First Amendment

A sidewalk in Rochester

Walter Chaplinsky was handing out literature for the Jehovah’s Witnesses on Wakefield Street in Rochester, New Hampshire, near the entrance to City Hall. A crowd gathered and an officer took him toward the police station. On the sidewalk Chaplinsky told the city marshal that he was a racketeer and a fascist, and that the government of Rochester was made of fascists or their agents. A New Hampshire statute forbade addressing an offensive, derisive, or annoying word to anyone lawfully in a public place. Chaplinsky was convicted under it, and in 1942 the Supreme Court affirmed without a dissent.

The interesting part is how the Court got there. It did not weigh the insult against the marshal’s dignity. In Chaplinsky v. New Hampshire, 315 U.S. 568 (1942) (opens in a new tab) it said the right of free speech has never been absolute. Some speech falls into “well-defined and narrowly limited classes” the government has always been free to punish, and insulting or fighting words were one of them. So the case was decided by sorting rather than by balancing. That is still how most First Amendment cases work, and this reading walks the sort in order. The activity has you make an AI produce the history of one clause and then audit it, which goes faster once you know which step a summary skipped.

First question: is the law aimed at what was said?

Gilbert, Arizona regulated outdoor signs by what they said. Its sign code gave ideological signs one set of limits, political signs another, and signs pointing people toward a nonprofit event a third. Clyde Reed’s Good News Community Church rented space for Sunday services and put out signs telling people where to go. Those signs could be no larger than six square feet, no more than four on a property, and up no more than twelve hours before the service and one hour after. An ideological sign the same size faced none of those restrictions. To know which rule governed a sign in Gilbert, you had to read the sign.

That is what content neutrality forbids. A law is content-based when it draws its lines by subject matter or message, and in Reed v. Town of Gilbert (2015) the Court said such laws are presumptively unconstitutional. The government may still defend one, but only by proving that it is narrowly tailored to serve a compelling interest. Gilbert could not.

Now change the ordinance. Suppose no sign in the residential district may exceed six square feet, whoever posts it and whatever it says. An official enforcing that rule never has to read the sign. It is a time, place, and manner restriction, and it draws a gentler test, because it leaves every message equally free to compete. The neutrality question comes first for a practical reason. Its answer picks the test, and the test usually picks the winner.

Second question: how hard does the government have to work?

The levels of scrutiny are the three settings a court can use. Strict scrutiny demands a compelling interest and a law narrowly tailored to serve it. Intermediate scrutiny asks for an important interest and a law that substantially serves it. Rational basis asks only for a legitimate interest and a reasonable fit.

These are not degrees of judicial enthusiasm. They are allocations of the burden of proof. Under rational basis the challenger has to show the law is irrational, which is nearly impossible. Under strict scrutiny the government has to make its own case from scratch against a presumption that it will lose. Naming a restriction content-based is what moves that burden across the table, and it is why so much First Amendment argument is really argument about classification.

You will meet this framework again. Activities 9 and 11 apply it to due process and equal protection rather than rebuilding it, so the vocabulary you learn here is the vocabulary those pages assume.

What sits outside, and what counts as speech at all

Chaplinsky named the classes of unprotected speech. Later Courts shrank almost every one of them, and that pattern is worth knowing, because an AI summary will often hand you the old, wider version.

Advocacy of violence is the clearest example. Brandenburg v. Ohio (1969) held that a state may not forbid advocacy of force or law violation unless the advocacy is directed to inciting imminent lawless action and is likely to produce it. Both halves matter: intent and likelihood, and both aimed at something imminent. That test displaced the looser clear-and-present-danger approach the Court had used since Schenck v. United States (1919), which had let the government punish words for their tendency to obstruct the draft.

The other boxes narrowed the same way. Miller v. California (1973) confined obscenity to three findings made together. Taken as a whole, the work must appeal to the prurient interest under contemporary community standards and lack serious literary, artistic, political, or scientific value. It must also depict sexual conduct specifically defined by state law, in a patently offensive way. New York Times Co. v. Sullivan (1964) did the same for defamation. A public official suing over a falsehood about his official conduct must now prove the statement was made with knowledge of its falsity or reckless disregard for the truth. And in New York Times Co. v. United States (1971) the Court said any system of prior restraints arrives bearing a heavy presumption against its validity. That is why the government almost never stops a publication before it happens.

The category system also runs in reverse. Conduct with no words in it can be speech, and symbolic speech is the doctrine that covers it. Tinker v. Des Moines (1969) protected students wearing black armbands, because the school could point to nothing that would materially and substantially interfere with its discipline. Texas v. Johnson (1989) protected a burning flag, because a government may not forbid an idea for being offensive.

Two more rules turn on who is speaking and where. Garcetti v. Ceballos (2006) held that when public employees make statements pursuant to their official duties they are not speaking as citizens, so the First Amendment does not shield those statements from employer discipline. Nieves v. Bartlett (2019) held that a plaintiff claiming an arrest was retaliation for speech generally has to plead and prove there was no probable cause for it. Neither rule changes what speech is; both change who can bring the claim.

The religion clauses run the same sort

The two religion clauses use the same habit of mind, and each has a sorting question of its own.

For the Establishment Clause, courts spent decades running the test from Lemon v. Kurtzman (1971): a secular purpose, a principal effect that neither advances nor inhibits religion, and no excessive entanglement between government and religion. In Town of Greece v. Galloway (2014) the Court upheld a town board’s practice of opening meetings with prayer and said the clause must be interpreted by reference to historical practices and understandings. Then came Kennedy v. Bremerton School Dist., 597 U.S. 507 (2022) (opens in a new tab). The Court said plainly that it had abandoned Lemon and the endorsement test that grew out of it, and it protected a high school football coach who knelt alone at midfield to pray after games. So the first thing to check in any Establishment Clause summary is which test the writer is using.

The Free Exercise Clause has its own break. Sherbert v. Verner (1963) and Wisconsin v. Yoder (1972) required a state to show a strong interest before it could substantially burden religious practice. On that reasoning Yoder freed Amish families from compulsory schooling past the eighth grade. Then Employment Division v. Smith, 494 U.S. 872 (1990) (opens in a new tab) changed the default. The right of free exercise, the Court held, does not relieve anyone of the obligation to obey a valid and neutral law of general applicability.

That moved the whole fight onto two words. Church of Lukumi Babalu Aye v. City of Hialeah (1993) shows what happens when a law fails them. Hialeah’s ordinances reached ritual animal sacrifice and left almost every other killing of animals alone, so the Court found them neither neutral nor generally applicable, applied strict scrutiny, and struck them down. A law that is genuinely neutral and genuinely general stands under Smith; a law that only looks that way goes to the top tier and usually falls.

There is a footnote in Chaplinsky that ties the two halves of this reading together. The Court noted that even if his literature distribution counted as religious activity, it would not immunize what he said afterward. The clauses cover different questions, and a case can raise one without raising the other.

Taking this into the activity

You now have the order of operations. Is the law aimed at content, which tier follows from that, and does the speech fall in or out of a historic category? For religion, is the law neutral and generally applicable? The activity has you pick one clause, run the Clause Historian for its history and landmark cases, and then check the result.

Two errors are worth watching for. Models attach the right test to the wrong case, and they state a holding wider than the Court went. Both are easy to catch once you know the sequence. The Constitution Annotated essays on the First Amendment (opens in a new tab) will settle most of them in a minute. Oyez (opens in a new tab) will get you oriented on a case, but the holding is confirmed in the opinion itself. Start with the tests. That is where the errors cluster.