Limitation periods are customarily justified by the repose of the defendant: at some point an individual should be able to stop retaining records and stop anticipating a claim. The justification is real and it is secondary. Treating it as primary generates a distorted account of how the rules actually work.
The primary function is evidential. A court determines what happened by hearing witnesses and reading documents, and both deteriorate -- witnesses forget, move, and become unavailable, and documents are destroyed on schedules that nobody established with litigation in contemplation. A claim brought twenty years late is not merely inconvenient to defend; it is one the court is poorly positioned to decide, and it will be decided anyway, which is the difficulty. The limitation rule eliminates from the system a class of disputes in which the machinery for finding facts can no longer be relied on.
Read that way, the exceptions stop appearing to be concessions to sympathy. Time does not run against a claimant who could not reasonably have known of the claim. The repose account struggles with this, for the defendant's position is identical whether the claimant knew or not. The evidential account addresses it directly: a rule aimed at preserving the quality of adjudication has no reason to bar a claim during a period in which nobody could have brought it, and every reason to start the clock at the point where proceedings become possible.
The same account explains why fraud or deliberate concealment postpones the period. On a repose theory this is a penalty, and penalties are incongruous things to incorporate into a rule about time. On an evidential theory it is not a penalty at all: a defendant who has concealed the material has himself degraded the evidence, and the reason for closing the door -- that the court cannot now find out -- is a reason he manufactured.
Nor is the distinction merely theoretical. Where a legislature extends a period, the repose account predicts objection from defendants and nothing else. The evidential account predicts a different and more awkward consequence: courts will be asked to adjudicate more claims on worse material, and the visible effect will be not injustice to defendants but a rise in the number of cases that turn on which party's account the judge finds more plausible, which is what adjudication looks like when the evidence has gone. That is testable, and where the extensions have been sufficiently long to study, it is what has been observed. The repose account is not thereby refuted. It is demonstrated to be a description of one effect of a rule adopted for a different reason, which is a common enough relationship between a doctrine and the explanation that grows up around it.
