Suppose someone opens an old website preserved in an archive many years from now. The files are intact. There are papers, dates, version numbers, and an essay about how ideas survive. The reader finds several pages interesting and others overstated, records the address, and closes the window. Something has unquestionably survived. But what?

There has been a reading, and the documents remain available. Neither fact establishes a continuing intellectual practice. A historian can reconstruct what a theory once attempted without taking up its commitments. A curious stranger can understand an argument, identify a mistake, and decline to become its successor. Understanding need not culminate in allegiance.

That is not a failure of generosity. Readers do not owe a theory a future. An account of enduring traditions becomes more interesting when it admits that its own tradition might never emerge. The Long-τ Consensus paper takes this step by including SAE itself among the candidates whose continuation remains uncertain.

Preserving a document is not preserving a practice

An archive matters. Identifiable versions and reliable copies allow future readers to establish what an author actually said. Without them, a later success story can quietly replace an earlier prediction. A dated record gives people something to compare with events, including events its writer did not anticipate or would have preferred not to see.

But access is an opening, not an outcome. Downloads need not become readings. Readings need not become understanding, and an accurate paraphrase need not demonstrate an ability to judge an unfamiliar case. A phrase may travel much farther than the work required to use it. That is a reason to distinguish the achievements, not a reason to dismiss publication or readers who do not continue further.

The paper asks whether someone could encounter the written work without direct contact with its author, develop an independent understanding, and help another person do likewise. This comes closer to a living practice than a citation count. Yet it still requires evidence. Departing from the author's conclusion is not sufficient proof of misunderstanding. Reproducing the author's conclusion is not sufficient proof of understanding either.

The criteria must also be available for questioning. If a founder can always decide afterward that a dissenter never really understood, independence becomes impossible to demonstrate. Every unwelcome example can be excluded. A theory would then be protecting its membership boundary rather than testing a claim about transmission.

An external reader must therefore be able to examine the record without first becoming a participant. Even in a future where no one continues SAE as a practice, someone could investigate what its papers predicted. The population of people able to inspect an archive and the population sustaining its ideas need not coincide. The disappearance of the latter does not make the former logically impossible.

A clue is not the whole thing we want to know

The difficulty becomes sharper when we ask whether conditions for sustained understanding are improving or deteriorating. Book production, records of reading, educational opportunities, and histories of collaboration can each disclose something. They do not become measurements of one comprehensive human capacity merely because all can be expressed as numbers.

More books may mean easier access without establishing that more people have time to read them closely. A reading assessment measures performance on specified tasks; it cannot silently supply scores for trust, commitment, and the ability to live with uncertainty. A school closing is an observable event. Its effects on learning across generations need further investigation. The people who taught there do not lose their knowledge when the door is locked.

The source calls these kinds of evidence proxies: ways of approaching particular aspects of something not directly observed in full. Their partial character matters. Less reading in one dataset cannot justify a verdict that an age has produced worse people. Nor should the features easiest to count within one form of life become a scale for ranking cultures.

Historical interpretations need similar restraint. A tradition's expansion may involve changes in communication, livelihood, teaching, political support, and much else. A useful structural lens can suggest where to look without defeating every alternative explanation. Comparing explanations requires asking what each would lead us to expect and then examining records that might distinguish them.

A precise-looking number can obscure this unfinished work. The paper explores a roughly five-to-one relationship between an extended early period and a shorter later transition. That proposal is stronger than the ordinary observation that initial accumulation can be slow. Transferring a geometric relation within a model to elapsed historical time requires comparable cases and independent tests. It does not authorize entering a group's founding year into a formula to calculate when maturity will arrive.

The paper also assigns subjective percentages to several possible futures for SAE. These record an internal assessment, not measured frequencies from a statistical forecast. They can help preserve what the author expected, provided their status remains explicit. Printing a percentage does not supply the missing observations, and readers need not use it to decide whether the next essay is worth their attention.

The distinction is not a ban on mathematics or ambitious interpretation. It keeps different accomplishments visible. A conceptual account can suggest a mechanism. A simulation can explore behavior under chosen assumptions. A historical study can examine records. None needs to pretend to have completed the others in order to be useful.

A mistake needs somewhere to appear

An instructive moment occurs when the paper reports simulations that did not discriminate as hoped. Group size and network arrangements were varied, but a selected observable did not reliably distinguish some of the changes under the tested conditions. A tempting response is to interpret this immediately as proof that a deeper theory was right all along.

The more defensible first statement is narrower: in this model, with these parameters and this observable, the expected distinction was not detected. Perhaps the observable misses something relevant. Perhaps the model's assumptions limit what it can show. The result can motivate looking for another factor without proving, on its own, that a particular additional factor determines real human communities. The detailed setup, code, and data still need to be available for others to check.

Willingness to revise is not itself a test either. We need to say what would count as agreement or disagreement, how long observation must continue, and which records can reasonably be compared. People outside the framework must be able to dispute the judgment. Some claims cannot yet be assessed honestly. Giving each one a date and a decimal would not cure that limitation.

The paper addresses readers in 2046, 2076, and 2126. These are invitations to look back, not dates announcing inevitable vindication. They preserve occasions for asking which expectations met events. An important distinction remains: SAE's failure to continue as a particular project would not, by itself, refute every structural claim about transmission. Its continuation would not establish the whole model either. A meaningful assessment identifies the particular claim and the observation that bears on it.

This is less spectacular than a theory that wins under every outcome. Popularity cannot always count as confirmation while obscurity means only that the time is not ripe. Opposition cannot automatically prove a necessary stage; disappearance cannot become a subtler form of success. If no result could ever count against the account, interpretive ingenuity has replaced exposure to evidence.

There is an ethical consequence too. Endurance is not a gift readers owe an author. Someone can understand the work and choose a different direction. That encounter remains real. A theory concerned with people as ends cannot quietly treat every reader as raw material for its own continuation. It has to allow understanding that does not become succession.

The future visitor opens the archive again. Perhaps a passage helps with a question nobody now knows how to ask. Perhaps another passage receives a note identifying a confident mistake. We cannot write the visitor's conclusion in advance. We can make the argument legible, preserve its grounds, and show where certainty stops. Whether to carry the work further belongs to the people who come afterward.