Epistemic Class Injustice: Class Composition and Industrial Action

ABSTRACT

Writings on epistemic injustice have assessed how people can be harmed in their capacity as knowers when they are a racial minority, a woman, disabled and so on. But what about when they belong to the working class? This paper is an initial attempt to understand why class has so far received limited attention within writings on epistemic injustice and to respond to these reasons. It focuses on how testimonial and hermeneutic injustices specifically harm workers in ways distinctive from the harm one might suffer due to other social identities. It does this by drawing attention to the special case of industrial strike action and the play of conceptual resources and credibility assessments that influence the action’s success. Additionally, it provides a first-time exposition for social epistemologists on what I term the ‘class compositional approach’, derived from 1960-70s Italian labour struggles. This approach, I argue, succeeds in evading the criticism of class reductionism while developing recent philosophical work on class-based injustice.

KEYWORDS:

• Class
• epistemic injustice
• autonomism
• Marxism

Introduction¹

Race, gender and class, together abbreviated as ‘RGC’, are today among the most frequently recognised dimensions of oppression. However, as Kandal (1995, 143) argues, in RGC studies, class is regarded as ‘the weak link in the chain’. Although Kandal noted this phenomenon in 1995, and RGC studies is a now-defunct field of research, replaced by other fields more sensitive to other features by virtue of which people are harmed (including ethnicity, disability, nationality and more), the disparagement of class persists 28 years later, the poor being what Jones (2020, 2) terms ‘the one group in society that it is still acceptable to sneer at, ridicule, even incite hatred against’. Rather than attempting to understand the origin and persistence of this disparagement across the humanities, this paper proposes to make a local intervention into debates surrounding the concept of ‘epistemic injustice’ as it exists within the field of social epistemology. To my knowledge, only two papers have approached a treatment of this exact theme; however, neither is exactly satisfactory.²

The structure of my intervention here is as follows. In §1 of this paper, I explain epistemic injustice, and especially its two main, original variants: testimonial and hermeneutical injustice. In §2, I explain why the lack of attention given to class specifically in the literature around epistemic injustice is surprising, followed by two possible background reasons which may justify this lack of attention. §3 responds to the first of these background reasons by showing that it is possible to focus on class-based injustices, as opposed to race-based or gender-based ones, without succumbing to class reductionism. I show this in the context of what I term the ‘class compositional approach’. I introduce this approach for three reasons. Firstly, because it is consonant with recent research into class-based injustices completed by Cicerchia. Secondly, to introduce the approach to a broader audience of social epistemologists who may benefit from its use. And thirdly, because it provides an understanding of class that cannot be accused of class reductionism. §4 responds to the second background reason by showing that there are epistemic injustices that workers suffer qua workers for which there is no analogy for epistemic injustices that target other social identities.

Testimonial and Hermeneutical Injustice

According to Fricker’s (2007, 1) memorable wording, an epistemic injustice is ‘a wrong done to someone specifically in their capacity as a knower’. Thus, to better understand what ‘epistemic injustice’ means, it must be understood how it is possible for people to be wronged specifically in their capacity as knowers. For this purpose, Fricker distinguishes between two kinds of epistemic injustice. These are testimonial injustice and hermeneutical injustice; however, other scholars have since identified more kinds of epistemic injustice, such as distributive injustice (e.g. Coady 2019). For the sake of this article, I will focus only on the original two kinds of epistemic injustice, testimonial and hermeneutical.³ I do this because they provide the most straightforward illustration of epistemic injustice in the special case on which I rely later to show that there is something distinctive about epistemic injustice as it applies to workers: industrial action.

The first kind of epistemic injustice for Fricker is testimonial injustice. A testimonial injustice can be said to occur when someone’s testimony is discounted because of presiding prejudices on behalf of the hearer that concern the social identity of the speaker. Take a simple case in which a woman claims to have been hit on by a colleague at work but is responded to by her employer with disbelief. In this case, her testimony might be discounted because she is a woman, and her employer is subject to a culture in which women are portrayed as unreliable sources of testimony. She may be described as having undergone a testimonial injustice because she experiences a credibility deficit associated with her social identity as a woman. A credibility deficit occurs when someone is treated as less credible than they otherwise ought to be. Or, in Fricker’s (2007, 1) elegant phrasing, ‘testimonial injustice is caused by prejudice in the economy of credibility’. The second kind of epistemic injustice for Fricker is hermeneutical injustice. By contrast with testimonial injustice, which occurs within the economy of credibility, or how credibility is differently distributed among speakers, hermeneutical injustice ‘is caused by structural prejudice in the economy of collective hermeneutical resources’ (Fricker 2007, 1). One way of understanding what Fricker means by the economy of collective hermeneutical resources is as follows. There is an important difference between our experiences and the understanding of our experiences, and it is by means of our hermeneutical resources that we are able to understand our experiences. Looking at the understanding of our experiences in this way need not commit us to saying that any experience is such that there is only one way of understanding it. Indeed, the fulcrum of hermeneutical injustice is the contestation between different ways of understanding an experience – or, in more radical cases, being able to understand an experience at all.

When this unequal distribution of hermeneutical resources is a product of structural prejudice, it causes a hermeneutical injustice. What might it mean for the distribution of hermeneutical resources to be caused by structural prejudice? In Fricker’s original formulation, this distribution concerns either the presence or absence of a concept sufficient for allowing us to understand our experiences.⁴ However, others have added greater complexity to the ways in which these resources can be unequally distributed. Important for my purposes is Medina’s (2022, 323) recognition that, even in cases where a concept has already been developed, it can be ‘blocked and excluded from the dominant epistemic systems’. This can be demonstrated well using Fricker’s example of the concept of sexual harassment. In the early 1970s, a group of American women coined the term ‘sexual harassment’ to describe the reasons for which Carmita Wood, an administrative assistant at Cornell University, was forced to leave her job (Fricker 2007, 149–50). Before ‘sexual harassment’ as a term was coined, women lacked a hermeneutical resource with which to understand and communicate their experiences of unwanted sexual attention. They were thus, to use Fricker’s (2007, 153) term, ‘hermeneutically marginalized’.

Why Not Class?

It should be clear by this point that epistemic injustice identifies a serious issue related to people’s status as knowers, especially when that status is denigrated for prejudicial reasons. In her own work, Fricker frequently speaks of gender, race and class in the same breath (e.g. Fricker 2007, 16, 23, 32, 159 and 160). Despite this, publications concerning epistemic injustice and class have so far been a rarity. I begin this section by explaining why the lack of attention given to class is surprising. I then offer two possible background reasons for this lack of attention; these are ‘background reasons’ because they are not usually articulated when class is overlooked, and academics overlooking class may not even be conscious of their reasons for this. The first reason I term the indistinctiveness reason, and the second the harmfulness reason. Concerning the second of these reasons, I provide evidence of its presence in the extant literature on epistemic injustice. The remainder of this paper responds to these two reasons.

Class and the Origins of ‘Epistemic Injustice’

Marx’s name does not appear even a single time in Fricker’s 2007 book.⁵ Despite this, in a 2017 retrospective on her work, Fricker situates her concept of epistemic injustice ‘at one remove from the long shadows of both Marx and Foucault’ (Fricker 2017, 56). This is notable just because, while the Foucauldian dimension of epistemic injustice has received attention from authors working in the aftermath of Fricker’s book, the Marxist dimension has been almost entirely ignored.⁶ In the context of Fricker’s 2017 comment on her work, the extent of the discussions of class in Epistemic Injustice may indicate the implicit presence of the German theorist’s ideas. Accordingly, Fricker acknowledges class as a dimension of epistemic injustice and, in the direct origins of epistemic injustice as a concept lies perhaps the most committed theorist of class injustice, Karl Marx. This is why the lack of commentary on epistemic injustice and class is surprising. However, more should still be added to Fricker’s way of being concerned with class. After all, it may be the case that she does deal with it, but in a different way than Marx and Marxists do.

To see how Fricker deals with class as a dimension of epistemic injustice in her own work, it should be noted that she deploys an intersectional conception of injustice. As evidence for Fricker’s commitment to an intersectional conception of injustice, the following warning from her 2007 book is noteworthy: ‘the speaker may be a woman, but the fact that she is white and middle-class may mean that there is no hermeneutical gap depriving her of the expressive resources she needs, in the context, to render herself intelligible’ (Fricker 2007, 171). This warning indicates at least a subtle commitment to an intersectional conception of injustice because it recognises that it is insufficient from the perspective of anticipating epistemic injustices to identify possible victims based solely on their belonging to one social identity group. Instead, not all women will suffer epistemic injustices because the likelihood of one’s testimony being believed and one’s capacity to participate in collective hermeneutical resources are sensitive to one’s belonging to other social identity categories as well. This recognition may allow us to make sense of the fact that class-based epistemic injustices are not treated distinctively relative to gender-based and race-based ones. This is so because according to the intersectional view, someone is never just a worker, but is also a member of various other social identity groups. Because nobody is just a worker, it is unlikely that the injustices that any individual faces will be solely class-based, and therefore to treat class separately from gender and race risks making our understanding of any injustice needlessly abstract, or worse, reductive. Be that as it may, this fact does not enable us to make sense of the inverse phenomenon: that gender and race receive greater attention in the literature on epistemic injustice than class does.

Justifying the Disparagement of Class

In the above discussion, I have been concerned with outlining why it is surprising that class has been given less attention in writings on epistemic injustice than have race and gender specifically. However, it is not necessary to let this phenomenon remain surprising, and it is possible to identify good reasons for this lack of attention that may be operating as background reasons against discussions of class. It is worth emphasising that these are likely to be background reasons because, to my knowledge, not only do scholars working on epistemic injustice not tend to offer to class the same attention they do to race and gender, they also do not tend to explain their reasons for this lack of attention. Because of this, identifying these reasons requires some degree of speculation. Accordingly, it should be noted that the reasons I offer below are not likely to be an exhaustive list, but only to be those that I suspect are the most likely to be operative whenever class fails to receive the same attention as race and gender. These reasons are as follows:

[Indistinctiveness]

There is nothing distinctive to be gained by giving specific attention to class-based epistemic injustices.

[Harmfulness]

Specific attention to class in the context of epistemic injustice may in fact be harmful.⁷

In the following, I will explain both of these reasons. I will focus on [Indistinctiveness] first. Because someone endorsing [Indistinctiveness] believes that nothing distinctive is to be gained from investigating epistemic injustice and class as opposed to race, gender, disability or anything else, such investigations to them will appear naturally unnecessary. If one believes that there is nothing to be gained by researching class-based epistemic injustices, and that it is therefore unnecessary, what reasons might they have for believing this? One might think that there is nothing distinctive to be gained by giving specific attention to class-based epistemic injustices because these injustices are already sufficiently well understood.⁸ However, someone who believes these are already sufficiently well understood may be uncommitted on the value of this already-existing research, and therefore need not offer resistance to anyone further pursuing this research apart from emphasising that it has already been done. Another reason for which one might believe that this research is unnecessary could be that there is simply nothing distinctive about class-based epistemic injustices. There might be nothing distinctive about injustices of this sort if there are no features of it for which there are no analogies with other injustices perpetrated on people with respect to their participation in other social groups. For instance, workers’ testimonies are sometimes discarded because of prejudices existing in the social imaginary that are also used to discredit the testimonies of women or racial minorities. When workers’ testimonies are discarded due to these shared prejudices, the injustice that workers suffer is not distinctive relative to the injustice suffered by women and racial minorities. Because the injustice in question is not distinctive when applied to workers, one might be led to believe that there is nothing distinctive to be learned by focusing on workers as opposed to women or racial minorities.

[Harmfulness] is more strongly opposed to research into epistemic injustice and class and is difficult to construe otherwise than negatively. Along these lines, two construals of the harm done by research into epistemic injustice and class recommend themselves. On the first construal, attention to class is harmful simply because it occurs to the detriment of attention to other dimensions of epistemic injustice. The harm in this sense might be referred to as a distributive harm because it challenges the allocation of resources, specifically the attentional resources of researchers. On the second construal, attention to class is instead an intrinsic harm, reflecting the ways that research into class across the humanities is often argued to be harmful to women, racial minorities and others. This second construal is especially important because variations on it are used by authors writing on epistemic injustice. In the 2017 retrospective on her work, Fricker (2017, 56) comments that ‘as regards Marxism, for my purposes the monolithic social ontology of class – or its gender or race counterparts – remained at that time riskily insensitive to other dimensions of difference, even if it was recognised to be an abstraction rather than an empirical generalisation’. Additionally, Alcoff (2015, 36) writes that ‘class reductionist approaches that deny the relevance of whiteness for the white poor do not pay sufficient attention to the psychic benefits of whiteness in racially hierarchical communities, or the social practices of interaction, both formal and informal, that are structured by race’.⁹ I will give Alcoff’s objection close attention in the next section. But clearly, a satisfactory treatment of epistemic injustice and class must carefully avoid exacerbating the harm suffered by people due to existing epistemic injustices.

Class and Class Composition

In this section, I will deal with [Harmfulness]. To reiterate, [Harmfulness] states that specific attention to class-based epistemic injustices may be harmful. I specified earlier that this harm can be construed distributively or intrinsically. On the distributive construal, or [Harmfulness_dist], the harm reflects the limited attentional economy of researchers. This limited economy creates a distributive harm when insufficient attention is paid to the important ways in which people are or can be harmed. If one believes that class is not one of these important ways, then research into it as an epistemic injustice may be construed as a harm. On the intrinsic construal, or [Harmfulness_int] the harm reflects the nature of research specifically into class. The next section will focus on [Indistinctiveness]. There I show that workers can be distinctively wronged, specifically in their capacity as knowers, and therefore that the indistinctiveness reason should be resisted. This will also provide a way of resisting [Harmfulness_dist]. For that reason, in this section I will focus on the intrinsic construal. I reply to it by offering a treatment of class that I believe avoids class reductionism, and which provides adequate place for other forms of oppression that people might suffer. In short, the argument this section provides is the following. Research into class is intrinsically harmful when it is class reductionist; research into class does not need to be class reductionist; therefore, research into class does not need to be intrinsically harmful. As this formulation implies, I do not claim that class reductionist approaches do not exist, only that alternative approaches to class-based injustice exist.

To avoid class reductionism, the following discussion of class must recognise that there are ways in which people suffer that are not reducible to their class-status while still focussing on class specifically. To achieve this, I propose to understand class, following Mohandesi, as defined by the conflict between two competing characterisations. As will be seen below, treating class this way avoids class reductionism, grants class overall its distinctiveness relative to other forms of oppression¹⁰ and connects to recent philosophical work being done on class injustices. Mohandesi’s characterisation of class, specifically the working class, is directly inspired by Marx. As Mohandesi (2013, 74) explains, in Marx one finds a definition of class as firstly an ‘economic category’ and secondly a ‘political subject’. This dual-treatment of class is fundamental to the class compositional approach, first theorised by scholars and activists involved in labour struggles in Italy from the 1960s–70s (Mohandesi 2013, 84–5).¹¹ In the following, I will explain the class compositional approach and how it can be used to speak of class without falling to [Harmfulness_int].

Class Composition

Class compositional approaches traditionally attempt to relate Marx’s two definitions of class in terms of technical composition and political composition in the following way. The working class is first an economic category, where members are grouped because of a broadly shared situation, referred to as their technical composition. Then, by means of what class compositionists¹² call the ‘political leap’ (Mohandesi and Haider 2018), the working class ‘is no longer bound to certain positions in the production relations of a given historical conjuncture but signifies instead a process of subjectification through the articulation of opposition’ (Mohandesi 2013, 76). As a result of the political leap, the working class moves from being defined by its technical composition, or shared situation, to its political composition, or capacity for action. Class composition is thus an attempt to relate the working class’s technical composition, or shared situation, to its political composition, or unification as a subject capable of organised action. Just what the shared situation of the working class is, is a topic of some dispute, especially since certain explanations reproduce just what Fricker (2017, 56) calls a ‘monolithic social ontology of class’, that is, an understanding of class that does not sufficiently recognise the internal differentiation of the working class into fragmented identity groups. In response to this, Cicerchia (2021, 617) recommends that ‘internal divisions within the working class are constitutive of class formation’. In other words, in contrast to class reductionist approaches, which begin with the unity of the working class, for Cicerchia what constitutes the working class is the exact way in which its members are separated by gender and race. Members of the working class are internally divided specifically because ‘from the point of view of workers, unemployment, underemployment, or a diminished position within the labor market is a perpetual threat’ (Cicerchia 2021, 618). This leads to the form of racialised division between workers identified by Alcoff, while adding that it is actually because of this form of division that white and black workers can be said to share a class-status.

What is remarkable is that Cicerchia’s recommendation, despite being a fairly recent one, can also be interpreted as the reason for which the class compositional approach was first proposed by Romano Alquati (Wright 2017, 42–5). It was out of the sense that the Italian activists and scholars had failed to grasp capitalism’s unique ability to reorganise the relations of production that the class compositional approach was born. This ability to reorganise relations of production is also an ability to impose class oppression on groups historically excluded from the world of industrial labour, such as women who have tended to be consigned to household labour. Just how this reorganisation is to be described within the framework of the class compositional approach is a matter of some disagreement. The description the British group Notes from Below have offered is perhaps the most elegant and resistant to charges of reductionism among those that exist today. For this reason, I will focus on their approach. Notes from Below suggest that, to fully capture the distinctiveness of how the political composition of the working class develops from various sources, and in response to disparate struggles, the concept of social composition should be introduced as well (Woodcock 2021, 91). The introduction of social composition responds to the fact that ‘previous analysis of class composition has based workers and their resistance almost exclusively on the workplace’, providing a way for accommodating struggles ‘over the conditions of state-provided social services, migration and borders, housing and rent, and a wide range of other issues’ (Notes from Below 2018). In other words, social composition, unlike technical composition, focuses on the features of working-class life outside of spaces of immediate production, to encode struggles in which workers might be engaged beyond demands concerning the workplace.

A final feature of this treatment of class deserves recognition. By construing class-membership specifically as a relation to the labour market, it becomes possible to perceive people who do not work in for-profit industries, or who occupy varies strata of the wage ladder as working-class. Note that, according to Cicerchia’s (2021, 617) definition, ‘if one confronts dependency on the labor market as a problem and one depends on collective action to address it, then one is working class’. This definition commits her to no particular motivation for the purchasing of workers’ labour-power.¹³ For this reason, even employees of nationalised services, such as the British healthcare system on which I focus later, can be working-class in her sense. Defining class by dependence on the labour market as a ‘problem’ seems particularly apt, also because it permits the assumption that this dependence can be a greater or lesser problem for different people. Jobs in high-demand may entail a lesser problem for workers since this demand reduces the means for employers to threaten their employees. Workers in high-demand jobs are not only in a stronger position against unemployment by their direct employer, but should they lose their jobs, they will experience less difficulty in finding employment elsewhere.¹⁴ It also accommodates wage-stratification within the labour market, with those who are better compensated for their labour experiencing the exchange of this for a wage as less of a problem than those who are inadequately compensated.

Therefore, instead of using a class reductionist theory, the class compositional approach provides an account of the internally differentiated constitution of the working class. It does this by requiring researchers to study geographically local and historically specific forms of oppression, including those related to race and gender, while innovating responses to these structures without relying on older mechanisms of class struggle, such as activism emanating from political parties or trade unions, which have both in their broad histories sacrificed workers for the sake of survival.¹⁵ By refusing to rely on older mechanisms of class struggle and developing responses based on what injustices can be observed at a specific time and place, a safeguard is set up against approaches that would reduce the specificity of these injustices to forms of class-based oppression. By treating class as internally differentiated and related to non-class struggles against injustice, as it is treated within the class compositional approach, [Harmfulness_int] can be undermined. It can be thus undermined because each of the possible justifications underlying [Harmfulness_int] concern forms of class reductionism, whereas those following the class compositional approach cannot be accused of class reductionism. Naturally, this does not deal with [Harmfulness_dist], which proffers an additional means of resisting further research into epistemic injustice and class. Notice, however, that [Harmfulness_dist] relies on either [Harmfulness_int], or [Indistinctiveness]. It relies on these because, for [Harmfulness_dist] to hold, it must be agreed that our resources are unjustly distributed if any of them are allocated to researching class specifically, and one might agree with this either because they believe that such research is intrinsically harmful or because nothing distinctive is gained from undertaking it. One might also think that allocating resources to the study of epistemic injustice and class is unjust because, although people are genuinely epistemically harmed due to their class status specifically, this harm is less important than the harms that people suffer due to gender, race, etc. I do not intend to deal with this third possible justification for [Harmfulness_dist] because I do not believe that ranking the social identities by virtue of which people suffer injustices is productive, provided that the harm created by the injustice is ongoing and genuine.

Epistemic Class Injustice

In this section, I will address [Indistinctiveness] as a reason for avoiding research into epistemic injustice and class. For this purpose, it will be worth returning to the definition I previously proposed for ‘distinctive harm’. I earlier stated that some harm is distinctive if it is one for which there is no analogy with other injustices perpetrated on people with respect to other social identities. An example of such a harm is the way in which women’s testimonies are discarded on account of their supposed intuitiveness rather than rationality. This harm is distinctive because, while the testimonies of other groups, including workers and black people, are often similarly discarded on prejudicial grounds, the prejudice concerning intuitiveness rather than rationality is unique to the treatment of women. By contrast, consider cases in which women’s testimonies are discarded because they are prejudicially supposed to be ignorant. Such cases do not involve a wrong distinctive to women, because the testimonies of workers and racial minorities are frequently discarded on the same grounds. Such cases might be termed ‘normatively distinctive’ because they concern the justification offered for the perpetration of certain wrongs. This is one form that the distinctiveness in question can take.

Another form it can take concerns wrongs perpetrated in situations in which one not only participates by virtue of their belonging to one social identity category rather than another, but also in which one cannot participate by virtue of belonging to any other social identity category. Whenever this is the case, the epistemic injustice might be termed ‘situationally distinctive’ because it concerns the kind of situation in which the injustice occurs. The example on which I will elaborate below as an example of a situationally distinctive injustice concerns industrial action, specifically industrial strikes. Epistemic injustices that occur while assessing the credibility of an industrial strike are situationally distinctive because an industrial strike is something that workers uniquely have the ability to carry out. By contrast, an example of a situationally indistinct injustice might involve how one is treated in a restaurant. If someone is in a particularly fancy restaurant, any number of facts about a person might be involved in their receiving exceptionally poor service. It may be because they are not dressed well enough, which encodes specifically class-based prejudices, or because they are perceived to have the wrong skin colour or gender to be dining in this restaurant. Because this injustice may be suffered because of any of these reasons, it is situationally indistinct.

I have already recommended that assessments concerning the credibility of an industrial strike may perpetrate a situationally distinctive wrong on workers. This is a good example for the present discussion as well since it involves an assessment of credibility, making it epistemic in character. Wrongs perpetrated in this context are therefore not only wrongs that workers distinctively suffer, but are also wrongs that they suffer in their capacity as knowers. I specify ‘industrial strikes’ because there are other types of strikes that concern the non-class group membership of those participating in them. For instance, in 1974 the Italian Marxist Mariarosa Dalla Costa (2019, 52) described the possibility of fomenting a general strike as a part of the Wages for Housework campaign, which is based on the recognition that ‘we all do housework; it is the only thing all women have in common, it is the only basis upon which we can gather our power, the power of millions of women’ (emphasis in original). The fact that such strike action would protest work that women disproportionately complete makes the distinction between their status as women and their status as workers less precise. Indeed, this ambiguity reflects the way in which real political struggle often involves overlapping group memberships. However, that is beside the point. I only wish to stress that there are forms of strikes that involve non-class group membership, not that in any of these strikes the participants’ class membership is irrelevant.

Epistemic Injustices Around Strike Action

With that aside, I will now focus on the ways in which epistemic injustices might be said to influence assessments of industrial strikes. As Ovetz (2020, 109–12) shows, a large number of factors influence the success or failure of an industrial strike, including whether the strike itself even succeeds in beginning. Luxemburg (1999) quotes Engels in her famous 1906 pamphlet The Mass Strike as saying that perhaps the most important factors are having ‘a perfect organisation of the working-class and a full war chest’. By this, she means that enough employees in a workplace must be committed members of the organisation orchestrating the industrial action, that the organisation itself must be sufficiently well developed and managed to supplement the work of the organisers of the action, that the organisers must be competent enough to facilitate the action and that the organisation controls a sizeable enough strike fund to offset the majority of the immediate monetary costs for the workers of the industrial action.¹⁶ Ovetz adds several factors to this as well, creating a checklist for evaluating the credibility of an industrial strike threat (see Ovetz 2020, table 4.2). These factors are ultimately compiled into the credibility assessments of workers as they determine whether they will be able to perform the industrial action, and employers as they determine whether to take the industrial action seriously.

Whether employers take industrial action seriously is important because it is a precondition of the workers’ demands being met. Credibility assessments can also change as the industrial action continues. For instance, should the employers underestimate the commitment of a large enough portion of the workers in their company to the action, causing the strike to persist longer than expected, incurring for the business a substantial loss in profits, then they may choose to negotiate with the workers’ organisation to reach a compromise. That the employer has entered negotiations does not indicate that they are taking the action seriously, since they might do so only to convince them to return to work without satisfying any of their demands. The only indication that an employer has evaluated an industrial action as credible is when they recognise the genuineness of the workers’ demands and understand that threats of legal action or dismissal to the workers will not dissuade them.¹⁷ The strike may dissolve before the employer has reached this stage in their credibility assessment of the industrial action; in that case, the most plausible explanation is that the action lacks credibility in the eyes of the workers. Such cases indicate a worker-side credibility failure because the employer may have been prepared to enter negotiations having assessed the strike threat as credible even if the workers do not share this assessment.

With this description of the role of credibility assessments in industrial strike action aside, how might these credibility assessments be said to perpetrate epistemic injustices? On a fairly neutral definition, an injustice occurs whenever there is an infraction of some standard of justice. Therefore, for an epistemic injustice to be identified here, some standard of justice must be infracted, and that standard of justice must concern people’s capacities as knowers. A first way of describing the epistemic injustice operative when credibility assessments are being made of industrial strike threats is as a testimonial injustice. As I explained earlier, a testimonial injustice occurs when someone’s testimony is discarded for prejudicial reasons, and these prejudices can cause a hearer to challenge on false grounds whether the speaker knows what they claim to. The testimony concerning which a potentially unjust credibility assessment is made here is the workers’ testimonies concerning their reasons for and intention to strike. Their reasons for striking may include, as in the recent wave of National Health Service (NHS) strikes in the UK, overwork, understaffing and failure to raise pay in line with inflation.¹⁸

Because the NHS is a nationalised industry, any epistemic injustice inflicted on workers in this case is inflicted on them primarily by the government. One form that this has taken is a denial of the workers’ knowledge of their inadequate working and living conditions, epitomised in Prime Minister Rishi Sunak’s insistence that a pay rise of 4–5% is ‘appropriate and fair’ (The Guardian 2022). However, although this outright dismissal of the nurses’ testimonies regarding their reasons for striking is startling, it does not itself constitute an injustice. Sunak may, after all, sincerely believe that a pay rise of 4–5% is adequate for satisfying the workers’ demands, for instance if the Health Secretary Steve Barclay provided him expert testimony of this that Sunak interpreted as having greater credibility than the contrasting testimony he was receiving from the striking workers. This observation makes good of Medina’s (2013, 60) addition to Fricker’s concept of testimonial injustice, that testimonial injustices are the product of a combination of credibility deficits and credibility excesses. In this case, Medina’s addition allows us to see a credibility excess for expert testimony that contradicts and therefore produces a credibility deficit for the testimony of the people directly suffering the consequences of working conditions in the NHS.

That aside, for the assessment of the striking workers’ testimonies to be unjust, the credibility deficit they suffer must be unjust (Fricker 2007, 22). The injustice becomes apparent if we observe that, in the UK, ‘in spite of the structural facts of mass unemployment and deepening social inequalities, issues such as poverty and worklessness continue to be framed by governments as cultural or behavioural issues’ (Frayne 2015, 100). Following this, one might shift attention from structural causes of poverty, such as government mismanagement of the NHS, to cultural and behavioural issues by deploying negative identity prejudicial stereotypes that depict workers as ‘thick. Violent. Criminal. “Breeding” like animals’ (Jones 2020, 4). To this list of stereotypes targeting workers might be added greedy, which operates more clearly in denunciations of strike action for better pay (Bland 2022). As I noted earlier, these reasons for discarding workers’ testimonies may not be normatively distinctive, insofar as they are similarly used to target other oppressed groups. However, their use in this instance is situationally distinctive, because when they are used to produce a credibility deficit for people involved in an industrial strike, they are being used to target these people in a situation in which they must be workers. Popular depictions of industrial action that operate on prejudiced depictions of workers can be expected to affect employers as well as workers by lowering the latter’s self-confidence. Lowering the self-confidence of striking workers may succeed in breaking the strike if it either creates internal division among the workers, weakening the strength of their organisation, or if it causes the workers to lose faith that their demands can be met if they internalise some part of the stereotype. As a consequence of this, the workers may be more willing to reach a compromise more unfavourable to them with their employers than they might have otherwise been able to achieve.

More deserves to be said of the NHS here, which might otherwise undermine the account I have been offering.¹⁹ It may be significant that, according to the most recent surveys, although people with black ethnic backgrounds comprise 4% of the UK’s total population, they make up 7.4% of all NHS employees (National Health Service 2022; Office for National Statistics 2021). For comparison, people with white-British backgrounds comprise 74.4% of the total population and are 74.3% of all NHS employees. While white Brits are proportionally represented within the UK’s single-largest employer, black people of all backgrounds are over-represented in this generally low-paying occupation. Because of this, it may seem unwise to separate the recent NHS strikes from racial factors, treating the injustices NHS employees face as distinctive to class in any capacity. Acknowledging these facts, I believe it is possible to still see NHS strikes as manifesting situationally distinctive epistemic class injustices. Although any injustice will occur at the intersection of multiple, overlapping group memberships, on occasion one membership may be a stronger factor than others. Bohrer, for example, refers to the murder in 2014 of Eric Garner, an African American man, by the police. The preponderance of evidence concerning the importance of race in such cases as Garner’s permits us to afford his race a strong explanatory role in his murder. But it does not permit excluding all other factors. In Bohrer’s (2019, 198) words, ‘even though, in this particular instance, race has an important emphasis, a full or complete picture of Eric Garner’s death would not be complete without understanding the gendered, sexual, and class politics of his murder by the NYPD’.²⁰

It is not my intention in focussing on the NHS to suggest that factors apart from class are never important for explaining injustices within the workplace, only that, as far as industrial action is concerned, references to class will feature in any explanation of the injustices faced therein, unlike references to other social group memberships which may feature only in some (or many). To illustrate this, consider that, while an organisation that primarily represents black employees in a workplace may announce a strike, no comparable organisation could organise a strike among black employers, just for the reason that industrial strikes are uniquely a tool usable by workers.²¹ Against an organisation which specialises in facilitating industrial action among black employees, the use of negative identity prejudicial stereotypes which target black people, or even black workers specifically, could be expected. Given considerations such as this, situational distinctiveness should not be mistaken for situational exclusivity. Insofar as someone is on strike, it should not be thought that only their class identity is operational. But the contrasting reality that one could not be involved in an industrial strike unless certain facts about one’s class-membership held licenses us to treat class distinctively in this scenario. To go on strike is always to run the risk of being targeted by class-specific prejudices, but it is only sometimes to run the risk of being targeted by prejudices which are not class-specific. This, I hope, is reason enough to agree that epistemic injustices arising within credibility assessments of industrial strikes are class distinctive. However, I welcome disagreement with my conclusions concerning industrial action.

The epistemic injustices that workers face are not only of the testimonial variety. While the case of industrial strikes provides us a way of identifying testimonial injustices as they are directly inflicted on workers, a subtler, hermeneutical injustice carries on in the background that the striking workers must constantly resist the influence of. To see this, let us consider Anderson’s description of hermeneutical injustices around the Black Lives Matter (BLM) movement. In her article, Anderson analyses the epistemic injustice produced when the phrase ‘Black Lives Matter’ is met with post-racialist responses. Post-racialist responses all propose that someone’s race should make ‘no moral, legal, political, or social difference to how people are treated’, although they may differ on whether the concept of race should be eliminated or merely excluded from certain domains of decision-making (Anderson 2017, 142). What is consistent across post-racialist responses to BLM is that in each the hearer fails to grasp the intended force of the utterance ‘Black Lives Matter’, and this by ‘ceasing to think, act, and feel in terms of race’ (Anderson 2017, 144). In this case, a hermeneutical injustice is produced because the hearer refuses to acknowledge a hermeneutical resource that is crucial for the articulation and appreciation of the wrong suffered by those sharing a social identity with the speaker. This refusal to acknowledge a hermeneutical resource causes a hermeneutical injustice here because that resource, the concept of race, is necessary for a full understanding of the nature of the wrong being described.

The principle that can be taken from Anderson is that when someone is harmed due to their social identity, and others refuse to accept the existence of this social identity, they suffer a hermeneutical injustice. By comparison, in a 1992 article for Newsweek, Margaret Thatcher stated that ‘class is a communist concept. It groups people as bundles and sets them against one another’ (Thatcher 1992). Although this claim was made in 1992, Kundnani (2000, 4) shows that New Labour today carries on ‘the same anti-working-class hostilities which characterised Thatcherism’. When employers or the government refuse to recognise the concept of class, a crucial component of the injustice that workers face is also dismissed. When nurses are denied fair wages, it cannot be unimportant that they are financially dependent upon access to work, in which they exchange their labour-power for a wage. The importance of this relation manifests in numerous facts surrounding their strike action. For reasons of space, I will elaborate upon only one before concluding. In going on strike, workers at the same time refuse themselves access to the wage on which they are financially dependent. If it is not understood that they are striking as members of the working class, the depth of the personal consequences that they are willing to risk cannot be grasped. Acknowledging the personal hardship that striking workers are willing to bear provides a counterweight to negative identity prejudicial stereotypes that are present in evaluations of their testimonies. It is unlikely that anyone would risk access to their means of basic subsistence unless they had very good reason to do so.²² This point is emphasised when the risk, not only of temporarily losing access to a wage, but of losing one’s job altogether, is leveraged against the workers. In this sense, the hermeneutical injustice of refusing the concept of class its place in understanding strike action contributes to the testimonial injustices surrounding such action as well.

Conclusion

In this paper, I have attempted to explain the dearth of research around the concept of epistemic injustice that concerns class specifically. For this purpose, I began by outlining Fricker’s two original kinds of epistemic injustice. Following this, I explained why one might find it surprising that class has received diminished attention in writings on epistemic injustice and two possible background justifications for this diminished attention. It is surprising because Fricker deploys a broadly intersectional conception of injustice, which demands recognition of the fact that any individual is always the member of multiple social groups, while she also recognises class as a social identity because of which one might suffer injustices. Not only this, but Fricker recognises a relation between her work and Marx’s, the latter of whom gives sustained attention to class. Even so, one might omit to give class specific attention in writing about epistemic injustice for two reasons which I termed [Harmfulness] and [Indistinctiveness]. I also distinguished between two possible construals of [Harmfulness], one intrinsic and one distributive, both of which specify the precise way in which research into epistemic injustice and class might be harmful to those on whom epistemic injustices are inflicted.

After this, I focussed on responding to [Harmfulness_int] by showing that research into class specifically need not be reductionist. Because [Harmfulness_dist] relies on either [Harmfulness_int] or [Indistinctiveness], I did not treat it separately. I responded to [Indistinctiveness] by showing how workers specifically can be harmed in their capacity as knowers. I did this by focusing on the epistemic injustices that arise during industrial strike action. In these circumstances, workers suffer a distinctive epistemic injustice when negative identity prejudicial stereotypes that target poor workers are used against them to either dismiss their knowledge of their reasons for striking or to undermine their credibility assessment of the strike. Additionally, workers suffer a distinctive hermeneutical injustice under these circumstances when the concept of class fails to be acknowledged by hearers.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Notes on contributors

Kenneth Novis

Kenneth Novis is a DPhil candidate in Philosophy at the University of Oxford, St. Hugh’s College.

Notes

1. For their help in discussing this paper with me, I would like to thank Milan Ney, Dante Philp, Clark McAllister and Mary Peterson. Additionally, thanks to Miranda Fricker for providing me with recommendations for this paper at an earlier stage. Thanks also to my reviewers from Social Epistemology for their many insightful remarks.

2. Kwok (2021) focuses on how workplace hierarchies create and reproduce epistemic injustices for workers. However, the focus of Kwok’s paper is narrower than mine, and I propose to begin by recognising that the injustices workers suffer cannot be reduced to those undergone in the workplace but extend across all of society (cf. Tronti 2019). Ferreira’s (2022) paper is curious because the phrase ‘epistemic injustice’ does not appear anywhere in it except the title. However, in her description of the presumption that poor Brazilians do not have the understanding necessary to make politically important administrative decisions, it is clear that this is a paper on epistemic injustice. Ferreira rejects Marxism on the basis of its class reductionism; I deal with the class reductionist complaint in §3.

3. Fricker has since come to refer to these kinds of epistemic injustice as ‘discriminatory epistemic injustice’ (Fricker et al. 2017, 53, emphasis in original).

4. Fricker has since added that hermeneutical injustices can be caused by ‘insufficiently widely shared hermeneutical practices’ (Fricker 2016, 166).

5. Although ‘Marx’ does not appear in the book, the word ‘Marxist’ does have a single instance on page 147.

6. On Foucault and epistemic injustice, see Allen (2017), and Lorenzini (2022).

7. A third reason might be added: that, because epistemic injustice and class is already sufficiently well understood, there is no need to study it any further. Because in my view this reason is effectively neutral about research into epistemic injustice and class, I do not consider it an issue, and therefore do not discuss it in the following. A fourth possible reason may be that, because prejudiced perception is an integral feature of epistemic injustice, if class is not perceived in a way comparable to race and gender, it may simply not be suitable for consideration in theories of epistemic injustice. Whether class is thus perceived seems to me a complicated and fascinating discussion; however, I will put it aside for the sake of this paper, recognising this as a potential weakness of my account.

8. See fn. 7.

9. Cf. Alcoff (2021).

10. The importance of this point is highlighted by Cicerchia (2021).

11. The definitive work for those interested in the history of class composition remains Wright (2017).

12. I borrow this term from Roggero (2023, 52).

13. This is a further point of agreement between Cicerchia and class compositionists, the latter of whom stress the importance of analysing class, not primarily in terms of the organisation of the capitalist class, but in terms of the experiences of the working class. In Tronti’s words, ‘we too saw capitalist development first and the workers second. Now we have to turn the problem on its head, change orientation, and start again from first principles, which means focusing on the struggle of the working class’ (Tronti 2019, 65, emphasis mine).

14. These two facts may be related: the threat of unemployment is less of a threat to workers when employment can easily be found elsewhere.

15. Independence from older parties and trade unions was a key concern for these Italian activists: see Wright (2017, 13–9).

16. I use the phrase ‘organisation’ rather than ‘union’ because strikes may take place without the mediation of a union specifically, such as in wildcat strikes.

17. It is not necessary that the workers’ demands are genuine. Conceivably, workers could strike opportunistically, perceiving the inability of their employers to resist their demands, and not because of an injustice which the workers face. In my opinion, however, genuinely motivated industrial action is the more common occurrence. On the question of justifications for strike action, see Gourevitch (2018).

18. For an excellent and recent class-based inquiry into the NHS, see further the anthology compiled in 2022 by the Anarchist Communist Group and Angry Workers of the World, titled Sick of It All! Work, Inquiry, & Struggle in the NHS (Anarchist Communist Group and Angry Workers of the World 2022).

19. I thank my reviewers for their generous remarks, which prompted the following discussion.

20. Bohrer abandons this notion of the contextual salience of certain social group memberships in favour of what she terms the ‘equiprimordiality’ of exploitation and oppression (see Bohrer 2019, 203–5).

21. In 1869, black workers in the U.S.A. established the Colored National Labor Union to represent them largely due to the fact that the National Labour Union had failed to respond adequately to their concerns as black workers (Foner and Lewis 1978, 1–34),

22. See fn. 17.