Ethics without will

(English). Western ethics is built on the will: intention, maxim, consent, culpability. This is why it structurally fails before machines – systems that determine without willing. A classifier, a diagnostic system, a language model performs thousands of determinations per second; each is a cut that produces a remainder – that which the determination could not carry – and none has a will behind it. Classical ethics reacts with two equally false exits: it either ascribes intentions to the machine or denies it any ethical relevance because it is not a subject. This note proposes an ethics without will: the first-order question of machine responsibility is not what a system intended, nor only whether its cuts are fairly distributed, but whether those it affects can return. Returnability replaces culpability as the basic concept; it generalizes no-fault liability from tort law into a philosophical criterion applicable to everything that cuts, human or machine. Three minimal indicators make returnability buildable, testable and contestable: contestability, revisability, and the declared remainder – an architecture that outputs not only what it decided but what it did not take into account (the remainder field). The remainder grounds no moral guilt of the machine, but a form debt of the system that produces and deploys the cut: ethics becomes an architecture of return, not an ethics of the will. The note marks its difference to fairness-based approaches – even the fairest cut leaves a remainder – and warns against exoneration by correctness: the most dangerous machine is the demonstrably good one whose quality becomes the argument for closing its door. A falsification clause states the conditions under which the proposal fails.