Reasonable Restrictions and the Categorical Approach

The costs of each

The Indian method is flexible and unpredictable. Reasonableness varies with the bench, and a doctrine that asks about prevailing conditions will produce different answers in different conditions, which is the point and also the objection.

The American method is predictable and brittle. Everything turns on which category the speech falls into, so the argument moves to definition, and a new problem that fits no category is either fully protected or forced into an ill-fitting one.

The British method sits between them. Necessary in a democratic society is a proportionality standard applied case by case, with a margin of appreciation, so it has the Indian flexibility and the Convention's structured four-stage analysis.

The comparison in one table

A worked example

A statute forbids a speech that promotes disaffection towards the government.

United States. It is content-based, and the speech is not within an unprotected category unless it is incitement under Brandenburg. Strict scrutiny applies and the statute fails.

India. Does it fall within a ground? Security of the State and public order are the candidates. Is it reasonable? The Court will look for a proximate connection with actual disorder, and will read the provision down to advocacy amounting to incitement if it can.

United Kingdom. Article 10(2), prescribed by law, legitimate aim of national security or public safety, necessity and proportionality.

The Indian and British routes can uphold a narrowed version of the statute; the American route cannot.

The cases

Facts. Shreya Singhal v. Union of India, (2015) 5 SCC 1, is also the leading case on method. Section 66A of the Information Technology Act 2000 punished grossly offensive or menacing electronic communications and false information sent to cause annoyance, inconvenience, danger, obstruction, insult, injury, criminal intimidation, enmity, hatred or ill will.

Held, by two judges. The section was struck down in its entirety. The terms used were open-ended and vague, so a person could not know what was prohibited; the section had no proximate relation to public order; and it could not be saved by reading down, because the vice went to the whole. The Court distinguished discussion and advocacy, which are protected, from incitement, which may be restricted.

Why it matters here. It shows the Indian method doing what the American method does structurally: vagueness and overbreadth defeating a statute rather than case by case balancing.

Facts. Brandenburg v. Ohio, 395 U.S. 444 (1969), concerned an Ohio criminal syndicalism statute punishing advocacy of violence as a means of political reform.

Held, per curiam. Advocacy may be proscribed only where directed to inciting or producing imminent lawless action and likely to incite or produce it; a statute punishing mere advocacy is unconstitutional.

Why it matters here. It is the clearest example of a category defined in advance: everything outside it is protected regardless of how bad it is thought to be.

Distinctions

What beginners get wrong

"Article 19(2) governs all of article 19." Clauses (2) to (6) each carry their own grounds for their own freedom.

"Reasonable means the court thinks it a good idea." It is a legal standard with settled factors: nature of the right, mischief, urgency, proportion, prevailing conditions, and procedural fairness.

"American courts balance free speech against the public interest." They do so only within the levels of scrutiny; the threshold question is categorical and expressly not a balancing exercise.

"Reading down always saves a statute." Shreya Singhal refused, because the vagueness infected the whole provision.

Limits and criticism

Enumerated grounds age. Eight grounds written in 1951 and 1963 must now be applied to platforms and algorithms.

Categories are unstable at the edges, and "true threat" and "fighting words" have both required repeated definition.

Proportionality can hide the work. A four-stage test can be recited and the conclusion reached in the last stage without the reasoning being visible.

And the deepest difference is who bears the risk of error: the American design accepts that some harmful speech will go unpunished, and the Indian design accepts that some valuable speech will be restrained.

Quick revision

1. India: a limitation clause per freedom, clauses 19(2) to 19(6), applied by a reasonableness test.
2. Factors: nature of the right, mischief, urgency, proportion, prevailing conditions; substantive and procedural.
3. Proximate connection with the ground, not remote; least restrictive alternative; proportionality in the modern cases.
4. United States: unprotected categories first, historically defined and not expandable by balancing.
5. Then tiers: strict scrutiny for content-based, intermediate for content-neutral time, place and manner.
6. Shreya Singhal: vagueness and want of proximity; no reading down where the vice is general.
7. Brandenburg: the model of a fixed category, directed to imminent lawless action and likely to produce it.

Test yourself

1. Set out the Indian test of reasonableness. There is no abstract standard; the Court examines the nature of the right alleged to be infringed, the underlying purpose of the restriction, the extent and urgency of the evil sought to be remedied, the disproportion of the imposition and the prevailing conditions. Both substantive and procedural reasonableness are examined, so an arbitrary procedure for imposing a restriction defeats it. The restriction must have a proximate connection with the enumerated ground relied on, and the modern cases apply the four-stage proportionality analysis of legitimate goal, rational connection, necessity and balancing.

2. Explain the categorical approach and its justification. Certain classes of expression are outside First Amendment coverage altogether, defined historically rather than by weighing costs against benefits: incitement to imminent lawless action, true threats, fighting words, obscenity, child pornography, defamation and speech integral to criminal conduct. Everything else is covered, and content-based regulation of it must survive strict scrutiny. The justification is that a case-by-case balance gives the speaker no notice, allows the value of speech to be assessed by the officials it criticises, and is at its weakest exactly when unpopular speech most needs protection.

3. Where do the two methods reach different results? On speech that is inflammatory but not imminent. American doctrine protects it absolutely, because it fits no category: Brandenburg protects advocacy of violence that is not directed to imminent lawless action. Indian doctrine asks whether a restriction on it is within a ground and reasonable, and a legislature may restrain it if the connection with public order is proximate. The same divergence appears with offensive speech, hate speech and blasphemy, all of which India may restrain within article 19(2) and the United States may not.

4. What does each method cost? The Indian method costs predictability: reasonableness varies with the bench and with prevailing conditions, and eight enumerated grounds are a standing invitation to legislate. The American method costs adaptability: everything turns on classification, so a new problem that fits no existing category is either fully protected or forced into an ill-fitting one, and the categories cannot be enlarged by showing that the speech is harmful. Behind both lies a choice about who bears the risk of error, the public or the speaker.