The Underrepresentation of Women in Philosophy

Philosophers have long-argued that the profession is not as equally welcoming of everyone as it should be. For decades, philosophers have documented patterns of systematic exclusion based on social group,^{Footnote 1} philosophical projects or methods,^{Footnote 2} as well as canons of work studied.^{Footnote 3} Within the past decade, philosophers have begun investigating the underrepresentation of certain social groups as part of this larger project of addressing systematic exclusion. While there are a range of underrepresented social identities in philosophy, identities around race and gender are the most-studied. The literature on the underrepresentation of people of colour in philosophy is slowly growing,^{Footnote 4} however, the overwhelming focus in the profession has been on the status of women in the field.

One of the main questions that work on underrepresentation addresses is: just how underrepresented are women in philosophy? The profession has something of an answer to this question, though some of the following numbers stem from a single source or one whose generalizability is uncertain. Women constitute 20-30% of professional philosophers in Canada, Australia, the U.K., and the U.S.^{Footnote 5} Not only is that less than their representation among the general population (51%),^{Footnote 6} but it is also less than their representation among Ph.D.’s earned in all fields (45-50%)^{Footnote 7} and Ph.D.’s earned in the humanities (50-55%).^{Footnote 8} While in Canada, women constitute around 30% of full-time, tenure-track faculty,^{Footnote 9} in other countries that proportion is lower. Kathryn Norlock reports that in the U.S., women constitute only 16.6% of full-time philosophy professors.^{Footnote 10} While the proportion of women among graduate and undergraduate students in the U.S. is higher (30% for graduate students and 30-35% for philosophy majors),^{Footnote 11} that is still less than their representation among students in general, since women constitute 55% of new college enrollments^{Footnote 12} and 57% of recent B.A.s earned.^{Footnote 13}

Although some may think of the gender gap as being the gap between women’s participation in philosophy (20-35%) and demographic gender parity (51%), another measure of underrepresentation is the gap between women’s participation in philosophy and their participation in academia or the humanities more generally (50-65%).^{Footnote 14} Additionally, as Meena Krishnamurthy et al. point out, the ‘gender gap’ is really a set of multiple gaps in various places—from philosophy majors to women publishing in journals.^{Footnote 15} These numerous gaps are not obviously all part of the same, single phenomenon. In some places, women’s underrepresentation in philosophy seems at odds with trends in the humanities and academia in general (e.g., women earning M.A.s in philosophy), but in other places, the trends in philosophy match those in academia more generally (e.g., women being made associate or full professors). Furthermore, there are different places in the pipeline of higher education where underrepresentation of women increases or decreases. There does not seem to be a further dip in underrepresentation between women earning Ph.D.s and getting jobs in philosophy, but there may be a significant one between women taking their first philosophy classes and majoring in philosophy.

To summarize, underrepresentation is not necessarily a single, unified phenomenon. That is part of the reason that, as Louise Antony and others have argued, underrepresentation may be the result of multiple factors, some independent and some intertwined.^{Footnote 16} Some of those factors could play a significant role in places, a marginal role in others, and be completely absent in yet others. As a result, rectifying the different instances of underrepresentation (in need of rectifying) might require multiple, diverse solutions.^{Footnote 17} Thus, what I discuss in this paper is not the causes of and solutions for all instances of underrepresentation. I focus on women taking undergraduate courses in philosophy and majoring in philosophy, since those are places where interests may come into play the most.

Is Investigating Women’s Interests Worthwhile?

The larger question of why women are underrepresented in various areas of society is not new. Cora Castle asked in her 1913 dissertation, “A Statistical Study of Eminent Women,”

Castle’s question references two of the most popular hypotheses regarding the cause of women’s underrepresentation: gender differences in abilities and widespread discrimination towards women. However, a third hypothesis, that there are gender differences in interests, has been recently gaining attention in philosophy^{Footnote 19} and has been popular in science, technology, engineering, and math (STEM) fields for a while.^{Footnote 20} Rebecca Jordan-Young observes that, “Both scientists and laypeople are increasingly likely to attribute sex differences in these fields to a variation in interests, rather than in abilities (not to mention discrimination).”^{Footnote 21}

There has been no formal, full rejection of the differences-in-interest hypothesis, but some philosophers have argued against or expressed scepticism towards aspects of it. Sally Haslanger, for instance, argues that although,

Haslanger does not reject the entire differences-in-interests hypothesis, but expresses a degree of scepticism based on the evidentiary and theoretical hurdles such an explanation faces in explaining the differing degrees of underrepresentation in philosophy subfields.

Another form of scepticism is expressed by Kathryn Gines in reflecting on being asked why she would be interested in philosophy. She remarks, “Why wouldn’t I (or shouldn’t I) because of (or in spite of) my embodied existence—that is, my embodiment as a Black woman—be interested in philosophical reasoning and fields of inquiry?”^{Footnote 23} At least for some women, and members of other marginalized groups, being uninterested in philosophy does not speak well to their experiences. Some of Gines’ scepticism and frustration might be related to a point raised by Michelle Bastion. She observes that, in advocating for the differences-in-interests hypothesis, some unfairly put the onus

A final reason to doubt the differences-in-interests hypothesis is that, until very recently, there was not much evidence that suggested that women actually are less interested than men in philosophy. Therefore, for some who investigate underrepresentation, especially as a facet of systematic exclusion in philosophy, the differences-in-interests hypothesis has not seemed fully plausible or very insightful.

Given more recent evidence and analysis, however, the differences-in-interests hypothesis is indeed a plausible explanation of some forms of underrepresentation, so, in that sense, asking what women want is a fruitful avenue of further inquiry. I argue this by reviewing the limitations of the two other popular explanations for underrepresentation (the discrimination hypothesis and the differences-in-abilities hypothesis), and then explaining how the differences-in-interests hypothesis can address those limitations.

The Discrimination Hypothesis

The discrimination hypothesis claims that underrepresentation is due, at least in part, to the fact that women’s abilities and accomplishments are often unfairly evaluated in comparison to men’s abilities and accomplishments. This can happen consciously (if someone insists that a female candidate cannot be as good as a male candidate because of her gender) or unconsciously (if hiring committees do not realize they are collectively more sceptical of ambitious research plans in applications that bear a female name.) There are currently no controlled experiments or surveys that show that discrimination or bias exist specifically in professional philosophy. Some, however, take the evidence for bias in general and in other fields and professions,^{Footnote 25} coupled with personal experience, the large volume of anecdotal reports, and data on the difficulty of recognizing one’s own biases^{Footnote 26} as compelling reasons to presume that these phenomena do likely exist in philosophy. Others disagree, and there is a further disagreement over who has the burden of proof—those claiming that discrimination and bias likely does exist in the profession, or those claiming that there likely is no or little discrimination and bias.

Even if there is some discrimination and bias within philosophy (disclosure: I think there is), there is still a limitation to the discrimination hypothesis. Some instances of underrepresentation are not well explained by discrimination and bias, even if they are at play in the profession. Current data suggest that women are underrepresented in introductory philosophy classes—at least in some departments. Women constitute 43% of introductory classes but 55% of new college enrollments and 57% of all undergraduates.^{Footnote 27} If this form of underrepresentation proves a robust finding, spanning across departments with different reputations and records concerning their treatment of women, the discrimination hypothesis would likely not explain this instance of underrepresentation. After all, students signing up for their first philosophy class likely do not know much about the field in general.

If this form of underrepresentation exists across departments with different reputations, then the expectation of likely discrimination would not be a factor either. Some initial evidence to this point is Christopher Dobbs’ work that looks at survey results from two million college freshmen. Dobbs finds that from 2004-2009, only 34% of incoming college freshmen who already intended to major in philosophy were women.^{Footnote 28} While it is not impossible for direct discrimination to play a role in creating these numbers, it does not seem overly plausible.

Therefore, while some suspect that discrimination is an important part of the puzzle for explaining at least some instances of underrepresentation (such as women getting tenure, getting promoted to full professor, or getting hired at very prestigious institutions), it plausibly or even likely does not explain other instances, such as women taking introductory philosophy courses and, perhaps, majoring in philosophy.^{Footnote 29} This is a limitation that advocates of the discrimination hypothesis, such as Haslanger, have acknowledged since the beginning of the conversation on underrepresentation.^{Footnote 30} In fact, Antony’s ‘perfect storm’ model, which in part argues that there are likely multiple, different factors in play, is presumed by many who work on underrepresentation in philosophy.^{Footnote 31}

The Differences-in-Abilities Hypothesis

The differences-in-abilities hypothesis claims that the underrepresentation of women is largely caused by differences in (presumably innate or fixed) cognitive ability or talent. This hypothesis faces a different set of limitations. It has overall waned in plausibility, both in philosophy and STEM fields. Some might wonder whether comparing underrepresentation in philosophy and STEM fields is warranted. The main reasons for thinking that work on underrepresentation in STEM is applicable to philosophy is that the degree of underrepresentation is similar in both areas. From 2011-2012, women earned 51% of all Ph.D.’s in the U.S., and more than 50% in many fields, such as area studies (63%), biology (53%), education (68%), English (62%), foreign languages (60%), pharmacy (62%), physical therapy (69%), veterinary medicine (78%), nursing (87-95%), psychology (70%), anthropology (64%), and sociology (63%).^{Footnote 32} Women earned only 26% of Ph.D.’s in philosophy during that time, however. Fields where women earned a similar proportion of Ph.D.’s are theology (26%), engineering (22%), and mathematics (24%). Some areas in STEM have even lower proportions of women earning Ph.D.’s, such as physics (19%), computer science (18%), mechanical and aerospace engineering (15%), and nuclear engineering (14%).^{Footnote 33} Based on the numbers, philosophy has more in common with some of the STEM fields than it does with most of the humanities.

Another reason to apply work on underrepresentation in STEM to philosophy is that some of the hypothesized causes are the same—such as differences in abilities, discrimination, and implicit bias. Lastly, STEM fields are one of the few places where systematic work on women’s underrepresentation in higher education is being done at all, so it is one of the few places that philosophers can draw from for potential insight into their own field. These reasons, taken together, make the case for why philosophers should apply research on underrepresentation from STEM fields.

The differences-in-abilities hypothesis claims that women and men, as groups, have different average cognitive abilities that are biologically innate and/or fixed. As Lawrence Summers infamously surmised, something like ‘IQ,’ ‘mathematical ability,’ or ‘scientific ability’ may have different standard deviations from the mean when it comes to gender.^{Footnote 34} Thus, at the extreme end of the spectrum—such as, according to Summers, physicists at a top 25 research university^{Footnote 35}—we should not be surprised to find a significant gender imbalance there. While this hypothesis is not fully dead and remains possible, it is not plausible enough, given the totality of research on underrepresentation and cognitive sex differences. For starters, the sort of bell-curve picture that Summers and others offer requires a host of non-obvious assumptions, such as there being a single thing (or correlated cluster of things) that constitutes scientific ability, that hiring processes in academia actually select for this thing (scientific ability), and that the most prestigious departments consistently hire people who are in the top 1-10% of scientific ability.

Even if those premises were plausible, or unnecessary for Summers’ conclusion, the differences-in-abilities hypothesis faces another hurdle: showing that differences in ability cannot be (easily) overcome by practice and training. After all, one of the things that humans are very good at is learning new skills and practicing them. If the relevant abilities constituting IQ or ‘scientific’ or ‘mathematical’ ability can be trained (somewhat easily), then persistent gender gaps are not likely due to differences in these abilities—since such differences can be changed with a reasonable amount of effort. Research suggests that this is the case, since one of the most studied proxies for mathematical ability—spatial ability—is trainable. David Uttal et al. conclude that spatial abilities are in general malleable and do not seem to be rigidly fixed. Their findings suggest that, “a wide variety of training procedures can lead to meaningful and durable improvements in spatial ability.”^{Footnote 36} One such procedure is video games, which Uttal et al. report may be as helpful as formal classes in honing spatial abilities.

That playing video games may be as effective as formal education is relevant, given another study on observed sex differences. Jiang Feng, Ian Spence, and Jay Pratt argue that, regarding spatial attention (another proxy for mathematical ability), there is an observed gender difference. However, they also found that “playing an action video game can virtually eliminate this gender difference in spatial attention and simultaneously decrease the gender disparity in mental rotation ability.”^{Footnote 37} This research suggests that, if individuals in STEM fields are receiving adequate training, such training should be able to overcome any innate gender differences in ability.

Even if the issue of practice and training could be overcome, observed differences in ability face another problem in that they are simply too small in themselves to account for the rather large gender gaps that exist in STEM fields.^{Footnote 38} Granted, some suggest that, although the average difference may be too small to account for underrepresentation, differences at the ends of a bell-curve distribution could be large enough—hence the picture that Summers described about underrepresentation at prestigious universities. Elizabeth Spelke argues that this hypothesis is implausible, because it relies on the premise that “males are more variable in their cognitive abilities” in order to explain why they “predominate at the upper reaches of mathematical talent.”^{Footnote 39} Much of the evidence for this ‘differences-in-variability’ hypothesis comes from boys outperforming girls on certain standardized tests. Girls and women do equally well as, if not better than, boys and men, however, when it comes to high grades and earning degrees in mathematics. Spelke therefore argues that the differences-in-variability hypothesis must first settle the controversial issue of which measures track advanced mathematical ability better—standardized tests or grades—before it can use standardized test scores as evidence that men do, in fact, have greater mathematic abilities than women.

Spelke concludes that,

This does not mean that the differences-in-abilities hypothesis is completely dead or disproven, especially if we shift the hypothesis to a ‘differences-in-observed-skills’ hypothesis instead of ‘differences-in-innate-abilities.’ However, regarding differences in skills, there is no compelling evidence in either STEM or philosophy that such differences exist, or that if they did they would be large enough or intractable enough to account for the demographic trends they are meant to explain.

The Differences-in-Interests Hypothesis

The differences-in-interests hypothesis claims that women and men, as groups, have different interests and/or life goals. As a result, the underrepresentation of women in philosophy is due to women choosing not to take philosophy classes or major in it or go to graduate school for it. This is what women want. The first thing to note is that, like the other explanations, this hypothesis does not claim to be the single factor that contributes to underrepresentation. For instance, differences in interests would not plausibly explain underrepresentation of women at the level of full professor—if that proves to be a robust finding.

While the differences-in-interest hypothesis is not entirely new to philosophy, it has recently gained more sustained attention. Katrina Hutchinson and Fiona Jenkins note in their edited anthology, Women in Philosophy: What Needs to Change?, that gender seems to still be correlated with “likely success or failure, as well perhaps as interest or disinterest, in the pursuit of philosophy.”^{Footnote 41} Adriane Rini notes in the same anthology that it may be the case that, “men’s interests in philosophy are allowed to define what is important in philosophy.”^{Footnote 42} Such anecdotal observations about differences in interests are gaining support from surveys and experiments. Sam Baron, Tom Dougherty, and Kristie Miller report in their 2015 study that, “at the first lecture, female students were less interested in philosophy.”^{Footnote 43} Morgan Thompson et al. found in 2016 in introductory philosophy classes that on average women reported “that they enjoyed the course and that they find philosophy interesting […] to a significantly lower degree than men.”^{Footnote 44} Dobbs also reports that, from 2004-2009, “significantly more men than women indicated that they intended to major in philosophy.”^{Footnote 45} Taken together, this data suggests that women who enter college may be, on average, less interested in philosophy than are men.

To summarize, the differences-in-interests hypothesis is gathering evidence in its support and is especially fruitful given the limitations of the other two popular hypotheses. It cannot explain all instances of underrepresentation, but it does potentially explain some better than the differences-in-abilities hypothesis or the discrimination hypothesis—though, whereas the differences-in-abilities hypothesis struggles to explain any instance of underrepresentation, the discrimination hypothesis can still plausibly explain some. Overall, investigating the differences-in-interests hypothesis and asking what women want is a worthwhile project for those who work on underrepresentation and exclusion in philosophy.

What Are the Implications of Asking What Women Want?

Among those who agree that the differences-in-interests hypothesis is a good lead to investigate, there is stark disagreement about what its likely implications are, especially regarding how and whether philosophers should address underrepresentation in practice. Sharon Crasnow suggests that, if we want to better help cultivate women’s interests in philosophy, having more professors and role models in the profession who are women may have a small effect, by way of helping women project themselves into the role of philosopher.^{Footnote 46} Norlock argues that students’ interest in philosophy can likely be improved by having more inclusive curricula.^{Footnote 47}

However, others find such suggestions to be hasty and unwarranted. Criticisms of such reforms are often due to doubt over the discrimination hypothesis, but these criticisms also are sometimes bolstered by the plausibility of the differences-in-interests hypothesis. Nevan Sesardic and Rafael De Clerq, for instance, take issue with the decision by two editors of the Stanford Encyclopedia of Philosophy to address the potential under-citation of women by encouraging subject editors and authors to include more subjects about and citations of women who may have been left out.^{Footnote 48} Sesardic and De Clercq argue that, insofar as the editors are assuming that underrepresentation is problematic, and thus needs to be alleviated, their effort “resembles an attempt to cure a disease without knowing its cause.”^{Footnote 49} Furthermore, they contend that this kind of reform might have inadvertent adverse effects, making it further unwarranted. Their criticism of these reforms stems largely from their criticism of the discrimination hypothesis, though they do mention the differences-in-interests hypothesis as a possible alternative. The existence of possible alternative explanations presumably adds to their reasons for thinking that one should not jump to the conclusion that underrepresentation is a problem in need of certain interventions.

David Papineau argues that underrepresentation may not be a moral problem in all of its manifestations and that reforms that treat it as such might prove unnecessary or even harmful. He, unlike Sesardic and De Clercq, concludes that discrimination is likely or plausibly causing some amount of underrepresentation, and that the profession should address that through its hiring practices, for instance. However, Papineau contends that discrimination may not explain all instances or degrees of underrepresentation, and that, if something like gender differences in interests also play a role, that does not necessarily create a moral problem similar to discrimination. In emphasizing the possibility that gender differences in interests could be benign and not something that calls for intervention, Papineau emphasizes the possibility that such differences in interests could be caused by something other than pernicious social forces, such as sexism. Papineau concludes that, after addressing the more straight-forward issues of discrimination, the profession might want to ultimately “let the gender numbers fall where they may”^{Footnote 50} refraining from any further intervention into the field’s demographics, since further interventions could prove unnecessary, or even harmful.

These sorts of concerns are not limited to just philosophy. Steven Pinker, in writing for a more general audience about underrepresentation in STEM fields, also discusses the potential risks of too hastily accepting a certain explanation for underrepresentation.^{Footnote 51} Pinker does not deny that underrepresentation is likely or plausibly a moral problem, but he does caution his readers about the risks of enacting certain reforms advocated by those who believe it is likely such a problem. The call for encouraging more women to enter underrepresented fields carries significant risks to both women and the profession. He warns that “Overestimating the extent of sex discrimination is not without costs.”^{Footnote 52} Some interventions meant to fix a gender gap could unintentionally make it worse or do further harm to women. Specifically, “Young women may be pressured into choosing lines of work they don’t enjoy.”^{Footnote 53} On his reading of the proposed interventions, the cures may be worse than the (misdiagnosed) disease.

These criticisms and conflicting takeaways from the differences-in-interests hypothesis raise two issues about underrepresentation’s perniciousness and uncertainty, which I will take up respectively in the next two sections.

Stereotypes, Belonging, and Self-Identity

If underrepresentation is caused by (in part or largely) pernicious or unfair factors, such as sexism, then those factors warrant redress, and that is a reason to consider enacting reforms such as more explicit mentorship for women and more socially inclusive curricula. If underrepresentation is instead largely caused by benign or immutable factors, such as individual choices that do not contribute to harmful cultural norms, or biological features whose influence cannot be overcome no matter how hard we try, that is a reason to favour not enacting reforms such as increased mentorship. I argue that the potential role that stereotypes play in influencing interests is a reason to suspect that, if underrepresentation is due in part to gender differences in interests, those in turn are due (at least in part) to pernicious social forces, such as sexism, which warrant redress.

There are a number of ways that sexist ideals and norms could potentially influence interests. One way is that stereotypes may create consonance or dissonance with certain self-identities, specifically via ‘stereotype threat’ and ‘belonging uncertainty.’ Stereotype threat is a phenomenon that occurs when individuals become aware (consciously or not) of an environmental cue that implies that a negative stereotype about their social group could apply to them or threaten their standing in the current environment. For instance, a woman who walks into a math class where she is the only woman in the room, or who is told that a test she is about to take is harder for women, might experience stereotype threat. If she is already aware that there is a culture-wide stereotype that women are bad at math, then certain environmental cues (such as the demographics of the room or the description of the test she is taking) can create a sense of potential threat to her standing or self-worth, via this stereotype.^{Footnote 54}

Research on stereotype threat often emphasizes that the phenomenon causes people anxiety and lowers their performance on certain kinds of tasks such as tests.^{Footnote 55} This is called the ‘underperformance effect’ of stereotype threat. It is an important part of the phenomenon, but, as I argue elsewhere, it is not stereotype threat’s only effect or even core feature.^{Footnote 56} Stereotype threat can also influence individuals’ interests.^{Footnote 57} Furthermore, not everyone who experiences stereotype threat will underperform. Therefore, stereotype threat is not necessarily or primarily threatening to individuals’ performance; I argue that it is primarily threatening their self-identity and social standing—which sometimes has the additional effect of influencing their performance. This claim is consistent with research on stereotype threat, but is often overlooked or underemphasized.

On this broader account of stereotype threat, the underperformance effect is a side-effect of some individuals’ reactions to (or coping strategies in the face of) a potentially threatening stereotype. That is, some who experience stereotype threat become motivated to disprove the stereotype—in order to protect their social standing or self-identity. A woman in a math class might double-down on the test she is taking, applying extra effort to make sure she does well. In doing so, she is trying to protect her standing in the math class or her identity as a good math student. If the test is an easy one, she is actually likely to do better on it than a woman taking the same test but not having been reminded about the potentially threatening stereotype.^{Footnote 58} However, if the test is hard and requires a lot of attention and focus, she is likely to underperform and do worse than a woman who is not under a stereotype threat condition. Not only does the underperformance effect require a difficult task, but it also requires that individuals care enough about the domain they are in (e.g., math) to want to disprove the stereotype. There seems to be no underperformance effect (or at least there is less of one) when women who do not care about math—who are not invested in it or identify with it—are primed with the potentially threatening stereotype.

Some individuals’ responses to stereotype threat is not to try to prove the stereotype wrong, but to simply disengage or distance themselves from the domain in which they are stereotyped.^{Footnote 59} For instance, a woman who notices that all of her upper level math classes are dominated by men, but who does not particularly care about math or excelling at it, might respond to the potentially threatening stereotype that women are bad at math by putting less effort into her math classes, or by choosing a major that has nothing to do with math. She might even start to think of herself as specifically ‘not-a-math-person.’ This is one way that stereotype threat might influence individuals to dis-identify with domains in which their social group is negatively stereotyped.

This ‘dis-identification effect’ is presumably stronger when individuals do not already identify with the stereotyped domain, but it may influence even those who do identify with the domain, such as women who are majoring in math.^{Footnote 60} Even individuals who are vulnerable to the underperformance effect specifically because they are highly motivated to succeed in the stereotyped domain, may experience decreased interest in the task itself or the domain more generally,^{Footnote 61} as well as decreased motivation to improve their skills in that domain.^{Footnote 62} There is a lot of further research needed to understand more fully how and why individuals react to stereotype threat in these various ways. But, overall, individuals respond to stereotype threat by protecting one of their self-identities—either by dis-identifying with one of the identities involved in the stereotype, or by attempting to demonstrate that the stereotype does not apply to them. If this is right, then cultural stereotypes can create a kind of dissonance between two self-identities (such as ‘woman’ and ‘math person’ or ‘woman’ and ‘philosopher’), which individuals need to then resolve by challenging the stereotype or by downplaying one of those identities.

In this way, stereotype threat could help flesh out a number of different claims made by philosophers investigating underrepresentation and the differences-in-interests hypothesis. Cheshire Calhoun suggests that women “have difficulty envisioning themselves as philosophers and that this produces a weaker attachment to the discipline.”^{Footnote 63} Interestingly, Calhoun distinguishes between women having an ‘interest’ in philosophy, whereby they enjoy the activity, and having an ‘attachment’ to it, whereby they put more weight on the option of doing it, and thus would be less likely “to see other options as live ones and to choose one of those other options.”^{Footnote 64} Whereas women who are already ‘attached’ to philosophy may be more likely to respond to stereotype threat by trying to disprove the negative stereotypes about women—making them more vulnerable to the underperformance effect on difficult tasks—women with less attachment to philosophy may be more likely to respond by distancing themselves from philosophy, even if they enjoy philosophy or have some nascent interest in it.

Stereotype threat’s potential effects on interests also help flesh out Thompson et al.’s hypothesis regarding interests and belonging.^{Footnote 65} Thompson et al. observe that, “For women more than men, the extent to which a student has brilliance-based beliefs about philosophy leads them to be less interested in philosophy and less willing to continue in philosophy.”^{Footnote 66} A “brilliance-based belief” is a belief that, to do well in a certain domain, one must have a kind of innate brilliance or talent. One reason that women with brilliance-based beliefs may be less interested in philosophy than men with brilliance-based beliefs is that brilliance-based beliefs may exacerbate ‘belonging uncertainty’ for women in philosophy, but not for men.

Belonging uncertainty is a sense that one’s social identity might make one a poor fit for a given domain. One common environmental cue for belonging uncertainty (and stereotype threat, which can then lead to belonging uncertainty^{Footnote 67} is observing low numbers of individuals who share one’s social identity in a given domain, or being the only individual present with that social identity.^{Footnote 68} If a woman enters a high-level college math class, and sees that there are no other women in the class, this could create (or exacerbate) a sense of belonging uncertainty. In a more general sense, belonging uncertainty might be similar to the feeling of walking into an event and realizing that one is dressed very differently from everyone else. One might have a vague uncomfortable feeling that one might not fit in or belong in that situation. Belonging uncertainty, however, is likely not just a one-off feeling, but a more chronic experience, and it is one that is often tied to social-identity.

One effect of belonging uncertainty is that individuals often become more vigilant for potential clues as to whether or not they indeed belong or are valued in the domain. One potential piece of evidence could be that one needs a kind of innate brilliance to do well in that domain, coupled with a stereotype that people with a certain social identity are less likely to have that kind of innate brilliance. Thus, belonging uncertainty coupled with cultural stereotypes might be why brilliance-based beliefs tend to lower women’s interest in philosophy, but not men’s. Women with brilliance-based beliefs—who perhaps are experiencing higher levels of belonging uncertainty—might be more likely to respond to stereotype threat by dis-identifying with the field in which they are uncertain they belong, rather than expending the effort to repeatedly disconfirm the relevant stereotype about women.

An interesting implication of work on belonging uncertainty is that it suggests that stereotypes might function more like hypotheses or seeds of doubt, and less like fully-fledged beliefs or claims. Gregory Walton and Geoffrey Cohen argue that a negative pressure on one’s self-worth, such as that from stereotype threat, can be “especially pernicious”^{Footnote 69} when it takes the form of a hypothesis, rather than a belief or definite perception. This is because hypotheses can be as attention-directing as beliefs—i.e., making one more vigilant for evidence for or against it—but more resilient as well. A doubt is actually often harder to discount than a straightforward claim. Walton and Cohen found that students who were worried about whether they belonged in college were more likely to notice and pick out events that were consistent with and potentially explained that hypothesis, such as having few friends. Furthermore, because people are often uncomfortable with ambiguity, with not knowing where they stand, they will be motivated to seek out evidence that can confirm or deny their suspicions.^{Footnote 70}

If this account of stereotype threat and belonging uncertainty is right, it suggests that we should find underrepresentation where we find barriers to social groups identifying with a field and/or experiencing belonging uncertainty in that field, via negative stereotypes and other factors. This account could explain why the underrepresentation of women has changed in some fields and remained the same in others. In computer science, for instance, the drop in women majors in the U.S. in the 1980s corresponds with when computers were marketed as ‘toys for boys.’ That marketing may have created or exacerbated a stereotype that hindered women’s ability or willingness to identify with (or be ‘attached to’) computer science in the face of also being less likely to have the relevant experience with home computers. The ‘toys for boys’ marketing created a one-two punch for women: it made them less likely to have experience with computers and less likely to identify with them. Similarly, if women are underrepresented in philosophy, this may be due in part to them having difficulties identifying with it. Some work in philosophy is finding it to be the case that women are less likely to identify with being ‘philosophers’ or ‘philosophy majors.’^{Footnote 71} Underrepresentation may in part be caused by, or mediated by, being able to (easily) identify with a field, and whether there are cultural stereotypes that create dissonance with that identity.

Although this work on stereotype threat has not confirmed that this is, in fact, what is happening in philosophy and what is, in fact, causing underrepresentation, it is worth taking seriously as a potential explanation. Many who discuss gender differences in interests, especially when pointing out the limits of hypotheses that posit pernicious social forces as responsible for underrepresentation, do not acknowledge the ways in which stereotypes might influence interests in a manner separate from discrimination and perhaps even implicit bias, but which ultimately still poses a serious moral problem worthy of redress.

Some might have a specific reason for ignoring or downplaying this work on stereotypes, such as being sceptical of the quality of this research, in part due to the ‘replication crisis’ in social psychology and other fields.^{Footnote 72} I have two responses to such scepticism and one acknowledgement. The acknowledgement is that there is likely some publication bias in many scientific fields, including social psychology. That is, seemingly significant effects may turn out to be ultimately noise in the data, and truly significant effects may have their effect size over-estimated. However, the existence of publication bias in general does not mean that stereotype threat is likely non-existent. The meta-analysis conducted by Paulette Flore and Jelte Wicherts, for instance, shows that there is likely publication bias and over-inflated effect sizes in stereotype threat research, but it does not show that the basic effect is not real.

Specifically, Flore and Wicherts conclude that one is unwarranted in generalizing from current research to the claim that stereotype threat certainly or likely causes women to leave or “stay clear from occupations in the STEM domain.”^{Footnote 73} That is, Flore and Wicherts conclude that one should not (as of yet) claim that stereotype threat is likely causing underrepresentation. They do not argue further that stereotype threat is an unlikely or implausible explanation—presumably because they conclude that it does exist as a general effect, though its true effect size is uncertain. This distinction is important, especially for the philosophy profession that has even less evidence to work with than do STEM fields. The philosophy profession is not in a position to say with any certainty what is or is not causing underrepresentation.^{Footnote 74} Currently, what the profession can decide is what are plausible or implausible hypotheses. This general level of uncertainty is important for assessing which hypotheses should have a greater burden of proof, which is due in part to the risks of acting or not acting in certain ways.

Acting in the Face of Uncertainty: Underrepresentation and Inductive Risk

As mentioned in the previous section, some are sceptical of research on stereotypes, and do not think it gives one reasons for enacting certain kinds of reforms, such as increasing mentorship for women or making curricula more inclusive. Furthermore, one might agree that underrepresentation is likely due to pernicious factors such as discrimination, bias, and stereotype threat, yet still be sceptical of some of those proposed interventions. How should the profession assess which reforms are warranted, and does a focus on differences in interests change that calculus?

Yann Benétreau-Dupin and Guillaume Beaulac argue that there currently is not enough evidence to support any hypothesis as being a definitive cause of underrepresentation, which seems right.^{Footnote 75} Benétreau-Dupin and Beaulac still argue in favour of certain interventions meant to increase the proportion of women in philosophy, though, despite this empirical uncertainty. They appeal to independent reasons for adopting those interventions. Anonymous reviewing, for instance, will likely help the profession in other ways, such as increasing fairness for everyone. Such interventions may mitigate pernicious forces, should they be present, but they do not rely on those forces being present to be useful or worthwhile. I find Benétreau-Dupin and Beaulac’s reasoning compelling, but I will suggest an additional line of argument that may support certain interventions, even in the face of uncertainty regarding underrepresentation’s causes. Essentially, implementing those interventions may be less risky than not doing so.

Some of the philosophers whom I have discussed in this paper have brought up the issue of risk when assessing whether certain interventions meant to reduce underrepresentation and exclusion are warranted. Some of those interventions risk being unnecessary or inadvertently harmful. To assess these risks in an even-handed manner, it is helpful to think about them through the framework of inductive risk. Inductive risk is, as discussed by Heather Douglas^{Footnote 76} and Kevin Elliot and Ted Richards,^{Footnote 77} the ever-present possibility that inductive claims (e.g., scientific hypotheses) could be wrong. Not only is there epistemic risk involved—the risk of falsely believing something to be true—but there are also usually practical risks if one is wrong about an empirical matter. A standard example is someone deciding whether a certain chemical should be greenlit for commercial use, if it is uncertain whether that chemical is toxic in small doses. If scientists or the Food and Drug Administration endorse a certain dosage as safe, and they are wrong, their mistake could potentially lead to people getting sick or dying. However, if they claim that dosage is unsafe, but they are wrong, their mistake could lead to needless inefficiency in an industry, a slowdown of research that depends on that chemical, or even an adverse effect on public health, if that chemical is used in medicine.

A takeaway from work on inductive risk is that, when dealing with uncertainty and inconclusive evidence, individuals and institutions sometimes need to reflect on what they value (e.g., public health; economic efficiency) to help determine which hypotheses should be favoured or acted upon until more evidence is produced. Practical risks, weighted by background conditions and values, are, and should be, a key factor in some decisions to endorse or use a specific hypothesis or not. Applying this framework to underrepresentation means that, in the face of uncertainty regarding the causes of underrepresentation, the profession should choose which hypotheses it will treat as the default ones to be acted on until further evidence based in part on what are the risks of choosing one hypothesis and being wrong. Papineau, Pinker, and Sesardic and De Clercq have done part of this calculus in laying out some of the potential risks of treating the discrimination hypothesis as the default one (though they often frame this, unfairly I would argue, as individuals ‘assuming’ that discrimination is a major cause.) They do not give a full or even-handed account of all of the risks involved, however, which is crucial for making an informed decision about what policies and interventions should be undertaken or rejected. I do not give a full account here either, but I do sketch how one might fill in the gaps in risk assessment and explain why a more complete assessment may favour interventions that presume certain pernicious causes, even if evidence for those causes is still inconclusive.

First, one should consider the general risks that may be in play for both accepting and rejecting a hypothesis. As mentioned by others, there are risks if one posits discrimination or another sexism-related hypothesis (such as stereotype threat) as a major cause of underrepresentation, and one is wrong. Those risks are wasting resources on unneeded efforts to increase the proportion of women in philosophy, certain interventions inadvertently backfiring and decreasing the proportion of women in philosophy, unfairly favouring women in certain selection processes, and decreasing the profession’s time and energy on projects and issues more worthy of its attention. Some of these risks, such as efforts backfiring, are more-so risks of hastily implementing these interventions, and not slowing down or taking more time to assess the potential impacts of one’s actions.

Second, if the profession rejects or chooses not to act on the hypothesis that underrepresentation is significantly caused by sexism-related factors, and is wrong, it faces other risks. Primarily, it risks having a potential problem continue unabated, which may lead to individuals being unfairly excluded from philosophy. This could have serious material impacts, such as financial and institutional resources being unfairly distributed. It could also have serious psychological impacts, such as students unfairly receiving biased assessments of their work, experiencing more belonging uncertainty, not being encouraged to identify with philosophy, or not having sexist stereotypes about them rejected by their professional community. It could even potentially impact individuals’ health, insofar as stereotype threat and belonging uncertainty may lead to or exacerbate stress.^{Footnote 78}

Does the differences-in-interest hypothesis change this calculus? In some ways, not so much, but in other ways, yes. Even before the differences-in-interest hypothesis received attention, many of these potential risks were acknowledged. However, some of the risks regarding interventions aimed at influencing individuals’ interests now may deserve increased weight. Similarly, the risks involving stereotype threat’s various effects on interests, motivation, and self-identity might also deserve increased consideration. This is another reason that the differences-in-interests hypothesis is worth investigating further—doing so might give us a better understanding of what practically may be at stake, even if it does not immediately give us conclusive answers regarding the various theories being proposed.

Lastly, I suspect that a broader, more even-handed assessment of all the potential risks in play, when combined with background values that the profession does and should have—such as fairness and a desire not to exacerbate systematic exclusion—will favour enacting certain reforms over not enacting them, even if the phenomena that they are meant to mitigate are not confirmed by current evidence. Essentially, it is better to be safe than sorry (about one’s profession continuing to be sexist). Some reasons I suspect that an assessment of inductive risk would turn out this way is that some of the risks proposed by those such as Sesardic and De Clerq, Papineau, and Pinker, are easy to mitigate.

For instance, some suggest that encouraging more women to pursue philosophy might (a) backfire and scare them away instead or (b) end up ‘pressuring’ them to pursue a field against their interests and preferences.^{Footnote 79} Worry (b) is pretty easy to neutralize, by stipulating that philosophers should encourage women who seem potentially interested in philosophy to pursue it. One should not aggressively encourage every single woman in their class or that they meet on the street, regardless of their individual interests and goals, to pursue professional philosophy. For instance, a practice I engage in is taking time once a semester in my introductory courses, after the first or second paper of a class has been graded, to write personal letters or emails to students who have shown interest in philosophy. I tell them that their interest is noted, I think they might enjoy and excel in other philosophy classes (regardless of their grades in my class, since grades are something that one can improve), and that I would be happy to talk with them more about it. I also emphasize when their grades have improved over the course of the semester, in order to explicitly encourage those who might think they are not already good enough at philosophy to pursue it further.

I do not only write these letters to women and members of marginalized groups, but I do take extra time to consider students with marginalized social identities, and whether they have shown interest and engagement with the material. Doing this risks potentially failing to notice an openly straight white man who has shown some interest in the material, but I think that is an acceptable risk, given background circumstances which suggest that such a student is more likely to receive other forms of encouragement in his intellectual and professional pursuits.

Worry (a), that one might unintentionally scare away women in one’s attempt to mentor them, is a somewhat more substantial worry, but also one that is somewhat easy to mitigate. First, those philosophers who are not skilled at mentoring could seek out training to get better at it, reducing the likelihood of their mentoring attempts having inadvertent consequences. Second, this risk might be trumped by other risks. Telling students that there are significant levels of discrimination in the field, if that turns out to be mistaken, could adversely impact their understanding of and sense of belonging in philosophy. But their sense of belonging will be impacted even if it’s true that there is widespread discrimination, and likely even if a philosopher does not explicitly tell them this. Furthermore, not giving students any indication that there may be significant forms of discrimination against them, when there are, runs its own set of risks, such as setting those students up for disappointment, frustration, anger, and hurt—as well as not helping those students protect themselves against potential discrimination and bias.

Another version of this worry, particularly regarding stereotype threat and bias, is that, if one encourages students, but one seems to be encouraging them based solely on their social identity, those students might discount that encouragement, and actually take it as further evidence that no one believes that they have what it takes to do well on their own merits in philosophy. This is a valid worry, but one that again is somewhat easy to mitigate. When encouraging students, one should not indicate that one is encouraging them specifically because of their social identity, and one should cite individuals’ specific achievements, behaviour, and participation when possible. This might require some training for some philosophers, but it is overall not that hard a skill to develop.

Others still might have the additional worry that, even if it is true that discrimination and stereotype threat is widespread in the field, it is not a good idea to scare students with this information. However, we might owe it to our students to do so, especially if they may be the targets of bias, discrimination, or harassment. If the profession being honest about its state of affairs turns women away from philosophy, that is not a sign that we should tone down how we talk about the field. That is a sign that there is serious work the profession needs to do, to make itself worthy of future generations of scholars. Kristie Dotson observes, some people who leave philosophy “with a sigh of relief” have come to realize not, “I have failed at philosophy,” but rather, “philosophy has failed me.”^{Footnote 80} As Anita Allen argues, we should “shift the burden to the discipline to explain why it is good enough for us; we should be tired of always having to explain how and prove that we are good enough for the discipline.”^{Footnote 81}

Lastly, this sort of worry about discussions of discrimination intimidating or putting off women might be unfounded. If our students who are likely to be the targets of discrimination and bias know this, or already think that there is widespread discrimination and bias against them in society more generally, then they might not take a philosopher’s acknowledgement of potential discrimination in the field as a reason to steer clear of philosophy. They might, in fact, appreciate the honesty and one’s attempt to mitigate such discrimination. Therefore, talking openly about the potential existence of discrimination and bias might further encourage students who have marginalized social identities. When one tallies up all of these risks, I suspect that it turns out that increased efforts at mentoring and encouraging women will be worth any of the risks they entail—again, even if it is currently still inconclusive whether discrimination, bias, and stereotype threat are widespread in the field or significant factors in underrepresentation.

What Do Philosophers Want (Philosophy to Be)?

Another sort of worry or risk that is sometimes raised in discussions on underrepresentation is the risk of damage to the integrity of philosophy if philosophers broaden their curricula to be more inclusive, but the new and more diverse material turns out to be ultimately less worthy than the old, traditional subject matter. Papineau, for instance, argues that, if women and men are interested in different topics of study, then we may be tempted to expand the scope of philosophy so that it includes more topics that are of interest to women. Such a response, Papineau argues, could lead to the content and boundaries of philosophy being dictated arbitrarily by a set of special interests, i.e., women’s.

Even if one grants this as a potential risk, one should also take into account the implications and risks that Papineau does not mention. For instance, if Papineau is right about women’s interests serving as a potentially arbitrary influence on philosophy, then the boundaries of philosophy are already likely determined in part by arbitrary interests—men’s. Women were culturally shut out of philosophy for centuries, so if the genders really do have innately different interests, how could that claim entail anything but, as Margaret Crouch puts it, “an admission that philosophy as a discipline is dominated by […] male interests.”^{Footnote 82}

More generally, as things stand, philosophers cannot yet factor in all of the risks involved in changing philosophy’s curricula, because there is deep disagreement about what the boundaries of philosophy are and should be. Thus, if the profession is going to continue investigating whether women are interested in philosophy—and I think it should—it should also ask what philosophers want, in terms of professional norms regarding the boundaries of the profession and what counts as having a ‘philosophical’ interest. How broadly or narrowly we choose to define ‘philosophy’ might have a significant impact on whether men and women are ultimately interested in philosophy to similar degrees.

As part of this question, the profession needs to continue thinking about who has a say in what counts as a philosophical interest. Papineau presents one account, but it is a non-obvious one that would need an argument before being accepted as a profession-wide norm. Papineau claims that, if philosophers offer more courses that appeal to women (as women), they risk taking away resources from topics that potentially are more deserving of “sustained philosophical attention.”^{Footnote 83} A better approach, according to Papineau, is “first to decide which topics are really worth studying, and then to see who wants to study them.”^{Footnote 84} Papineau’s proposed method would prioritize what current philosophers are interested in studying, at the expense of what future philosophers, or not (yet) philosophers are interested in. Perhaps Papineau thinks that such disagreements would not happen, and that philosophy is a kind of timeless and universal endeavour that can be settled once and for all. Those premises are, to say the least, somewhat controversial. Thus, to fully address the risks involved in expanding curricula or keeping it the same, the profession needs to continue thinking about who has a say in shaping philosophical research and the content taught in classes.

These are the sort of questions that Dotson has in mind when she urges philosophers to consider the disciplinary power and politics behind the question: “How is this paper philosophy?”^{Footnote 85} There are issues of power and control within that question, as well as in judgements of whether certain projects are “‘philosophically uninteresting,’ ‘unphilosophical,’ ‘obscure,’ ‘radical,’ ‘counterintuitive,’ ‘implausible,’ ‘misguided,’ ‘trivial,’ ‘irrelevant,’ or ‘irrational,’” as Carmen Maria Marcous argues.^{Footnote 86} If one suspects that reforms to philosophy curricula might damage or undermine the integrity of philosophy, whose vision of the field is being undermined? Whose control over philosophy would wane if those reforms are enacted?

I want to conclude by thinking about George Yancy’s suggestion (taken from Paul Weiss) that one could view philosophers as a kind of “troublemaker”^{Footnote 87} On Yancy’s account of philosophy, black philosophers would be, counter-stereotypically, a paradigm of being a philosopher, because “Doing philosophy in Black skin is always already a site of troublemaking; it is to trouble the mythos of white superiority.”^{Footnote 88} This suggestion is interesting for two reasons. First, it suggests a way to frame philosophy that might counteract some of the stereotypes about the field that may form a barrier for certain groups identifying with the profession. If we take the troublemaker as a paradigm of doing philosophy, more people of colour and women may be able to better identify themselves with philosophy because they will have a sort of exemplar or role into which they can project themselves. Kate Manne argues that some of the core activities in philosophy, such as “disagreeing with, challenging and correcting people” are “inherently hierarchical” and that “[o]ne is typically not supposed to break the ranks of gender, race, or class in doing so, in the course of ordinary social life.”^{Footnote 89} Thus, philosophy has “the potential to be exceptionally liberating for members of non-privileged groups”^{Footnote 90}—especially if individuals can understand their role as a philosopher as one of social troublemaker. This theme of finding a role with which one can identify is already present in Yancy’s and Dotson’s work, but it can be fleshed out further by work on belonging uncertainty and stereotypes. Individuals might be able to turn experiences of belonging uncertainty and stereotype threat into experiences of further identifying with philosophy. An initial sense of discomfort could be taken, not as evidence that one does not belong, but as evidence that one is fulfilling a role as philosophical gadfly.

Second, thinking about philosophy as troublemaking suggests that the troublemaker might have different goals and values, as well as a different calculus of risks, than non-troublemakers. To give a full account of inductive risk, one needs to ask whose interests are informing what counts as a risk. We should consider whether the ‘risks’ of which some are so afraid, such as damaging the integrity of philosophy, might instead be a form of philosophical troublemaking that the profession badly needs. In summary, the profession should continue thinking about what women want, but it should also think about what men want and have historically wanted, what philosophers want, what the troublemakers want, and what those who are not (yet) philosophers want.