Racial Fatigue

An exhaustion has become common in people who have not been harmed by anything under discussion. It is not fear, which would track exposure, and it is not hostility, which would track interest. It is depletion, and it appears at every distance from the events being argued over, including in people on other continents who will never meet anyone involved. Whatever produces it sits in the argument rather than in the events.

The subject here is that affect and the sequence that produced it. Figures are taken as given; the question is whether the steps between them hold. The sequence runs from a rule about who qualifies for assistance, through a change in household composition, into a set of outcomes, into the moral warrant the assistance required, and into a rule about what can count as evidence. The affect is what that link produces. A final step then attributes the affect back to the change in household composition already in the chain. That step is asserted rather than shown, and it is the step the account’s own conclusion is stated from.

Standing

A statistic about a place carries an interpretation, and the interpretation requires knowing what the place is like. Someone who has never been inside a community can read its rates correctly and still be unable to say which of several accounts the rates are consistent with, because the discriminating information sits in the texture rather than in the numbers. Daily exposure supplies evidence that a table does not. Raised against an outsider commenting on a community’s internal condition, that is a real objection about epistemic access rather than an attack on the speaker.

It is answered by relocation. The analysis being objected to did not originate with the person delivering it. It was the work of a member of the population it described, restated by someone outside that population, so the objection applied consistently admits the content it was raised against.

The answer works, and it works by paying what the objection asked for. The general form of this objection is not defeated here; it is satisfied. Nothing establishes that an outsider could have conducted the analysis, and the sequence below inherits the limit. Every step that turns on what a rule felt like to the people living under it is a step where the objection retains purchase.

The eligibility rule

Assistance was made conditional on the absence of a father in the household, at a level that could exceed what a low-wage job paid. A rule of that shape does not need anyone to intend a result. It prices one household arrangement above another and leaves the arithmetic to the people doing it.

The timing sits against a baseline that makes the usual explanation difficult. In the decades before the rule, under formal segregation and far higher poverty, roughly 78 per cent of black children lived with both parents and about 15 per cent were born to unmarried mothers. The figures now are under half and between 70 and 78 per cent. The severity of discrimination was greater on the earlier side of the change and the family structure was more intact there.

This is the strongest link in the chain, and the reason is specific. Two things moving together licenses nothing on its own. This rule named the variable in advance: the payment was conditioned on the presence or absence of a father, and the presence of fathers is what subsequently moved. An alternative cause has to explain not only the timing but why that particular variable was the one that shifted.

The same decade, differently explained

The decade supplied other candidates, and they are not weak ones. Heroin arrived in quantity in the same cities over the same years. The leadership of the political movements then organizing was killed, imprisoned, or dispersed under a federal program designed to do that. Lending and housing policy had already concentrated the affected population into the neighbourhoods where all of this landed. Any one of these would be expected to raise violence, and they arrived together.

Set against the homicide figures, the correspondence is close on both accounts. Black male homicide victimization stood near 42 per 100,000 around 1960, roughly doubled by 1970, peaked at 278, and now sits near 50, still above the 1960 level and against a long international decline.

What separates the two accounts is thinner than either side treats it as being. The eligibility rule specified its variable beforehand and the alternatives do not; a drug epidemic predicts more violence without predicting fewer marriages, and the same is true of the loss of a political leadership. That is a genuine asymmetry and it establishes a mechanism rather than a magnitude. It gives no figure for how much of the change the rule accounts for, and the accounts are not competitors in the first place. Several causes operating on one decade is the ordinary case.

What the concession costs

The material concedes multi-causality outright. The eligibility rule is presented as a major contributor and explicitly not as the only one.

Taken seriously, that concession reaches further than it is allowed to. An incentive changes behaviour at a margin, and where the margin sits is set by everything that came before it. The same conditional payment offered to a population holding accumulated property, intact inheritance, and ordinary legal recourse produces one response; offered to a population from which those were removed inside living memory, it produces another. On that account the rule is the trigger and the prior removals are what loaded it, and identifying the trigger settles nothing about who loaded it. Contributory causes operating at different depths are not rivals, and agreement about the shallow one is not agreement about the account.

This is not answered. The reply offered is that the substantive claims were granted, therefore nothing was refuted, which treats agreement on a contributory cause as agreement on the whole causal story. The second reply is a prescription: remove the rule, since that is the part fixable now. A prescription about what to do next is compatible with any distribution of causal weight and settles none of it. The deep-cause account survives the chain intact, and the concession the material makes is the reason it does.

Arrests and offenses

The share of crime attributed to a population is usually derived from arrest data, and arrest data measures enforcement contact. Enforcement contact is produced jointly by offending and by policing intensity, and policing intensity is not distributed evenly. Where police are denser, a fixed amount of offending generates more arrests. An arrest-derived share is therefore an upper bound on the offending share, and the size of the gap is not measured by the same instrument that produced the number. Where a population also accounts for 53 per cent of exonerations, the gap has some evidence behind it.

Part of this has an answer. Homicide victimization data does not pass through arrest, because bodies are counted regardless of whether anyone is charged, and the series above is built from it. The figure raised in objection is itself victim-derived: of roughly 13,000 murder victims in a year, 55 per cent were killed by black offenders. On homicide specifically the measurement problem is largely closed, and homicide is the category the argument turns on.

For the broader crime categories it is not closed, and nothing here closes it. What happens instead is that the standard is dropped. The objection is raised against one figure and then abandoned within the same argument, which later attributes 69 per cent of all crime to white Americans on exactly the arrest basis just rejected. That inconsistency establishes something about the selection of conclusions and nothing about the standard. A measurement objection is not weakened by the failure of the person raising it to hold to it, and this one is correct.

The base rate

The argument from magnitude runs cleanly. Roughly 7,660 murders per year attributed to a population of 22 million men works out to 0.035 per cent, which is a very small fraction, and a public conversation organized around a behaviour that 99.96 per cent of a group does not engage in has misallocated its attention.

The quantity is per year and the conclusion is drawn about people. A man does not live one year. Multiplied across a 72-year life the annual figure compounds to about 2.52 per cent, which is a different number answering the question actually asked.

The corrected figure is also not the right instrument, and the material says so. It ignores age structure, which concentrates offending in a narrow band, and it ignores repeat offenders, which inflates the count of distinct people. Correcting an error does not produce a good number, it produces a number of the right kind.

The reassurance the small percentage was offered as depends on something nobody has. A share of a male population committing homicide carries no meaning without a comparison class, and there is no intuition available for what fraction is ordinary. The figure was presented as low because it looked low next to zero, which is the only comparison the reader had.

The rarer category

Attention should follow magnitude, and a category of harm affecting a very small number of people should not organize national argument. That principle is correct, and it is worth holding.

Applied to the whole field it removes more than it was aimed at. The category of killing that produced a national protest wave is smaller by orders of magnitude than the category the principle was invoked to shrink, and in the one city whose figures are given, roughly 90 per cent of the victims share a population with the offenders. Applied evenly, the principle directs attention toward the larger category and away from the movement that invoked it.

What that shows is limited and worth stating precisely. The principle stands and the argument would be improved by anyone applying it. What fails is its use as a reason to close one discussion while another continues, because the selection between the two is being made by something the principle does not name.

Totals and the counter-charge

Two moves in the exchange do not bear on the claim at all.

Raw counts drawn from populations of different sizes settle nothing about representation. A larger group generates more of everything, including enrolment in assistance programs; the relevant figure is the rate, and on the rate the direction reverses. The other move introduces a different offense category committed by a different population. Whether that figure is accurate has no bearing on the first, because two claims about two populations can both be true and neither answers the other.

A third move attributes the argument to the traffic that discussing the subject generates. An account of why someone holds a position leaves the position exactly where it was, and the form is available against anyone, including against the imputation, later in the same exchange, that the objections are motivated by wanting the payments continued without conditions.

The warrant

The program had to be paid for by people who would not receive it, and that required a reason. The reason given was that the condition being addressed had been produced by what those people’s forebears did, which made the transfer restitution rather than charity.

That warrant has a property worth tracing. It makes the causal account load-bearing. While the transfer is justified by external causation, every internal cause admitted is a subtraction from the justification, so the account of why the outcomes exist cannot be revised without the entitlement becoming revisable. Everyone standing under the warrant acquires a reason to reject internal explanations that is independent of the evidence for them.

And on the chain above, the transfer is among the causes of the condition it compensates. Whether that closes depends entirely on how much weight the eligibility rule carries, which is the quantity the previous sections could not supply. The structure is available rather than established.

What could count against it

The account that survives under the warrant is stated in a specific form. Street names commemorating slaveholders. Schools named for the same men. The existence of suburbs, the covenants of homeowner associations, the custom of tipping. Each is offered as an instance of the same cause.

Taken together they leave nothing outside. An account that absorbs the naming of streets, the shape of housing, and the practice of leaving a gratuity has no arrangement of the world it would fail to fit, and a claim compatible with every observation is not being supported by any of them.

The same position also objects to being described as calling everything racist, and the enumeration is what it offers in answer. Stated at full strength that is not a contradiction: the reply is that the charge is accurate item by item rather than reflexive, and a reply of that kind is the right one to make. What it cannot do is separate the two possibilities, because a list reaching street names, housing, and gratuities is what a reflex and an accurate description would each produce.

Agreement does not retire it

By the end of the exchange the three substantive claims had been granted. The outcome disparities are real. Fatherlessness contributes to them. The eligibility rule contributed to the fatherlessness, a point volunteered before it was argued for.

The characterization did not move. The same position that supplied the agreement continued to describe the argument as racist and its author as an outsider without standing to make it.

If a set of claims can be granted in full and the label survives unchanged, the label is not resting on those claims. The ground given for it in the same exchange is a content ground: that the condition is described without the history that produced it. Part of that ground is the deep-cause objection above, which agreement on the three claims does not retire and which nothing here answers. The rest is the enumerated account, which reaches street names and gratuities and which no arrangement of evidence reduces. The label was also in the exchange before the first claim had been heard. What follows is narrower than a claim about tone: agreement on what was argued leaves the label where it was, because the label was never resting on what was argued.

Why the demand exhausts

Two properties have now accumulated. No evidence reduces the account, because the account fits every arrangement of evidence. No agreement satisfies the demand, because the demand rests on that account rather than on anything an interlocutor is in a position to concede.

Three responses remain. Endorsement is available to anyone who thinks the account is right. Endurance costs attention every time the demand is made, and it is made continuously, on a schedule set by whoever is making it. Withdrawal of attention is the third, and it costs whatever relationship the argument was embedded in.

The exhaustion is the third response arriving. It requires no view about any population and does not produce one. It appears in people who have never been harmed by the events under discussion because those events are not what is producing it, and it appears at long distance because an argument travels and the events do not.

The last link

The closing move of the account is a relocation. The affect is attributed to the household condition described earlier rather than to any population: an adult raised with no resident authority who assigns responsibility grows up locating causes outside himself, and the argumentative pattern above is that disposition operating at scale. On that reading the exhaustion and the thing producing it have one origin, and the chain closes on itself.

Stated at full strength the account is coherent. A household in which nobody holds a child to an account is a household in which nothing installs the habit of asking what the child did, and an adult without that habit reaches outward first because nothing ever made him reach anywhere else.

Nothing connects the developmental claim to the argumentative pattern. No figure relates the two, and the population holding the pattern is not shown to be the population raised without fathers. Every other link in the sequence carries either a named variable or a dated change. This one carries an analogy between an individual upbringing and a collective habit, and the analogy is the entire connection. The sequence stops being shown here, at the link its own conclusion rests on.

What the chain establishes

The eligibility rule is the strongest link and narrower than the use made of it. It named the variable that then moved, which is more than co-occurrence and short of a magnitude. The outcome link is close in timing and unsettled in weight against the other causes that arrived in the same decade. The warrant link holds as a description of an incentive, and the loop it suggests is available rather than demonstrated. The rule about what can count as evidence holds, and it is the link that matters here, because it is the one that makes the argument unendable. The closing step does not hold at all: nothing carries the account from a household without a father to a pattern of argument, and that is the step the conclusion is stated from.

What breaks are the objections that change standards partway through: a measurement standard applied to one figure and dropped for the next, a magnitude standard applied to one category of killing and not to a rarer one. Those failures show that the standards are not selecting the conclusions. They leave both standards correct.

The deeper account is untouched. That prior removals set the margin the incentive then operated on is consistent with every link the chain establishes, and it is not addressed anywhere in the sequence. What the sequence does establish is the mechanism of the affect rather than the merits of the dispute: an account that no evidence reduces and no agreement satisfies imposes a cost with no stopping point on everyone inside it, and withdrawal is the only response that ends the cost. Withdrawal is available regardless of what a person concludes about the underlying question, which is why the exhaustion spreads independently of the evidence on either side.

Links

• Validity and Truth — the causal-chain form and its characteristic failure, a link asserted rather than shown.
• White Guilt — the debt the warrant invokes, and whether the arguments establishing it hold.
• Per Capita — the rate-versus-total step this argument turns on repeatedly, and the test for when its omission is deliberate.
• Racial Egalitarianism - The Disparity Inference — the question of what produces group disparities, which the warrant assumes an answer to.
• Moral Language as Leverage — why the warrant’s vocabulary works on the people who hold it, and why exposing the asymmetry does not retire it.

Open questions

If an account cannot be reduced by evidence and cannot be satisfied by agreement, what would distinguish it from a correct account whose opponents are unwilling to concede it?

The eligibility rule named the variable that moved and the alternative causes did not. What evidence would establish how much of the movement the rule accounts for, given that all of the candidates arrived in the same decade?

Sources

This YouTuber Gave Me Black Fatigue, 2026-04-27 — https://www.youtube.com/watch?v=7TSJkkY-D5c. Supplied the causal sequence from conditional welfare to family structure to crime, the homicide and household figures, the per-year and arrest-versus-offense corrections, and the closing relocation of the affect.