Glossary

The Constitutive Obligation Principle

Also called: COP

Definition

Where an institution produces its subjects’ operative standing as entitlement-bearers, its legitimacy depends on discharging non-waivable obligations of legibility, non-abandonment and remediation, not on their authorisation.

The Constitutive Obligation Principle (COP) holds that when an institution constitutes its subjects as operative entitlement-bearers, its legitimacy depends not on their authorisation but on discharging non-waivable obligations that keep them recognisable as such: obligations of legibility, non-abandonment and remediation. It is the positive proposal of Beyond Consent, and it is read off the structure of constitutive dependence.

Each obligation answers to one feature of that dependence. Because the institution produces recognition, it owes legibility: criteria of recognition responsive to the material reality of the people it must reach. If a readable fingerprint is the criterion and it systematically fails those whose labour has worn their fingers, the institution must redesign it. Because recognition is revisable from inside, it owes non-abandonment: it may not withdraw what it produced, through database errors, technical glitches or policy changes, without meeting justificatory burdens. And because the subject has no standpoint of refusal outside the institution, it owes remediation: a substantive pathway of appeal inside it, in languages, registers and places accessible to those most likely to be failed. An online grievance portal offered to someone whose fingerprint did not read redoubles the failure.

Three features matter. The obligations are non-waivable, because a subject whose standing the institution produces cannot coherently release it from the conditions of that standing. They are borne by the institution, not correlative to rights she could waive by consenting. And they are a necessary floor, not a complete theory of legitimacy: an institution can discharge them and remain unjust in ways the principle does not address.

The principle takes from O’Neill the point that obligations, unlike rights, say who must do what, and it fills a gap in recognition theory. Recognition theory tells us the labourer is wronged when the system fails to read her; the COP says what the system owed her for having made her readable. It is not paternalistic. It applies only where authorisation is not available in the first place, and it requires the voice of those governed in deciding which criteria and which appeals are appropriate.

My essays and papers on it

Questions that turn on it

In the library

Further reading

  1. Saxena, A. (2026). Beyond consent: Algorithmic welfare, constitutive dependence, and the limits of liberal AI ethics. AI and Ethics 6: 414.doi.org/10.1007/s43681-026-01270-5
  2. O’Neill, O. (2002). Autonomy and Trust in Bioethics. Cambridge: Cambridge University Press.doi.org/10.1017/CBO9780511606250
  3. Fraser, N. and Honneth, A. (2003). Redistribution or Recognition? A Political-Philosophical Exchange. London: Verso.