Case Treatment Taxonomy

The framework

The taxonomy answers one practical question: can I still rely on this case, and if so, to what extent? It comes from the Case Treatment Classification Guide, and its first commitment concerns units. Classifications attach to specific holdings, not to whole cases. One later opinion can overrule one holding of an earlier case, distinguish a second, and follow a third, and a single flag at the top of the case cannot say that.

Three dimensions

Every classification records three things:

1. Scope: which part of the cited case the treatment touches, whether the entire case, a specific holding, a factual application, or dicta.
2. Severity: how much the treatment damages or supports the cited case's authority.
3. Mechanism: the legal process through which the treatment operates, such as appellate reversal, legislative supersession, or doctrinal shift.

Scope keeps a narrow treatment from being read as a verdict on the whole case. Mechanism matters because different processes reach differently: a reversal on direct appeal ends the judgment, a statute overrides a rule, and a doctrinal shift erodes authority without any single decisive event.

Four families

Governing philosophy

"When in doubt, over-flag rather than under-flag." The costs are asymmetric. An unnecessary caution costs a lawyer a few minutes of reading. A missed negative treatment can put a dead holding into a filed brief.

Key distinctions

• Holdings versus dicta. Dicta cannot be overruled in the strict sense, because they never bound anyone. Treatment of dicta is recorded at dicta scope.
• Procedural posture. A ruling under Rule 12(b)(6) and a ruling on summary judgment resolve different questions. A later court's treatment at one stage may say nothing about the earlier holding at the other.
• Distinguished versus limited. Distinguishing is factual and neutral: the rule stands but does not reach these facts. Limiting is legal: the rule itself is narrowed.
• Same topic is not treatment. A later case on the same subject does not treat an earlier one unless it engages it substantively.

The full guide adds decision flowcharts, a catalog of common errors, worked examples, and a scholarly apparatus.

How to apply it

A constructed example, with invented cases. A state supreme court decides Alpha, which contains three propositions: H1, that a limitations period runs from discovery of the injury; H2, that a certain contractual exclusion is unenforceable; and a passing remark about fee-shifting. Years later the same court decides Beta. It applies H1 as binding and decides the case on it. It then turns to H2 and holds that the unenforceability rule applies only to consumer contracts, not to contracts between businesses. It never mentions the fee remark, although Beta also involves fees.

Three lessons follow. First, any single case-level indicator for Alpha would mislead: a caution flag overstates the risk to H1, and a clean signal hides the narrowing of H2. Second, the over-flag rule settles close calls. If Beta's language on H2 could be read either as distinguishing Alpha on its facts or as narrowing the rule, the classifier records Limited, the more cautious label. Third, the record tells the lawyer what to write: cite H1 from Alpha with Beta as reinforcing authority, and cite H2 only as limited by Beta, stating the narrower rule.

Known limitations and critiques

Classification takes judgment, and classifiers disagree. The line between Questioned and Criticized, or between Undermined and Abrogated, is not self-evident. The citator literature shows how much this matters. In Hellyer's study, reported in Good Law for What?, Shepard's, KeyCite, and BCite agreed that negative treatment existed in only 53 of 357 citing relationships that at least one of them had labeled negative.

Over-flagging has a price. A system that flags too often trains its users to ignore flags. The philosophy is right about the asymmetry, but it does not say where excess caution starts to erode attention.

The implicit categories are the hardest. Abrogated and Undermined require inferring an incompatibility or an erosion that no court announced. They are the categories most likely to be missed, and also the most likely to be over-applied.

Treatment events are not the whole story. The taxonomy records what later courts did to a case. It cannot record a shift in a doctrine's practical direction that involves no treatment event at all (the Kisor class described in Good Law for What?), and it says nothing about forum, posture fit, whom a holding helps, or arguments courts have already rejected. It is one input to the Proposition-Usability Model, not a substitute for it.

Severity is ordinal, not measured. The families rank damage roughly, but the taxonomy has no scale for comparing a Questioned treatment from a high court with a Criticized one from a peer court. Treatment from a court outside the forum's hierarchy is recorded like any other, yet its force depends on where the reader is filing.

Neutral labels can hide a trend. Cited, Discussed, and Explained take no position individually, but a long run of them may show a holding drifting out of use. The taxonomy classifies each treatment; it does not aggregate them.