Not So Irrelevant: The Epistemic Significance of Social Identity

Abstract

Our social identity affects what we believe. But, how should we epistemically evaluate this doxastic impact? Achieving a robust picture of the epistemic significance of social identity requires us to explore the understudied intersection of irrelevant influences and standpoint epistemology, which leads us to cases of double higher-order evidence. Reflecting on social identity through the lens of irrelevant influences gives us higher-order evidence of error, while reflecting through the lens of standpoint advantage gives us higher-order evidence of accuracy. We must weigh the strength of each piece of higher-order evidence case by case to epistemically evaluate the doxastic impact of social identity.

Résumé

Notre identité sociale influe sur nos croyances. Mais comment évaluer cet impact doxastique d'un point de vue épistémique ? Pour obtenir un portrait fiable de l'importance épistémique de l'identité sociale, nous devons explorer la jonction peu étudiée des influences non pertinentes et de l’épistémologie du point de vue, ce qui nous amène à des cas de double preuve d'ordre supérieur. Le fait d'envisager l'identité sociale sous l'angle des influences non pertinentes nous donne une preuve d'ordre supérieur d'erreur, tandis que l'examen selon l'avantage du point de vue fournit une preuve d'ordre supérieur d'exactitude. Nous devons donc soupeser la force de chaque élément de preuve d'ordre supérieur au cas par cas afin d’évaluer épistémiquement l'impact doxastique de l'identité sociale.

1. Introduction

It's uncontroversial that features of our social identity — gender, race, religion, socio-economic position, nationality, sexual orientation, and (dis)ability status — affect what we believe. But, what should we make of the epistemic significance of this doxastic impact? This is where the controversy comes in.

Two branches of literature in contemporary epistemology seem to recommend conflicting views here: the literature on irrelevant influences suggests that the doxastic impact of social identity is epistemically problematic, while the literature on standpoint epistemology suggests that it's epistemically beneficial. To bring this tension into focus, consider the following case:

Abortion: Abby is a sexually active woman of reproductive age living in Texas, where access to abortion care is under threat. As a woman, Abby is highly attuned to a range of considerations bearing on the moral status of abortion: the heavy physical and emotional tolls of pregnancy (especially an unwanted one), the deep impact that carrying and bearing a child has on one's life path, the various stages of fetal development, and all the things that go into women making reproductive decisions with their doctors. After reflecting on the issue — and in large part based on considerations like these — Abby comes to believe that abortion is generally morally permissible, and, thus, that abortion should be legal. When she discusses her view with her conservative friend Conor, he questions whether Abby believes what she does about abortion just because she's a woman.

We are all familiar with the sort of accusation Conor levies against Abby: You just believe that because you're a woman … a Muslim … an American … because you're Black … or rich … or trans.

These familiar accusations point us back toward the uncontroversial fact with which we started: many of our beliefs have been influenced by features of our social identity. This fact about us leads to a worry: features of our social identity are epistemically irrelevant, and so beliefs that have been influenced by them are epistemically problematic and, perhaps, in need of revision. To the extent that beliefs should be rational and true, they should be based solely on arguments and evidence, not on things like our gender and race.

The worry becomes even more troublesome when we recognize not only the pervasiveness, but also the depth, of such irrelevant influences. Many of our most dearly held beliefs — the religious, moral, political, and philosophical beliefs around which we centre our lives — are far from immune from, and may be even most susceptible to, the influence of social identity. If irrelevant influences are epistemically problematic, and if rationality calls for doxastic revision when we learn of a problematic irrelevant influence, then we'd be left with the troubling verdict that rationality requires mass scepticism.

Epistemologists have addressed these worries at length in the contemporary literature on irrelevant influences (Avnur & Scott-Kakures, 2015; Cohen, 2000; Dipaolo & Simpson, 2016; Mogensen, 2016; Vavova, 2018; White, 2010). Many of them seek to sort out which seemingly irrelevant influences are epistemically problematic and which are innocuous.¹ The thread running through this treatment of irrelevant influences is that the familiar sort of accusation above — that we believe something because of some feature of our social identity — is prima facie worrisome, and the best possible result is that the influence turns out to be epistemically insignificant.

But, as Abby's case highlights, this treatment of the epistemic significance of social identity doesn't tell the whole epistemic story: Abby may be in an excellent position to get it right about the moral status of abortion, in part because, as a woman, she is highly attuned to relevant truth-related considerations.² This suggests that sometimes features of our social identity are epistemically advantageous, rather than problematic (or, at best, innocuous). In other words, instead of being either insignificant or significant but bad, the doxastic impact of social identity is sometimes significant and good. This idea is developed at length in the literature on standpoint epistemology.³ In this literature, the epistemic significance of social identity is celebrated, rather than fretted over. Features of our social identity are seen as putting us in a better position to get it right about certain matters.

So, where does this leave us? From an epistemological standpoint, should we worry about or celebrate the doxastic impact of our social identity? Should Abby worry about or celebrate the influence her gender has had on her beliefs about abortion?

The goal of this article is to develop an answer to these questions by exploring the understudied intersection of irrelevant influences and standpoint epistemology. After briefly looking at the treatment of the epistemology of social identity through the lens of each branch of the literature, I set out to identify and resolve the tension between them. I argue that — for some believer S's belief that p, and for some feature of S's social identity F — reflecting on the epistemic significance of F on the relevant belief can give S conflicting pieces of higher-order evidence. Looking at the epistemic significance of F through the lens of irrelevant influences gives S a piece of higher-order evidence of error, while looking at the epistemic significance of F through the lens of standpoint advantage gives S a piece of higher-order evidence of accuracy. Whether S should celebrate or worry about her belief that p in light of F, I conclude, depends on the respective strength of each piece of higher-order evidence. Along the way, I explore the various considerations that determine the strength of each piece of higher-order evidence, and the various types of cases when the epistemic significance of social identity should be celebrated, and those when it should be worried about.

2. Delineating Irrelevant Influences

Let's start with irrelevant influences.

There are two main ways of framing irrelevance in the literature on irrelevant influences: first, as irrelevant to the truth of the propositional content of the belief in question, and, second, as irrelevant to the believer's epistemic position with regard to that content.⁴ For our purposes, it's best to combine these two frameworks and characterize an irrelevant influence like this: a factor that influences (causes, gives rise to) S's belief that p but bears neither on the truth of p nor on S's epistemic position with respect to p.⁵

This characterization of irrelevant influence is broad, leaving lots of room for a variety of irrelevant influences. To direct our focus, it'll be helpful to follow Yuval Avnur and Dion Scott-Kakures (2015) and look at a particularly prominent type of irrelevant influence: desire, broadly construed to include hopes, fears, aversions, and wishes. Not all desires that influence beliefs are irrelevant on Avnur and Scott-Kakures's view. Some desires put us in positions to attain and appreciate (or ignore and overlook) certain evidence relevant to our beliefs. For example, my love of sweet treats, and my corresponding desire to buy and consume them, puts me in a good epistemic position with regard to the location of local bakeries. Avnur and Scott-Kakures call these “positional” desires. Positional desires do not necessarily constitute irrelevant influences since they help determine what evidence we have and, by extension, what we are in a position to rationally believe.

Directional desires, in contrast, are more worrisome. These desires incline us to handle evidence in ways that favour a particular outcome, thus exerting a “directional” force on our beliefs. A paradigmatic example is of a proud parent's evaluation of his child's talent. Because the proud father wants to see his daughter succeed and stand out amongst her peers, he will be prone to handle the relevant evidence in ways that favour the view that she is especially talented. These directional desires are clearly irrelevant influences: they influence our beliefs but do not bear on the truth of their content, or on our epistemic position.

Avnur and Scott-Kakures argue that directional desires are often behind the type of familiar accusation in the introduction: “You just believe that because you're X,” where X is a placeholder for various potential social labels (Catholic, woman, Black, rich, trans). On their view, these labels serve as a proxy for relevant directional desires, and the worry encapsulated in these accusations is that X implicates directional desires on the beliefs in question. For example, consider the accusation that Cathy believes that the pope is infallible just because she's Catholic. The thought here is that, as a Catholic, Cathy has certain desires — a desire to be a “good” Catholic, a desire to remain fully psychologically and emotionally invested in her religious community, a fear of lacking faith — that exert a directional force on her doctrinal beliefs.

Directional desires are epistemically pernicious; this means that, when we learn that a belief of ours has likely been influenced by a directional desire, we should worry, and probably revise our belief. This thought can be cashed out in a general way and a specific way. Generally, irrelevant influences are epistemically problematic because they manipulate reasoning. If our reasoning has been manipulated, then it's been sensitive to things besides the relevant evidence. And, if our reasoning has been sensitive to things besides the relevant evidence, then the resulting beliefs are irrational and less likely to be true. Specifically, directional desires are epistemically problematic because directionally influenced reasoning is less reliable than non-directionally influenced reasoning, since what we desire doesn't reliably track with the truth. If a desire that doesn't track with the truth sways our reasoning process, then the resulting beliefs are irrational.

3. Social Identity and Irrelevant Influence

Features of social identity are often appealed to as central examples of irrelevant influences in the contemporary literature (Bogardus, 2013; Mogensen, 2016; Vavova, 2018). For our purposes, social identity tracks membership in certain social groups, where membership in the relevant groups tends to play a substantial role in members’ individual self-conceptions. Paradigmatic features of social identity include gender, race, religion, socio-economic class, nationality, sexual orientation, and (dis)ability status.

At first glance, it's easy to see why features of our social identity are often considered central examples of irrelevant influence. After all, these features are poised to give rise to desires that may exert a directional force on many of our beliefs. That's in part because our social identity is intimately intertwined with the conative components of our psychology. What's more, much of our social world is structured around social identity, and what's in our own best interest is often heavily dependent on the surrounding social structures and circumstances in which we find ourselves. It is no surprise, then, that what is in our best interest often significantly depends on features of our social identity. Because our desires tend to closely track our interests, we should expect that our social identity often implicates desires that risk posing a directional force on our beliefs.

We've already seen one example of this: a Catholic having desires that exert a directional influence on her doctrinal beliefs. But, there are many more examples. Imagine a Black high school student being accused of defending affirmative action initiatives in university admissions just because he's Black; a trans man being accused of accepting a psychological account of gender just because he's trans; or a Palestinian being accused of rejecting Israeli settlements in the West Bank just because he's Palestinian.

The epistemic worry is similar in all these cases: the social identity of the accused believers — Catholic, Black, trans, Palestinian — implicates a vested interest in the truth of certain beliefs about the topic at hand. This vested interest, in turn, implicates desires that threaten to exert a directional force on these beliefs. When accusations with the form “You just believe that because you're X” are levied against these believers, it seems that they have good — though defeasible — reason to worry that their beliefs are mistaken. After all, their beliefs may be based on things that do not bear on their truth, and rationality may require these believers to revise their beliefs in light of this worry.

4. Not So Irrelevant: Lessons from Standpoint Epistemology

Things look very different when we evaluate the epistemic significance of social identity through the lens of standpoint epistemology. Through this lens, the same features of our social identity that seemed epistemically pernicious when thought of as irrelevant influences seem, instead, epistemically advantageous. Standpoint epistemology suggests that features of our social identity are not irrelevant in many cases; rather, they're very relevant and, often, in a good way.

“Standpoint epistemology” refers to a family of views according to which features of our social identity can be epistemically beneficial. One of the central tenets of standpoint epistemology is the Situated Knowledge Thesis, which says that features of our social identity determine what we are in a position to know (Toole, 2021, p. 340). This thesis can also be phrased in terms of epistemic justification or rationality: features of our social identity help determine what we are in a position to justifiably or rationally believe.⁶

The thought behind the Situated Knowledge Thesis is that social facts related to our social identity shape our circumstances, interests, motives, and experiences (CIMEs) and these socially shaped CIMEs then go on to bolster our epistemic position when it comes to some subject matters (Toole, 2021, p. 340). First, our CIMEs shape our conceptual and linguistic resources, which help determine what we are in a position to rationally believe. Second, our CIMEs “open us up” to evidence available in our socio-epistemic environment that those with similar conceptual and linguistic resources will overlook or otherwise fail to grasp. So, facts about our social identity ultimately help determine what we are in a position to rationally believe by way of influencing the very evidence we have and the conceptual and linguistic resources that enable us to make sense of this evidence.

To get a better grasp on the Situated Knowledge Thesis, it'll be helpful to walk through an example. Consider Benoît, a chef and restauranteur who loves making and serving delicious meals to his customers. Social facts related to Benoît's career will end up being epistemically advantageous for him. These social facts shape Benoît's interests (learning about unfamiliar ingredients, discovering complementarity among surprising flavours, running a profitable business), his motives (producing yummy food, making his customers happy, increasing revenue), and his culinary experiences (many hours in the kitchen, tasting different foods, arranging tables in a dining room). These social facts also determine the types of people and things that populate Benoît's day-to-day life (herbs and spices, various cuts of meats and cheeses, sous chefs, servers).

From here, Benoît's CIMEs help determine the very evidence he has, as well as the conceptual and linguistic resources that allow him to make sense of that evidence.

First, Benoît's CIMEs put him in a position to notice, understand, and grasp information available in his socio-epistemic environment that others may overlook. Given his experience and interests as a restauranteur, Benoît may notice that the tables at a restaurant he is visiting are set about 5 feet apart, that the server seems overworked, and that the food takes 30 minutes after ordering to come out. His dining companion (who works outside of the restaurant business) may have the relevant concepts of 5 feet, understaffed, and 30 minutes but — given her lack of interest and experience in the restaurant world — she may not pick up on the same things. The upshot is that Benoît has a different and better body of evidence about the restaurant than does his dining companion despite their shared socio-epistemic environment.

Second, Benoît's socially shaped CIMEs determine his cooking-related concepts and language. Given his CIMEs, Benoît has concepts and language for rosemary and tarragon, for parmesan and pecorino, for sautéing and simmering. These concepts enable him to identify and distinguish between ingredients and actions-types. Other people — those who have little experience in the kitchen or interest in cooking — will lack these resources. They may have the concept cheese but not the concepts or language to distinguish between different varieties.

Importantly, Benoît's conceptual and linguistic resources are epistemically significant in that they determine what he is in a position to rationally believe. Equipped with his cooking-related concepts and language, Benoît is in a position to know propositions like “that is Havarti” by looking at a charcuterie board; “this soup has rosemary and thyme” by tasting the soup; and “she is pressing the garlic” by observing someone cook. Those who do not share Benoît's array of cooking-related language and concepts are not in a position to rationally believe these propositions since they don't have the relevant epistemic tools.

5. Social Identity and Standpoint Advantage

This example sets us up to grasp the epistemic advantages of social identity — call it “standpoint advantage.” Facts related to our social identity shape our CIMEs. These socially shaped CIMEs, in turn, shape our conceptual and linguistic resources, which enable us to rationally believe propositions involving these concepts and language. They also “open us up” to evidence available in our environment that others will fail to pick up on.

One important upshot of this is that members of marginalized social groups will tend to have an epistemic advantage when it comes to propositions related to their marginalization⁷: Black people will tend to have conceptual and linguistic resources related to racism, and will be more open to evidence of racism in their environment, than white people; women will tend to have conceptual and linguistic resources related to sexism, and will be more open to evidence of sexism in their environment, than men; and so on.⁸

To help get a handle on standpoint advantage, it will be helpful to consider a particular case. This one is from Briana Toole:

Vertical Campus: Elsabeth and Janie are both students at Bovie College, a vertical campus located in a busy district in the heart of Manhattan. Elsabeth and Janie both have classes on the top floors of the building, which requires they take stairs, escalators, or elevators. Elsabeth, who uses a wheelchair, finds the trek especially tedious, as she can use neither the stairs nor escalator, and must wait for the elevator. Particularly problematic for her is the lack of “activation switches,” or ADA push buttons, that open the heavy doors she must pass through en route. As there are no such buttons on the floor where her class is located, she often must wait for someone else to open the door for her. When the college sends around an accessibility survey to its students, Elsabeth responds that the building is not accessible to people with disabilities, and offers specific steps the college can take to improve. Janie responds that she doesn't know if there is any accessibility problem on campus, but notes that things seem fine to her (Toole, 2021, pp. 338–339).

Elsabeth's social identity as a disabled person puts her in an excellent epistemic position to rationally believe certain things about the accessibility of her campus. Her socially shaped CIMEs help determine her conceptual and linguistic resources — for example, her concept of and language for ADA ⁹ push buttons. Her social identity also “opens her up” to evidence that non-disabled people who have similar conceptual and linguistic resources will be prone to overlook. Elsabeth will be much more likely than Janie to pick up on the fact that some floors in her campus building are missing those ADA push buttons.

6. Identifying and Resolving the Tension

We have now explored the epistemic significance of social identity through two lenses: the lens of irrelevant influence and the lens of standpoint epistemology. When seen through the lens of irrelevant influences, features of our social identity are epistemically worrisome because they often implicate desires that exert a directional force on our beliefs. Forming beliefs under the influence of directional desires tends to be inaccurate, and rationality calls for revision of inaccurate beliefs. But, when seen through the lens of standpoint epistemology, features of our social identity often put us in an excellent epistemic position that enables us to rationally believe things that we otherwise wouldn't be able to rationally believe. By shaping our CIMEs, our social identity helps determine our linguistic and conceptual resources and “opens us up” to available evidence. In some cases, we have reason to think that directional desires and standpoint advantage are in play when it comes to the same beliefs.

So, where does this leave us? From an epistemological perspective, should we worry about or celebrate the doxastic influence of our social identity on our beliefs? Perhaps unsurprisingly, I think the answer is: It's complicated.

What makes things so complicated is that reflecting on the epistemic significance of our social identity can give rise to two conflicting pieces of higher-order evidence. Higher-order evidence is evidence that a believer has about either the quality of her first-order evidence or how she has responded to her first-order evidence.¹⁰

With that in mind, consider the epistemic significance of some feature of social identity F on S's belief that p. On the one hand, by reflecting on the lesson from irrelevant influences — that F may implicate a directional desire on her belief that p — S has a piece of higher-order evidence of error: her social identity makes it more likely for her to get it wrong about p. Given F, either S's first-order evidence about p is more likely to be bad or otherwise deficient, or else S is more likely to have reacted improperly to her first-order evidence. On the other hand, by reflecting on the lessons from standpoint epistemology — that F puts her in an excellent epistemic position with respect to p — S has a piece of higher-order evidence of accuracy: her social identity makes it more likely that she got it right about p.¹¹ Given F, S's first-order evidence is more likely to be good or otherwise strong, or else S is more likely to have reacted to her first-order evidence properly.

So, it seems, reflecting on the epistemic significance of one's social identity gives rise to an interesting sort of case involving higher-order evidence — namely, cases of double higher-order evidence. These cases involve (i) a body of first-order evidence, (ii) a piece of higher-order evidence of error from irrelevant influences, and (iii) a piece of higher-order evidence of accuracy from standpoint advantage.

In these cases, we should worry about the epistemic significance of our social identity when the higher-order evidence of error is stronger than the higher-order evidence of accuracy; but, we should celebrate when the higher-order evidence of accuracy is stronger than the higher-order evidence of error. To put it another way, when the epistemic effects of our social identity are net negative — when, all higher-order things considered,¹² being F makes it more likely that the believer got it wrong about p — rationality requires revision. When the epistemic effects of our social identity are net positive — when, all higher-order things considered, being F makes it more likely that the believer got it right about p — rationality recommends remaining steadfast.¹³ Rationality also recommends remaining steadfast when the epistemic effects of the believer's social identity are net neutral — when, all higher-order things considered, being F make it no more likely that the believer gets it right or wrong about p.

As we will see, one consideration that can help us determine the valence of the net epistemic effects of F is whether an otherwise similar believer S* with a different social identity in the same category F* (for example, man when F picks out women) has a different sort of epistemic advantage when it comes to p — whether, for example, being F* puts S* in a position to appreciate another set of truth-related considerations that bear on p. If being F* puts S* in an epistemically comparable position to S regarding p, then this casts doubt on the claim that being F makes it more likely for S to get it right or wrong about p.

7. The Good Independent Reason Principle

Thankfully, there are theoretical resources already available in the literature on higher-order evidence that can help us evaluate the various epistemic considerations in play in double higher-order evidence cases. In particular, I have in mind Katia Vavova's Good Independent Reason Principle (GIRP): “To the extent that you have good independent reason to think that you are mistaken with respect to p, you must revise your confidence in p accordingly — insofar as you can” (Vavova, 2018, p. 145).

Good independent reason: GIRP is formulated with an independence condition — we must evaluate the epistemic status of our belief that p independently of the first-order evidence and reasoning that led to our original position (Christensen, 2007, 2011; Elga, 2007). In other words, a good reason for thinking that we're (not) mistaken about p cannot rationally be based on the first-order evidence and reasoning we used to arrive at our belief that p. In Vavova's words, “[Y]ou must evaluate your epistemic situation in a way that doesn't stack the deck in your favor. This means you shouldn't assume exactly what has been called into doubt” (Vavova, 2018, p. 145). GIRP, then, is a principle about higher-order evidence. GIRP tells us to revise if we have higher-order evidence of error. Importantly, it is sensitive to any sort of higher-order evidence that gives us good independent reason to think that we're (not) mistaken with respect to p. This higher-order evidence may come from a variety of sources: peer disagreement, irrelevant influences, a message from the epistemology oracle or — we might add — reflections on standpoint advantage.

Mistaken with respect to p: We may be mistaken with respect to p if our belief that p is irrational or false.

Revise accordingly — insofar as you can: When we have good independent reason to think that we're mistaken about p, GIRP directs us to revise our belief that p accordingly and insofar as we can. Vavova purposefully leaves these directions imprecise because how we should revise and to what extent (slightly or drastically or somewhere in between) depend on the contingencies of a given situation. It depends, for example, on how strong our higher-order evidence is, how likely it is that our belief is mistaken, and the extent to which it is likely to be mistaken.

Although GIRP is explicitly concerned with higher-order evidence of error, it clearly has implications for higher-order evidence of accuracy. Whether we have good independent reason to think that we're mistaken with respect to p depends, in part, on whether we have competing higher-order evidence of accuracy. If we have higher-order evidence suggesting that we've gotten it right with respect to p that is stronger than any higher-order evidence suggesting that we've gotten it wrong, then — all higher-order things considered — we do not have good independent reason to think that we're mistaken with respect to p and GIRP does not recommend revision (we've already gotten it right, after all!). In this case, our higher-order evidence of accuracy rebuts our higher-order evidence of error.

8. Applying GIRP: Social Identity and Double Higher-Order Evidence Cases

GIRP can help us evaluate the epistemic status of the beliefs in question in double higher-order evidence cases. In these cases, we have some good independent reason to think that our social identity has put us in an excellent evidential position with respect to p (higher-order evidence of accuracy), but also some good independent reason to think that it has given rise to a vested interest in the truth value of p, implicating directional desires on our p-related beliefs (higher-order evidence of error).

When it comes to these cases, the question is whether, all higher-order things considered, we should worry about or celebrate the doxastic impact of our social identity. First, we'll consider celebratory cases — those cases when the higher-order evidence of accuracy outweighs the higher-order evidence of error. Second, we'll consider worrisome cases — those cases when the higher-order evidence of error outweighs the higher-order evidence of accuracy. Third, we'll consider more complicated cases when it's less clear whether worry or celebration is called for.

8.a Celebratory Cases

There are at least two types of celebratory double higher-order evidence cases.

In the first type of case, the believer's social identity gives rise to a primary vested interest in getting it right about p (whatever the truth may be). This type of case is familiar. After all, it is usually prudentially best for us to have true beliefs and avoid false ones. Thus, there are going to be many times when our social identity leads us to have a vested interest in having true beliefs about p (whatever the truth may be). Consider the following case:

Career Deliberation: Ben, a Black employee, feels limited in his current position at his company. He is deliberating about whether to switch companies or stick around in hopes of a promotion. As a part of his deliberation, he wants to take into account any racial disparities in promotions at his current company.

Given his race, and in order to make the best career move, it is in Ben's best interest to have true beliefs about any racial disparities in promotions at his current company. What's more, given his CIMEs, Ben is especially attuned to the existence and implications of such disparities, in part because they disproportionately negatively affect people like him. In this case, Ben has a vested interest in getting it right about the matter at hand, which makes the relevant piece of higher-order evidence of error from irrelevant influences trivial. It is easily outweighed by his piece of higher-order evidence of accuracy from standpoint advantage.

In the second type of celebratory double higher-order evidence, a believer's social identity gives rise to a vested interest in a particular outcome (rather than a vested interest in getting it right), but the higher-order evidence of accuracy from standpoint advantage still manages to outweigh the higher-order evidence of error from irrelevant influences. It's plausible that Vertical Campus falls into this category of celebratory case. We considered this case above as an example of social identity giving rise to standpoint advantage: being disabled gives Elsabeth a robust evidential edge when it comes to beliefs about campus accessibility. But, we can add a few plausible details about Elsabeth in order to make Vertical Campus a double higher-order evidence case. Suppose that, as a disabled person, Elsabeth has a general vested interest in it being true that her campus is accessible to people like her — perhaps it would be moderately disappointing for Elsabeth if her academic community failed to value students like her enough to implement appropriate accommodations. The worry is that this vested interest puts Elsabeth at risk of believing that her campus is more accessible than it is.

Reflecting on this interest would give Elsabeth a piece of higher-order evidence of error, while reflecting on her standpoint advantage would give Elsabeth a piece of higher-order evidence of accuracy. Upon weighing these two pieces of higher-order evidence, though, it seems reasonable to conclude that the piece of higher-order evidence of accuracy outweighs that of error. Elsabeth's standpoint advantage is very strong — someone with her (dis)ability status is very likely to have excellent first-order evidence about the accessibility of frequented buildings. In contrast, the vested interest in it being true that her campus is accessible — because of its generality and lack of emotional intensity — is relatively weak.

Before moving on to worrisome cases, let's consider an interesting subset of the second type of celebratory cases. In these cases, the believer has a vested interest in p being true — where p describes some state of affairs that is bad for the believer — because, if p is true, then changes will be made (or at least called for) that improve the relevant state of affairs for the believer. Consider the following case:

Departmental Climate: Fatima is a professor in a prestigious philosophy department, where men vastly outnumber women. Although there are no complaints of sexual harassment within the department, Fatima finds the department's climate a bit troubling. In particular, she notices the small ways in which some of the departmental members treat women differently than men: some professors close the door when meeting with men grad students but not with women; go out for beer with men but not women; and discuss openly the need to invite “at least one woman” colloquium speaker for the sake of diversity. She also notices how the women in her department are implicitly expected to take on the majority of the department's emotional labour. An independent investigation team is evaluating the department's climate. If the investigation finds that the department is less-than-friendly to women, the department has committed to implementing the measures recommended by the team to improve the climate for women. When Fatima is asked by the independent investigation team to rate the department's friendliness to woman on a scale of 1 (very unfriendly) to 10 (very friendly), she rates it a 4.

The striking thing about this case is that Fatima, as a woman, seems to have a pronounced vested interest in it being true that the department is unfriendly to women because, if true, then changes will be made that would improve her workplace situation.

Being a woman also puts Fatima in an excellent evidential position when it comes to evaluating departmental climate: her CIMEs have given her conceptual and linguistic resources like emotional labour and tokenism. They also open her up to available evidence that those who lack similar CIMEs tend to overlook. For example, as a woman, Fatima is more likely to be sensitive to expectations regarding emotional labour since those expectations are likely to fall on her. What's more, it seems that there is nothing about being a man that gives men a different but similarly pronounced evidential advantage when it comes to evaluating departmental climate. So, being a woman does seem to give Fatima an overall evidential boost here when it comes to the matter at hand.

Suppose that Fatima recognizes the conflicting ways that her gender has influenced her epistemic position. In effect, she has two pieces of higher-order evidence. One piece of higher-order evidence (arising from reflections on irrelevant influence) suggests that her rating of the department's climate is irrationally low. Another piece of higher-order evidence (arising from reflections on standpoint advantage) suggests that her rating of the department's friendliness to women is correct. So, all higher-order things considered, do these reflections on the epistemic significance of her gender identity give Fatima good independent reason to worry or celebrate?

Let's first evaluate the piece of higher-order evidence from irrelevant influences. The implicated desire is the desire for the recommended climate improvement measures to be implemented. How likely is it that this desire exerts a directional force on Fatima's belief about departmental climate? It's plausible here that Fatima's desire to have the relevant measures implemented is contingent on the truth of the belief in question: the belief is explanatorily prior to — and grounds for — the desire.¹⁴ After all, if the departmental climate were fully friendly to women, then Fatima wouldn't have a pronounced vested interest in the measures being implemented. Given this, it is unlikely that a desire to have the measures implemented exerts a directional force on the belief in question. It's because the departmental climate is unfriendly that Fatima has a vested interest in having the measures implemented; it's not because Fatima wants the measures to be implemented that she believes that the department is unfriendly. The upshot here is that Fatima's higher-order evidence of error from irrelevant influences is not very strong, and is outweighed by her higher-order evidence of accuracy from standpoint advantage.

The epistemic dynamics in Departmental Climate can be seen in many other double higher-order evidence cases. For another example, consider a revised version of Vertical Campus in which Elsabeth knows that her university will take steps to make the campus more accessible — by, say, installing more ADA push buttons — depending on the results of the accessibility survey. It might be thought that this gives Elsabeth a pronounced interest in it being true that the campus is inaccessible because, if true, then changes would be made (or at least called for) that would improve her situation. But, like Fatima's belief that her department is less-than-friendly to women, Elsabeth's belief that the campus is inaccessible is prior to the relevant desire for things to change. And, when the relevant belief is prior to the relevant desire, rather than the other way around, it is unlikely that desire exerts a directional force on the belief.

In both the first and second type of celebratory case, believers have better independent reason to think that their social identity has made it more likely for them to get it right than wrong about the matter at hand. All higher-order things considered, they do not have good independent reason to think that their belief is mistaken, and so GIRP recommends remaining steadfast. These believers should celebrate, rather than worry about, the doxastic impact of their social identity on the relevant beliefs.

8.b Worrisome Cases

Let's now turn to worrisome double higher-order evidence cases.

These cases lack the features discussed in relation to celebratory cases that trivialize or otherwise weaken the higher-order evidence of error from irrelevant influences. First, in these cases, the relevant feature of the believer's social identity gives rise to a vested interest in the truth of p (rather than a vested interest in getting it right about p, whatever the truth may be). Second, in these cases, the relevant belief is not prior to the desire — and, thus, it's more likely that the desire grounds the belief, instead of the other way around. Consider one such case:

Tax Break: Veronica works as a venture capitalist. Her personal net worth reaches into the hundreds of millions of dollars. The government is considering a bill that will give super rich people like her a substantial tax break. If the new tax bill is passed, Veronica will be allowed to keep millions of dollars that she would have otherwise had to pay in taxes. The government hopes that this tax break will encourage people like Veronica to invest in local business, thereby stimulating the economy in ways that benefit everybody. Given her elite vantage point, Veronica has seen firsthand, time and again, how her personal investments (and those of her super rich peers) can save fledgling companies, create and preserve high-paying jobs, and fund outstanding products that consumers enjoy. Veronica believes that the new tax proposal would be universally beneficial for all, not just for economic elites like herself.

Reflecting on the epistemic significance of her epistemic position gives Veronica two pieces of higher-order evidence: a piece of higher-order evidence of accuracy from standpoint advantage, and a piece of higher-order evidence of accuracy from irrelevant influences. How strong is the piece of evidence from standpoint advantage? Well, it's plausible that being a member of the economic elite has shaped Veronica's CIMEs in ways that put her in a good evidential position to appreciate some considerations that bear on the truth of the belief in question — for example, the relevant incentives enshrined in the proposed tax legislation and the benefits of increased investment in business.

That said, it seems that having a different social identity in the same category (that is, being in a lower socio-economic class) comes with a different sort of evidential advantage when it comes to the subject matter at hand. After all, it's plausible that the CIMEs of people in the working class have been shaped in ways that put them in a good evidential position to appreciate some of the other relevant factors bearing on the truth of the proposition in question — for example, the importance of welfare programs funded by the taxes of people like Veronica, and the failure of past trickle-down policies to actually trickle down. This makes it less clear that Veronica's socio-economic status makes it more likely that she got it right about the matter at hand. In the end, then, it seems like evidential boost from standpoint advantage is non-negligible, but modest. Veronica's higher-order evidence of accuracy, then, is middling at best.

In contrast, Veronica's piece of higher-order evidence of error from irrelevant influences is strong. After all, Veronica has a vested interest in p being true: if it is the case that the proposed tax break will be universally beneficial, then it should presumably be passed, and, if it's passed, Veronica stands to gain substantial financial benefits.

What's more, unlike Departmental Climate, it isn't clear that the relevant belief in Tax Break is prior to, or is grounds for, the relevant vested interest. A good test of whether the belief is prior to the implicated desire is to consider whether the believer would still have the desire if p were false. If the believer wouldn't have the relevant desire if p were false, then that is a reliable indicator that the belief is prior to the desire and, by extension, that the desire isn't epistemically worrisome; if the believer would still have the relevant desire if p were false, then this is a good indicator that the desire is prior to (and potentially exerts a directional force on) the belief. In Tax Break, it seems likely that Veronica would still have a pronounced interest in the tax bill being passed even if it were not universally beneficial. Whether or not the tax bill is universally beneficial, she personally has a lot to gain from it being passed. Her higher-order evidence of error from irrelevant influences, then, is quite strong — strong enough, it seems, not to be defeated by the middling higher-order evidence of accuracy from standpoint advantage. Given all of this, Veronica has good independent reason to worry about the epistemic status of her belief given the vested interest implicated by her social identity. According to GIRP, rationality requires her to revise.

8.c Complicated Cases

In light of the previous section, it might be concluded that we should worry in any double higher-order evidence case where we have a pronounced interest in p being true even if p were false. But, this is too hasty. In double higher-order evidence cases, the higher-order evidence of error from irrelevant influences must be weighed against the higher-order evidence of accuracy from standpoint advantage. Whether we should worry or celebrate (or neither) depends on how strong each piece of higher-order evidence is. This, in turn, depends on particular features of the case: how strong the desire is for p to be true; how likely it is that the desire exerts a directional force on the belief; if exerted, the extent to which the desire sways the belief; the degree of the relevant evidential advantage corresponding to the feature of social identity in question; and, relatedly, whether those who have a different social identity in the same category have different forms of epistemic advantage when it comes to the topic at hand.

With that in mind, let's return to Abortion — the case we considered in the introduction. In this case, Abby is a sexually active woman of reproductive age living in Texas who believes that abortion is generally morally permissible and, thus, should be legal.¹⁵ How should Abby evaluate the epistemic significance of her gender on her belief that abortion is generally morally permissible?

Start with her higher-order evidence of error from irrelevant influences. Abby does have something to gain from p being true. After all, if abortion is generally morally permissible and should be legal, then the state should make it legal. What's more, it seems that Abby would still have a vested interest in abortion being legal even if it were false that it should be; whether or not abortion is morally permissible and should be legal, it remains to Abby's benefit to have access to legal abortion care. That said, the benefits that Abby stands to gain seem less pronounced than the ones featured in Tax Break. Suppose it's not clear to Abby if she will ever need access to abortion care or how difficult it would be for her to access it if abortion were banned in Texas. Given all this, it is plausible to say that the strength of Abby's higher-order evidence of error is significant but not overwhelming.

Now consider her higher-order evidence of accuracy from standpoint advantage. It's plausible that being a woman has put Abby in a position to recognize and appreciate some important relevant moral considerations bearing on the truth of p, like the ones detailed in the case — for example, the heavy physical and emotional tolls of pregnancy (especially an unwanted one). What's more, it's doubtful that men qua men tend to have different sorts of evidential advantage when it comes to the subject matter at hand. In other words, it's doubtful that being a man tends to put men in a position to grasp another set of considerations that bear on the truth of the moral status of abortion. Given all of this, it is plausible to say that Abby's higher-order evidence of accuracy is fairly strong — strong enough, by my lights, to rebut the higher-order evidence of error from irrelevant influences. If this is right, then, all higher-order things considered, Abby shouldn't worry about the epistemic impact of being a woman on the belief in question.

But, similar cases may warrant different verdicts. Consider a variation on Abortion — call it Abortion (Constitutionality) — in which Abby believes that the constitutional right to privacy enshrined in the 14^th amendment entails a constitutional right to abortion care and, thus, that Supreme Court's majority opinion in Dobbs v. Jackson striking down Roe v. Wade was incorrect. In this variation, the higher-order evidence of error from irrelevant influences is relevantly similar to the one in the original case. After all, if abortion is protected by the 14^th amendment, then access to abortion is constitutionally protected across the U.S., and Abby has some to gain from having access to legal abortion care. However, the higher-order evidence of accuracy from standpoint advantage seems substantially weaker. That's because — while being a woman may give Abby an evidential advantage when it comes to moral considerations regarding abortion — it isn't clear that being a woman gives her an evidential advantage when it comes to the intricacies of American constitutional law.¹⁶ So, when it comes to the belief that the right to privacy enshrined in the 14^th amendment entails a right to abortion, it seems that Abby lacks strong enough higher-order evidence of accuracy to rebut the competing higher-order evidence of error. If this is right, Abby has reason to worry about the epistemic impact of being a woman on this belief about constitutionality, and, according to GIRP, rationality may require her to revise it.

This point brings out a more general lesson: the weight of each relevant piece of higher-order evidence depends heavily on the particular content of the belief in question. And, importantly, GIRP may render different verdicts about similar beliefs on the same topic. In other words, rationality may recommend that we maintain one belief on some topic but may require us to revise another belief on the same topic.

So far, we have been considering complicated double higher-order evidence cases when, upon careful consideration, it seems that one of the competing pieces of higher-order evidence is stronger than the other. But, there are some cases when the respective weight of the competing higher-order considerations is difficult to determine, or else is roughly equally strong. Consider the following case:

Reparations: Jaren is an African American university student who has recently been considering slavery reparations. After looking into the issue extensively, Jaren comes to believe that the U.S. government ought to make substantial direct cash payments to current generations of African Americans who — like himself — are decedents of formerly enslaved people.¹⁷ When Jaren discusses his view with his white friend Jake, Jake wonders aloud if Jaren believes in direct cash reparations because he's Black.

This is a hard case, and it isn't entirely clear how Jaren should evaluate the epistemic significance of his race on his beliefs about reparations. Of course, things will depend heavily on how the details of the case are spelled out. We can imagine some details that make it so that Jaren's higher-order evidence of accuracy is stronger than that of error, and vice versa. We can also imagine details that make it so that the competing pieces of higher-order evidence seem roughly equal strong.

With these complications in mind, here's one way that Jaren's evaluation might go. On the one hand, Jaren might recognize that he has a vested interest in p being true — if it is the case that the government ought to make direct cash payments to descendants of formerly enslaved people, then legislation to this effect ought to be passed, and, if passed, Jaren and his community stand to gain financially in significant ways. On the other hand, Jaren might also recognize that his race gives him an evidential boost when it comes to the beliefs in question. Moving through life as an African American has shaped Jaren's CIMEs in ways that put him in an excellent evidential position when it comes to the moral dimension of reparations and related policy proposals. For example, given his race, Jaren is able to recognize and viscerally grasp the extent and degree to which the social injustices and anti-Black racism that are the legacy of slavery continue to harm African American communities.

Following this line of evaluation, it may be reasonable for Jaren to conclude that the epistemic advantages and disadvantages of his race on his belief about reparations come out in the wash: he has a fairly strong piece of higher-order evidence of error from irrelevant influences but also a fairly strong piece of higher-order evidence of accuracy from standpoint advantage. So, in this case, it might be that the epistemic impact of his racial identity on his belief about reparations is neutral.

What does rationality require of us in neutral cases? According to GIRP, rationality recommends that we maintain our original belief since, all higher-order things considered, we lack good independent reason to think that we are mistaken in these cases. That said, it might be epistemically good for us to approach these beliefs with an added dose of caution, recognizing that we may have miscalculated the weight of each piece of higher-order evidence. What's more — although this is not be something that rationality requires — it might be epistemically good for us to “do further homework” into the relevant issues in neutral cases, either by gathering more first-order evidence or by looking into other sources of higher-order evidence. We may consult epistemic peers or experts on the subject matter at hand — for example, Jaren may chat with fellow Black university students who have also carefully considered reparations — and then re-evaluate the epistemic status of our belief, but this time without the epistemic pressure to reflect on the impact of social identity.

9. Concluding Thoughts

To return to the motivating question, should we celebrate or worry about the doxastic impact of our social identity? The answer: It's complicated. Although not especially satisfying or surprising, the answer is, nevertheless, significant and illuminating.

The epistemic complexity of social identity is at risk of being downplayed when social identity is considered in isolation as an irrelevant influence or as a source of standpoint advantage. The literature on irrelevant influences tends to place disproportionate emphasis on the epistemic risks, framing the doxastic impact of social identity as epistemically worrisome — in the best-case scenario, it's innocuous. The literature on standpoint epistemology tends to place disproportionate emphasis on the epistemic benefits, framing the doxastic impact of social identity as epistemically advantageous — features of social identity are not only epistemically relevant but are also beneficial. Of course, these respective emphases make sense given the specific epistemological questions and puzzles undertaken in each branch of the literature. But, achieving a fuller, clearer picture of the epistemic significance of social identity requires us to explore the tension at their intersection.

Diving into this tension, I have argued that reflecting on the epistemic significance of social identity gives rise to an interesting family of cases: cases of double higher-order evidence. Reflecting on social identity through the lens of irrelevant influences gives us a piece of higher-order evidence of error, while reflecting through the lens of standpoint advantage gives us a piece of higher-order evidence of accuracy. To determine whether to worry or celebrate — to revise or remain steadfast in our belief that p — we must weigh the strength of each competing piece of higher-order evidence. GIRP can serve as our guide here: to the extent that we have good independent reason to think that we are mistaken with respect to p, we should revise accordingly, insofar as we can.

In double higher-order evidence cases when we have good independent reason to think that we are mistaken with respect to p, we should revise. There will be worrisome cases like Tax Break and Abortion (Constitutionality) when the piece of higher-order evidence of error outweighs the one of accuracy. In double higher-order evidence cases when we do not have good independent reason to think that we are mistaken, we should remain steadfast. These include not only celebratory cases like Career Deliberation, Vertical Campus, Departmental Climate, and Abortion, but also neutral cases like Reparations. In celebratory cases, the higher-order evidence of accuracy outweighs the higher-order evidence of error. In neutral cases, the higher-order evidence of error and of accuracy are roughly equal in strength. In both types of cases, we lack good independent reason to think that we are mistaken with respect to the belief in question and so rationality does not require revision; we should retain whatever belief we originally arrived at based on our first-order evidence. In celebratory cases, that's because the relevant higher-order considerations suggest that our social identity makes it more likely that we got it right about the matter at hand; in neutral cases, that's because the relevant higher-order considerations suggest that our social identity has ultimately not affected our chance of getting it right or wrong.