Opening proposition
The idea in brief
Clock time places an event in the past as soon as it has occurred. Yet many events remain operationally unsettled: a diagnosis awaits confirmation, a transaction can be reversed, a verdict can be appealed and a memory can be reconsolidated.
The proposal defined pastness as a frontier in a dependency graph. A node becomes genuinely past only when no still-active process can invalidate it, demand that it be replayed or change the consequences derived from it.
Different aspects of the same event could cross this frontier at different times. A payment may be economically usable, legally contestable and cryptographically final on three distinct schedules.
Full exposition
Occurrence is not closure
Chronology answers whether an event happened before or after another event. It does not tell us whether the event is settled. A court judgment can be issued yet remain appealable. A transaction can be posted yet reversible. A memory can be old yet change when recalled. A scientific measurement can be published yet await a calibration result capable of invalidating it.
The Pastness Frontier separates occurrence from closure. An event occurs at a point in clock time, but it becomes past relative to a system only when no active dependency in that system can force the event to be revised, replayed or reinterpreted. Pastness is thus not only a date; it is a relation between an event and the processes that still have jurisdiction over it.
Events live inside dependency graphs
Imagine decisions and observations as nodes in a graph. An arrow from A to B means that B depends on A: if A changes, B may need to be recomputed. While some authorised process can still change A, every downstream node remains exposed to revision. The pastness frontier lies behind the oldest nodes that are still reachable by such revision.
This frontier moves as appeals expire, transactions finalise, experiments are replicated, contracts are relied upon and memories are integrated. It can also move backward. New evidence may reopen a diagnosis or historical judgment that appeared closed, restoring live dependencies to an event that had been treated as past.
There is more than one past
Because different systems maintain different dependencies, an event can be past in one sense and open in another. A bank transfer may be irreversible in the payment network but contestable in law. A relationship may be socially over but psychologically active. A software event may be immutable in the log while its effects remain recomputable.
This pluralism is not merely subjective. Each frontier can be specified by a set of authorised revision operations. Legal pastness depends on courts and limitation periods; technical pastness depends on rollback and replay mechanisms; personal pastness depends on the processes capable of changing current commitments and interpretations.
Living with open and closed events
The framework encourages institutions to state what remains open. Instead of one ambiguous status such as “complete,” a system could distinguish operational, evidential, legal and social finality. People could then see which consequences are safe to build upon and which may still be withdrawn.
At a personal level, the question “what can still force this to be reopened?” may be more useful than “why am I not over it yet?” Some experiences remain active because present life still depends on an unresolved interpretation, apology, diagnosis or decision. Closure does not require forgetting. It means the event no longer holds a live veto over the structures built after it.
Three senses of the past
An event can be past in at least three senses. It may have occurred earlier in clock time. It may no longer be actively experienced or remembered. And it may be closed to revision because later arrangements depend on treating it as settled. These senses often coincide, which makes them easy to confuse. A meal yesterday is temporally past, psychologically faint and practically closed. A disputed election, diagnosis or contract may be temporally past while remaining institutionally open for years.
The Pastness Frontier concerns the third sense. It asks where revision is still live. An event becomes practically past as the routes by which it could be reopened lose their ability to alter accepted downstream states. This is not a metaphysical claim that history itself changes. It is a claim about the organisation of consequences: whether present systems are still prepared to propagate a correction.
Events are bundles of claims
Ordinary language packages an event as one thing, but institutions often settle its components separately. After an accident, the physical occurrence is fixed immediately, the medical consequences may develop for months, legal responsibility may remain disputed, compensation may be reopened on appeal, and the event’s place in a relationship may never reach shared closure. There is no single date on which all of these become past.
The frontier is therefore granular. It runs through claims, records and dependencies rather than around whole episodes. One part of an event may be final for one purpose and open for another. A court may close liability while medicine revises causation; a family may accept what happened while an archive later corrects the public record. Mapping pastness means specifying which proposition, for which community and through which downstream dependencies.
Closure grows through reliance
When later decisions rely on a fact, revising that fact becomes more expensive. Payments are made, roles assigned, memories narrated, databases copied and further conclusions built. Each dependency creates work that a correction must undo or reinterpret. Enough accumulated reliance can make a proposition effectively final even when everyone agrees it may be false. Conversely, an old event can remain open if institutions preserve a cheap and authoritative correction path.
This explains why elapsed time is only an imperfect proxy for pastness. Time often allows dependencies to accumulate, witnesses to disappear and reversal costs to rise, but deliberate design can interrupt that process. Versioned records, appeal windows, provisional classifications and reversible allocations allow later evidence to travel. A system can keep an old decision young by keeping its consequences corrigible.
Power determines whose past can reopen
The ability to reopen an event is unevenly distributed. An institution may revise a benefit years later while the recipient cannot reopen the original denial. A platform can rewrite a record that users have already acted upon, while users cannot revise the platform’s classification. In such cases, pastness is asymmetric: one party enjoys finality against challenge but retains optionality for itself.
This asymmetry reveals that closure is not simply a neutral property of complex systems. It is allocated. Statutes of limitation, retention policies, audit rights, pardon powers and data-correction procedures decide whose evidence can cross the frontier. A fair system must balance the value of settlement against the risk that finality freezes an error for those least able to bear it.
Living with unresolved history
Not every dependency should remain reversible. A world in which every settled fact could be reopened without limit would make planning impossible. People need to rely on titles, judgments, promises and completed transitions. The idea does not oppose closure; it makes the cost and location of closure visible. It encourages explicit decisions about when the value of finality exceeds the value of corrigibility.
At a personal level, the frontier can clarify why some experiences feel present despite their age. The issue may not be rumination alone. A live dependency—a pending apology, an unresolved obligation, an uncertain diagnosis, an identity organised around the event—continues to make present choices conditional on how the event is interpreted. Closure may require changing that dependency rather than trying to make the memory feel older.
A thought experiment
A scientific result is published on Monday. Its timestamp is fixed, but downstream conclusions remain dependent on a calibration check. If the instrument fails, analyses must be reopened. On this view, the result has occurred but has not fully become past.
Consequences
What the idea changes
Time would become partly architectural. Instead of asking how long ago something happened, one would ask which live systems can still reach backward and change what must now be done.
Ways to think with it
- It explained why some old experiences feel unfinished while recent events can feel conclusively over.
- It supported multiple pastness frontiers—legal, technical, biological and personal.
- It made finality a visible design choice rather than a vague feeling.
Further reading
The intellectual neighbourhood
Notes on the idea’s provenance and editorial review are kept separately in the editorial appendix.