A Citation Finds a Decision; It Does Not State the Holding

Citator graphs.

Natural-language search.

A research platform in a navy blazer offering a confidence dial for “treatment,” as though the common law had finally agreed to get a dashboard.

Somebody can draw a family tree from R v Jordan to twelve later cases, colour the nodes by sentiment, and call the orange ones “procedurally spicy.”

Then there is the judgment.

It has a court, a date, a disposition, reasons, a vote line, perhaps a concurrence, perhaps a dissent, and a very low tolerance for being described by a sentence somebody found in a headnote during a sandwich.

It would like you to keep its parts separate.

The citation can get you to its door.

It cannot come inside and announce what happened.

Not wrong.

Just smaller than it looks.

That’s easy to miss, because a citation has the concentrated seriousness of a luggage tag.

A string such as 2012 BCCA 206 looks as though it has already done the legal thinking.

It hasn’t.

It identifies a decision through three elements: year, court code, ordinal.

It’s a very good address.

An address doesn’t tell you which room contained the argument, who entered it, or whether the argument mattered to the result.

The citation has excellent stationery

Your citation has several jobs, and none of them is “summarise the court.”

Its six template slots: case name, decision date or year, core address, parallel addresses, pinpoint, and court parenthetical.

The core address identifies the decision as a whole.

The pinpoint identifies material inside a source.

Put the second function into the first field and your useful record becomes a small administrative swamp.

For a report cite, the core address is volume, reporter abbreviation, initial page.

Cornell supplies the tiny sturdy example:

671 P.2d 1085, 1092 (Idaho 1983)

Page 1085 tells you where the opinion begins.

Page 1092 tells you where the relied-on material sits.

The second page isn’t a second reporter.

It’s a location inside the first one, where the proposition has been left to wait.

Pinpointing has three principal units.

Pages, numbered paragraphs, and source-defined subdivisions.

The version supplies your unit and the destination accepts it.

You don’t choose it because one notation looks especially judicial on a Friday afternoon.

Canadian guidance gives R v Jordan, 2016 SCC 27 at para 32, and ranges as “at paras 37-38.” The English Criminal Practice Directions use bracketed paragraphs, [30]-[35].

A page from an electronic viewer does not quietly become a court paragraph because both happen to have numbers.

You need five pinpoint properties: unit, marker, value or range, source version, citation position.

That source-version field is what stops you copying a pinpoint from one report into a parallel report, where it acquires a new page number through optimism.

Your locator belongs to the version you actually consulted.

A slip opinion and a later United States Reports version can differ in pagination, typographical corrections, and citation form.

The paper has changed clothes.

Its page number may have changed with it.

There are four kinds of address stability.

Court-assigned identity.

Report-edition location.

Publisher independence.

Display persistence.

A neutral citation is strong on the first.

A report cite is strong on a particular volume-and-page location. A URL is convenient and terrible at the last one.

ECLI carries jurisdiction, court, year, and unique identifier in a five-part design. ECLI:EU:C:1998:27 is not improved by taking scissors to it.

All exceptionally useful.

Still not a holding.

The outcome is not the rule

The judgment tells you what the court did.

Affirm, reverse, vacate, remand, grant, deny.

The reasons tell you why a legal proposition was needed.

A decision can reverse for a narrow procedural defect. It can affirm on a ground that avoids the larger question.

It can remand for fact finding.

An order granting or denying an application produces an immediate result without extended reasons.

The Supreme Court describes orders as short summary rulings, unlike opinions that set out judgment and reasoning.

You are reading at least four levels of legal commitment.

Judgment.

Holding.

Supporting reasoning that may not be necessary.

Dicta.

The holding is the legal resolution necessary to the operative act on the material issue.

Supporting reasoning can be important without being indispensable.

Dicta are observations or hypotheticals the outcome didn’t need.

A phrase beginning “we conclude” can introduce any of the four.

Grammar has never volunteered to be a reliable clerk.

Separate opinions make the file more honest, not more communal.

The record has three parts: what the court did, the reasoning that drew enough agreement, and alternative reasoning supplied by individual judges.

A concurrence can share the result and reject the route.

A dissent can preserve the losing position with considerable flair and no majority support.

If two judges join only Parts I and II, Part III does not become the opinion of all three because it contains an excellent sentence about federalism.

Publication labels cause a related misunderstanding.

“Unpublished” is a label with research consequences, not a trapdoor under the decision.

The Fourth Circuit posts published and unpublished opinions, and unpublished ones can still give reasons or adopt lower-court reasoning.

Federal Rule of Appellate Procedure 32.1 permits citation of designated federal judicial dispositions issued on or after 1 January 2007.

The rule doesn’t prescribe the weight a court must give them.

That later question belongs to the relevant court and rule, wearing its own jurisdictional paperwork.

Meanwhile the headnote waits near the photocopier with an overdeveloped sense of accomplishment.

A conventional headnote is ordinarily a publisher’s summary.

Not part of the judgment, not binding precedent.

Even a Supreme Court syllabus is prepared for reader convenience and isn’t part of the Court’s opinion.

It can locate an issue beautifully.

It cannot do the judicial act on the court’s behalf.

The routine fits on a card

The remedy isn’t a larger graph.

It’s a small record that makes your inflated verbs uncomfortable.

Your litigation version has six fields.

Court and date.

Procedural posture.

Judgment.

Issue resolved.

Holding or reason.

Publication or citation status.

Then pinpoint pages and source URL.

The editorial version asks you eight questions before a legal claim gets a verb stronger than “said.”

1.

What court issued the document?

2.

What date and case identifier does it carry?

3.

Is it an opinion, order, or judgment?

4.

What did the court do?

5.

What issue produced the proposition?

6.

Who joined the relevant reasoning?

7.

Was the statement necessary to the result?

8.

Does the source show decision text, not only a summary?

Your card opens with a citation-format record of nine fields.

Preferred label, jurisdiction, issuing or editorial body, source-object scope, citation template, locator model, pinpoint convention, reporter relationship, current source.

Under current source you preserve the exact URL, the issuer, the version or effective date where printed, the access date, and the relevant part.

Court, court system, and destination are separate data.

“England and Wales, EAT, later judgment” does a different job from “UK academic writing,” despite both being capable of buying the same tea.

For a reporter, you capture seven properties.

Full title, official abbreviation, jurisdiction, court coverage, series or edition, authorised status, preferred-use status.

For a parallel cite, you record four relationship fields.

Primary address, additional address, ordering rule, shared-decision verification.

Check party name, decision date, court, and decision number before you treat two handsome strings as the same decision.

A parallel citation is an edge between addresses. Not a second decision that happened to dress similarly.

Then the three-pass check does ordinary work.

First, you use the headnote, syllabus, or database result to locate the issue.

Second, you read the cited passage in the opinion.

Third, you compare it against the judgment and the vote line.

A proposition that survives all three can be described within the boundaries the documents show.

One that survives only in a headnote has found an editor’s account, not the court’s holding.

Citation rules are not rules of precedent

Your citation also has to arrive in the format its destination accepts.

That’s a separate problem, and it has four authority tiers.

Binding court rule or practice direction.

Official reporting manual.

Publisher or academic standard.

House style.

The Employment Appeal Tribunal direction, New York’s official reporting manual, OSCOLA, and a law-firm style sheet do not all govern the same destination.

Precision in the wrong tier is still you being wrong.

Just in polished shoes.

Publication preference has five levels in AGLC4’s stated order.

Authorised report.

Generalist unauthorised report.

Specialist unauthorised report.

Unreported medium-neutral judgment.

Unreported judgment with no neutral citation.

The Australian High Court submission rule runs a related sequence: authorised report, then another report, then an internet citation where no report exists.

Those rules decide which address belongs in your document.

They don’t decide whether the sentence at that address was necessary to the result.

Your final check is retrieval.

A citation should support two paths: human recognition and system resolution.

A case name and a readable reporter cite help the first.

A structured identifier helps the second.

The CJEU displays date, name, case number, and ECLI, and ECLI links can reach a relevant paragraph.

That’s good document design.

It remains document design.

No identifier, however elegant, has been delegated authority to turn a concurrence into a majority holding.

What the card is for

Call it the case card.

It doesn’t decide whether a case controls a later court. It doesn’t settle an issue the disposition left open.

You record the decision, the version, the locator, the operative act, the joined reasons, and the proposition your source actually carries.

The citation gets you there.

The card makes you look.