Opening proposition
The idea in brief
Advice about difficult choices often distinguishes one-way doors from two-way doors. That distinction treats reversibility as a property of the external situation: can the decision, in principle, be undone?
Owned reversibility proposed a stricter test. A choice is reversible only if the choosing person will retain the information, resources, authority, health and motivation needed to activate the return path. A nominal exit controlled by an employer, doctor, platform or radically changed future self would not qualify.
The proposal joined feasibility and identity. It asked not only whether a path exists in the future state space, but whether control of that path remains continuously attached to the agent whose freedom it is supposed to protect.
Full exposition
Reversibility is usually defined from outside
A choice is commonly called reversible when the world contains a route back. A subscription can be cancelled, a job can be left, a procedure can be reversed, or data can be exported. This description takes the perspective of an outside observer who can see all technically possible transitions between states.
Owned Reversibility changes the unit of analysis from the state transition to the agent-path pair. The relevant question is not simply whether state B can return to state A. It is whether the person who entered B will still possess a usable, authorised and recognisable path back. An exit that exists only for an administrator, a wealthy customer, a medically different patient or an institution is not the chooser’s reversible option.
The components of an owned exit
An exit can fail along several dimensions. The agent may lose legal authority, practical skill, money, time, information, bodily capacity or access to the necessary infrastructure. The path may also become cognitively invisible: a future self no longer remembers why the option mattered or cannot recognise the earlier preference as its own.
Ownership therefore has to persist, not merely exist at the moment of choice. A genuinely reversible design maintains the instructions, credentials, resources and decision rights required for reversal. It also makes the path legible at the moment it may be needed. A buried export function or nominal appeal process is weaker than an exit whose cost, duration and controller are known in advance.
The future self problem
Some decisions alter the very person who would later evaluate them. A career changes skills and social ties; a medication can change mood; a platform shapes habits; a relationship changes commitments. It is too simple to say that the future self can reverse the decision if that self has been produced by the decision and no longer has access to the earlier standpoint.
Owned Reversibility does not require preferences to remain frozen. It requires continuity of jurisdiction: the evolving person must inherit meaningful control over the escape route. That might involve advance directives, cooling-off periods, periodic re-consent, trusted delegates or protected resources that cannot be consumed by the initial commitment.
Designing for maintained agency
The framework turns reversibility from a label into an obligation. A service claiming portability should test whether ordinary users can complete an export years later. A medical intervention described as reversible should specify who can request reversal, what changes may have become irreversible, and whether the patient will retain the capacity to choose. An organisation offering an appeal should preserve evidence and deadlines needed to use it.
The practical test is compact: who will control the undo, under what future conditions, and with whose resources? If the answer shifts away from the chooser as the commitment deepens, the option is becoming reversible in name but not in ownership. Freedom then depends less on the number of doors than on whether the keys travel with the person.
Reversible for whom?
A process is often called reversible when some route exists from the later state back to the earlier one. That definition is adequate for chemistry or an abstract state machine, but it can be misleading for human decisions. The return route may be available only to an employer, a platform, a wealthy buyer, a clinician or a future institution. If the person who bears the consequences cannot initiate and complete the return, the option is reversible in a diagram but not in their life.
Owned Reversibility adds an agent index to the concept. Instead of asking whether state B can reach state A, it asks whether this person, after arriving at B, retains the authority, knowledge, resources, health, time and recognition needed to produce that return. The same transition can therefore be reversible for one participant and irreversible for another. Reversibility becomes a relation among a state, a route and an agent rather than a property of the state alone.
A path must remain executable
Merely knowing that an exit exists is insufficient. A usable reversal has several components: the person must be able to recognize that reversal is warranted, decide to pursue it, invoke the relevant procedure, withstand the transition costs and remain eligible for the destination. A missing component breaks the path. A refund policy is not owned if only a hidden support channel can activate it; a medical implant is not meaningfully removable if removal is unaffordable or unavailable where the patient lives.
The components also interact over time. The original choice may deplete the money needed to undo it, weaken the skills needed to return, transfer legal authority to someone else or make the old social role unwilling to receive the person back. A decision can begin with an owned exit and gradually lose it. Reversibility should therefore be monitored as a maintained capability, not certified once at the moment of consent.
The identity condition
The hardest case is a decision that changes the decision-maker. Education, parenthood, medication, ideology, institutionalisation and long immersion in a platform can reshape preferences and capacities. A later person may be psychologically continuous with the earlier person yet no longer value the promised exit. Does that show successful adaptation, or that the choice disabled its own future reversal?
Owned Reversibility cannot solve every philosophical problem about personal identity, but it exposes what must be specified. One may require that the future person retains the ability to understand the former standpoint, that reversible trials occur before preference-changing effects become entrenched, or that an independent advocate preserve the earlier person’s claim. The aim is not to freeze identity. It is to prevent “your future self will agree” from serving as automatic evidence that the initial option remained free.
Exit, voice and dependence
The concept sits between two familiar protections. Exit allows a person to leave an arrangement; voice allows them to change it from within. An owned reversal may require both. A worker can formally resign yet lack a viable return to the profession because their credentials have lapsed. A platform user can export data yet be unable to reconstruct social relationships elsewhere. In each case, the nominal exit exists while the substantive restoration path has eroded.
Dependence is not itself objectionable. People routinely enter relationships and institutions that transform what is possible, and many valuable commitments would be shallow if they carried no cost. The concept instead distinguishes chosen commitment from one-sided option destruction. A commitment is better owned when the person understands which exits will decay, has influence over that decay and retains protected routes for reconsideration during predictable moments of vulnerability.
Designing a reversibility budget
Institutions could treat reversal capacity as a budget that is spent over time. Long notice periods, portable records, retraining rights, cooling-off windows, escrowed funds, staged consent and interoperable formats all preserve different parts of the path. The relevant question is not whether every choice can be undone at zero cost. It is which capabilities must remain with the chooser for the option to count as theirs.
This framing changes evaluation. A convenient service may look less attractive if its convenience is financed by the slow loss of exportability. A risky treatment may become more acceptable if follow-up, removal and restoration are guaranteed independently of the original provider. A career experiment can be made more genuinely reversible by maintaining credentials and relationships rather than merely telling oneself that returning is possible. Reversibility becomes something designed and provisioned, not a reassuring adjective.
A thought experiment
A cloud service advertises that customers can always export their data. After years of use, export requires expensive specialists, the original file formats are obsolete and only an administrator can authorize the process. The migration is technically reversible, but the customer no longer owns the reversal.
Consequences
What the idea changes
Freedom would be measured by maintained control over escape routes, not by the mere presence of alternatives on a diagram. This changes how one evaluates consent, medical interventions, careers, subscriptions and delegated authority.
Ways to think with it
- It exposed the difference between an exit existing and an exit remaining usable.
- It turned reversibility into a property that must be maintained over time.
- It would force designers to assign ownership, cost and expiry to every promised undo mechanism.
Further reading
The intellectual neighbourhood
Notes on the idea’s provenance and editorial review are kept separately in the editorial appendix.