Truth, Records, and Knowledge

From an occurrence to a record

An occurrence is something that concretely takes place. A foot presses into sand, a latch closes, or an instrument responds to a signal. Its happening does not depend on someone observing it or writing it down.

An interaction may leave a mark: a retained difference in what it encounters. The sand is displaced; the latch punches a notch into a strip; the instrument changes a stored value. If that difference becomes sufficiently stable and preservable to remain available in principle for later comparison, it counts as a record.

Recordhood therefore depends on what an arrangement can retain and make available. A brief electrical state may last long enough for another circuit to compare it, while an archive needs a much longer period of preservation. A record need not last forever or be intended for human use. It also need not carry a date: a footprint can preserve something of a passage before anyone measures when it occurred.

The distinction concerns roles within a history. Producing a record is itself an occurrence, and a record is a real arrangement that can undergo further change. The notch is something the latch closure produces; it is not the closure continuing to happen inside the strip.

Nor must every interaction complete this path. A difference may disappear before it becomes a stable record. A preserved mark may never be found. A readable inscription may survive while its connection to a particular source becomes uncertain.

A record can be real and wrong

Imagine a test that produces a slip marked PASS when a sample satisfies a specified threshold. Beside the original slip are a faithful photocopy and a training facsimile made to look identical. The facsimile was prepared separately and has no source connection to the test.

All three sheets are real. Each can be read, copied, filed, and acted upon. Yet their appearance alone cannot establish which test, if any, they record. The photocopy inherits its content from the original; it does not repeat the measurement. The facsimile has a production history of its own, but that history does not connect it to the trial.

Four questions help separate what is at stake.

Suppose the sample failed, but the facsimile is mistakenly accepted as its result. The sheet may trigger release of the sample and coordinate several people’s actions. Changism calls a stabilized record integrated into such a procedure an operational determination. It can guide action successfully while conveying a false claim about its alleged source.

Here, the sheet’s existence and its effects are facts. The claim that the sample passed is false. For Changism, a fact is an obtaining feature of reality: something that is actually the case. Truth concerns whether what a claim asserts obtains. Institutional acceptance does not make every accepted assertion true.

How records become evidence

A record becomes evidence for a claim through assessment of its relevance, fidelity, and inferential weight. Evidence is always evidence for something. The same sheet may strongly support a claim about what a release procedure received while offering little support for a claim about the sample’s actual performance.

One part of that assessment concerns provenance: where the record came from and how it reached its present form. For the photocopy, this means tracing it to the original slip and the original to the trial. A label or an official-looking format can help identify a record, but either can also be mistaken. We need enough of the source history to support the connection being claimed.

Another part concerns fidelity: how well the record tracks the relevant feature under the stated conditions. A genuine result can preserve an instrument error. A faithful copy can reproduce that error perfectly. Authenticity and careful copying matter, but they cannot repair a faulty measurement at the beginning of the route.

The interpretation also needs checking. What did PASS mean in this procedure? Which threshold and tolerance applied? Does the surviving record distinguish this trial from another? A record can be accurate in one respect and insufficient for a more demanding question.

Finally, assessment asks what else could account for the record and what might expose an error. Calibration, comparison, and investigation of possible substitutions can address different weaknesses. The appropriate checks depend on the claim. Establishing that a meeting was scheduled and establishing that everyone attended it require different evidence, even when both inquiries begin with the same calendar entry.

This does not make every ordinary judgement a laboratory project. It makes the standard answerable to the question. The more a conclusion depends on a particular detail, the more care that detail requires.

Why agreement can preserve a shared error

Ten photocopies of one slip provide ten places from which to retrieve its content. They provide one underlying test result. Their agreement can demonstrate successful copying without supplying independent confirmation of the original claim.

The difficulty can also affect repeated measurements. Suppose several instruments use the same faulty calibration reference. Their readings may agree because the same error enters each route. Repetition can test some sources of variation while leaving that shared dependency untouched.

Useful independence is therefore specific to a possible failure. A comparison gains force when it can reveal an error that the first method could miss. Two inquiries need not be completely isolated from one another; they need sufficiently different dependencies for the question being assessed.

This is also why invoking an unspecified possibility of bias is inadequate. A serious challenge identifies a plausible source of error and explains how it could affect the result. Inquiry advances by testing those possibilities, rather than treating either agreement or suspicion as decisive by itself.

When every usable trace disappears

Return to the latch and the wax strip. In a numbered trial, the latch closes once and punches a notch. Before anyone retrieves the strip, a cleaning cycle melts it and resets the apparatus.

Now stipulate the limit of the example. Within the declared apparatus, scale, tolerances, and available methods, no copy, witness, log, or other accessible evidence distinguishes closure from non-closure. This is a thought experiment about lost access. It does not claim that melting wax annihilates every microscopic consequence throughout the universe. Washing away a footprint does not establish that stronger claim either.

Keep the historical statement fixed: “The latch closed once in that trial.” If the closure completed as specified, Changism holds that the statement remains true after the record disappears. The melting is a further occurrence. It changes what remains available; it does not change the completed trial into one in which the latch stayed open.

This is called historical non-retractability. Once the specified occurrence has completed, later transformations do not turn its having happened into its never having happened. Reversing an effect is also different from preventing its cause: resetting the latch can restore its position without removing the earlier closure from its history.

The assessor’s position is quite different. Under the stipulated conditions, no available evidence favours closure over non-closure. A guess may happen to be correct, but correctness alone would not make it knowledge. Withholding judgement is the responsible response.

The distinction permits a history to exceed what inquiry can recover. It gives no evidential advantage to a particular unsupported story about that history.

What makes an unrecorded historical claim true

The disappearing record raises a philosophical question. If the occurrence no longer continues and nothing preserves it for inspection, what makes the claim true?

Changism’s answer is that the completed occurrence satisfies the claim’s content. The specified latch closure took place, so the statement that it took place is true. This is what historical truth-grounding means here. It concerns the relation between what a statement asserts and what happened, rather than a causal signal transmitted to a later reader.

The proposal does not require the occurrence to remain as an enduring object, a hidden inscription, or part of a permanently stored past. It rejects the further requirement that something available where a claim is assessed must stand in for every occurrence the claim concerns.

That is a philosophical commitment with work still to do. Changist research has developed a candidate formal account for specified cases, under explicit assumptions about the occurrence, the fixed claim, and the continuations considered. A general account of historical truth remains unfinished. Difficulties include vague event boundaries, reference after source connections disappear, and claims that something never occurred.

The basic distinction between truth and surviving evidence is shared by other philosophies. Changism’s proposed account of their relation must be compared with those alternatives. The erased record alone does not decide between competing theories of time or establish that the Changist explanation is the best one.

When missing evidence matters

The possibility of an unrecorded occurrence does not make every absence of evidence irrelevant. What matters is whether evidence should have been available if the claim were true.

Suppose a functioning counter should register every latch closure within a defined trial, and its output is preserved and checked. Finding no registration may count against the claim that a closure occurred. That inference depends on the counter’s coverage, sensitivity, and reliability, as well as the preservation of its output. If the counter was disconnected or the log destroyed, the same absence has a different evidential meaning.

The question is therefore how well the available inquiry distinguishes the proposed histories. “No record was found” and “the event did not occur” are different claims. Moving from the first to the second needs an account of what would have produced, preserved, and revealed the expected record.

We should also distinguish a missing answer from an unclear question. We may understand exactly what “the latch closed once in trial T” asserts while lacking evidence to decide it. A claim that “the organization became successful” may instead leave the meaning of success, the relevant period, or the threshold unspecified. More records will not by themselves settle a question whose content remains unclear.

Knowledge as organized correction

Most inquiries have more to work with than a wholly erased trace. Records survive in different forms, methods can be checked, and new comparisons become possible. Knowledge grows through the organization of those opportunities.

Objectivity, on this account, involves making claims answerable to what they concern. A record’s source can be investigated. A measurement can be calibrated. An alternative explanation can be tested. When a mistake becomes visible, the interpretation or method can change without making the earlier occurrence depend on that correction.

Communities and institutions help by preserving records, sharing standards, and enabling others to examine the same claims. They can also preserve a common mistake. Their contribution depends on how well their practices expose relevant errors, including errors that an accepted explanation makes easy to overlook.

Responsible confidence can be strong while remaining open to revision. Discovering a fault may defeat a previously warranted judgement without showing that the original inquiry was arbitrary. Conversely, practical success does not establish every assumption behind it. A procedure can work under familiar conditions while failing outside the range that has been tested.

Changism therefore joins realism about what happens with discipline about what we may claim to know. The useful questions are concrete: which claim does this record support, how did it reach us, what could make the inference fail, and which further comparison could reveal that failure?

Read further

Begin with Changism, Volume I, Chapter 17, “What Happened and How We Know”. It develops these distinctions through records, copies, shared error, and lost traces. Chapter 3, “Before and After Without Time”, introduces the difference between productive history, retained records, and historical truth. Chapter 13, “Where Things Meet”, explains interaction and registration.

For the philosophical arguments, visit Research. “What Happened Without a Surviving Trace?” develops the conceptual account of historical truth. “Truth Without a Surviving Ground” offers a conditional formal development. “Dependence Across Registers” examines why producing an effect, satisfying a claim, and warranting belief require different accounts.

Continue exploring: Causation and Dependence follows how differences are produced. Time and Change distinguishes occurrence order from its measurement. Physics and Cosmology examines how physical proposals become answerable to observation.

Chapter numbers refer to the second edition of Changism, Volume I.