditions-conditions like reduced implicit bias, reduced outgroup bias, and positive diversity mindset. When these epistemic conditions are absent, diversity actually tends to decrease task performance. This analysis sheds light on how diversity is related to objectivity, and what we can do to enhance objectivity in philosophy: increase both women's representation and receptivity to women's contributions.
This strategy can be implemented individually and collectively. In section 4, I discuss cognitive biases that undermine individual receptivity, and useful remediating strategies for addressing these biases. Then I outline institutional safeguards that collectives can implement to reduce the effects of cognitive biases and distorting group dynamics, borrowing here from Helen Longino's criteria of 'epistemic effectiveness ' (2001).
Then, in section 5, I elucidate what this means for philosophy departments. Specifically, I offer suggestions for how we can achieve epistemic effectiveness (objectivity) by taking such measures as appreciating contributions from those at the margins of the discipline, implementing specific affirmative action policies, and cultivating alternative research communities, while ensuring that their output is eventually integrated into mainstream discourse. These aren't the only possible strategies for increasing gender parity in philosophy, but they're supported by Harding's and Longino's proposals, in conjunction with the empirical research on cognitive distortions, group dynamics, and group reasoning that I cite in sections (1)-(4).
Before delving into this argument, a couple of preliminary remarks would be useful. Although this paper is on the status of women in philosophy, everything I say here should be taken to apply, mutatis mutandis-and perhaps even a fortiori, since women are a relatively privileged subgroup-to other underrepresented groups. Second, when I cite women in this paper for the first time, I'll include their full name, and I'll also 'star' their name with an asterisk in the references section, to make it easier for readers to identify them and cite them in future publications. This is consistent with the ethos of this paper, i.e., boosting women's representation in the field.

How Are Women in Philosophy Underrepresented?
There's ample evidence that women are underrepresented in philosophy. Rather than citing reams of data here, I'll just describe some of the more illuminating studies.
Eric Schwitzgebel and Carolyn Dicey Jennings recently conducted a quantitative analysis of faculty lists from various sources, and found that (1) gender disparity remains large in mainstream Anglophone philosophy; (2) ethics, construed broadly to include social and political philosophy, is closer to gender parity than are other fields in philosophy; and (3) women's involvement in philosophy has increased since the 1970s . . . [but] by most measures, women's involvement and visibility in mainstream Anglophone philosophy has increased only slowly; and by some measures there has been virtually no gain since the 1990s. (2017,83) This is a recent finding, which is corroborated by other research. Reports place the percentage of women in postsecondary instructor positions at 19%-21% (Molly Paxton, Carrie Figdor, and Valerie Tiberius 2012;Julie Van Camp 2015Sally Haslanger 2008Kathryn Norlock 2006). In addition, women comprise 26% of parttime instructors but only 16.6% of full-time instructors (Van Camp 2015). The gender gap in philosophy is much worse than in the rest of the Humanities (Norlock 2006;Paxton, Figdor, and Tiberius 2012), so we can't take solace in the idea that this is a universal problem-it's very much a philosophy problem.
Women are also less cited and credited than men: Only nineteen of the five hundred most-cited articles in four top-tier journals (JP, Mind, PR, and Nous) were by women; a single male philosopher-David Lewis-had twice as many citations as all of the women combined (Healy 2015). Women receive less 'extended discussion' than men, as measured by citations appearing more than once in a paper's abstract; only 13% of the philosophers receiving extended discussion in the Philosopher's Index since 2010 were women (Schwitzgebel 2015). According to Helen de Cruz (2014), this means that philosophy journals wouldn't pass a philosophical version of the Bechdel Test, which requires that two women in a film talk to each other about something other than men. It seems that women aren't talking to or about each other in philosophy very often, and men aren't helping much.
The upshot is that there's a gender gap in philosophy, and it's not a trivial one. This is wrong for all kinds of reasons, including that it's unfair and harmful to women. These might be the more obvious reasons for increasing women's representation. In this paper, I want to focus on a (perhaps) less obvious, but in a sense more pressing, problem: that the lack of gender parity undermines objectivity in the field. This claim might not be obvious because it relies on some non-obvious assumptions about the nature of knowledge and its production-assumptions that we find in standpoint feminist epistemology; but it's an especially pressing (indeed, inescapable) worry because it's of interest to all philosophers, no matter one's disciplinary focus, since we're all inherently committed to objectivity by virtue of our disciplinary commitments. All philosophers try to produce objective knowledge; so, if the gender gap is undermining objectivity, this is something that we all have reason to care about. We all have an inescapable reason to do what we can to remediate this problem.
In what follows, I'll elucidate what we should and shouldn't do to improve the situation. It might seem as if weak objectivity is the result of implicit value judgments, and so we should try to eliminate all values from philosophical inquiry. Although intuitive to many, this thought is mistaken, because all inquiry is inherently value laden, and some values are actually useful. In section 3, I'll explain why inquiry is necessarily value-laden, and in section 4, I'll explain what 'strong objectivity' amounts to, why it's hard to obtain, and what we can do to promote it.

All Knowledge Is Value-Laden: Why Not to Pursue Value-Neutrality
It might seem like a good strategy to expunge all values from philosophical inquiry, thereby making it impartial and 'objective.' This is an especially tempting strategy given that implicit bias is a type of value that undermines objectivity (more on which shortly). Yet this seemingly intuitive strategy is mistaken: all knowledge is inherently value-laden, as standpoint feminist philosophers have taken pains to show.
As Anne Jaap Jacobson points out, even basic perceptual representations are inherently normative; the mind constructs a "somewhat fictive" and "highly selective" image out of successive scenes, encoded as the eyes saccade, and memory then stores a version of this selective impression (2008,597). Thus, even perception isn't value-neutral; it's informed by an individual's (typically implicit) selection criteria. While visual encoding is subjective, memory adds yet another layer of subjectivity: only certain features of the initial visual impression are stored, and these are typically altered in subsequent retrievals. This is one reason for thinking that inquiry is value-laden-or as holists put it, that 'theory is underdetermined by data.' Even our basic models of the world are informed by subjective (implicit) norms, and they can be altered depending on our shifting needs, preferences, and experiences.
This observation is connected to a more general claim advanced by normative philosophers and feminist standpoint epistemologists, namely, that the 'view from nowhere' (Nagel 1989) or 'God trick' (Donna Haraway 1988) is impossible. That is, we can't observe the world from a value-neutral, disembodied perspective, since all knowledge is 'situated' and incomplete (Lorraine Code 1991;Harding 1993;Longino 1993). The kernel of this idea can be traced back to Quine's view that all knowledge is underdetermined by evidence, though it has been further developed in recent years by the likes of Lynn Hankinson Nelson (1993), Elizabeth Potter (1993), Longino (1993Longino ( , 2001, and Harding (2015). Quine's view is that all knowledge is partly based on implicit norms, which comprise the 'background assumptions' of a research community. These norms include priority, simplicity, elegance, and conservativism.
To this basic thesis, we can add three contributions from standpoint epistemologists.
First, standpoint epistemologists submit that a broader range of norms than Quine envisioned make up part of our background assumptions. Salient examples include material and cultural values, or what Harding calls the "local material realities" of a particular culture (2015,125). These norms help guide research projects in such a way that they are both useful and epistemically relevant to human beings. As Rawls famously observed, we don't expend time and energy counting blades of grass on well-trimmed lawns (1999,432). This would be a waste of time and resources. We ought to take into account our needs and interests as situated human beings before undertaking any research project. These normative values shouldn't be expunged from inquiry-they often shape research projects in positive ways. But we should take care to ensure that our normative assumptions are robustly justified, not of excessively narrow or merely academic interest.
Second, feminist epistemologists such as Nelson (1993) and Potter (1993) have observed that the central producer of knowledge is not, as traditionally conceived, the epistemic agent-historically, the 'Cartesian ego,' who discovers knowledge through unassisted rational reflection, without consulting anyone else. Rather, the primary source of knowledge is the epistemic community in which the individual is ensconced. Potter holds that the community is "epistemically prior" to the individual, since it's the source of many of the individual's beliefs; thus, it partly "underwrites" individual knowledge claims (1993,165). So, if the community is epistemically flawed, the individual probably is too. Since individual beliefs are often implicit and inherited from the community, community values often constitute unrecognized evidence for those beliefs. This evidence can't be wholly explicated and evaluated from an individual perspective.
Third, individual beliefs are (at least partly) the result of interpersonal reasoning and could not exist apart from this process; so, engagement with other people is necessary for certain kinds of knowledge to exist. In line with this, Potter shows that research communities' standards are virtually always the result of "epistemic negotiations" amongst two or more members (1993,169). To illustrate this, she transcribes negotiations amongst scientists verbatim-for instance, a debate between two scientists about what quantity of peptides suffices to prove a psychobehavioural hypothesis (1993, 168)-and she parses them for normative content. These 'microexchanges' prove to be inherently normative, but their normative character may be opaque to the researchers engaged in them: they're part of the group's implicit 'background assumptions.' Normative negotiation is a critical, and often helpful, part of research activities, and we should not seek to eliminate it; but we have a responsibility to assess the broader value of these norms-perhaps by soliciting feedback from other researchers or the public.
This illustrates that research is inherently value-laden, and trying to expunge it of all values would be pointless and possibly detrimental. But not all values are equally justified: we need a way to discriminate between productive values and unproductive ones (such as implicit biases based on someone's demographic attributes). And we obviously can't use Descartes's standard of 'indubitability' to do this.
In the next section, I explain Harding's theory of 'strong objectivity' and her proposal for how to achieve it (via diversity), and then I suggest refinements to this proposal in light of emerging empirical research. In particular, I emphasize that diversity must be paired with 'uptake' or proper receptivity in order to be effective.

Objectivity as Diversity Plus Uptake-Conducive Background Conditions
Harding argues that insufficiently diverse epistemological communities can't adequately reflect on their shared assumptions, and this undermines the objectivity of their research output. Too often, we simply reproduce the biases of our community, even if we have internal methods for weeding out substandard work: The procedure of [researchers] repeating each other's research processes certainly does work well to identify those values, interests, and assumptions that differ between individuals or teams of researchers. But in cases where social values, interests, or assumptions are shared by all or virtually all researchers in a given field-as has been the case for male supremacy, white supremacy, or Eurocentrism, for example-repeating observations within such a field will not bring shared social commitments into focus. (Harding 2015, 34) Not all of these biases afflict contemporary academic contexts, but many do. Haring notes that economists typically define 'work' so as to exclude predominantly female labour, such as housework, 'caring labour,' volunteer work, and sex work (2015,28). This exclusionary definition distorts the reality of labour in our society, particularly the fact that 'female' labour is essential to modern capitalism-the economy would be unsustainable without it. This economic assumption may also support the existing wage gap (women make about 80 cents for every dollar men make, according to the US Census Bureau), by giving it the false veneer of intellectual respectability. This is just one of many examples of objectivity-undermining bias offered by Harding, pointing to a fairly sweeping problem.
Does implicit bias affect philosophical research? There's a case to be made that top-tier philosophy venues have exhibited (and perhaps still do) a bias against feminist content and critical race theory. Sally Haslanger notes that from 2003 to 2008, "Mind, Philosophical Review, and Philosophy and Phenomenological Research did not publish any articles with feminist or race content," and the journals Ethics, Journal of Philosophy, and Philosophy and Public Affairs collectively contained only 2.36% feminist content (2008, 220). It's not plausible that feminist philosophers did so little valuable work during this period. The likelier explanation is that implicit bias and stereotype threat played a role, as Louise Antony (2012) suggests in her 'perfect storm' account of gender disparity in philosophy. These factors might affect not only hiring selections but publishing selections, with editors and referees harbouring implicit bias against certain kinds of content, and feminist philosophers exhibiting some degree of stereotype threat or lack of confidence in their publishing ability. Another possible factor is status quo bias, the tendency to favour the familiar; if feminist philosophy is less familiar to most people, it's likely to seem inherently less plausible (regardless of its objective value). It's well known that achievement is partly a matter of 'standing on the shoulders of giants'; unfortunately, feminist philosophers and critical race theorists just don't have as many shoulders to stand on. This is a historically rooted problem, but not a problem that we're helpless to change.
Returning to Harding's proposal: she holds that non-diverse groups lack 'strong objectivity.' Strong objectivity, in precise terms, consists of fairness to the evidence, to one's critics, and to "the most severe criticisms one can imagine" (Harding 2015, 33). I'll call this the principle of epistemic fairness (PEF). Epistemological communities that fail to satisfy PEF (pro tanto) lack strong objectivity. They achieve at best 'weak objectivity'-objectivity relative to a narrower range of perspectives.
Harding's notion of strong objectivity provides a valuable tool for evaluating the epistemic authority of a given research community: we can ask if it satisfies PEF. But her proposal has some loose ends. Tying these up would help to strengthen her position, lending it stronger objectivity. In what follows, I'm going to address 5 potential sources of scepticism or confusion about SOT. A sceptic might ask one of the following questions: (1) Why is it so hard to achieve strong objectivity? (2) What's the nature of the relationship between objectivity and diversity? (3) What degree of diversity should we strive for? (4) How do we effectively imagine 'the most severe criticisms,' and is using our imagination enough? (5) How do we implement PEF in institutional contexts? I expect that these kinds of questions are likely to occur to people who are sceptical of SOT-both those who think it's too strong and those who think it's too weak. Answering these questions is in the spirit of Harding's work, since it increases the intelligibility of this thesis to a broader audience.

Why aren't we objective? The role of cognitive distortions in human reasoning.
Harding nowhere in her work explicitly mentions 'cognitive distortions' or 'cognitive biases,' although she talks about "fundamental assumptions" (2015,19), "dominant assumptions" (28), "assumptions, attitudes, aspirations, and antipathies" of research communities (x), and so on. When we view Harding's proposal through the lens of earlier work in feminist epistemology (especially Potter 1993), it's natural to see these 'fundamental assumptions' as part of the shared background norms of a given research community. Seeing them in this light lends clarity to Harding's proposal; but we can gain still more clarity by identifying some of the more pernicious types of 'background assumptions'-those that we need to be especially wary of. It's reasonable, in light of contemporary psychological research, to think that some of these assumptions are implicit states, that is, relatively unconscious and automatic features of judgment and behaviour. These states aren't directly introspectible, which explains why they're so persistent. And some of these states have epistemically (and morally) problematic content. We can see this by looking at three examples and considering their likely effects on reasoning. The three states I'll examine are (i) implicit bias, (ii) confirmation bias, and (iii) self-serving bias.
(i) Implicit bias: Implicit biases are relatively unconscious, automatic states that can motivate overt behaviour. The kinds of implicit biases typically studied by researchers have such contents as race, gender, age, disability, sexuality, and weight. The Harvard Implicit Association Test (HIAT), which can be taken online, 2 finds that most people score in the moderate-to-strong range for these biases. These implicit biases are morally problematic because they favour privileged social groups, such as European Americans, men, and thin people, and disfavour disadvantaged social groups, such as African Americans, women, and overweight people. The contents of these biases are thus culturally informed, not 'innate' or 'inborn.' It's not just White people who implicitly favour White people; it's everyone. While more White and Asian people show automatic White preference, 50% of Black people do too, showing that implicit bias is culturally mediated. (Favouritism is gauged by reaction times to stereotypic words [e.g., lazy, diligent] following exposure to social group words [e.g., Black, White]. There's typically a shorter latency period between stereotypical pairings, showing an implicit association in unconscious cognitive processing). Since members of disadvantaged groups disfavour their own ingroup members on these tests, the contents of these attitudes can't be construed as innate (evolutionarily based) preferences (though the form might be innate); yet the contents of implicit bias are the result of a pervasive Patriarchal-Colonial cultural narrative that favours certain groups.
Since implicit biases are acquired, it's not surprising that they're somewhat flexible; in particular, they're moderately responsive to such measures as counter-stereotypical exposure, implementation intentions, and belief conditioning. 3 Research shows that, although we don't have direct insight into our implicit states, we can indirectly mitigate them by taking such steps as spending time with diverse social groups (counter-stereotypical exposure), and focusing on weakening stereotypical associations (belief conditioning). Implicit biases can also be weakened sheerly by accident-for example, if we just happen to be exposed to counter-stereotypical images or outgroup members in our habitual environment. But even in a fairly heterogeneous environment, we can deliberately alter our motivational profile using cognitive strategies.
In sum, implicit biases are pervasive but not uncontrollable. We can mitigate them by using intentional strategies and placing ourselves in felicitous (bias-reducing) circumstances. Notably, these strategies are social in the following sense: they're designed and recommended by epistemological communities like the HAIT researchers, they're supported by social networks, and they're often directly constituted by interpersonal engagement-socializing with non-stereotypical groups. This supports the sociality dimension of SOT: optimal human reasoning is social. But in a moment, I'll show that diversity is also critical to objectivity: we not only need sociality, but sociality that includes diverse members.

(ii)
Confirmation bias is the tendency to seek out and favour evidence that confirms our prior assumptions. If we already favour privileged groups and belief systems due to implicit bias, then confirmation bias makes the problem worse by giving these perspectives additional presumptive credibility. We think they're (even more) plausible because we already favour them.
(iii) Self-serving bias is the tendency to favour one's own beliefs and values. This bias further compounds the problem by disposing us to favour our own judgments (many of which are biased) while discounting other people's judgments.
We can see how these implicit biases can (if left unchecked) work together to distort our judgements, giving privileged groups and dominant conceptual frameworks excessive credence. This is how objectivity is undermined at the level of individual reasoning.
Another possible obstacle to strong objectivity is harmful group dynamics. Group interactions can undermine reasoning outcomes if checks and balances aren't implemented to regulate them.
On the simplest interpretation, collectives harbour biases because they're comprised of individuals, and individuals harbour biases. So, at the very least, collectives aggregate the implicit biases of individuals. But there's also a case to be made that collectives amplify these biases, making the whole greater than the sum of its parts. This is because people tend to conform to majority opinion and the opinions of authority figures, and these group dynamics can exacerbate individual biases.
This idea is supported by research on group dynamics and obedience. For example, Solomon Asch famously showed that individual subjects solve a task correctly 100% of the time, but when placed in a group, many subjects (32% in his 1951 trial) conform to the group's answer even when it's patently false. This exemplifies the dynamic of group conformity. Stanley Milgram famously showed that we're also susceptible to obedience to authority, by prompting subjects to punish a mock learner (actually his confederate) with increasing electrical shocks; in a normal trial, all of the subjects exceeded the expected threshold of 300 volts, and 65% continued on to the hypothetical point of death-the maximum of 450 volts (1963). (The 'victim' complained of heart problems and pounded on the wall, to no avail). In a similar vein, Philip Zimbardo (1973) primed subjects to take on the role of a prison guard, modelled on a notoriously brutal character in the film 'Cool Hand Luke.' The 'guards' quickly subjected the 'prisoners' to physical abuse and sexual humiliation-similar to what later happened at Abu Ghraib, where real guards physically and sexually tortured real prison inmates, and later distributed photographs to their friends, thinking it was a funny joke (see Zimbardo 2007). These scenarios show that we're highly susceptible to social conformity and obedience to authority, and these tendencies can have the effect of solidifying orthodox beliefs and suppressing dissent, particularly from outnumbered members.
That said, there are outliers who refuse to conform. Not all of Milgram's subjects obeyed the experimenter-some defected right away. People are more likely to defect if they're perceived as experts, high in status, or they've benefited the group in the past (Moscovici 1974). Unfortunately, marginalized groups are less likely to fit this description, so they're less likely to influence the group's dynamics. In addition, groups tend to manifest 'black sheep effect,' the tendency to judge perceived ingroup members more charitably and 'deviant' members more harshly (Marques, Yzerbyt, and Leyens 1988). So, marginalized members are both less likely than others to defect from established norms and practices, and less likely to be taken seriously by the majority. This is worrisome because it seems to imply that groups are doomed to perpetuate their dominant biases ad infinitum.
But this isn't the end of the story. Group dynamics can also enhance objectivity if the right background conditions obtain. For starters, collaborative reasoning, in which individuals cooperate toward a common goal, tends to enhance objectivity. John Doris compiles a robust inventory of findings in support of this view-a position that he calls 'collaborativism' (2015). To give some examples, studies show that groups outperform individuals at reasoning tasks (e.g., Laughlin 1965;Davis 1969;Maier 1970), groups excel at error correction (Gayle W. Hill 1982, 524), and groups are quicker at finding non-obvious solutions to difficult tasks (Schwartz 1995). This suggests that, although groups can undermine objectivity, they can also enhance it when they are willing to cooperate on a joint task.
Doris's work shows that optimal reasoning is social, but it doesn't yet show that it requires diversity, that is, that it involves not only cooperative aggregates, but aggregates composed of different demographic backgrounds (or 'intersectionalities,' as feminist epistemologists sometimes put it). Yet the diversity proposal receives empirical support elsewhere. Although research on demographic diversity over the last fifty years has been equivocal, the most recent research indicates that diversity enhances task performance given suitable background conditions.
The importance of background conditions, such as group mindset and managerial strategies, is essential because group performance is influenced by moderating variables that researchers are only beginning to properly appreciate. Traditional workgroup diversity studies operated on a 'black box' model that ignored intervening process variables-variables that support theoretical constructs like 'decision-making' and 'conflict' (Lisa H. Pelled, Kathleen M. Eisenhardt, and Katherine R. Xin 1999). Recent studies indicate that diversity tends to enhance workgroup performance only under favourable epistemic conditions, such as low 'subcategorization' (e.g., stereotypical perceptions, implicit bias), high cooperative interdependence, and having a positive 'diversity mindset'-a positive understanding of the value of diversity (Van Knippenberg and Michaela C. Schippers 2007).
Similarly, a recent comprehensive literature review indicates that workplaces that foster positive attitudes toward diversity, fairness, and inclusion enhance workplace performance (Guillaume et al. 2014). The emerging consensus, when these variables are taken into account, is that diversity enhances cognitive task performance under favourable epistemic conditions. The most likely reason is that workgroup diversity is "associated with differences in information, knowledge, and perspectives, and . . . this diversity may benefit group performance"-on the condition that negative subcategorization, emotional conflict, and other negative epistemic factors are minimized (Van Knippenberg and Schippers 2007, 527). This makes good intuitive sense: different perspectives contain different kinds of information (different 'knowledges'), and if they receive proper uptake, they enhance reasoning outcomes.
This research lends support and adds nuance to SOT: diversity enhances objectivity, provided that uptake-conducive background conditions obtain. I'll return to these considerations when needed while clarifying other aspects of SOT in what follows.

The relationship between diversity and objectivity.
What's the relationship between diversity and objectivity? Harding never specifies this. Depending on how one answers this question, SOT will appear more or less plausible. One possibility is that diversity and objectivity are conceptually related; that is, there's no possible world in which objectivity obtains without diversity. This interpretation isn't feasible. As the research shows, diversity can impair task performance when background epistemic conditions are unfavourablewhen uptake is low. The right way to think about the diversity-objectivity relationship is as a contingent relationship, dependent upon multiple intervening variables. Like most complex concepts, objectivity is multifactorial and fragile. But it's best served in conditions that enhance both diversity and receptivity to diverse speakers and views. The background conditions, group dynamics, and attitudes of the reasoners involved in an epistemic project all affect the outcome of the project.

Degree of diversity?
Harding never specifies what degree of diversity is required for objectivity. How many diverse members are needed for strong objectivity? This is a complicated question on which empirical research is (again) divided. Some studies find that beyond a certain degree of diversity, group performance is impaired because groups lack a "common frame of reference" (Van Knippenberg and Schippers 2007, 532). That is, they can't empathize with each other. But again, this effect is mediated by intervening factors such as subcategorization, cooperation, and diversity mindset. Given favourable epistemic conditions, people can empathize better with others. So, if these uptake-conducive conditions are in place, there shouldn't be an upper limit on the benefits of diversity. Now, while one diverse group member could in principle suffice to maximize objectivity, in practice this isn't likely to happen because of facts about human psychology. Specifically, groups are taken more seriously than isolated individuals. In this connection, Miranda Fricker (2007) writes that underrepresented groups tend to be discredited on the basis of their demographic attributes (i.e., they suffer 'testimonial injustice'), and as a result, they lose confidence in their beliefs and competency. This creates a positive feedback loop, in which epistemic injustice undermines the confidence of minorities, which in turn undermines public trust in their beliefs and abilities, limiting their ability to contribute effectively to group reasoning.
This epistemic theory is supported by research on corporate innovation. When women are token minorities on a corporate board, they have trouble influencing board strategic task decisions (i.e., the most important decisions in the firm), and they are more likely to conform to the majority's ideas and behaviours, as predicted by standpoint epistemology theory (Mariateresa Torchia, Andrea Calabro, and Huse 2011). Token minorities also tend to be closely monitored, stereotyped, and excluded by dominant groups' members, which limits their ability to contribute to strategic decision-making. However, when at least three women are on the board, they achieve the "critical mass" required to influence board decisions, and their contributions enhance organizational innovation-that is, they improve the firm's performance (ibid.). This shows that critical mass is epistemically valuable. But critical mass is the lower limit of epistemically valuable diversity; a higher proportion of women on the board could potentially enhance innovation even more. Notably, the average number of board members in the study was 7.12, so three women was a large proportion of the average board. In philosophy, which consists of thousands of people, far more women would be needed to achieve critical mass, and if we are committed to ideal objectivity, we should strive for more than the bare minimum.
As we saw, only 19-20% of postsecondary philosophy instructors are women, only 19 of 500 select top-tier journal articles are by women, and few of those articles received any extended discussion. If quantity matters, it follows that we need more women in philosophy. So far, our efforts since the 1990s have yielded negligible results. This means that we should be investing as much effort as possible in recruiting more women, for the sake of enhancing the epistemic value of our work.

'Fairness to the evidence'?
Harding says that objectivity is served by responding to "the most severe criticisms one can imagine" (2015,33). I don't disagree with this prescription, but I would suggest an important caveat. While we should try to imagine relevant criticisms, we shouldn't think that this is sufficient for the most robust objectivity. Rather, imagination should be seen as a last resort, to be used when we can't actually solicit feedback from people who hold relevant criticisms. The reason is that we don't, in fact, have very strong powers of imagination, partly because we're susceptible to cognitive distortions that significantly limit our deliberative possibilities.
We've already seen that Descartes's naïve theory of the (transparent) mind was fundamentally mistaken. So, we shouldn't repeat Descartes's mistake by thinking that cogitating diligently will reveal the truth of the matter to 'the light of (individual) reason.' It won't. We need to collaborate with other people, including, ideally, demographically diverse reasoners, and we need to foster uptake-conducive epistemic conditions, consistent with PEF plus the findings on workgroup task performance.
Doing this can be seen as a matter of epistemic responsibility; individuals and groups that collaborate with others and seek to enhance uptake are epistemically responsible. Epistemic responsibility might be seen as a virtue that enhances epistemic agency (and perhaps also moral agency, insofar as attending to relevant moral testimony enhances moral agency). Honing our sensitivity to relevant perspectives and criticisms-or cultivating proper vigilance-makes us better reasoners. 4

Diversity and uptake.
My penultimate clarification is that, on the basis of the empirical work, we should add uptake to the requirements for objectivity. That is, in addition to diversity, we also need felicitous epistemic conditions-those that facilitate uptake to diverse perspectives. We can easily imagine a robustly diverse group in which a historically marginalized subgroup isn't taken seriously. This is, in fact, the standard result of many earlier reasoning studies that didn't take account of relevant intervening variables: there was diversity without optimal reasoning outcomes. But in conditions of optimal uptake, groups do achieve optimal cognitive performance.

Institutional standards.
My final qualification emphasizes the need for institutional standards of epistemic effectiveness, not merely individual ones. Harding doesn't explicitly discuss these criteria. And PEF sounds (to me, at least) like a recommendation for individuals: it's not hard to see how an individual can take steps to be fair to a diversity of perspectives, but how do we implement PEF on an institutional level?
Fortunately, Longino (2001) offers a set of conditions for enhancing 'epistemic effectiveness' in institutions, and these can be annexed to Harding's proposal, making it easier to apply on a large scale. Longino's criteria for epistemic effectiveness are (1) fair venues, (2) uptake, (3) public standards, and (4) tempered equality. In more precise terms, these conditions entail that (1) there must be public venues for criticism, these venues must be the same venues in which 'original research' is presented, and criticism must be given the same or nearly the same weight as original research; (2) the community must effectively respond to criticism and dissent; (3) there must be publicly recognized standards (not just private ones) for evaluating theories, hypotheses, observations, etc.; and (4) there must be equality amongst equally capable members. Rather than assessing every member of an epistemological community, we can use these conditions to evaluate whether the community satisfies (1)-(4).
As philosophers, we can evaluate the epistemic state of our own discipline using Longino's criteria, and we can also apply some of Harding's proposals-both of which I'll do in the next section. This should be taken as an abbreviated assessment: a much more thorough assessment is possible, but it exceeds the scope of a single article.

Recommendations
Here, I'm going to make recommendations based on both Longino's and Harding's work.
First, we can apply Longino's institutional criteria. On scrutiny, philosophy doesn't fully satisfy any of (1)-(4). Regarding (1) fair venues, most philosophy journals don't publish discussion notes, those that do are heavily weighted in favour of original research (Leiter Reports Blog 2007), and feminist content, as we saw, is relatively scarce.
Regarding (2) adequate uptake, since feminist philosophy tends to be published in specialized venues like Hypatia and Feminist Philosophy Quarterly, and presented at specialized conferences-and, in addition to this, there's little extended discussion of women's work in generalist venues-it's reasonable to infer that uptake for women's work, especially feminist contributions, is not up to the ideal standard.
Concerning (3) public standards, while some journals now have triple-blind review (due in large part to the urging of feminists), many don't; and blind review doesn't eliminate content bias (against certain fields or points of view). This could perhaps be rectified by creating more diverse editorial boards. In terms of hiring, some philosophers have argued that 'pedigree bias' is a major hurdle in the fair evaluation of job candidates (de Cruz 2014). The interview, too, permits hiring committees to (implicitly or explicitly) judge candidates on the basis of demographic attributes, enhancing the risk of biased outcomes. In a competitive job market like ours, where there are more qualified candidates than positions, implicit bias is more likely to be a factor. So, every effort should be made to minimize these effects (more on which in a moment).
Finally, (4) tempered equality isn't met because a preponderance of instructors and published authors are men (more specifically, White men), and there's no reason to think that they're substantially more talented or credible than other candidates, in defence of their over-representation. Indeed, if the basic premise of standpoint epistemology is right, there's reason to think that underrepresented groups have underappreciated insights that could enhance objectivity in the field, so their perspectives are especially epistemically valuable.
Overall, then, philosophy appears to have systemic problems that deserve our attention.
Two possible solutions to these problems are (i) diversify evaluators (on editorial boards, hiring committees, referee panels, etc.), and (ii) reduce the role of implicit bias. The University of Florida offers guidelines for reducing implicit bias in interviews, such as doing phone interviews first, preparing questions beforehand, and asking all candidates the same questions. 5 A more radical step would be to eliminate interviews altogether. This policy would probably be met with resistance, but it's reasonable to think that a candidate's right to a fair (unbiased) assessment trumps a department's right to choose a candidate they like on the basis of personal preference.
Next, let's consider the situation from the perspective of Harding's work. I want to home in on three of Harding's recommendations, which I think have important practical applications for philosophy. These recommendations are (1) include voices from outside and below, (2) implement affirmative action, and (3) promote alternative research communities. I'll elaborate on these recommendations in turn.
(1) Harding says that we should solicit input "from outside" and "from below" (2015,(34)(35)(36). What does this mean for women in philosophy? Women are 'outside' in the sense that they have unique insights rooted in their lived experience, which the standpoint epistemologists refer to as epistemic privilege or situated knowledge. This privileged insight may be embodied in the 'different voice' described by Carol Gilligan (1982), or the 'ethics of care' described by Nel Noddings (1984, 2002, or simply the embattled perspective of someone afflicted by the 'perfect storm' of gender prejudice identified by Louise Antony (2012). If any of these accounts is right, women have situated knowledge that's valuable to the field of philosophy.
Women are 'below' in the sense that they're generally less senior, less cited, less discussed, and less well-compensated than their male peers. In other words, they occupy a position of lower status. This position can confer epistemic privilege in the sense that it can confer a high degree of epistemic freedom-the freedom to evaluate the norms of the field from an impartial, unencumbered perspective. Lowstatus members of a community are 'epistemically freer' because they're less invested in the community's established practices: they haven't had a hand in a designing those practices, popularizing particular schools of thought, or promoting certain people-at least, not to the same extent as more senior members. This lack of prior investment creates a high degree of 'free will' with respect to evaluating epistemic practices in the field: low-status members are freer to implement new and innovative methods, and to question the validity of institutional practices accepted by the majority.
The next two recommendations-(2) implementing affirmative action and (3) promoting alternative research communities-are mentioned by Harding in a single page (2015, 35), but they deserve more attention, since they're controversial but potentially very effective.
(2) I take affirmative action to mean a policy that favours a member of an underrepresented group. This is a controversial strategy because it's seen by some to be unfair. But it's not unfair if there's a preexisting bias in the hiring process, and the policy counteracts that bias. Dan Kelly and Erica Roedder (2008) advise that professors should skew underrepresented students' grades slightly upward to correct for the expectable effects of implicit bias. If this is a good policy for graders, it's likely a good policy for hiring committees. This 'moderate' kind of affirmative action can serve to mitigate the implicit bias that plays a predictable role in epistemic judgments in philosophy.
A similar policy that's not often discussed but makes sense in the current climate is to slightly favour certain content or research orientations, to correct for institutional content bias. This is something that hiring committees, editors, and referees can do in their various gatekeeper roles. This method could also be used to improve the objectivity of public ranking lists, like the Brooks Blog 6 and the Leiter Reports 7 rankings. For starters, the Leiter ranking system could be expanded to include more than 20 journals, and more than just generalist journals. (This would be helpful because Leiter's rankings are fairly influential). Furthermore, all rankings systems, which glean data from respondents through surveys, could include a notification in their surveys stating that implicit biases may affect people's judgments of certain types of content, particular non-standard content like feminist philosophy, critical race theory, and non-Western philosophy, and respondents should be aware of this. This notice could also be included on the websites that publish the results. This would be a valuable step because it would both alert people to the role of implicit bias in their assessments, prompting them to reflect on those biases, and would increase the cognitive salience of nonstandard journals, priming people to reflect on them more carefully and (hopefully) more fairly. This notice would be especially helpful to those who might not think about these journals at all.
Regarding (3) alternative research groups, I fully agree with Harding that these groups are a valuable academic resource, but I would emphasize that the output of these groups must be integrated with mainstream philosophical discourse for the condition of uptake to be met. If feminist discourse remains sequestered in disciplinary venues, then we're losing much of the value of these alternative frames of reference. One example of 'sequestration' is the existence of 'feminist perspectives on ' [science, globalization, disability, autonomy,' etc., etc.] in the Stanford Encyclopaedia of Philosophy, alongside 'mainstream' entries on these same subjects. Ideally, the two subjects would be integrated into a single comprehensive entry, or, failing that, some of the citations from the 'feminist perspectives' entry would be added to the generalist entry, enhancing the broad recognition of those authors.

Concluding Remarks
I've argued that the underrepresentation of women in philosophy jeopardizes the attainment of strong objectivity, a goal to which all researchers are explicitly committed. I've suggested strategies for ameliorating this problem, rooted in the core idea that we need to boost women's numbers while simultaneously increasing uptake-conducive background conditions, that is, a hospitable epistemic climate for women's contributions, to enhance objectivity. We can foster this climate individually (e.g., by focusing on minimizing our stereotypical associations) and collectively (by implementing and enforcing institutional standards of epistemic responsibility). Some of my proposals may be controversial, but they're merely suggestions and far from exhaustive: there are many possible constructive steps that we can take. One very simple but constructive step is talking about the problem, which puts us in a better position to decide on effective and mutually agreeable strategies. This paper is part of that conversation. My hope is that it will help spur constructive solutions that will eventually close demographic gaps in philosophy.