Reconstructing Testimonial Knowledge: A Coherentist Critique of Epistemic Injustice — Epoche C1
The exchange at issue The example Miranda Fricker uses to fix her central concept is the jury in Harper Lee's To Kill a Mockingbird (1960), which discounts the evidence of Tom Robinson because he is a black man in Alabama in the 1930s, not because of anything in what he says. In Epistemic Injustice: Power and the Ethics of Knowing (2007) Fricker names this testimonial injustice : a hearer, in the grip of a prejudice attached to the speaker's social identity, assigns the speaker less credibility than the evidence warrants. She distinguishes it sharply from a second wrong she calls hermeneutical injustice , in which a person's experience cannot be made intelligible even to themselves because the shared stock of concepts has no name for it — her illustration is the situation of women subjected to conduct at work before the term 'sexual harassment' existed to describe it. That distinction has to be drawn at the outset, because the earlier version of this essay ran the two together. It described the historical devaluation of the testimony of indigenous communities in the Philippines and of ethnic minorities in Myanmar, and the dissonance produced when such accounts are finally admitted, as instances of testimonial injustice. The credibility deficit in a particular hearing is testimonial; the absence from the collective record of the categories in which such an experience could be stated is hermeneutical. Both are at work in the cases discussed below, and the argument depends on keeping them apart. A second correction concerns the target. The view opposed here was described as the received reading of Fricker, on which testimonial injustice is essentially a failure of reliability — prejudice makes our information channels less truth-conducive. That is a reading of the epistemic cost that some reliabilists have offered; it is not Fricker's own position, which locates the primary harm in the wrong done to the speaker in her capacity as a knower and treats the remedy as a virtue of the hearer. The claim of this essay is therefore narrower and more defensible than before: the reliability diagnosis, whoever holds it, understates the damage, and a coherentist account of justification shows why. Why testimony cannot be checked item by item The first step is to establish something the earlier version asserted without argument: that testimonial knowledge is constituted communally rather than assembled by individual verification. C. A. J. Coady's Testimony: A Philosophical Study (1992) makes the case against the reductionist programme, which holds that trust in testimony is justified by each individual's inductive record of finding reports true. Coady's objection is that the record is far too thin to carry the weight. The overwhelming majority of what any person believes — the date of a battle, the existence of Antarctica, the safety of a medicine, the name of one's own mother — has never been checked against anything except further testimony, and the small sample anyone has personally verified is not remotely representative of the vast and varied population of reports being underwritten. The inductive base cannot support the conclusion. If reduction fails, then trust in testimony is not a conclusion drawn from private evidence but a condition of having a body of belief at all. That is the sense in which testimonial knowledge is communal. It also means that the exclusion of a class of speakers does not simply remove some items from a stock of information; it alters what the remaining items are supported by. Coherentism, and the objection that decides the case Coherentism enters as an answer to a problem any undergraduate meets: the regress of justification. If a belief is justified by another belief, and that by another, the chain must terminate somewhere, circle back, or run on forever. Foundationalism takes the first option, positing basic beliefs that are justified without resting on others. Coherentism takes the second and denies that the circularity is vicious: on Laurence BonJour's account in The Structure of Empirical Knowledge (1985), justification is a property of a system of beliefs as a whole, conferred on individual beliefs by their membership in a set that is consistent, inferentially connected, and free of unexplained anomalies. No belief is justified alone. The standard objection to coherentism is the one that matters here, and it is not a weakness of the position so much as the point of applying it. A system of beliefs can be perfectly coherent and yet have no contact with the world; a sufficiently careful fabrication hangs together as well as the truth does. BonJour, who was alive to this, added what he called an observation requirement: a coherent system must, over time, contain beliefs that its own holders regard as arising from observation rather than from inference, and must accommodate them. Coherence without input is not justification. Without such a constraint the theory permits what may be called sealed coherence — internal harmony purchased by the absence of anything that could disturb it. What the injustice actually damages Put the two results together and the diagnosis follows. If testimony is not reducible to individual checking, then in the social case testimony is the input channel: a community's beliefs about its own past and present are answerable to what its members report, and to nothing else that is available in the relevant sense. Systematic testimonial injustice therefore does not merely add noise to a channel. It closes the channel selectively, and it closes it precisely where the reports would be least congruent with what is already believed — since the prejudice that discounts the speaker is normally part of the same structure of belief the speaker's report would embarrass. The result is a body of collective belief that satisfies coherentism's positive condition and violates its constraint. It is consistent; its parts support one another; the anomalies have been r