What other intellectual resources were there to hand? I looked to the English-language philosophy of the kind I had read as an undergraduate, and there was nothing that spoke directly to any of these themes. Indeed, the whole self-conception of analytic philosophy as such (at that time) had things sewn up so that any questions designed to prompt thoughts about how relations of social identity and power might impinge upon an aspect of epistemology could only sound misplaced, like vaguely embarrassing, untutored overspill from a neighbouring discipline such as sociology, or political theory. The fascinating and challenging feminist writings I had more recently been exposed to on my MA were barely registering in mainstream analytical quarters.

This made me wonder: how might one try to bring theoretical articulation to some of these themes in the idiom of analytic epistemology? If one could start with describing an everyday on-the-ground experience of being wrongfully undermined or disadvantaged in some aspect of epistemic function, then maybe that would give rise to some new questions and perspectives. So I tried to do that, and the obvious place to start was with the experience – had by so many, and described by so many in different ways and other theoretical and literary contexts – of having one's word underestimated (genuinely misjudged) because of prejudice. I called that phenomenon Testimonial Injustice – the experience of receiving a level of credibility reduced by prejudice. I analysed it as wrongfully undermining or insulting the person specifically in their capacity as an epistemic subject – as a knower – thus casting it as one basic type of ‘epistemic injustice’. I conceived it as typically disadvantageous, though of course there can be, sometimes empowering, exceptions where the particular circumstances mean it turns out advantageous to be wrongfully underestimated (think of Miss Marple).

Yes, for sure – I mean, there are as many as we want there to be, depending on how many fine-grained differences may strike us as worth distinguishing. In defining Testimonial Injustice I thought it was most useful to circumscribe the category with reference to the basic ‘kind’ of credibility reduced by prejudice; but that is a broad category that naturally permits a certain internal heterogeneity. One dimension of heterogeneity is simply the degree of deficit, so that Testimonial Injustices can range from slight depressions of credibility – which may or may not take the credibility level below the threshold for being believed, and may or may not be seriously disadvantageous to the person concerned – to relatively severe depressions of credibility – which are very likely to press the credibility level below the belief threshold, but again may or may not be seriously disadvantageous, depending on the practical context. To see that even severe depressions of credibility need not be damaging, imagine a scenario in which a very angry customer demands to see the Manager. Unaware of their intention to cause the Manager bodily harm, you reveal that you are the Manager, but a prejudice on their part leads them to doubt that you are in that position of authority, and so you are saved from their violence. They did you a Testimonial Injustice, and a Testimonial Injustice is an intrinsic epistemic wrong, but one which on this occasion brought you a serious practical advantage – your personal safety.

The most severe sorts of Testimonial Injustice will be forms of silencing. Imagine a situation in which the prejudicial deficit of credibility is so severe that it's as if the person hasn't said anything at all. Their word doesn't even register with the hearer, and so the speaker receives zero credibility, not because a credibility judgement of zero is made but rather for the reason that no credibility judgement is made. I styled this as a total epistemic silencing: something is said, yet it is as if nothing were said. Or, again, imagine a situation where people like you are prejudicially overlooked as a relevant source of testimony or information about a given topic, so you don't even get the opportunity to contribute. I styled this latter kind of silencing as ‘pre-emptive’ Testimonial Injustice, because the prejudice is such that the opportunity for testimonial contribution is pre-empted altogether. Literally, nothing gets said. These are of course well-known phenomena, and they can be theorized in different ways. But I hope constructing them as cases of extreme Testimonial Injustice, so that we see them as forms of maximal prejudicial credibility deficit, was useful for some purposes.

There are other dimensions of internal heterogeneity besides the degree of credibility deficit, and the most important one is whether or not the prejudice depressing the person's credibility is local to a specific social context or whether it is what I call a ‘tracker’ prejudice, which follows the person through a range of areas of social activity (educational, political, sexual, professional …). Tracker prejudices are generally structural – the obvious cases are central kinds of identity prejudice such as those of gender, race, class, sexual orientation, religion, and so on – and so the Testimonial Injustices that are produced by tracker prejudices are of central interest to any account of how epistemic injustice is integral to wider structural social injustice and oppression. I called these ‘systematic’ testimonial injustices; and they contrast with cases of the localized kind, which are driven by non-tracking prejudices, and which I classed as ‘incidental’. It's worth remarking that an incidental Testimonial Injustice might possibly be disastrous for a particular individual's life, yet without intersecting with broader structures of social injustice. But if we are interested in painting a picture that reveals the ways that epistemic injustices are woven into the fabric of social injustices more generally, then it is the systematic kind of Testimonial Injustice in which we will be most interested.

The distinction between incidental and systematic forms, then, was for me highly important. They belong, however, to a single normative genus – the genus ‘Testimonial Injustice’ – in virtue of the fact that they each exemplify the core epistemic wrong that is the constant in both incidental and systematic forms: the prejudicial credibility deficit that expresses a wrongful underestimation of one's standing as a Giver of knowledge. I believe this focus on the core ethical-epistemic wrong has perhaps facilitated the application of Testimonial Injustice in areas of social-institutional life – such as healthcare, for example. Imagine a patient as such finding themselves on the receiving end of some negative assumptions about the credibility of their complaints of increased pain. Depending on the social identity of the patient, this might be an incidental Testimonial Injustice or a systematic one. If the prejudice attaches to the patient because of some aspect of their strictly individual eccentric personal style, then it would be an incidental epistemic injustice. But if it attaches to them as a member of a structurally disadvantaged social group – as when a black woman complains of chronic pain and finds her word is not taken seriously because of the intersectionally racialized and gendered way in which she is perceived – then this constitutes an epistemic injustice of the systematic kind. I hope this sort of application illustrates the usefulness of a conception of epistemic injustice that can span situations of both structural injustice and one-off moments of injustice, for the same intrinsic wrong is at work in both kinds of case. It is the same epistemic wrong, but occurring in importantly different ways and with very different resonances both in terms of the phenomenology and in terms of the causal connection with wider injustices. That was perhaps the main value of starting to talk in terms of something distinctively cast as epistemic injustice.

The second kind of epistemic injustice I offered an account of presents a different sort of wrongful epistemic disadvantage, as it relates not to a wrongful deficit of credibility but rather to a wrongful deficit of intelligibility – intelligibility to others, or sometimes even to ourselves. I called it Hermeneutical Injustice. Once again, it comes in many different shapes and sizes. A maximal case might be one in which a person has a social experience that is not fully intelligible even to them; and a minimal case might be one in which a person fully understands their own experience, and can readily communicate it to many others, confident in their capacity to understand it properly, and yet the particular parties to whom they need to make it intelligible – perhaps an employer, the police, a partner – do not share the concepts needed to fully grasp it. Of course there can be many cases between maximal and minimal too, as they are just poles of a continuum.

Hermeneutical injustices happen, on my account, when the explanation for why there are insufficient, or insufficiently shared, concepts for an experience to be intelligible is that the person in question is a member of a ‘hermeneutically marginalized’ group. That is, they are socially positioned such that they under-contribute to the stock of concepts and social meanings that are shared and usable by all. I called this universally shared stock of concepts and social meanings the ‘collective hermeneutical resource’, and it designates that subset of concepts and meanings that everyone can use with confidence that they will be understood by anyone else – it's the intersection of the many overlapping sets of hermeneutical resources operative in a given society at a time.

Manifestly, in most societies there are many different and overlapping sets of social meanings. Some social groups routinely use some concepts and meanings that only they fully understand, and this exclusivity means these concepts are not part of the collective hermeneutical resource. Mundanely, any expert group is like this, for instance; and, as others have rightly emphasized, so are some oppressed groups. That was indeed the point. The generation of concepts and meanings that swiftly become shared by all is a social practice in which unequal power means there can be over-contributors and under-contributors. If I belong to an under-contributing group then I am ‘hermeneutically marginalized’, and that places me at a wrongfully raised risk of having an experience I need to render intelligible to myself and/or to others, but cannot, either because I lack the concepts myself, or because even while I and my community may have long possessed the relevant concepts and used them to make sense of many experiences, still relevant others lack them, and may like it that way, and so I cannot render my experience fully intelligible across social space. Either sort of case exemplifies Hermeneutical Injustice.

As with Testimonial Injustices, in the case of Hermeneutical Injustice there can be both systematic and incidental cases. If the hermeneutical marginalization is localized to a particular context of social activity (perhaps, for instance, the experience of being a new father who finds himself attending something still called a ‘mother-and-babies’ group, which operates with a set of concepts and meanings presupposing all primary carers are women), then any hermeneutical injustice that he experiences in that local context need have no implications for his likely susceptibility to other kinds of injustice. Indeed he might be a man whose social positioning has him at the advantaged end of many power relations – class, race, gender-identity, and so on – so that this experience may be a first for him, a situation in which he learns what it can be like to find oneself largely outside the only operative set of meanings, without shared resources in the context to give proper expression to what he is experiencing. The point is that, for him, the Hermeneutical Injustice he may experience in that context will be something of a one-off – incidental.

By contrast, in a case where group members' hermeneutical marginalization is more widely spread across different areas of social activity – for instance, they under-contribute to shared concepts and meanings regarding gender identity, access to organized religion, to education, to healthcare, to a range of career paths – then any given hermeneutical injustice resonates across these different domains to render the injustice more systematic in nature. Imagine, for instance, a father who is a trans-man arriving with his baby at what he had reason to believe was an inclusive parents-and-babies group, only to find that the group running today is ‘mothers-and-babies’. In such a situation, any Hermeneutical Injustice he experiences at the group will resonate with hermeneutical marginalization of a non-local kind, insofar as his under-contribution to shared concepts and meanings is not local to activities of parenting but rather to many areas of social activity – thus it is systematic in kind. My aim in carving out this distinction between incidental and systematic Hermeneutical Injustice was, as ever, to identify the core epistemic wrong that is constant through the different sorts of case, while attending nonetheless to the differences, and thereby offering an explanation of what makes only the systematic cases resonate with other dimensions of social injustice.

It's exciting to see the categories of Testimonial Injustice and Hermeneutical Injustice being used by others in new and distinctive ways. Other philosophers and writers are exploring applications in healthcare, education, business, AI, aesthetics, recognition theory, and other areas. Speaking personally, an especially rewarding experience for me in this connection has been having the opportunity to collaborate with women victims and campaigners in Colombia regarding the systematic sexual violence that was meted out against women in the five decades of civil war. (I use the word ‘victim’ rather than ‘survivor’ because that is the preference of the women concerned – they are demanding to be recognized as victims of crimes.) If people who have suffered systematic criminal assault and brutality find it helpful to be able to protest not only against the original crimes but also against the secondary injustices that occur when their accusations are not properly heard, then the theory will have had some valuable use of a kind I would not have dared envisage when I wrote the book.

My purpose in writing it was not to point out the empirical fact that prejudice often entails that people are wrongfully dismissed or under-estimated – I make almost no direct claims in the book about which groups are or have been historically wrongfully epistemically underestimated, but rather I explore illustrations drawn from fiction and memoir in order to try to furnish the philosophical imagination with a sense of the multiplicity of possibilities in this regard. They indicate both the ordinariness of epistemic injustice, and its immense structural power. My purpose was to create a theory of two basic forms and to do so in the idioms of both moral philosophy and social epistemology (especially virtue epistemology), so that we had a way of theoretically characterizing what was distinctive about the wrong, situating it in relation to other kinds of wrong, and naming two of its broad forms so that they might be discerned in life and protested in their own right. It was in this sense a narrowly academic and analytical project, which tried to portray the border between the feminist philosophy that had inspired it, and the analytic epistemology that I was aiming to expand. The book also tried to limn a different border – one of philosophical style. I have always felt it must be possible to write philosophy in a way that is porous with other kinds of writing – novels, memoirs, screenplays. The text of my book was a patchwork in this regard, using long quotations simultaneously as illustration and source material. I don't know how far I succeeded in pushing any boundaries, but I think perhaps the extended use of these non-theoretical kinds of text did make the book more readable both to philosophers and to non-philosophers. I hope so anyway.

To be honest, I've never agreed with that famous line from Marx … For a start, he must have had a very a narrow conception of philosophy in mind, as there is a lot of valuable philosophy that has precious few practical implications for how human affairs should be arranged. And of the philosophy that could in principle be implemented in practice as part of a political system, well, it's an open question whether it should aim to change anything in the sense of being put into practice. It all depends. There is surely philosophy that would have a good effect in a given context at a given time, if implemented; and there is philosophy that would have a disastrous effect. In the case of Marxism, which probably has the greatest claim to being philosophy that brought changes to the world, it is a matter of opinion whether the historically epic, monolithic vision of permanent revolution that he bequeathed modern political thought (and without which our critical understanding of capitalism, exploitation, class, gender and race as we know it would not have got out of the starting blocks) should be thought of as flowering in something good or something frankly horrifying. The totalitarian implementations of communism that history has witnessed do not make a good advert for philosophy that changes the world. This suggests to me that the primary value of historical materialism is critical, explanatory, imaginative, corrective, argumentative …; but not implementational.

However, the more general drift of Marx's comment surely remains. ‘Changing the world’ need not be understood in terms of direct implementation, after all, but rather something more indirect, discursive and contested. The version that I would certainly endorse is the implied observation that culture, including political culture, can be more or less intellectually rich or bankrupt, so yes, let there be philosophy that can help stave off the bankruptcy. I believe philosophy has a role to play in substantiating humane values and continually reshaping our sense of what various kinds of justice demand of our institutions. Therefore, it is highly valuable that enough philosophy is done in a manner that might ultimately contribute to public debates. Not all philosophy need be like that, for sure; but it is socially valuable that, where appropriate, some of it can find a way to make that sort of contribution. Universities should be autonomous, but porous. I believe that is how philosophy, and the humanities in general, may continue to make an essential, though essentially indirect, contribution to the polity and to public life more generally.