6.1The claim
Hume's argument ends with a requirement: every chain of inference about matters of fact must end in some fact present to memory or the senses. If the calm is to be broken, something present must break it. With that requirement in hand, the two records' revisit lines, where WSTiger asks for a condition under which the judgment would be reopened, can be read for what they name. Both sides of the debate stand to learn from the result, which is an asymmetry in what the records make observable. They make the custom observable, and they name what would relax the caution, but they do not name the fact that would end the calm.
6.2Some fact must be present
Hume insists that all inferences about absent things must start from something present:
If I ask why you believe any particular matter of fact, which you relate, you must tell me some reason; and this reason will be some other fact, connected with it.
The chain cannot run back forever; it must stop at a fact present to the memory or senses, or the belief has no foundation. For the operator at the plant, the present fact is the alarm, and the customary conjunction is alarm-then-nothing. For the approver at the desk, the present fact is the profit, and the customary conjunction is breach-then-profit. To break either custom, a different present fact is needed: one that the customary conjunction does not cover.
The child at the candle in Section IV makes the same point from the other side:
When a child has felt the sensation of pain from touching the flame of a candle, he will be careful not to put his hand near any candle; but will expect a similar effect from a cause which is similar in its sensible qualities and appearance.
One burn breaks a custom of reaching. The child does not need a hundred burns. A single present fact, vivid enough, establishes a conjunction at once. That is the hope behind any revisit condition in a risk record: that the condition names a fact which, if it appears, will be vivid enough to break the custom the run has built. It is also the danger, since the fact that breaks a custom of calm in a chemical plant may be the accident itself.
Hume's description of the harmony between nature and the succession of ideas states what is at stake:
Here, then, is a kind of pre-established harmony between the course of nature and the succession of our ideas; and though the powers and forces, by which the former is governed, be wholly unknown to us; yet our thoughts and conceptions have still, we find, gone on in the same train with the other works of nature.
The harmony holds while the course of nature continues as before. The whole question of a revisit condition is what present fact would show that the harmony has broken, before the accident shows it.
6.3The plant's revisit line
WSTiger's record for the plant states when its judgment would be revisited:
If a cause investigation finds that a warning was unrelated to an actual risk, or if the inspection cost is excessive, the matter is re-judged.
Read slowly, the line names two conditions, and both point the same way. If an investigation finds that a warning was unrelated to actual risk, the cautious judgment is revisited. If inspection costs too much, the cautious judgment is revisited. Both conditions describe circumstances in which the calm would be confirmed or in which the price of doubting it would be too high. Neither describes the circumstance in which the calm would be shown false: a narrowing margin, a warning whose cause is found to be real, a near thing. The example judgment asked the plant to record how close to the limit the equipment came each time; the revisit line does not say what trend in those records would reopen the decision in the direction of more caution.
This is not a flaw that a reader can lay at the door of a synthetic scene. A revisit condition naturally names what would make the present judgment wrong, and the present judgment is the cautious one. What would make caution wrong is evidence that the warnings were noise. What would make the calm wrong is not, in this frame, a condition on the judgment, because the judgment has already set the calm aside. Still, the result is that the line where the record writes its future leaves the end of the calm unnamed. In Hume's terms, it names the facts that would let custom resume its course, and not the fact that would break it.
The words "excessive" and "unrelated to an actual risk" also leave much to judgment. Neither carries a number, a date or a named reviewer. An investigation can find that a particular warning was unrelated to risk only by an inference from experience, so the first condition reintroduces at the point of review the very step Section IV says has no ground. This is not a contradiction in the record; it is Hume's point that practice must go on by custom, written into the form of a revisit condition.
6.4The desk's revisit line
The desk record's revisit line reads:
A re-judgment occurs if the independent review confirms that the new limit falls within the risk budget, or if the limit is upheld without retroactive approval.
The pattern repeats. Both conditions describe circumstances in which the cautious judgment would no longer be needed: the raised limit turns out to fit the budget, or the limit comes to be kept without after-the-fact approvals. Neither names the circumstance that would show the run of profitable breaches had ended, such as a breach that lost money or a recalculated worst case that exceeded what the desk could bear. The second condition is worth reading twice in Hume's light. Its condition that "the limit is upheld without retroactive approval" is itself a description of a run: a stretch of time in which the limit holds. The condition would reopen the judgment on the strength of a new run of good behaviour, which is the same kind of evidence that produced the problem in the first place, only pointing the other way.
The line does name an independent review, which is one of the record's options. That gives the condition a role, in the sense that a reader can see who would perform the review, and the coders of the parent study both read it so (see the appendix). It does not say when the review would happen or what figure in it would count as falling within the budget.
6.5An asymmetry
Set side by side, the two records show a consistent shape. Each writes, in its example judgment, the distinction Hume drew between a calm record and a safe future. Each proposes, in its example judgment, a way of changing which facts are present: the plant would record ignored warnings and margins; the desk would recalculate the worst case apart from the outcome. Those proposals make the custom itself observable. But each record's revisit line names what would relax the caution, not what would end the calm. The future the records write is the future in which the doubt turns out unnecessary.
This article does not claim the asymmetry is a fault. A record of a cautious judgment that names what would undo the caution is doing what a revisit line is for. And the event that would end the calm in a plant may not be something anyone should wait to observe. Still, if the question is whether the records write what would break the custom they describe, the answer from these two is that they write the means of seeing it and not the event that would break it.
6.6What the case adds
The third case adds to both sides of the debate a point about documents. Sagan found that learning from near-accidents was limited by secrecy and blame. The high-reliability school holds that practice can keep failure in view. The two records suggest a third place where learning can stall: in the form of the revisit condition itself, which tends to name the facts that would let a decision relax and to leave unnamed the facts that would show the calm had ended. Hume's demand that every belief rest finally on some present fact gives a way to state what is missing: the records name the present facts that would confirm the run, and not the one that would contradict it.