Daimon. Revista Internacional de Filosofía, nº 84, 2021 pp. 115-129

ISSN: 1130-0507 (papel) y 1989-4651 (electrónico)


Licencia Creative Commons Reconocimiento-NoComercial-SinObraDerivada 3.0 España (texto legal). Se pueden copiar, usar, difundir, transmitir y exponer públicamente, siempre que: i) se cite la autoría y la fuente original de su publicación (revista, editorial y URL de la obra); ii) no se usen para fines comerciales; iii) se mencione la existencia y especificaciones de esta licencia de uso.


I am large, I contain multitudes:
Epistemic pragmatism, testimonial injustice and positive intersectionalism


Soy grande, contengo multitudes.
Pragmatismo epistémico, injusticia testimonial e
interseccionalismo positivo





Abstract: We explore the compatibilities and incompatibilities between two highly successful approaches to knowledge: Brandom’s epistemic pragmatism, [EP], and the view that derives from Fricker’s seminal work on the ethics of knowing, [EK]. [EP] and [EK] are complementary approaches that put forward aspects of the application of the concept that deserve to be preserved. Nevertheless, their mere cumulative superposition produces dysfunctions that call for certain readjustments. We propose a positive kind of intersectionalism, [PI], that accounts for the fact that individuals simultaneously belong to diverse groups with variable epistemic conditions, some advantageous, some disadvantageous. [EP], [EK] and [PI] make a rich and coherent picture of subjects as full epistemic agents.

Key words: Brandom, Fricker, intersectionality, knowledge, licensing, McGowan


Resumen: En este trabajo exploramos las compatibilidades e incompatibilidades entre dos enfoques del conocimiento distintos y de gran éxito. El primero es el análisis del conocimiento que ofrece el pragmatismo epistémico de Brandom [PE]. El segundo es el punto de vista que se deriva de la obra seminal de Fricker sobre la ética del conocimiento [EC]. [PE] y [EC] plantean concepciones complementarias de la aplicación del concepto de conocimiento que merecen preservarse. Sin embargo, su mera superposición acumulativa produce disfunciones que exigen ciertos reajustes. Proponemos un tipo de interseccionalismo positivo, [IP], que da cuenta del hecho de que los individuos pertenecen simultáneamente a diversos grupos con condiciones epistémicas variables, algunas ventajosas y otras desventajosas. [EP], [EK] y [PI] conforman un cuadro rico y coherente sobre los sujetos como agentes epistémicos plenos.

Palabras clave: Brandom, conocimiento, Fricker, interseccionalidad, licencia, McGowan


Recibido: 31/05/2021. Aceptado: 22/06/2021.

* The following people have commented previous versions of this paper, helping to improve the final result: Alba Moreno, Eduardo Pérez, Claudia Picazo, and Andrés Soria. Two anonymous referees have also contributed with their suggestions. To all of them, we are deeply indebted.

** Professor of Logic and Philosophy of Science (Department of Philosophy I, University of Granada). Contact: frapolli@ugr.es. Research interests: philosophy of logic, philosophy of language, pragmatism. Recent publications: Frápolli, M. J. and Forero-Mora, J. A. (2021), “Show me: Tractarian non-Representationalism”, Teorema, 40(2), 63–81; and M. J. Frápolli (2019), “Propositions first: Biting Geach’s bullet”, Royal Institute of Philosophy Supplements, 86, 1–9.

*** Predoctoral fellow (Department of Philosophy I, University of Granada). Contact: llanosnavarro@ugr.es. Research interests: philosophy of language, metaethics, social epistemology. Recent publications: Navarro, L. (forthcoming), “Where are ethical properties? Predication, location and category mistake”, in N. S. Miras and M. Bella (eds.), Women in Pragmatism: Past, Present, and Future, Berlin: Springer; and Almagro, M., Navarro, L. & Pinedo, M. (forthcoming), “Is testimonial injustice epistemic? Let me count the ways”, Hypatia:A Journal of Feminist Philosophy.

**** The names of the two authors are in alphabetical order. Both authors have contributed equally to the paper.



1. The argument


Let us begin with a broad characterisation of [EP]. Being a knower is a social status whose possession requires the acknowledgement of other people. Discursive creatures, i.e. sapients in Brandom’s terminology, are participants in socio-linguistic games in which reasons are provided, evaluated, and consumed. Paying tribute to Sellars, Brandom identifies the game of ‘making assertions and giving and asking for reasons’ (Brandom, 2000, 26) as occupying the ground level. Paying tribute to Hegel, one of Sellar’s inspiring heroes (Brandom, 2019, 114), he assumes that normative statuses, socially instituted, take the form of ‘mutual or reciprocal recognition’ (Brandom, 2019, 12). This intuition shared by Hegel, Sellars, and Brandom is common to any pragmatist approach to language. Conceptual contents emerge—and communication occurs—in highly regulated linguistic actions where participants play specific roles for which they have the appropriate kind of licensing. We will take this point up in the next section.


This is the first claim of our overall argument. For our purposes, it can be reduced to the Hegelian-like slogan, [HS],

[HS] Knowledge requires acknowledgement.


The second claim is a development of the first: the role of knowledge, as a normative notion, is the expression of complex attitudes of the agent who attributes it and the subject to which it is attributed. Let us call this claim ‘epistemic expressivism’, [EE]. Brandom’s epistemic expressivism makes self-attributions of knowledge logically dependent on third-person knowledge attributions. Agents learn the game of attributing knowledge to themselves, i.e. they recognise themselves as knowers, by learning the game of attributing knowledge to others. This is a possible wording of [EE]:


[EE] Mastering the practice of self-attributing the property of being a knowledge possessor presupposes mastering the practice of attributing it to others.


[HS] and [EE] make conceptual remarks. They expose necessary connections between the concepts involved and the practices of applying them. Epistemic pragmatism, [EP], is the view that results of the combination of [HS] and [EE].

[EP] directly implies that people completely deprived of the status of knower and, as a consequence, deprived of the status of knowledge attributor, cannot develop the required skills to recognise themselves as knowledge possessors.

Let us now consider the basic tenets of [EK] as we will understand it in this paper. In Fricker’s analysis, epistemic injustice occurs when people are harmed in their condition of knowers because they belong to disadvantaged social groups. Other researchers (Medina, 2013; Cassam, 2019) follow Fricker’s lead in stressing that members of oppressed groups are prone to develop epistemic vices caused by their disadvantaged social positions. Fortunately, oppressive situations not only produce pernicious effects in people who suffer them. Some epistemic virtues are also produced by systematic epistemic harm. And the other way around. Situations of dominance contribute to the development of epistemic vices and, sometimes, of epistemic virtues too. Thus, the ethics of knowing and virtue epistemology make room for hope, articulating the possibility of overcoming their underprivileged starting point.

There is a sharp contrast between [EP] and [EK] in terms of their relative rigidity. [EP] points to certain necessary connections between concepts and their application, whereas [EK] explains the essential plasticity of individuals, which allows them to modify, adjust, and counterbalance their epistemic positions. We contend that both sets of intuitions are grounded and that there should be room for both in a mature approach to knowledge.

The relevance of [EP] conceptual analyses for the [EK] narrative is straightforward. Among the consequences of [EP] for [EK] we count the following three. First, [EP] implies that testimonial injustice, as defined by (Fricker, 2007), essentially produces an epistemic kind of injustice, harming the victims in their status as epistemic subjects (Almagro et al., forthcoming).

Second, related to hermeneutic injustice, disadvantaged subjects do not possess the status to enact the norms that would facilitate the circulation of those terms and narratives that they use to explain their conditions. This way, members of dominant groups will be unable to take over the conceptual resources to understand oppressive situations.

The third one is that [EP], because of its conceptual and hence necessary nature, precludes the possibility of victims of different kinds of epistemic injustice to escape their fate by their own means.

These two last consequences of [EP] are too strong, contested by the analyses of Fricker, Medina, Cassam, and others, and refuted by empirical data. Our aim in this paper is to show how [EP] and [EK] should be enriched to accommodate the empirical data and build up a consistent proposal that gathers together conceptual analyses and political approaches. Some hypotheses whose addition we suggest have been amply discussed, as the contextual nature of knowledge, and we will not insist upon them. Others seem to have attracted less attention, at least in the version required in this paper. In any case, our proposal is largely programmatic and in need of completion. Nevertheless, we believe that making the structure of the general argument explicit is a substantial step forward towards a general approach to knowledge that benefits from [EP] and [EK].


2. Epistemic pragmatism


Attributions of knowledge are the expression of hybrid attitudes. The notions of knowledge and belief are conceptually distinct because knowledge, unlike belief, incorporates the attributor’s perspective. Even if self-ascriptions of belief can sometimes be challenged from outside, considering e.g. the believer’s behaviour, in ascriptions of knowledge the external perspective is an essential ingredient (Brandom, 2009, 157; Frápolli, 2019).

In this picture, knowledge attributions always give complex information that concerns the attributee’s as much as the attributor’s mental states. When we attribute knowledge to a subject, S, concerning a particular issue, p, as represented in scheme (1),


(1) S knows that p,


we perform three intertwined actions: (i) we attribute to S a belief that p, (ii) we attribute to S the entitlement to her belief that p, and (iii) we express our justified endorsement of p. These conditions are the counterparts of the traditional analysis of knowledge as belief, (i), that is justified, (ii), and true, (iii).

Attributions of factive notions always involve the attributor’s attitudes together with the attributee’s states, they always expose the attributor’s states and reveal the attributor’s perspective. Because factives support the inference from their attribution to the independent assertion of their propositional contents, the attributor indirectly asserts the content she attributes to a third party. This is basically Brandom’s account of the meaning of the concept of knowledge (Brandom, 2009, 157–8). Some examples of inferences involving factives are the following:


(1) She knows that p, therefore p,

(2) Her claim that p is true, therefore p,

(3) She remembers that p, therefore p.


(1-3) are instances of (4),


(4) S ϕ s that p, p.


Moore-like paradoxes pop up whenever the assumption of the premise in (1-3) combines with the rejection of the conclusion.

The pragmatic layout just depicted explains Gettier-like phenomena. In Gettier’s example, Smith is unable to distinguish between his justified true belief that the candidate who will get the job has ten coins in his pocket and the counterfactual situation in which knowledge of this fact could be attributed to him (Gettier, 1963).

Normative pragmatism shows that, exclusively from the first-person perspective, the speaker cannot discriminate between her credentials to assert that p, to assert that p is true, to assert that she believes that p, and to assert that she knows that p. To attribute knowledge, some kind of social triangulation is required, since speakers on their own lack perspective—which involves access to the right kind of information—to ground a principled distinction between their knowledge and their mere justified belief (Frápolli, 2019).

Thus explained, knowing is something that others attribute to you, and that you can claim for yourself as a result of being attributed. This claim has widest scope and amounts to saying that the individual practice of knowledge self-attributions is parasitic on the social practice of attributing knowledge to others, i.e. that the former cannot be made sense of without the latter. Speakers learn the self-attribution game as a result of mastering the game of external knowledge attributions. This situation highlights the importance of the role of the others for the image that we have of ourselves as epistemic subjects.

Brandom follows Hegel’s core idea, [HS], i.e. that recognition (Anerkennung) is the notion that grounds social practices in which other people are treated as bearing normative statuses. In his book on Hegel’s Phenomenology, we read:


To be a self in the full normative sense is to have not only actual normative attitudes, but also actual normative statuses. It is not only to take one-self or others to have authority or be responsible, but actually to have authority or be responsible. To achieve such a status, a normative subject must participate in a general recognitive dyad: must actually be recognized by someone that subject actually recognizes. For only suitably socially complemented attitudes institute actual statuses. It follows that normative statuses, normative subjects of such statuses, and recognitive communities of normative subjects are all synthesized simultaneously by recognitive processes that have an appropriate structure: the structure of reciprocal recognition. (Brandom, 2019, 293–4)


Brandom’s normative pragmatism is an elaboration of this intuition, which will be developed from a different perspective in the next section.

We endorse the essentially social nature of knowledge. It might be argued that sometimes subjects are in the position to attribute to themselves the knowledge that others deny them. We agree. Consider the following example: Victoria, a successful psychologist, insists, against the opinion of her colleagues, that the patient is in high risk of engaging in self-harming behaviour. She knows, she says, because she has followed the patient’s health history in the past two years, and because she has been profusely exposed to similar cases. Does she know? The answer does not depend on the outcome, i. e. on whether the patient actually inflicts some harm to herself. It might happen that the patient engages in some kind of self-damaging behaviour without Victoria having prior knowledge of it, and the other way round, that the patient manages to continue with her life without any patent indication that she may cause herself any harm even though she was, at that point, a danger for herself. But still, this is a case in which we have reasons to attribute knowledge to Victoria, based on her expertise, socially acknowledged, although the attribution is subjected to the analysis of the arguments on one side and the other. Truth and justification do not suffice for the attribution of knowledge, which also needs the acknowledgement of the attributor who acts as referee of the different scores that this complex status involves (Brandom, 2000, 119). Subjective certainty is neither necessary (Wittgenstein, 1969, §§340-1; Williams, 2001, 18) nor sufficient (Ramsey, 1929, 256). And, as Gettier’s cases show, objective truth also falls short of grounding knowledge.

The role of the third-person perspective is common ground in speech act theory. Explaining some recent developments that support our argument is the aim of the next section.


3. Authority and licensing


The structure of situations that involve attribution of knowledge—or its refusal—parallels the structure of other types of speech acts to which several agents contribute with utterances that count as assertions. We follow McGowan (2019) in calling this general kind of speech exchange a ‘conversation’. McGowan argued that there are significant analogies between ordinary conversations and the performance of standard exercitives, in Austin’s sense. The CEO of a company uttering in an official meeting ‘From now on, holidays are suspended’ is an example of standard exercitive. A mother informing their children that dogs are not allowed inside the house is another, less formal, example. In exercitives, speakers enact norms that determine what is permissible in certain situations.

For the exercitive to be successful, the speaker who utters the permissibility norm has to possess a certain status. The secretary uttering exactly the same words as the CEO in the first example does not enact the norm; the child that pompously tells her siblings that dogs are not allowed inside the house does not produce a permissibility fact either. Only agents with the appropriate authority can enact norms in suitable contexts. As McGowan explains, the speaker who utters an exercitive must possess, besides authority, the intention to enact the norm, and his intention has to be recognised by the participants in the conversation. The success of the exercitive thus depends on the recognition of the speaker’s intentions, in the usual Gricean sense.

Contributions to ordinary conversations also enact permissibility norms—McGowan calls them ‘conversational exercitives’—that produce a change in the conversational scorekeeping, in Lewis’s terminology (McGowan, 2019, 28ff), or in the shared common ground, in Stalnaker’s terminology (McGowan, 2019, 39ff.). This is the main thesis of McGowan (2019). Hate speech and offensive or discriminatory linguistic actions cause harm by enacting permissibility facts that can potentially add oppression to members of already oppressed groups. The effect of changing the score—or the background—is a general feature of contributions to conversational exercitives, not restricted to situations of hatred speech.

Consider the following example: Victoria and Joan are discussing the best way to get to the concert hall to attend the performance of their favourite rapper, for which they bought tickets many months in advance. In the conversation, Victoria says: ‘My boyfriend, Mark, has a car.’ After this contribution, Mark becomes a salient individual in the conversation. The pronoun in Joan’s reaction to Victoria’s utterance, ‘Tell him to get us there’, unequivocally refers to Mark as a result of the permissibility fact enacted by Victoria’s sentence. After Victoria’s intervention, speaking of ‘the car’ is also permitted, etc.

Neither intentions to enact a norm nor the recognition of intentions is a success condition for conversational exercitives. By contrast with what happens in the case of standard exercitives, speakers do not need to have the intention to enact any norm when they contribute to ordinary conversations.

A further difference between standard and conversational exercitives concerns the role of authority. As we have mentioned, standard exercitives rely on the authority of the speaker to produce their effects, but McGowan is reluctant to extend the role of authority to conversational exercitives. She acknowledges that, in a sense or another, participants exert some authority over the conversation (McGowan, 2019, 66), although of a weaker kind that she calls ‘standing’ (McGowan, 2019, 67).

There is a complex network of social norms that regulate who can participate in a conversation, i.e. who has the appropriate standing. Not everybody is allowed to join an ongoing speech act. This is something that depends on a wide variety of subtle social facts: your position in the community, your relationship to the participants, specific aspects of the conversation at issue, etc. Adapting McGowan’s example, imagine that I approach a group of women at the Beverly Yacht Club—a club of which I am not a member. I friendly greet the ladies in the group and begin to talk with the patent intention to engage in a conversational exchange. Spite my efforts, they ignore me, not looking at me or even making any sign that they have heard me. They are making it clear that I am not welcome. As McGowan says, ‘I lack the requisite (in this case, social) standing’ (loc. cit.).

Possessing the appropriate standing is something that others confer upon you. One might think that my lack of standing rests on an objective condition; after all, I am not a member of this elitist resort. Had I owned the exclusive membership, their reaction to my efforts would have been different. A superficial reflection will show that this is not the case. The same group of ladies could have denied acknowledgement to members of the club, e.g. to a member who supported the Green Party in the last local elections, and could have welcome non-members, e.g. the elegant wife and daughters of the new governor.

Thus, the condition of being a suitable participant in a social exchange, linguistic or non-linguistic, does not depend on the candidate’s efforts. It is a normative status graciously conceded and maintained by people who already possess the standing. It is again an issue of Anerkennung, in Hegel’s and Brandom’s sense. Cases of bullying among teenagers and mobbing in working places, presently amplified by social networks, profusely illustrate how groups can ostracise people, denying them the access to participate of their activities.

Licensing is the kind of act by means of which the status to take part in a conversation is conferred upon a candidate. In standard cases, licensing is brought about by an explicit permission issued by others. Sometimes, nevertheless, the licensing is produced implicitly. A member of the group ostensibly speaking to the candidate enacts a permissibility fact to the effect of licensing the standing, thus permitting the candidate and the rest of the group to act as if she were a full member. In extraordinary situations, agents can even license themselves to act as authorities in specific contexts. In an emergency situation inside an aircraft, for instance, a passenger might assume the responsibility for controlling the passage, commanding other passengers to perform particular tasks, such as taking care of agitated passengers, calming children down, helping less able people to reach the emergency exits, etc. To do this, the accidental hero needs to be neither a physician, nor a policewoman, nor an emergency worker. Virtually anyone with the appropriate character and the will to help can take command of the situation. A further kind of self-licensing is produced by pretending: an agent can act in a situation as if she had the appropriate standing until it is recognised by the others participants (McGowan, 2019, 67). But even in cases of self-licensing, the others have to accept your authority. In the aircraft examples, the courageous passenger who steps up will only have authority if the passage accepts their commands.

Attributing knowledge to someone is considering her a full epistemic actor. The person who attributes and the person who is attributed must be licensed to participate in the epistemic conversation. People belonging to discriminated communities often lack the resources to assume and to pretend an authority role, and the possibility of being accepted by others as the bearers of authority. The effect is the silencing of discriminated subjects, who are most of the times ignored as interlocutors by members of dominant groups.

Licensing and silencing (Langton, 1993; Hornsby and Langton, 1998; see also Dotson, 2011) show that testimonial injustice applies to other types of speech act alongside with giving testimony. The concept of discursive injustice coined by Kukla (2014, 441) points out that those who belong to disadvantaged groups not only suffer in their capacity as knowers but also in their capacity as discursive peers. In a similar vein, Ayala and Vasilyeva (2015) argue that the range of speech acts that a speaker can perform within a conversation, what they call her range of speech affordances, is ‘adjusted in virtue of membership in intersecting social categories, more or less salient’ (Ayala and Vasilyeva, 2015, 132). Thus, Fricker is right when she connects epistemic injustice with other kinds of injustice to which members of disadvantaged groups are subjected. Testimonial injustice is the topic of the next section.


4. Testimonial Injustice


[EP] establishes a strict connection between the attitudes of other people towards an individual and her possibilities to become a full-fledged participant in social activities of a certain kind. The result is the impossibility for the subject to overcome a disadvantaged situation by her own means since self-attribution of knowledge and self-licensing are parasitic activities dependent on the intervention of others.

Alternative approaches to the analysis of knowledge, which pay attention to individuals, their social environments, and the character traits that they develop or inhibit because of their social position and experiences, make more flexible prognoses about people’s options to mend unfavourable conditions. These analyses deserve to be taken into account, if only because of their extraordinary success. In this section and the next one, we will touch upon the main lines of the phenomenon of testimonial injustice.

Fricker defines testimonial injustice as the credibility deficit received by a speaker due to prejudices against her social identity. As she says: ‘(T)he central case of testimonial injustice is identity-prejudicial credibility deficit’ (Fricker, 2007, 28). To qualify as a kind of injustice, the epistemic type must be connected to other kinds of injustice—social, political, economic—inflicted to the victim as a member of her identity group. For this reason, the phenomenon of testimonial injustice requires persistent and systematic mistreat. The feature of persistence calls for the transversality of epistemic offence across different context and situations; the feature of systematicity calls for its structural character. An isolated episode of credibility deficit does not count as testimonial injustice.

The scope of the harm caused by epistemic injustice, Fricker acknowledges (Fricker, 2007, 58), must be determined by empirical research. There is abundant literature that supports the predictions of [EK] at school (Ehrlinger and Dunning, 2003; Huguet and Regner, 2007; Reuben et al., 2012; Estes and Felker, 2012) and outside, in feminist and critical race studies.

The mere mismatch between the credibility received and the credibility a witness deserves, even if systematic and persistent, is not enough for testimonial injustice to arise. As Fricker defines it, no epistemic injustice derives from a credibility excess. Other authors have challenged this claim. Medina, for instance, widens the scope of epistemic injustice and includes the harm socially caused by the perpetuation of oppressive situations (Medina, 2013, 60–1). Excess of credibility can undermine the possibilities of the subject to become a sound epistemic agent by promoting the development of epistemic vices, as we will discuss in the next section. Luzzi (2016), in turn, argues that testimonial injustice may emerge even in cases of accurate credibility attributions.

The epistemic harm inflicted to a subject can be direct or indirect. When discussing hermeneutical injustice—a kind of epistemic injustice caused by the shortage of interpretive, i.e. conceptual, resources to make sense of certain situations that the agent undergoes—Fricker acknowledges a ‘secondary’ kind of epistemic injustice that concerns the practical consequences of the ‘primary’, direct, harm. She describes the case of Carmita Wood, a worker victim of harassment that had to leave her work and could not claim any kind of benefits for the lack of an acceptable way of justifying the reasons that led her to quit her job (Fricker, 2007, 149–50).

Hermeneutical and testimonial kinds of epistemic injustice erode the confidence of subjects in their epistemic abilities. The lack of confidence caused in a subject by her systematic and persistent valuation as an epistemically defective subject is a particularly significant consequence of the epistemic injustice of the primary kind (Fricker, 2007, 174). When the subject is not equipped to understand and explain the injustices inflicted to her, she is deprived of an essential instrument to conclude that she has reasons to feel what she feels and the right to seek reparation. When the subject is systematically negated the status of knowledge giver, she is seriously undermined in her possibilities to develop an appropriate degree of confidence in her epistemic competence (Fricker, 2007, 58, Roessler, 2015).

Cases of epistemic injustice produce the development of epistemic vices and block the development of epistemic virtues, both in victims and oppressors. [EK] thus offers an internalist approach that prescribes corresponding epistemic—and ethical—virtues to balance the effect of those epistemic vices responsible for and caused by epistemic injustices. Even if Fricker stresses that ‘the human condition is, necessarily, a socially situated condition’, and that her philosophical conception aims to analyse ‘what constitutes good epistemic conduct in the socially situated context’ (Fricker, 2007, 176), her approach relies on the development of the appropriate individual attitudes to overcome unjust individual situations.


5. Epistemic Friction


Being a victim of epistemic injustice is compatible with being epistemically benefited in some other features.

In the nuanced approached that Fricker (2007), Medina (2013), and Cassam (2019) offer in their discussion of the causes and consequences of the epistemic kinds of injustice, they acknowledge complex epistemic conditions to the members of different groups. Victims and oppressors develop epistemic vices and virtues, due to their position, that explain why victims may become epistemically apt to a certain extent and oppressors epistemically diminished.

People belonging to dominant groups are prone to be confident and assertive, but also arrogant, insouciant, and narrow-minded. People in disadvantaged groups, by contrast, sometimes develop beneficial character traits, such as humility, open-mindedness and intellectual curiosity (Medina, 2013; Fricker, 2007, 20; Cassam, 2019, 150), together with epistemically disadvantageous features derived from their social position, such as lack of confidence. Medina illustrates the point by mentioning the typical frame of mind of enslaved people, according to Tocqueville’s analyses, and the patent effects in women’s self-confidence as denounced by feminist literature. In both cases, situations of systematic epistemic humiliation may end up in a lack of epistemic cognitive control on the part of the abused subject, who may end up doubting that they possess any cognitive power or even a mind. Medina calls this extreme case of epistemic self-negation ‘ego-skepticism’: ‘a skepticism about the self, about its capacities and even about its very existence’ (Medina, 2013, 42).

Nevertheless, to draw the right conclusions from the theoretical and experimental work on undervalued populations, it is advisable to avoid both ends of an epistemic attitudinal continuous: neither should we ‘romanticized the predicament of the oppressed’ nor assume that they possess ‘impeccable epistemic characters’ (Medina, 2013, 40). And the same goes for the oppressing subject. The epistemic virtues and vices frequently detected in members of socially significant groups are neither exclusive of these groups, nor possessed by all individuals in the group, nor automatically developed by the mere belonging to the group (Medina, 2013, 43). Other individual traits and experiences may reshape the epistemic condition of particular agents.

Medina, following Wittgenstein, includes in his analysis the notion of epistemic friction. The epistemic character that a subject eventually builds up is substantially affected by the resistance that the world and other agents present to their beliefs. Note that our use of the term ‘resistance’ differs from Medina’s use. In his case, ‘resistance’ sometimes presents a political overtone that refers to the rejection by members of oppressed groups to accept the narrative with which dominant groups characterises them. In our case, it is a purely epistemic term that stresses the limits externally imposed for the development of a specific image of the world, understood in a broad sense. Not even mathematicians, who allegedly enjoy a maximal degree of freedom, develop their theories in complete absence of theoretical constraints (Frápolli, 2015, 337). Some friction is needed in order to maintain the subject’s web of beliefs grounded into the world. The willing exposition to the criticism of our own opinions reinforces epistemic responsibility (Medina, 2013, 54). People not accustomed to taking alternative epistemic perspectives into account, e.g. because of their dominant position guaranteeing their opinions to prevail, are less likely to develop the kind of humility and open-mindedness that would otherwise immunise them against what we might call ‘social blindness’. Dominant classes can afford some degree of social blindness since their survival, actual or metaphoric, does not depend on any kind of social negotiation.

Social blindness and a deficit of epistemic responsibility cause epistemic injustices according to Medina although not according to Fricker. The discrepancy is more apparent than real and rests on Fricker’s insistence in restricting epistemic injustices to those practices that cause harm to the speaker, whereas Medina widens the scope to cover also the harm socially caused in the form of perpetuation of injustices (Medina, 2013, 60–1). For our purposes, the debate is inconsequential.

In Medina’s global account, disadvantaged groups present a symmetric situation related to epistemic responsibility: because their members’ survival often depends on an understanding as accurate as possible of the oppressor’s viewpoint, it is in their own best interest to develop social lucidity (Medina, 2013, chapter 5).

The virtue of epistemic responsibility only needs a minimal amount of social and external friction, but it needs some amount. This minimal amount allows members of dominant groups, affected by some kind of social blindness, to develop some degree of responsibility and thus to be able to possess some knowledge about the life of others. On the other hand, the epistemic lucidity that members of oppressed groups develop also informs them of the lack of epistemic acknowledgement that they can expect. This meta-knowledge of the dynamics of social relations may counterbalance the effects of the different kinds of epistemic injustice that afflict disadvantaged subjects. Thus, their situation is not as desperate as it might be had them been ignorant of the dissimilarities defining opposing social groups. Besides, self-awareness of one’s capacity as a full epistemic subject can come from within your own marginalised group of belonging. An example are safe spaces, conceived as controlled communities, deliberately protected from oppressive epistemic practices, and aimed at reinforcing the epistemic agency of marginalised subjects. In safe spaces, women are treated as epistemic subjects by other women, racialised people by other racialised people. And, as we will see in the case of women reporting other people’s health conditions, even by members of dominant groups, although restricted to specific topics that derive from the social stereotype of women as carers. The effect of the variety and heterogeneity of epistemic situations within marginalised (and dominant) groups also covers hermeneutical injustice. Mason (2011), for instance, argues that the marginalised group of belonging is sometimes able to provide the hermeneutical resources that dominant communities lack. That they are able to export them to non-marginalised communities is a different story, as we have suggested in §1.


6. Positive Intersectionalism


It is now time to take stock. We have surveyed and endorsed the main lines of two contrasting approaches to knowledge. [EP] sheds an externalist stance on the complex information, partly descriptive and partly normative, that knowledge attributions convey. [EK], by contrast, insists on the effects that sustained epistemic mistreatment produces on individuals.

Unlike [EP], [EK] gathers together heterogeneous positions. Cassam, Fricker, Herbert, Medina, to name some of the authors mentioned in previous sections, do not completely share their overall backgrounds. What they have in common is their interest in epistemic kinds of injustice and their impact on subjects in terms of vices and virtues. Moreover, they promote individualist strategies to repair damages, strategies that, nevertheless, are compatible with denouncing structural injustices and supporting global policies to overcome social injustice.

Part of the theoretical and practical interests of [EK] consists in identifying ‘how vices of the mind are obstacles to knowledge or how, as José Medina, puts it, they “get in the way of knowledge (2013, 30)”’ (Cassam, 2019, 3). Moreover, [EK] presents a therapeutical derivation that [EP] completely lacks. Cassam (2019, chapter 8), for instance, discusses how self-improvement is possible. There, Cassam argues for what he calls a ‘qualified optimism’: ‘My aim isn’t that lasting self-improvement is always possible but that it sometimes is. Some epistemic vices are resistant to self-improvement strategies but such strategies are sometimes available and effective’ (Cassam, 2019, 171). A similar insistence on the importance of promoting epistemic virtues is detectable in Fricker (2007, 71) and Medina (2013, chapter 6).

Considered this way, the two positions operate at different levels and show different theoretical and practical concerns. In a relevant sense, their theoretical consequences are orthogonal and should be compatible. Nevertheless, we have been suggesting that this is not necessarily so. There are specific points at which the consequences of [EP] affect the picture developed by [EK]. On the one hand, [EP] offers a principled criterion to highlight the essentially epistemic character of testimonial injustice (Almagro et al., forthcoming), against those who have doubted that the victims of epistemic injustice actually are wronged in their epistemic rights. On the other, the conceptual priority of third-person attributions over first-person attributions seems to preclude that the capacity to overcome unfavourable epistemic statuses lies within the power of individual agents.

To accommodate the analyses derived from [EK] in a general overview respectful to the externalist perspective developed by [EP] we suggest the addition of a pluralist and dynamic approach to the composition and evolution of social groups. In doing so, we draw on the foundational idea of intersectional feminism (Crenshaw, 1989) to understand the complex nature of social identities. Discrimination does not occur along a single categorical axis, but a variety of mutually interacting or intersecting axes often combines to shape the individual’s social identity. Besides, Crenshaw argues, subjects defined by several axes of oppression suffer from a form of discrimination that goes beyond the sum of the basic oppressions.

We extend Crenshaw’s insight to cover not only those axes on which discrimination occur, a kind of ‘negative intersectionalism’, but also the positive epistemic situations that even members of multiply marginalised communities encounter. We call our proposal ‘positive intersectionalism’, whose leading insight is the acknowledgement that our position as epistemic agents is most of the times a mixture of advantaged and disadvantaged conditions. The complexity of social identities potentially has, we contend, a saving role. Otherwise, as it follows from [EP], we could not grow up as epistemic agents. Besides, as it follows from [EK], all of us present epistemic vices and virtues, and also character traits and epistemic habits that lead to knowledge and get in the way of it.

Positive intersectionalism puts forward two general points. First, social groups present fuzzy boundaries. This first aspect may be difficult to spot due to the underdescribed and schematic nature of the examples sometimes used. The excessive focus on identity aspects, disregarding political ones, illustrates the point. A migrant is only a migrant in her host country, and there are poor and rich migrants, although rich migrants are sometimes referred to by a different term: ‘expat’. Moreover, social identities are frequently named by vague predicates: ‘being working-class’, ‘being educated’, ‘being rich’, etc.

The second aspect that we want to highlight is that the relation between individuals and social groups is not one-to-one. Each of us belongs to an array of different communities, with varied epistemic statuses. This aspect has also been suggested by (Medina, 2012, 218–20). One can be a white middle-class educated woman or a wealthy homosexual black man. Positive intersectionalism intends to focus on positive conditions too. Within each group, individuals can occupy different power positions: not all members of disadvantaged groups are at the same risk of suffering from epistemic injustice. Wealthy people are always in better position than poor people, men better than women, white better than black, etc. Some black women also are physicians and mothers. With their patients and children, they are the authority figure and receive the right amount of credence to compensate the deficit they suffer when other roles are more salient.

To these two senses of [PI], one should add the context-dependence of knowledge attributions. Attributions of knowledge depend on contextual factors such as the relative epistemic position of attributor and attributee, the participants in the situation in which knowledge is attributed, but also the topic of attribution.

Manne offers some examples of the perspectival status of individuals as subjects of knowledge. Based on empirical research, Manne explains that a single subject can be respected as an expert about some contents and dismissed as unqualified about some others. This is the case with women who report on health issues of people at their care—elderly people or children—, about which they are considered ‘supremely competent’ (Manne, 2020, 89), while they are scorned when it comes to report on their own health. The studies that Manne discusses ‘suggest that when women try to testify to their pain, they are routinely dismissed by the medical establishment on both of these bases—impugned as incompetent and hysterical, on the one hand, or as dishonest malingerers, on the other’ (Manne, 2020, 86). She also presents results that confirm the diverse composition of groups, which are often treated as homogeneous: ‘According to recent estimates, black women in the United States are some three to four times more likely to die as the result of pregnancy or childbirth than their white counterparts’ (Manne, 2020, 77). Black women are seen as less credible than white women, independently of their social position or economic status, as Serena Williams’s case shows (Rankine, 2014, chapter 2; Manne, 2020, 78). Being a woman is not an epistemically homogeneous category.

[PI], in the sense that we have suggested, explains why members of disadvantaged groups are not condemned as epistemic subjects. Oppressed groups can present a sufficient degree of variability as to permit their members not to receive unjust epistemic treatment throughout all their epistemic interventions. Epistemic statuses should not be thought of as absolute states, granted once and for all and identical across all contexts, but as unstable, context-dependent properties, subject to permanent negotiation. Because epistemic injustice requires systematicity and persistence, as we seen in section 4, but not universality, [EK] can accommodate [PI].

That sound epistemic subjects can belong to marginalised groups does not neutralise the lessons learnt from [EP]. The necessary priority of third-person attributions still implies that subjects systematically deprived of any epistemic acknowledgement, i.e. victims of extreme epistemic violence, cannot become full epistemic agents. Fortunately, as positive intersectionalism suggests, absolute epistemic neglect is not frequent. The standard situation is the one in which most of us usually are, one in which we receive a variable degree of epistemic acknowledgement and epistemic contempt. Without this variability we would be condemned, as [EP] claims. Fortunately, as [EK] shows, some degree of variability that includes beneficial as much as harmful effects is (almost) always present in every human group.




Almagro, M., Navarro, L. & Pinedo, M. (forthcoming), “Is testimonial injustice epistemic? Let me count the ways”, Hypatia.

Ayala, S. & Vasilyeva, N. (2015), “Explaining injustice in speech: Individualistic vs. structural explanation”, in R. Dale, C. Jennings, P. Maglio, T. Matlock, D. C. Noelle, A. Warlaumont & J. Yoshimi (eds.), Proceedings of the 37th Annual Conference of the Cognitive Science Society, Austin, TX: Cognitive Science Society, pp. 130–135.

Brandom, R. (2000), Articulating Reasons: An Introduction to Inferentialism, Cambridge, MA: Harvard University Press.

Brandom, R. (2009), Reason in Philosophy: Animating Ideas, Cambridge, MA: Harvard University Press.

Brandom, R. (2019), A Spirit of Trust: A Reading on Hegel’s Phenomenology, Cambridge, MA: Harvard University Press.

Crenshaw, K. (1989), “Demarginalizing the intersection of race and sex: A black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics”, University of Chicago Legal Forum, 140(1), 139–167.

Cassam, Q. (2019), Vices of the Mind: From the Intellectual to the Political, Oxford: Oxford University Press.

Dotson, K. (2011), “Tracking epistemic violence, tracking practices of silencing”, Hypatia, 26(2), 236–257.

Ehrlinger, J. & Dunning, D. (2003), “How chronic self-views influence (and potentially mislead) estimates of performance”, Journal of Personality and Social Psychology, 84(1), 5–17.

Estes, Z. & Felker, S. (2012), “Confidence mediates the sex difference in mental rotation performance”, Archives of Sexual Behavior, 41(3), 557–570.

Frápolli, M. J. (2015), “Non-representational mathematical realism”, Theoria: An International Journal for Theory, History and Foundations of Science, 30(3), 331–348.

Frápolli, M. J. (2019), “The pragmatic Gettier: Brandom on knowledge and truth”, Disputatio: Philosophical Research Bulletin, 8(9), 563–591.

Fricker, M. (2007), Epistemic Injustice: Power and the Ethics of Knowing, Oxford: Oxford University Press.

Gettier, E. (1963), “Is justified true belief knowledge?”, Analysis, 23, 121–123.

Hornsby, J. & Langton, R. (1998), “Free speech and illocution”, Legal Theory, 4(1), 21–37.

Huguet, P. & Regner, I. (2007), “Stereotype threat among school-girls in quasi-ordinary classroom circumstances”, Journal of Educational Psychology, 99(3), 545–560.

Kukla, R. (2014), “Performative force, convention, and discursive injustice”, Hypatia, 29(2), 440–457.

Langton, R. (1993), “Speech acts and unspeakable acts”, Philosophy and Public Affairs, 22(4), 293–330.

Luzzi, F. (2016), “Testimonial injustice without credibility deficit (or excess)”, Thought: A Journal of Philosophy, 5(3), 203–211.

McGowan, M. K. (2019), Just Words: On Speech and Hidden Harm, Oxford: Oxford University Press.

Manne, K. (2020), Entitled: How Male Privilege Hurts Women, New York: Crown.

Mason, R. (2011), “Two kinds of knowing”, Hypatia, 26(2), 294–307.

Medina, J. (2012), “Hermeneutical injustice and polyphonic contextualism:

Social silences and shared hermeneutical responsibilities”, Social Epistemology: A Journal of Knowledge, Culture and Policy, 26(2), 201–220.

Medina, J. (2013), The Epistemology of Resistance: Gender and Racial Oppression, Epistemic Injustice, and the Social Imagination, Oxford: Oxford University Press.

Ramsey, F. P. (1929), “Last papers: Probability and partial belief”, in R. B. Braithwaite (ed.), The Foundations of Mathematics and Other Logical Essays, London: Kegan Paul, Trench, Trubner and Company, pp. 256–257.

Rankine, C. (2014), Citizen: An American Lyric. Minneapolis, MN: Greywolf Press.

Reuben, E., Rey-Biel, P., Sapienza, P. & Zingales, L. (2012), “The emergence of male leadership in competitive environments”, Journal of Economic Behavior and Organization, 83(1), 111–117.

Roessler, B. (2015), “Autonomy, self-knowledge, and oppression”, in M. A. L. Oshana (ed.), Personal Autonomy and Social Oppression: Philosophical Perspectives, New York: Routledge, pp. 68–84.

Williams, M. (2001), Problems of Knowledge: A Critical Introduction to Epistemology, Oxford: Oxford University Press.

Wittgenstein, L. (1969), On Certainty, Oxford: Blackwell.