LITERALLY LEGAL: The Installed Judge, the Ideological Activist and the Perfect Forgery of Justice — Why a Productivity Score Cannot Find Them, Which Sanctions Can, and Why No Court May Judge Its Own Impediments. What Aristotle, Radbruch, Merton, Cover, Hart, Fuller, Schauer and Kahan Already Hold, and What Remains

A judgment can be correct to the letter of a rule and unjust to the person before the court, in the same way that a sentence can be literally true and still deceive its hearer. This article separates three judicial failures and one success by two observations: whether a judge’s decisions vary with something that stands in for the reasons of the case — patrons, press, a faction, a score — and whether they follow the reasons the legal order makes decisive. The performer varies with an external substitute. The installed formalist holds the letter of an isolated rule whatever the case’s reasons. The installed activist holds an agenda whatever the case’s reasons, and dresses it in the language of principle. The sourced judge follows the reasons of the case, which sometimes require applying a harsh rule and sometimes require finding the exit the legal order provides. Seven results follow. The literally legal injustice is usually not legal at all but legal relative to one rule read outside the order that contains it; what is invariant is that the order contains an exit the judge did not use; its local form varies — in civil-law orders such as Brazil’s a tenant in a coma is not even in default, and in common-law orders the exit is procedural or equitable (T1). Literal legality stands to justice as literal truth stands to honesty (T2). Departure from the letter is not the mark of the just judge (T3). The ideological activist is the forgery of the sourced judge, as the formalist is the forgery of the faithful one, and a valence-swap test separates him from the judge he imitates (T4). Both are perfect forgeries of justice, and their harm grows with the trust their form earns (T5). A court governed by counts, speed and deviation alerts selects for installation (T6). And — the result that governs the article’s sanctions — any regime that punishes judges for the content of their interpretations converts activists into installed formalists rather than into just judges, and hands the punishing body a new channel of political control (T7). A second, normative part therefore proposes strong sanctions, including nullity with restoration, personal liability, loss of office and permanent disqualification, against every judicial failure that can be proven on the record — corruption, partisan activity, undisclosed interest, failure to answer a defeating argument, unexplained departure from one’s own precedent, unjustified inaction — and equally strong sanctions against anyone who disciplines a judge for the content of a decision. The same rule governs the hardest case: a supreme court whose members decide the rules of their own impediment, decide alone whether one of them is impeded, and can be investigated only at the initiative of a prosecutor who argues before them. In such a court a conflict of interest forbidden elsewhere can be made lawful by those it concerns, and a forgery of impartiality then passes the only inspection legality can run, because the inspected have rewritten it (T8). Brazil’s Federal Supreme Court supplies the instance: in 2023 it struck down, by six votes to four, the statutory rule that barred a judge from cases in which a party is a client of a relative’s law firm, and contracts and contacts between members’ families or members themselves and parties with business before the court have since become the subject of public allegations, which remain unproven. The article proposes a constitutional impediment rule that the court cannot strike down, disclosure placed on the firms and parties who know, the exclusion of firms engaged to manufacture an impediment, a ban on non-litigation contracts and on hospitality from interested parties, and the rule that no member of a court decides his own impediment or his colleagues’. Priority is conceded to Aristotle, Cicero, Radbruch, Pound, Merton, Cover, Hart, Fuller, Dworkin, Schauer, Wechsler and MacCormick (whose demands for neutral and universalisable rulings anticipate the valence sw