learnfun Books

learnfun Books · Essay

How to Think Like an Appellate Judge

By learnfun editors

Most people think judges decide who wins.

Trial judges and juries mostly do.

Appellate judges do something stranger. They decide what the rule is, who gets to decide the facts, and what the next thousand cases will look like.

The winner is almost a side effect.

If you learn to think this way, you'll read cases differently, argue differently, and write differently. You'll stop asking "who's right?" first. You'll ask four other questions before you get there.

Here's the ladder. Five levels. Most people stop at Level 1.

"It is emphatically the province and duty of the judicial department to say what the law is." (Chief Justice Marshall, Marbury v. Madison, 1803)

Level 1: Who should win?

This is where everyone starts. You read the facts, you feel something, you pick a side.

That instinct isn't useless. Judges have it too. But it's the least interesting thing an appellate court does, and it's the most dangerous place to stop.

Here's why. An appeal is not a do-over. The trial already happened. Witnesses testified, a jury or judge weighed them, and a judgment came out. The appellate court reads a cold record: transcripts, exhibits, briefs. It usually can't hear a witness, watch a face, or take new evidence.

So "who should win" is the wrong first question. The court isn't asking who it would have picked. It's asking whether something went wrong in a way the law lets it fix.

Hold your instinct. Write it down if you want. Then put it aside and climb.

"General propositions do not decide concrete cases." (Justice Holmes, dissenting, Lochner v. New York, 1905)

Level 2: What, exactly, is the question?

Great appellate thinking starts with a narrow, precise question.

Not "was the railroad negligent?" but "did the railroad owe this plaintiff a duty at all, given what its employees could foresee?"

That's the move Judge Benjamin Cardozo made in Palsgraf v. Long Island Railroad Co. (N.Y. 1928). Railroad guards helped a man onto a moving train. He dropped a package. It held fireworks. They exploded, and something at the far end of the platform fell on Mrs. Palsgraf.

You can ask a hundred questions about that case. Cardozo asked one: was harm to her within the range of risk the guards' conduct created? His answer turned on a sentence lawyers still quote:

"The risk reasonably to be perceived defines the duty to be obeyed."

Notice what framing did. Once the question was "duty to whom?", the outcome followed. The dissent framed it differently, as a question of how far liability should reach after a careless act, and got a different answer.

Your move: before you argue anything, write the question in one sentence. Then write the other side's version. If they're different, the case is really about which framing wins.

Level 3: Who gets to decide?

This is the level most non-lawyers have never heard of, and it decides more appeals than any other.

Every issue on appeal comes with a standard of review: how much the appellate court defers to the decision below.

Roughly:

The Supreme Court put the clear-error idea plainly in Anderson v. City of Bessemer City (1985): if the trial court's account of the evidence is plausible in light of the whole record, the court of appeals may not reverse it, even if it would have weighed the evidence differently. (470 U.S. 564, 573-74.)

Read that twice. An appellate judge can disagree with the result and still affirm. That's not weakness. It's the design. The trial judge saw the witnesses. The appellate judge didn't.

So the real question at Level 3 is: which box is this issue in? Lawyers fight hard over that, because the box often decides the case. Recast a fact question as a legal question and you get fresh review. Recast a legal question as a fact finding and you get deference.

Your move: for every issue, name the standard of review before you name the answer. If you can't, you don't understand the issue yet.

"The life of the law has not been logic: it has been experience." (O. W. Holmes Jr., The Common Law, 1881)

Level 4: What rule does this create?

Now we're above the case.

When an appellate court decides, it doesn't just resolve a dispute. It writes a rule that binds lower courts and guides everyone who reads it. Its reasoning becomes precedent.

That changes how a good judge thinks. The question isn't only "is this outcome right here?" It's "what rule am I announcing, and can I live with it in every case it will cover?"

Two habits come from that.

Decide on the narrowest ground that resolves the case. If a case can be decided on a statute, courts generally avoid reaching a constitutional question. If it can be decided on one element, they don't need to rule on all five. Narrow rulings make fewer mistakes, because they bind fewer future cases.

Respect what's already settled. Precedent (the doctrine called stare decisis) isn't sacred, but it's heavy. Justice Brandeis explained why in a dissent that is now cited constantly: "Stare decisis is usually the wise policy, because in most matters it is more important that the applicable rule of law be settled than that it be settled right." (Burnet v. Coronado Oil & Gas Co., 1932.)

That's a surprising sentence until you think about who relies on rules: people writing contracts, planning businesses, deciding whether to sue. A stable rule that's slightly wrong can serve them better than a "perfect" rule that changes every few years.

Your move: after you reach an answer, write the rule your answer implies in one sentence. Then test it on three hypotheticals: an easy one, a hard one, and a weird one. If the rule breaks on the weird one, narrow it.

"The prophecies of what the courts will do in fact, and nothing more pretentious, are what I mean by the law." (O. W. Holmes Jr., "The Path of the Law," 1897)

Level 5: What happens next?

The top of the ladder is prediction.

In "The Path of the Law," Holmes told students to look at law the way a "bad man" would: someone who doesn't care about morals, only about what a court will actually do to him. Holmes wasn't praising that person. He was giving lawyers a test. Strip away the moral language, and what's left is a forecast of consequences.

An appellate judge thinks about consequences in two directions.

Downstream: how will trial judges apply this rule tomorrow morning? Will it be clear enough to follow? Will it produce more appeals or fewer?

Outward: how will people change their behavior once the rule exists? Holmes urged lawyers to study history and economics for exactly this reason. A rule is an incentive, whether or not the court meant it as one.

This is also the level where honest judges admit uncertainty. Cardozo's 1921 Yale lectures, The Nature of the Judicial Process, are worth reading for that candor. He describes reaching, in hard cases, for logic, history, custom, and judgments about social welfare, and he doesn't pretend the method is mechanical.

Your move: before you finalize any argument, write two sentences: "If I win, trial courts will now..." and "If I win, people will now...". If either sentence scares you, a judge will feel it too.

Back to Level 1

Here's the twist. After all five levels, you come back to "who should win?"

But now it's a different question. You know what's actually being asked. You know how much deference the decision below gets. You know what rule your answer creates and what it will do in the world.

Often your first instinct survives. Sometimes it doesn't. Either way, you've earned the answer instead of assuming it.

That's the whole method. It isn't genius. It's order.

The ladder in one screen

  1. Who should win? Notice your instinct. Don't trust it yet.
  2. What's the question? One sentence. Then the other side's sentence.
  3. Who decides? Name the standard of review for every issue.
  4. What rule? Write it in one line. Test it on three hypotheticals. Go narrow.
  5. What happens next? Forecast trial courts tomorrow and people next year.

Try it today (15 minutes)

Pick one published appellate opinion you haven't read. Palsgraf is a good start, and it's short.

Before you read the majority, read only the facts. Write your Level 1 instinct.

Then read the opinion and fill in Levels 2 through 5 as the court did. Then read the dissent and fill them in again.

You'll find the two opinions usually disagree at one level, not all five. Finding that level is the skill.

General information about legal reasoning, for education. Not legal advice. Made with AI tools; every quote checked against the linked source before publishing.

Sources