Does the Constitution require Americans to accept Big Tech censorship? The claim is counterintuitive but the logic is clear: If you submit a letter to this newspaper, the editors have no legal obligation to publish it, and a statute requiring them to do so would be struck down as a violation of the Journal’s First Amendment rights. Facebook and Twitter, the argument goes, have the same right not to provide a platform to views they find objectionable.

Big Tech censorship has provoked interest in new civil-rights statutes—state laws that would bar the companies from viewpoint discrimination on their platforms and services. The First Amendment defense of this private censorship arose in a recent federal district court opinion expressing skepticism about a Florida anticensorship statute. It will come up again when other states, such as Texas, consider civil-rights statutes that focus more tightly on viewpoint discrimination.

With the possibility of multiple state statutes barring Big Tech viewpoint discrimination, it will be essential to understand the extent of the tech companies’ freedom of speech. For this, it is important to consider whether they are common carriers.

A statute limiting the ability of a Big Tech company to express its own views would almost certainly be unconstitutional. What about a law limiting viewpoint discrimination where the companies serve as a publicly accessible conduit for the speech of others?

This sort of distinction has long been ingrained in federal law—including Section 230(c)(1) of the 1996 Communications Decency Act, which distinguishes between information provided by an interactive computer service and “information provided by another information content provider.” Whatever the shortcomings of that statute, it draws a common and reasonable distinction between a company’s own speech and the speech of others for which it provides a conduit. This distinction doesn’t apply in the case of a newspaper. Its pages are not open to the public to post their views, and so it is speaking for itself when it makes editorial decisions about letters and other outside contributions.

Another reason to doubt the First Amendment claim: Antidiscrimination laws are familiar limits on speech. The U.S. has a range of local, state and federal antidiscrimination laws with significant speech consequences, and courts haven’t held that they violate the First Amendment. One has a First Amendment right to bigoted speech, but not, according to the courts, in circumstances that, for example, amount to discrimination in employment or public accommodations.

Yet another reason is that large tech platforms and services function as common carriers. The states and the federal government have the power to regulate common carriers, and this certainly includes the authority to ban discrimination. The common-carrier tradition can be traced to the common law, which viewed persons serving as common carriers as privileged by government. At the same time, it barred them from discriminating.

A similar common-law heritage regarding public accommodations underlay the Supreme Court’s decision in Heart of Atlanta Motel v. U.S. (1964), holding that Congress could adopt laws barring discrimination by private parties. The 14th Amendment empowered Congress to bar state discrimination, but it was the power over interstate commerce that let Congress prohibit private discrimination.

When a company offers its services to the public, it can qualify as a common carrier in one of two ways. It can serve a public function, so that even a small bus company can be treated as a common carrier. Or it can enjoy market dominance—when the services of one or a few companies are so prevalent as to leave the public with little alternative.

The large tech companies meet both definitions. They serve a public function, providing the public square or conduit for the information age. We meet and communicate on their services or platforms much more than on the grass of the village green.

These companies also enjoy a market dominance. Whether or not this amounts to a monopoly, it is enough to be an independent ground to recognize their services and platforms as common carriers.

One might protest that this doctrine penalizes private success. Why impose restrictions only on companies that are so innovative and efficient that they become dominant? Might that be invidious or at least irrational discrimination against the prosperous?

Not really. The point isn’t to restrict those that flourish, or that are large, but to recognize the reality that they provide publicly accessible conveyance for speech. The speech at stake is not the speech of the companies, but the speech of others that they convey. And if the physical conveyance of persons without invidious viewpoint discrimination is worthy of protection, so is the conveyance of their speech.

Nineteenth-century telegraph companies were treated as common carriers, and then telephone companies, and then some forms of television. Communications technologies change, and each is different, but the legal analysis still fits.

The recent court opinion questioning the Florida anticensorship statute noted that in censoring some speech, the tech companies are choosing what speech they will convey. From this, the court concluded that the companies’ platforms and services were more akin to newspapers than common carriers. But unlike a newspaper, these platforms and services are offered to the public for the conveyance of their speech. Unlike a newspaper, they serve the function of a common carrier. What is more, they enjoy market dominance.

That Big Tech is subject to common-carrier regulation is especially clear because Section 230 already recognizes the tech companies as akin to common carriers. Along these lines, Section 230(c)(1) protects Big Tech from being treated as “the publisher or speaker of any information provided by another information content provider.”

In other words, Section 230(c)(1) recognizes that the Big Tech platforms and services are conduits for information—that they serve the function of a common carrier. On this account, the section gives them the substantial statutory advantage of immunity from traditional publisher liability, and for a quarter-century they have not objected to it. On the contrary, they lobbied for recognition of their common-carrier-like function, and the corresponding statutory privilege, and they continue to defend it fiercely.

The problem with Section 230 is that it privileges the companies for serving the function of a common carrier without imposing the corresponding duties. This makes them uncommon carriers. They are so powerful as to avoid the burdens of common-carrier status while obtaining benefits for their role as conduits. State antidiscrimination statutes would merely impose a small portion of the common-carrier duties that Big Tech has thus far evaded.

Common-carrier status is particularly consistent with the First Amendment because the companies aren’t merely private actors. Yes, they are private. And they might protest that their dominance is simply a product of their private enterprise and superior product. But they have had profound government support, which helped secure their dominance.

The public privileging of these companies is extraordinary. Consider the fair-use doctrine in copyright law. A teacher can copy a small number of pages to show to a class. Google can copy whole books, and even when it shows only a snippet to the public, it can use the entire volumes to develop its algorithms and offer the public an online index. This appears to have been important for Google’s early enticement of the public into its services.

Section 230 privileges tech communication over print and in-person communication by excusing tech companies from liability in the courts. In contrast, paper and in-person communication is still fully subject to liability. The result has been to accelerate and accentuate tech dominance over other modes of speech.

So it isn’t true that the large tech services and platforms reached their dominance merely by private effort. Their dominance is partly the product of public privileging, and this reinforces the conclusion that tech dominance over speech is not only private enterprise. It is also the result of enterprising capture of government.

Not merely a generic subsidy for tech communication, Section 230 privatizes censorship. Section 230 relieves the large tech services and platforms of liability for restricting a congressionally specified list of materials, even when the materials are “constitutionally protected.” As one of us has explained in these pages, this is privatized censorship—a license to censor, free of concerns about ordinary legal duties that would apply to anyone else, including newspapers and individuals.

One might think it a newfangled mode of censorship, which the Constitution failed to anticipate. Not so. Rather than directly censor printing, the 17th-century English government mainly asked private entities, such as the universities and the Stationers Company (the printers trade guild), to suppress objectionable material. This was the primary example of what the First Amendment’s speech and press clauses prohibited.

The large tech companies are private, and the point isn’t that they violate the First Amendment when they censor users’ speech. But they have participated in the censorship secured by Section 230’s privileges. It therefore is not unreasonable for states to protect Americans from the tech company’s government-sponsored censorship.

The time has come to subject the Big Tech services and platforms to at least statutory barriers against viewpoint discrimination. They serve the function of common carriers or public squares and enjoy profound dominance. They achieved this dominance with substantial government privileges, including privileges they sought for serving as a conduit for information. So they can be regulated as common carriers, including with state antidiscrimination provisions. And this surely is just, for they have used their government-established dominance and privileges to enforce congressionally privatized censorship.