The Faces of Injustice

In the AI Story

The book's central methodological move is the analysis of what Shklar calls "passive injustice" — the failure to act against injustice when acting would have been possible. The concept allows her to identify political harm produced not by active malice but by institutional failure, by the willingness of the powerful to accept suffering as natural when treating it as natural serves their interests. This analysis extends Shklar's earlier work on cruelty into the specific domain of classification: the question is not only what cruelty is and how it operates but who decides what counts as cruelty and what is permitted to escape the category.

The book's sustained engagement with victims' perspectives anticipates much subsequent work in political theory on standpoint, testimony, and recognition. Shklar insists that the victim's perception of injustice is not merely subjective but carries specific epistemic authority — the authority of someone who has been in a position to observe what the classification from above cannot see. When the victim says "this is injustice" and the classifier says "this is misfortune," the disagreement is not between equally valid perspectives. It is a political contest whose outcome will be determined by power rather than by truth, and the framework's task is to equip the victim's perception with the institutional standing required to prevail.

The book's continuing relevance to the AI transition is direct and specific. The reclassification of the displaced expert's suffering as "failure to adapt," the reclassification of task seepage's depletion as "insufficient self-care," the reclassification of structural unemployment as the "cost of progress" — each of these follows exactly the pattern Shklar's framework identifies. The framework does not dispute that adaptation is difficult, that self-care matters, or that progress has costs. It disputes the classification that transfers the obligation of response from the institutions that produced the condition to the individuals who bear it.

The book's final chapters develop specific institutional implications. Injustice correctly identified generates obligations of structural remedy that misfortune does not — obligations that fall on the political order rather than on the victim. The institutional structures required include mechanisms for contested classification (courts, legislatures, public forums in which the classification can be disputed), protections for the perspective of those most likely to bear misclassification's costs (standing for the victim in formal proceedings), and procedural safeguards against the consolidation of classification authority in the hands of those who benefit from it. These recommendations have aged well, and their application to AI governance is one of the specific contributions Shklar's framework can make to contemporary debates.

Origin

The Faces of Injustice was based on the Storrs Lectures Shklar delivered at Yale Law School in 1988 and published by Yale University Press in 1990. The book followed Ordinary Vices (1984) and preceded her posthumously published work on political obligation. It is widely regarded, alongside Ordinary Vices and the "Liberalism of Fear" essay, as one of the three definitive statements of Shklar's mature framework.

Key Ideas

Classification is a political act. The distinction between misfortune and injustice is drawn by those with the authority to classify, in the interest of those who benefit from the classification.

Passive injustice is injustice. Failing to act against injustice when acting was possible constitutes a political wrong distinct from the wrong of the active injustice itself.

Victims have epistemic authority. The perception of injustice from below carries specific authority that the view from above systematically conceals.

Reclassification dissolves obligation. When injustice is reclassified as misfortune, the obligation of structural remedy disappears, leaving only the cheap coin of sympathy.

Contestation is structural. The framework requires institutional mechanisms through which misclassification can be contested, not merely cultural dispositions to listen to victims.