5.1The claim
Section V of the Enquiry turns from doubt to explanation. If no reasoning carries the mind from past to future, something else must, and Hume names it custom. He argues that custom is not a defect to be corrected but the condition of all practical life, and that it works with the force of instinct. Beside that argument stands WSTiger's 2026 trading desk, where a leverage limit has been raised after each of three breaches and each raise approved after the fact. Here the gain is for the second side of the debate, since Hume's account explains why the high-reliability school's practices must be institutional rather than individual, because the force they work against is one Hume describes as no more open to reasoning than love or hatred.
5.2The principle that is not reasoning
Hume begins Section V by reassuring the reader that the doubts of Section IV will not stop anyone acting:
Nature will always maintain her rights, and prevail in the end over any abstract reasoning whatsoever.
If reasoning does not carry the mind across, something does:
If the mind be not engaged by argument to make this step, it must be induced by some other principle of equal weight and authority; and that principle will preserve its influence as long as human nature remains the same.
He then imagines a person who has lived long enough to see events constantly joined:
Suppose, again, that he has acquired more experience, and has lived so long in the world as to have observed familiar objects or events to be constantly conjoined together; what is the consequence of this experience? He immediately infers the existence of one object from the appearance of the other. Yet he has not, by all his experience, acquired any idea or knowledge of the secret power by which the one object produces the other; nor is it, by any process of reasoning, he is engaged to draw this inference. But still he finds himself determined to draw it: And though he should be convinced that his understanding has no part in the operation, he would nevertheless continue in the same course of thinking.
The last sentence of this passage is the one a risk officer should read twice. Even a person who has been convinced that the inference has no ground "would nevertheless continue in the same course of thinking". Knowing that a run of profitable breaches proves nothing does not stop the run from shaping expectation. The understanding has no part in the operation, and so correcting the understanding does not stop the operation.
Hume names the principle:
This principle is Custom or Habit.
and explains why only custom fits the facts:
This hypothesis seems even the only one which explains the difficulty, why we draw, from a thousand instances, an inference which we are not able to draw from one instance, that is, in no respect, different from them. Reason is incapable of any such variation.
The conclusion of the paragraph is short: "All inferences from experience, therefore, are effects of custom, not of reasoning". Then the famous judgment on custom's place in life:
Custom, then, is the great guide of human life. It is that principle alone which renders our experience useful to us, and makes us expect, for the future, a similar train of events with those which have appeared in the past.
Without the influence of custom, we should be entirely ignorant of every matter of fact beyond what is immediately present to the memory and senses. We should never know how to adjust means to ends, or to employ our natural powers in the production of any effect. There would be an end at once of all action, as well as of the chief part of speculation.
These two paragraphs keep the reading honest. Custom is not the enemy of safe operation. A desk without custom could not trade at all; a plant without it could not run. The same principle that lets an operator expect the pump to start when the switch is pressed lets the operator expect the alarm to mean nothing. Hume gives no way of having the first without risking the second.
5.3Belief as feeling, and why it cannot be ordered
Part I closes with a description of belief that makes the point sharper still:
All belief of matter of fact or real existence is derived merely from some object, present to the memory or senses, and a customary conjunction between that and some other object.
This belief is the necessary result of placing the mind in such circumstances. It is an operation of the soul, when we are so situated, as unavoidable as to feel the passion of love, when we receive benefits; or hatred, when we meet with injuries. All these operations are a species of natural instincts, which no reasoning or process of the thought and understanding is able either to produce or to prevent.
Part II then describes what belief feels like:
I say, then, that belief is nothing but a more vivid, lively, forcible, firm, steady conception of an object, than what the imagination alone is ever able to attain.
And Hume closes the section by saying why nature would have arranged things this way:
It is more conformable to the ordinary wisdom of nature to secure so necessary an act of the mind, by some instinct or mechanical tendency, which may be infallible in its operations, may discover itself at the first appearance of life and thought, and may be independent of all the laboured deductions of the understanding.
The word "infallible" in this passage is worth pausing on. Hume says the tendency "may be infallible in its operations", meaning, on the most natural reading, that it reliably does what it does: it reliably carries the mind from the present object to the expected one. It does not say the expectation is reliably true. A reliable mechanism for forming expectations is exactly what turns a run of profitable breaches into a firm, steady conception that the next breach will be profitable too.
If belief is a feeling produced by a present object and a customary conjunction, and if no reasoning "is able either to produce or to prevent" it, then the remedy for a belief formed by custom cannot be an argument addressed to the person who holds it. It has to be a change in what is present to that person's memory or senses, or a change in who holds the decision. This is the point at which Hume's account speaks to the second side of the debate. Weick's practices, read through Section V, are ways of changing what is present: attention to failure puts small failures in front of people; staying close to operations puts the sensible facts in front of those who decide; deferring to expertise moves the decision to someone whose customs were formed on different experience.
5.4A limit defended by experience
Hume's long footnote to §36 contains a passage that bears on the second record with unusual directness. Discussing the common distinction between arguments from reason and arguments from experience, he gives an example from politics. The limits placed on government, he says, may be defended in either way: from reason,
which reflecting on the great frailty and corruption of human nature, teaches, that no man can safely be trusted with unlimited authority
or from
history, which inform us of the enormous abuses, that ambition, in every age and country, has been found to make of so imprudent a confidence.
He then argues that the distinction is shallow, because the argument from reason also rests on experience in the end:
The history of a TIBERIUS or a NERO makes us dread a like tyranny, were our monarchs freed from the restraints of laws and senates: But the observation of any fraud or cruelty in private life is sufficient, with the aid of a little thought, to give us the same apprehension; while it serves as an instance of the general corruption of human nature, and shows us the danger which we must incur by reposing an entire confidence in mankind. In both cases, it is experience which is ultimately the foundation of our inference and conclusion.
A risk limit is a restraint of exactly this kind. It exists because someone, drawing on wider experience than the trader's own, concluded that no one can safely be trusted with unlimited leverage. The limit encodes experience: other desks, other years, other people's losses. When the limit is breached and the breach turns out well, a second, narrower body of experience, the trader's own recent run, begins to argue against the first. Hume's footnote shows that both sides of that argument are arguments from experience. The question is not whether to trust experience but whose experience the custom of the desk is being formed on.
The footnote also contains Hume's warning about the details a run hides:
In every situation or incident, there are many particular and seemingly minute circumstances, which the man of greatest talent is, at first, apt to overlook, though on them the justness of his conclusions, and consequently the prudence of his conduct, entirely depend.
5.5The record: a desk that approves after the fact
WSTiger's 2026 record sets its scene on a synthetic trading desk. As with the plant, the desk, the trader and the approvers are roles in an invented scene, not real people or firms. The English translation of the scenario reads:
A synthetic desk raised a trader's leverage limit each time it was exceeded, even though the trader was making a profit. This is the third time this year, and each instance was handled as a retroactive approval.
The translation needs a correction before it can be read. The Korean original says:
가상의 데스크는 트레이더가 레버리지 한도를 넘었지만 수익이 나자 한도를 올려 줬다. 올해만 세 번째이며 매번 사후 승인으로 처리됐다.
The Korean first sentence says that the trader exceeded the leverage limit, but, once a profit came in, the desk raised the limit. The profit is the occasion of the raise, not something that happened in spite of it. The English "even though the trader was making a profit" reverses the relation and obscures the mechanism the record is about. This article reads the Korean meaning: a breach, a profit, and then a raise granted because of the profit.
So read, the scenario is the bread of §29 transposed to a trading floor. The breach that came before turned out well; does it follow that the next breach will turn out well, or that the limit was simply set in the wrong place? The desk has acted three times as if it did follow. The phrase "each instance was handled as a retroactive approval" describes something more specific than a drift in belief. It describes a procedure in which custom acquires a signature. Each time, someone with authority signed off on the raise after the profit had been made. The approval was given in the presence of the outcome. Fischhoff's finding that knowledge of an outcome changes judgment without the judge noticing, and Baron and Hershey's finding that good outcomes make the same decision look better, both apply with full force to a procedure that only ever approves after the outcome is known. Hume's account explains why the approvers might feel no strain in doing so: by the third time, the expectation that a breach will pay is a "firm, steady conception", present to them as a fact about this trader rather than as an inference.
The record asks:
Should the new limit be left as is, should the original limit be restored and retroactive approval be blocked, or should independent committee approval be required for any limit change?
Read against Section V, the three options differ in where they act on custom. Leaving the new limit as it is lets the custom stand and writes it into the rule. Restoring the original limit and blocking retroactive approval removes the occasion on which custom has been signing itself into policy: if no approval can be given after the outcome, the outcome cannot be what persuades the approver. Requiring an independent committee moves the decision to people whose expectations were not formed on this desk's run. That last option is, in Hume's terms, a way of changing whose experience is present when the decision is made, and in Weick's terms a form of deferring to expertise held elsewhere. None of the options is shown here to be correct. Each gains and gives up something the record does not price, and the people who would choose among them are not named in it.
The example judgment again draws the line where Hume draws it:
This is an example that distinguishes violations that had good outcomes from decisions in which the risk was appropriate. The worst-case loss possibility over the limit-exceeding interval is recalculated separately from the outcome.
The first sentence separates outcome from decision as Baron and Hershey separate them. The second proposes a procedure of the same family as the plant's: recalculate the worst-case loss over the interval in which the limit was exceeded, "separately from the outcome". The instruction is an attempt to take the outcome out of the room while the risk is judged. In Hume's language, it tries to bring before the mind a different present object, the calculated worst case, so that the conception formed is not only the vivid and steady one produced by three profitable breaches. Whether such a calculation can compete with custom is a question the record cannot answer and Hume's account suggests is open: a number on a page is a present object, but so is a profit already booked.
5.6The thread's own sentence
Both records belong to a thread to which WSTiger attaches a single sentence of interpretation. In the English translation stored with the plant record it reads:
When limit violations repeat without loss, the violation becomes normal, and the limit remains only in documents, disappearing in practice.
This is WSTiger's own statement of the normalisation of deviance, written for the platform. It says what Vaughan's term says, and it adds a picture of where the limit goes when it is normalised: into documents. That picture is worth setting beside Lee Clarke's argument that organisations write plans for catastrophes they cannot control, documents that could not work as written but persuade others and themselves that the risk is in hand (Clarke 1999). A limit that remains only in documents is not a fantasy document in Clarke's sense, since it once described a real constraint. But it ends in a similar place: a written assurance that no longer governs what happens. Hume's Section V explains how it gets there. The document records an inference from wider experience; the practice is governed by custom formed on narrower experience; and custom, being a feeling and not an argument, wins wherever the two meet without some arrangement that keeps the wider experience present.
5.7What the case adds
Read through Section V, the desk record shows custom acquiring an institutional form. The retroactive approval is the point at which the expectation formed by a run of profitable breaches becomes policy, signed by someone who has the outcome in view. For the second side of the debate, the case adds a reason why reliable practice has to be built into procedure rather than left to individual judgment: Hume's account holds that even a person convinced that the inference has no ground "would nevertheless continue in the same course of thinking". The record's own example judgment, and two of its three options, act on the circumstances in which the belief is formed rather than on the belief. That is the high-reliability school's wager stated in miniature. Whether it succeeds the record cannot say, because the record has no outcome.