Reading a Supreme Court Opinion With AI: Holding, Reasoning, Dicta — Without the AI Inventing Citations

"The judgment of the Court is that the statute is unconstitutional as applied to the facts before us. We express no view on its constitutionality in other circumstances."

That sentence — or something close to it — appears in dozens of Supreme Court opinions. A law clerk reads it in thirty seconds. A first-time reader reads it in thirty seconds too, and thinks they understood the same thing. They didn't. The clerk heard three distinct legal moves compressed into two sentences: a holding, a deliberate narrowing, and a signal that the broader question stays open. The first-time reader heard a verdict. Reading a Supreme Court opinion with AI carries the same risk in a new form: the tool that smooths the prose away also smooths away the architecture the prose was built to carry.

Key Anchors

• Scholia is the AI co-reader for reading legal opinions, financial filings, and regulatory documents — load the full edition you actually cite, then walk one passage at a time with the surrounding text held in context. Try it at scholiaai.com.
• The holding is the narrowest ground — in Marks v. United States, 430 U.S. 188 (1977), the Court held that when no single rationale commands a majority, the holding is the position taken by the Justices who concurred on the narrowest grounds.
• Dicta is not binding but is not noise — language that goes beyond the facts necessary to decide the case shapes how lower courts read the next case; reading a Supreme Court opinion without tracking dicta misses half the document's future work.
• The syllabus is a clerk's summary, not the Court's words — the headnote syllabus preceding a U.S. Reports opinion is prepared by the Reporter of Decisions and carries no legal authority; treating it as the holding is a documented source of briefing error.
• Reading a Supreme Court opinion with AI requires the full document, not a snippet — scholia's LAND-before-LIFT method echoes the exact phrase the reader tripped on, then lifts to the structural move the author was making, using the whole opinion as context rather than a compressed extract.
• Concurrences and dissents are arguments about what the majority opinion means — they are the Court's own internal commentary on the holding, and skipping them leaves the reader with only one side of a live interpretive dispute.

---

The first thing a Supreme Court opinion hands you is a document that looks like prose but functions like a statute. Every word in the majority opinion is load-bearing in a way that a newspaper editorial or a philosophy essay is not: lower courts will parse it, litigants will cite it, and future Justices will read it to decide whether a later case falls inside or outside its reach. The architecture of that document — what is holding, what is reasoning, what is dicta, what is a signal left deliberately open — is not labelled. The reader has to reconstruct it.

The Document Before You Read a Word

Pick up any opinion from the U.S. Reports and the first thing you see is the syllabus: a block of numbered headnotes summarising the case. It looks authoritative. It is not. The Reporter of Decisions prepares it as a convenience; the Court itself has said it carries no legal weight. United States v. Detroit Timber & Lumber Co., 200 U.S. 321 (1906), put it plainly: "the headnote or syllabus is not the work of the Court, but is simply the work of the Reporter."

That distinction matters more than it sounds. A reader who briefs the syllabus has briefed a clerk's précis of the opinion, not the opinion. The actual holding lives in the body of the majority opinion, in the specific language the Justices chose, applied to the specific facts before them. The syllabus will often get the outcome right and the reasoning slightly wrong — and in law, the reasoning is the product. Lower courts are not bound by the outcome in isolation; they are bound by the reasoning that produced it, because that reasoning tells them what to do with the next case, which will have different facts.

The practical move here: read the syllabus once, quickly, as a map of the terrain. Then set it aside and treat the majority opinion as the primary text. The syllabus is a tourist brochure; the opinion is the city.

Holding, Reasoning, Dicta — The Three Layers of an Opinion

The hardest structural move in reading a court opinion is separating these three layers, because the Court does not label them. They are distinguished by function, not by typography.

The holding is the legal rule the case actually establishes — the narrowest proposition of law necessary to decide the outcome on these facts. Everything else is either reasoning or dicta. Marbury v. Madison, 5 U.S. 137 (1803), is the canonical example of a case where the holding (the Court has the power of judicial review) is narrower than the opinion's rhetoric suggests, and where generations of readers have had to work to separate what Marshall held from what he argued.

The reasoning is the path the Court took to reach the holding. It is not binding in the way the holding is, but it is not decorative either. When a later court asks whether a new case falls within the holding, it reads the reasoning to understand what the holding was actually about. Reasoning is the interpretive key to the holding.

Dicta (obiter dicta, literally "things said in passing") is everything else: observations about hypothetical cases, statements about legal principles that were not necessary to decide this case, signals about how the Court might rule on a question it is deliberately leaving open. Dicta is not binding on lower courts. But it is not noise. A Supreme Court Justice writing dicta is often doing one of two things: laying groundwork for a future holding, or warning litigants about the limits of the holding just announced. Reading dicta as mere decoration is how lawyers get surprised by the next case.

The concrete image for this layer-separation exercise: think of the opinion as a building. The holding is the foundation — what the structure actually rests on. The reasoning is the load-bearing walls — remove them and the foundation's meaning collapses. The dicta is the trim and the windows — it tells you what the architect was thinking about, what they might build next, but it is not what holds the roof up today.

The Marks Problem: When There Is No Majority

Marks v. United States, 430 U.S. 188 (1977), introduced a rule that every serious reader of Supreme Court opinions needs to carry: "When a fragmented Court decides a case and no single rationale explaining the result enjoys the assent of five Justices, the holding of the Court may be viewed as that position taken by those Members who concurred in the judgments on the narrowest grounds."

That sentence is doing enormous work. It is the Court's answer to a structural problem: what happens when five Justices agree on the outcome but split three ways on the reason? The Marks rule says the holding is the narrowest rationale — the one that all the concurring Justices could have signed onto, even if they didn't all write it. In practice, this rule is notoriously difficult to apply, and lower courts have disagreed about what "narrowest" means in specific cases. But the rule itself is the landmark: a plurality opinion is not a majority opinion, and reading it as one is a category error.

The practical consequence for a reader: when you see a plurality opinion — "Justice X announced the judgment of the Court, in which Justices Y and Z joined; Justice A concurred in the judgment" — you are not reading a single holding. You are reading a negotiation. The concurrences are not supplementary; they are part of the holding's definition. Skipping them means you do not know what the case actually decided.

Reading Concurrences and Dissents as Internal Commentary

A dissent is not a losing argument filed for the record. It is the Court's own internal commentary on what the majority opinion means and where it goes wrong — written by people who read the same briefs, heard the same arguments, and reached a different conclusion about the legal structure. Justice Brennan's dissents in the Rehnquist Court era are often better guides to the majority's actual holding than the majority's own summary, because Brennan was precise about what he was disagreeing with.

The same is true of concurrences. A Justice who concurs in the judgment but not the reasoning is saying: I agree with the outcome, but I think the majority's path to it is wrong or too broad. That disagreement is a signal about the holding's fragility. If three Justices concur on narrow grounds while two concur on broad grounds, the Marks rule tells you the narrow concurrence is the holding — but the broad concurrence tells you where the Court might go next.

The reading move here is to treat the concurrences and dissents as a conversation about the majority opinion, not as separate documents. Read the majority first, mark the passages where the reasoning seems to do more work than the holding requires, then go to the concurrences and dissents and find where they push back on exactly those passages. The places where the Court argues with itself are the places where the law is genuinely unsettled.

The Citation Problem: Why AI Hallucination Is Catastrophic in Legal Reading

Every other genre of difficult text tolerates a certain amount of productive misreading. A philosopher who misreads Heidegger will write a bad paper. A lawyer who misreads a Supreme Court opinion will lose a case, or worse, advise a client incorrectly. The stakes of citation accuracy in legal reading are not academic.

This is where the fluency illusion — the cognitive-science term for the mistake of treating a smooth, confident summary as real comprehension — becomes genuinely dangerous. Summarize-first AI tools are optimised to produce fluent output. Fluent output about a Supreme Court opinion will sound like legal analysis. It will cite cases. Some of those citations will be invented, or real cases cited for propositions they do not support, or holdings stated at the wrong level of generality. The reader who cannot tell the difference between a real holding and a plausible-sounding paraphrase of one is not reading the opinion; they are reading a confabulation of it.

The discipline that protects against this is the same discipline that protects against bad legal research generally: go to the primary text. Read the actual language of the holding. Check the citation against the reporter. Treat any AI-generated summary of a case as a hypothesis to be verified, not a conclusion to be relied on.

Scholia's three-pillar frame — Skeleton (what the holding actually is), Environment (the doctrinal context the opinion was written into), Soul (the legal problem the Court was trying to solve) — is worth applying to every section of an opinion you read, even without the product. The frame forces you to ask: what is the load-bearing proposition here, and what is the Court doing with it?

Reading a Supreme Court opinion with AI is only safer than reading it with a bad research assistant if the AI refuses to replace the primary text. The tool that reads the opinion for you and hands you a summary has not helped you read the opinion; it has handed you a new document that you now have to read instead, with no way to check it against the original.

---

The question a Supreme Court opinion is always asking its reader is whether they can hold the difference between what the Court decided and what the Court said — because the law lives in the first, and the future of the law lives in the second.

Frequently Asked Questions

What is the difference between a holding and dicta in a Supreme Court opinion?

The holding is the narrowest legal rule the case actually establishes — the proposition of law necessary to produce the outcome on these specific facts. Dicta is everything the Court says that goes beyond that necessity: observations about hypothetical cases, statements of principle that were not required to decide this one, signals about where the doctrine might go. Dicta is not binding on lower courts, but it is not decorative either. It is often the Court's way of laying groundwork for a future holding or warning litigants about the limits of the one just announced.

How do I find the actual holding when reading a Supreme Court opinion with AI?

Start with the majority opinion's own language, not the syllabus. The syllabus is prepared by the Reporter of Decisions and carries no legal authority — United States v. Detroit Timber & Lumber Co., 200 U.S. 321 (1906), is explicit on this point. The holding lives in the body of the majority opinion, in the specific language the Justices chose, applied to the specific facts before them. Any AI-generated summary of the holding should be treated as a hypothesis to verify against that language, not a conclusion to rely on.

What is the Marks rule and why does it matter for reading plurality opinions?

Marks v. United States, 430 U.S. 188 (1977), established that when a fragmented Court produces no majority rationale, the holding is the position taken by the Justices who concurred on the narrowest grounds. In practice this means a plurality opinion cannot be briefed by reading only the lead opinion — the concurrences are part of the holding's definition, and skipping them produces a misstatement of what the case decided.

Why is AI hallucination especially dangerous in legal research and case briefing?

Because legal citations must be accurate in a way that philosophical or literary citations do not. A fluent AI-generated summary of a Supreme Court opinion will sound like legal analysis and will cite cases — some of which may be invented, or real cases cited for propositions they do not support. The fluency of the output is precisely what makes the errors hard to detect without returning to the primary text. The discipline is the same as it has always been: verify every citation against the reporter.

Should I read concurrences and dissents when briefing a Supreme Court case?

Yes, and not as supplementary material. Under the Marks rule, concurrences can define the holding itself when the Court is fragmented. Dissents identify the exact passages in the majority opinion where the reasoning is contested — which are also the passages most likely to be litigated in the next case. Treating concurrences and dissents as optional is treating the Court's own internal argument about the opinion as noise.

What is the syllabus in a Supreme Court opinion and is it authoritative?

The syllabus is a summary of the opinion prepared by the Reporter of Decisions before the opinion text. It is not written by the Justices and carries no legal authority. The Court addressed this directly in United States v. Detroit Timber & Lumber Co., 200 U.S. 321 (1906). Use it as a quick orientation to the case's facts and outcome, then set it aside and read the opinion body.

---

Stuck on the passage?

Scholia walks one passage at a time with the full-book context of the edition you uploaded. Open the PDF or EPUB you're reading at scholiaai.com and we'll land on the exact line you tripped on — then lift to mechanism.