Hume is usually caricatured as having no account of correcting a feeling. He has a whole apparatus — he just denies it works by refutation. This note lists the corrections, sources each one to the Treatise, and flags the two places the standard summary (including my own earlier draft) gets it wrong.
Bottom line
Hume’s corrections move you from a narrow human standpoint to a general one. None of them moves you from human standpoints to something outside them. They make sentiment consistent and shareable; they do not make it true.
Two things to keep straight, both of which the usual summary botches:
- These are not recommendations. Hume is not advising anyone — he is explaining why moral judgment is stable despite sympathy being variable. It is a defence of his theory against an objection, not a program.
- These are not emotional therapy. They fix your assessment of other people’s characters. They say nothing about managing your own anger, grief, or anxiety. → 5. Not a technique for fixing your own feelings
1. The problem Hume sets himself
Our raw sentiments are wildly partial. You feel more for your own child than a stranger’s, more for someone nearby than far away, more for a friend than a rival. If moral judgment were just raw sentiment, “good” would mean good from where I’m standing, and two people would never be talking about the same thing.
Hume takes this seriously — it is the standing objection that his theory cannot yield real moral distinctions at all. He states it himself at T 3.3.1.14, SBN 580–81: sympathy is stronger for those near us than remote, for acquaintances than strangers, for countrymen than foreigners — and yet the same moral qualities receive the same approbation in China as in England.
His answer: we do not judge from where we stand. We shift, deliberately, to a common point of view — what this person’s character does to the people who actually deal with them day to day. Their family, colleagues, servants, neighbours. Not what it does to me.
That is why we can call an enemy admirable, or praise a virtuous person in ancient Greece who can never affect us. And, on Hume’s account, it is why moral language works at all: without a shared vantage we would never converge enough to argue.
2. The three corrections
Corrected from an earlier four-item list
An earlier draft of this listed intimacy/connection as a separate distortion. That was double-counting. At T 3.3.1.14 Hume lists contiguity, acquaintance, and nationality together as one family of variations in sympathy, handled by one correction. Three, not four.
1. Situation — distance, acquaintance, nationality
The distortion: you feel more for what is near. A famine down the road moves you more than one across the world; a cruelty last week more than one last century.
The correction: disregard proximity. Judge the character by what it does to those in its own circle, whenever and wherever they lived. This is how we can praise Cato or condemn Nero.
Hume’s own illustration: a diligent servant may stir more affection than Marcus Brutus in the history books, but we do not therefore call the servant more laudable. His phrase — “correcting our sentiments, or at least, of correcting our language.”
T 3.3.1.14–17, SBN 580–83 — confirmed against the text.
2. Personal interest
The distortion: you approve of what benefits you.
The correction: we overlook our own interest in general judgments, and do not blame a man for opposing us in our own pretensions, because we allow a degree of selfishness as inseparable from human nature.
T 3.3.1.17, SBN 582–83 — confirmed.
Example to fix if you reuse this
The natural illustration — “you can call an enemy brave and mean it” — is not Hume’s case here. His case is declining to blame someone for thwarting you. The enemy appears only at T 3.3.1.23, in an aesthetic aside about admiring the fortifications of an enemy city.
3. Fortune — judge disposition, not outcome
The distortion: you approve of virtue that happened to succeed and discount virtue that got thwarted. A generous man who went bankrupt before he could give anything away.
The correction: judge the durable disposition, not the outcome. Good fortune makes a benevolent disposition more affecting, but not more virtuous, so we separate fortune from disposition as far as we can.
T 3.3.1.19, SBN 584— “Virtue in rags is still virtue.”T 3.3.1.21, SBN 585— fortune separated from disposition.T 3.3.1.4–5, SBN 575— the groundwork: character is durable; single actions matter only as signs of it.
Confirmed, and it is Hume’s own second heading — he himself draws the parallel to the correction for distance.
3. What the corrections run on
| Mechanism | Status | Source |
|---|---|---|
| Consistency — we fix on steady and general points of view in order to prevent continual contradictions and reach a more stable judgment | Confirmed; Hume’s stated motive | T 3.3.1.15, restated T 3.3.1.18 |
| Conversation — without the correction we could never converse on reasonable terms or use moral language at all; general language must be moulded on more general views because each person’s interest is peculiar to himself | Confirmed twice over | T 3.3.1.15–16; EPM 5.42, SBN 228–29 |
| Information — sympathy runs on what you know; learn the effects of someone’s conduct and the response shifts | ⚠️ Demoted. Not one of the corrections in T 3.3.1. Reason’s fact-supplying role is real Humean doctrine, but source it from EPM Appendix I, not here | EPM App. I |
Standpoint reached: “some common point of view” — T 3.3.1.30, SBN 590–91.
4. The limit Hume concedes — and it is his own concession, not a charity
This is the part to lead with, because it is repeated and explicit. The feeling often does not follow.
T 3.3.1.16— he corrects the sentiments, “or at least” the language, where sentiments are more stubborn.T 3.3.1.18, SBN 583— the corrections are not fully efficacious; men seldom love what is distant and unprofitable to them, and here we say only that reason requires impartial conduct, though we rarely manage it.T 3.3.1.21— “The passions do not always follow our corrections.”
His analogy: eyesight. A distant object does not stop looking smaller once you know it is not — but you correct for the distortion anyway and speak in terms of the real size.
So the corrections regulate judgment and vocabulary. They do not reliably move the heart.
”Recommendations” is the wrong genre
In T 3.3.1 Hume is not advising anyone. He is defending sympathy against an objection — the objection at T 3.3.1.14 that if approbation came from sympathy, and sympathy varies with distance, then moral judgment would vary too, and it does not. The corrections are his explanation of why it does not. His formulation: experience soon teaches us this method — a report of what competent moral speakers already do.
This distinction is live in the literature. Sayre-McCord’s argument is exactly that reading the general point of view as an Ideal Observer standard (fully informed, cleared of prejudice) misdescribes it, precisely because Hume is characterizing an ordinary human practice rather than an ideal.
The one normative flicker worth keeping as the wedge:
T 3.3.1.18— reason requires impartial conduct. That is what any normative reading has to lean on.
5. Not a technique for fixing your own feelings
Category error worth naming
Asking Hume “how do I fix my negative emotions?” is like asking a linguist how to write a better sentence: adjacent, but not the job.
The corrections above are answers to a question in moral epistemology — how can approbation come from sympathy when sympathy is so variable? Book 3 is a naturalistic explanation of a phenomenon. You could perform all three corrections flawlessly and still be miserable.
This is where the Stoic comparison misleads. Seneca and Epictetus are writing therapy — the Letters, De Ira, the Discourses are practical manuals for a person trying to live. Hume’s Book 3 is not that. → Can a Feeling Be Wrong - Stoics vs Hume and Hutcheson
What Hume does have on the practical question — scattered, and much weaker:
- Calm vs. violent passions (
T 2.3.3–2.3.8). Calm passions are not gentler feelings; they are ones that do not produce inner turmoil, and they can be quite strong. Strength of mind is the prevalence of the calm over the violent. Closest thing he has to a therapeutic ideal. - Custom and repetition. Passions weaken with familiarity and strengthen with attention. Practical upshot: change what you habitually attend to, since you cannot argue a passion away.
- Occupation and diversion. In the essay “The Sceptic” he is explicit that philosophy does very little to reform the temper — a fixed disposition is barely touched by reasoning, and useful activity and habit do more.
- Indirect leverage. Since reason cannot oppose a passion directly, the only handle is via belief: change what you take to be the case and the passion redirects itself. The same handle the Stoics use — without the claim that the belief was false.
The honest summary: Hume thinks temperament is largely given and hard to shift, and is skeptical that philosophy changes it much. The Stoics think it is fully remediable through correcting judgments about value, and built an entire pedagogy around that. On the practical question, the Stoics are ambitious and Hume is deflationary — a difference in what they thought was possible, not a disagreement between two competing techniques.
If the practical question is what you care about, the Humean material to read is Treatise 2.3 and the four essays on happiness — not Book 3.
6. Contrast with the Stoic, precisely
Both correct. The difference is the standard.
| Stoic | Hume | |
|---|---|---|
| Measured against | How things actually are | How a well-informed observer at the common standpoint would respond |
| Form of correction | Refutation — your fear of losing money is false; money is not a good | Reframing — from a narrow human standpoint to a general one |
| Can be done alone? | Yes — available to a person alone in a room | No — maintained by conversation with people standing elsewhere |
| Reaches outside human response? | Yes | No. Human all the way down |
Where the pressure sits
What if the general point of view is itself corrupted? A whole society sympathetically attuned to a cruel practice will find its sentiments beautifully consistent and widely shared — and will have performed all three corrections flawlessly.
Hume has no external appeal available; that is the price of the position. His partial answer is that the corrections keep going: more information, wider circles of people whose good counts, more collisions with those who see differently. Whether that is enough is exactly the live question — and it is where the neo-Aristotelian move tries to reinstate a standard (facts about what a creature like us needs) without going back to Stoic cosmology. → Philippa Foot · Moral Naturalism
Reference card
Three corrections (Treatise 3.3.1)
1. Situation — distance, acquaintance, nationality T 3.3.1.14–17, SBN 580–83
2. Self-interest T 3.3.1.17, SBN 582–83
3. Fortune — judge disposition, not outcome T 3.3.1.19–21, SBN 584–85
Why we do it
contradictions between our own judgments T 3.3.1.15, 18
moral language would be unusable otherwise T 3.3.1.15–16; EPM 5.42, SBN 228–29
Standpoint reached
"some common point of view" T 3.3.1.30, SBN 590–91
Limit Hume concedes
corrects judgment and language; passions often lag T 3.3.1.16, 18, 21
Provenance and verification status
Drafted from a Claude conversation, 2026-08-22 (share link below). Unlike the companion note, the Hume locators here were checked against the text in-session, and that pass is what produced the three corrections in §2 (down from four), the demotion of “information” in §3, and the example fix in §2.2. Treat
T/SBNnumbers as verified-by-web-search, not as personally collated against a printed Selby-Bigge — worth a spot-check before publication.The Sayre-McCord attribution in §4 is a report of the position, not a page-cited claim.
Source
Claude conversation, 22 August 2026 — https://claude.ai/share/54448ade-4c4e-46e8-ad5f-77a2817d3316