Profs. Adam Candeub & Philip Hamburger on "The Common Carrier Cure for First Amendment Uncertainty"

An interesting new twist on the argument, which particularly focuses on how social media quasi-common-carrier regulations can fight governmental pressure on platforms to censor certain material; it reminded me of Ian Samuel's The New Writs of Assistance, which similarly argued that some Big Tech companies should be legally required to limit the data they maintain about users, in order to fight governmental pressure on them to disclose it.

I'm not sure whether on balance such social media quasi-common-carrier rules are a good idea, but I thought the essay was worth passing along. (I have argued that some kinds of such rules are constitutional, and laid out some other arguments in their favor, but I can certainly see lots of arguments against them as well, some of which have been well laid out in this symposium.) I'd of course be glad to post or link to responses to this as well.

Recently leaked documents, published by Senators Grassley and Hawley, show the crucial importance of anti-discrimination requirements for Big Tech. Notably, they reveal the value of the Texas free speech statute currently under review in the U.S. Court of Appeals for the Fifth Circuit. The documents disclose planning by the Department of Homeland Security to work with its "partners" such as Twitter (and presumably the other dominant internet platforms) to monitor so-called "disinformation"—including through the infamous "Disinformation Board." This leads to the pressing constitutional question: Can the government sidestep the First Amendment by coordinating with private companies to shut down "misinformation" and "disinformation"?

The constitutional problem, interestingly, has a statutory solution. The First Amendment bars Congress from making any law abridging the freedom of speech, or of the press. But the courts have left unclear whether the Administration's coordination with the social media companies is unconstitutional. And that is why the Texas statute is so important.

The Texas statute, HB 20, finds the largest social media platforms to be common carriers—which is well justified by both their function and their size—and bars them from discriminating on the basis of viewpoint. It thus prohibits them from coordinating with the Biden administration to censor Americans in the ways described by the leaked documents. Just as AT&T cannot lawfully acquiesce to government-pressure to deny you a telephone, HB 20 makes it unlawful for the major internet platforms to go along with the Biden administration's pressure.

Government coordination with private organizations to censor individuals is not a new problem. The English government in the seventeenth century worked through the Universities and the Stationers Company (a trade guild) to censor printing. Although the government sometimes required them to do the censorship, it at times just coordinated with them.

This history reveals much about the reach of the First Amendment, because the seventeenth-century censorship was the primary example of what the amendment forbade. So there is no doubt that the First Amendment bars the government from privatizing its censorship—from using private companies to stifle dissent.

But can the Department of Homeland Security evade the First Amendment by simply avoiding force—by merely asking private companies to "cooperate." This is utterly creepy. But is it unconstitutional?

Suppose the Secretary of the Interior asked the Acme Construction Company (just as a favor to the government) to bulldoze your house. That would surely be as much a taking as if the secretary did it herself. Similarly, the government is abridging the freedom of speech when it coordinates with tech companies to silence your speech. This conclusion is suggested by the seventeenth-century history, the breadth of the term "abridging," and the spirit of the First Amendment.

But the courts have never clearly recognized this. So the administration thinks it can get away with asking for cooperation in silencing Americans without violating the First Amendment.

This failure of the amendment, as interpreted by the courts, to resolve the question is what makes common carrier anti-discrimination laws so important. By barring viewpoint discrimination by private companies, whether AT&T or YouTube, such statutes prevent government censorship through coordination.

The Texas law, for example, prohibits the largest social media platforms from engaging in viewpoint discrimination. Under this legal regime, the most massive social media platforms would have the legal basis to stand up to government pressure. Indeed, they would have to stand firm, lest the courts order them to cease their censorship.

There already is a compelling interest in recognizing the largest social media platforms as common carriers because the resulting duty against viewpoint discrimination protects freedom of debate. In addition, it limits the privatization of censorship through Section 230. But there is more. The anti-discrimination duty redresses a wide range of risks from government interference with speech, including interference that, under current precedents, is not clearly unconstitutional.

As Eugene Volokh has pointed out, current precedents permit government to urge private suppression of speech, as long as government "doesn't coerce the intermediaries by threatening prosecution, lawsuit, or various forms of retaliation." And as long as government doesn't "implicitly threaten[] retaliation, rather than simply exhorting or encouraging third parties to block speech."

The difficulty with the current state of these precedents is that, in shadow of its regulatory and contractual power, government can exert much pressure on private companies without being explicit about the threat of retaliation. It just has to hint at what it wants, and the companies will know that there may be consequences. So the First Amendment, as interpreted by current precedents, leaves a gap in the protections for freedom of speech.

One could consider this merely an evidentiary problem. But the First Amendment uses the term "abridging" rather than suppressing, forcing, or coercing. So the problem could also be viewed as a failure of current precedents to recognize the breadth of situations in which the amendment limits government. Reinforcing this point is that the Constitution was understood to give Congress no regulatory power akin to the commerce power over speech or communications.

In short, the doctrinal uncertainty in the space between government coercion and government persuasion leaves a profoundly dangerous opportunity for government to suppress speech. And the point here is that the common carrier anti-discrimination duty fills that gap.

Odysseus had himself tied to the mast to prevent him from heeding the sirens' song. Now, communications common carriers need to be tied to nondiscrimination, so they cannot conform to government's "coordination" and other subtle pressures that lie somewhere between overt coercion and mere criticism.

The companies respond that viewpoint discrimination is their freedom of speech. But information carriers have been treated as common carriers and barred from viewpoint discrimination since at least around 1700. It once was true of letter carriers on horse, it then was extended to the railroads and telegraph and telephone companies. Even to some degree to cable systems. And this has never been considered a suppression of their speech, because it merely limits them from discriminating when carrying the speech of others.

Of course, in barring viewpoint discrimination, the Texas common carrier law does not require firms to carry content they don't want on their platforms. So they can engage in any type of content discrimination—for example, by banning nudity, threats, or harassment, provided this is done without discriminating by viewpoint. And, nothing in the Texas law limits the freedom of social media companies to express their own opinions about any viewpoint or any statement their users make. They can disassociate themselves from any of their users' statements.

Even if the social media companies have a speech interest in suppressing others—which is improbable—states have a compelling interest in barring viewpoint discrimination by common carriers. This is how states since the founding have been able to make information carriers immune from government pressure for censorship. And now this is the only way to prevent the creepy and creeping censorship accomplished by "coordination."

The Texas statute, in other words, is of immense importance because it fills the gap left by a weak interpretation of the First Amendment. If government can coordinate massive private communications companies to discriminate according to its vision of "MDM"—misinformation, disinformation and malinformation—then dissent, minority viewpoints, and free speech will no longer inform public debate.

We will have returned to the time of Galileo—in politics and science. The Department of Homeland Security proposes that "Twitter become involved in Disinformation Governance Board Analytic Exchanges on Domestic Violent Extremism (DVE) and Irregular Migration"—these being code words that include discontent parents and critics of immigration policy. The Department also will counter "[d]isinformation related to the origins and effects of COVID-19 vaccines or the efficacy of masks." In other words, dissent on empirical medical questions will be stifled. And all of this will be done by providing "timely, quality information on disinformation threats … to …private sector partners." Again, creepy and unconstitutional. But absent clarity on the unconstitutionality, common carrier anti-discrimination laws are the only solution.

The documents include plans for what appears to be a meeting between Twitter executives and HHS Undersecretary Robert Silvers. Matters to be discussed will be "operationalizing public-private partnerships" to counter DVE and exploring "other ways the Department could be helpful to Twitter's counter-MDM efforts." Helpful for the government's agenda, too. But not for open debate.

Only the unprecedented dominance of a few internet firms in controlling information flows makes this public-private "partnership" possible. In earlier times, when there were thousands of independent newspapers and broadcasters, government coordination and control of information was difficult, expensive, and easily discoverable. Now, the consolidation of private communications carriers makes coordination easy for government. And the looming shadow of government retaliation (whether through regulation or the loss of government contacts) makes it wise for the companies.

Of course, it is possible that one day the Supreme Court will clarify that the First Amendment bars the government from coordinating with private entities to censor the rest of us. That would be true to the letter, spirit, and history of the Constitution. And it would be a good test of whether the allegedly Living Constitution has much life beyond personally important rights—such as those of sex, marriage, and abortion—to protect structurally important rights, such as those of speech and the press.

But unless the courts are willing to do such things soon, very soon, they need to recognize the profound interest in barring common carrier discrimination—if only to prevent government from evading the First Amendment, whether through open threats or mere "coordination."

There is a compelling public interest in barring discrimination—not least, in barring viewpoint discrimination in highly concentrated media environments. Under long-standing Supreme Court doctrine, this is dispositive as to the constitutionality of Texas HB 20 and other such anti-discrimination laws. But that's not all. The reality is that without such anti-discrimination laws, the First Amendment will become nearly meaningless—a mere formality, which government can evade simply by working through dominant private communications carriers.

Comments

[Michael Ejercito]

Conceding that laws have a tendency to have unintensded consequences due to poor wording, I support this law given what the alternative is.

[DaivdBehar]

This is going in the right direction. However, all proposals should be tested in small jurisdiction, like a state.

[Brett Bellmore]

“But the First Amendment uses the term “abridging” rather than suppressing, forcing, or coercing.”

There IS a persistent problem of the courts treating the word “infringe” as though it meant “completely abolish”.

[Lee Moore]

“abridge” and “infringe” seem to me to have subtly different connotations. Abridgement implies cutting short, or cutting a slice off, ie a lasting snip out of the fabric. Infringement implies any kind of stepping on the fringe of the fabric, even a small toe temporarily placed over the line.

This fits with the difference in the language of 1A and 2A. 1A is a prohibition on Congress passing laws – inherently a lasting snip, not a mere treading on the fringe. Whereas 2A is a a demand made to government generally, not just to Congess in its legislative function. The demand is that the government stays off your lawn altogether.

Thus one can envisage facial challenges getting a sympathetic 1 A hearing – here, look at this law, it’s sliced a chunk out of my freedom of speech ! But a cursory overstepping of the line in enforcement, a temporary as applied toe over the line – perhaps not so much.

Whereas with 2A, even a temporary as applied toe seems to be straight into infringement.

Which, roughly, seems to be pretty much the opposite of how the courts have developed 1A and 2A. 1A gets the infringement treatment – minor indiscretions by the government get corrected, 2A get the abridgement treatment – the government has to get the shears out big time, for the courts to notice, and by no means even then.

[jdgalt1]

The Texas statute, in other words, is of immense importance because it fills the gap left by a weak interpretation of the First Amendment. If government can coordinate massive private communications companies to discriminate according to its vision of “MDM”—misinformation, disinformation and malinformation—then dissent, minority viewpoints, and free speech will no longer inform public debate.

That is true but addresses only part of the problem. If the definition of MDM that these companies use to ban viewpoints is determined not by government but by a consensus of the companies, then the First Amendment ceases to apply but the effect of wrongfully stifling dissent still happens, and it is still morally imperative that those views be able to reach listeners.

What we need, I think, is not to appropriate some or all Big Social Media companies by making them common carriers, but rather to actively prevent any person, organization, or coalition from forming a monopoly in communications by creating a right for anyone to enter those fields of business and to use the common underlying facilities of the Internet, the banking system, and any other services the denial of which would enable a “consensus” group to silence dissent.

In other words, I propose a civil right to use any modern forms of communications without being subject to anything like the Chinese Social Credit Score system. Such as the one Apple and Google are trying to impose on us.

[mydisplayname]

It is true that “information carriers have been treated as common carriers and barred from viewpoint discrimination since at least around 1700. […] And this has never been considered a suppression of their speech, because it merely limits them from discriminating when carrying the speech of others.”

I am reminded of 16th and 17th century playing cards (which mush obviously be printed by someone) and the art on the face of such cards.

Is it possible to divide modern information purveyors into three groups — haulers, publishers, and chefs? The haulers have no privity to what they are hauling. The publishers take partial responsibility for what they publish: while they are not content creators, they do assert editorial control over their products and ascribe their names to them. The chefs take full responsibility over their creative concoctions made from harvested ingredients. But then what makes a publisher different from a chef; that is, what distinguishes a publisher from a (highly regulated) chef or a homebuilder?

As jdgalt1 notes above, monopoly is a factor. But even without monopoly, communications — including playing cards — are unique.

[mulched]

A single social media company can fit more than one role. The company curating, recommending, and promoting user content is the company’s own expression and should be strongly 1A protected.

Displaying a user’s content on the user’s personal page or to other users who have requested it is simply a pass-through and is very open to common carrier treatment.

[David Nieporent]

It is true that “information carriers have been treated as common carriers and barred from viewpoint discrimination since at least around 1700. […] And this has never been considered a suppression of their speech, because it merely limits them from discriminating when carrying the speech of others.”

The problem with this argument, even to the extent it’s historically and legally accurate, is that social media companies do not merely “carry the speech of others.” Never have. It might suffice to call gmail a common carrier. Google (like other email providers) holds gmail out — mostly — as serving to transmit private speech from one person to another without regard to its content. But social media companies have never done that. That’s not their role or function.

[David Nieporent]

The underlying post is a desperate attempt to eviscerate the first amendment by using a made up series of events: “Social media companies can’t be allowed to exercise their own 1A rights because in theory the government could coerce them into censoring other people.”

But there are already doctrines to address the situation in which the government does that. These people’s problem is that the things they want to prevent don’t actually fit within those categories.

[David Nieporent]

Suppose the Secretary of the Interior asked the Acme Construction Company (just as a favor to the government) to bulldoze your house. That would surely be as much a taking as if the secretary did it herself.

And this might be the stupidest analogy ever posted here, including in the comments’ section. Acme Construction Company has no pre-existing legal right to bulldoze your house. If the Acme Construction Company did this, it would be liable in tort, so it would not matter whether the government coerced it into doing so. But Twitter has every legal right to decide what speech to carry.

[Kazinski]

It’s not uncommon for states to regulate contracts, such as lending, rental, utility, etc., and proscribe what terms can be, or must be included.

I think giving people a property interest in their content and restricting the rights of social media companies to edit or delete or “throttle” the content of a users page, or Twitter thread in mandated consumer contracts or terms of use seem perfectly within the bounds of common state consumer regulations.

[David Nieporent]

No. Setting aside what on earth “a property interest in their content” even means — and how it can be squared with the Copyright Act, which actually determines the right of authors — such an interest, whatever it was, would not comprise an interest in the social media companies’ servers.

I think one aspect of either the FL or TX law — I forget which — required the social media companies to allow users to download their own content even if their accounts were suspended or deleted. That would probably fly as a business regulation.

But telling them that they must continue to distribute the content because the user “owns” it is exactly what the government can’t do.

[e34df38]

I appreciate your “property rights” view to social media content; a fecund suggestion.

[wolfefan]

Mike Dunford over on Twitter notes a couple of things, one being that this post doesn’t address the 5th amendment issues, and that Pruneyard was a 5th amendment case. I’d be interested to hear others thoughts.

[Archibald Tuttle]

That was a thoughtful and original post. Thank you.

It raises a point seldom mentioned in legal conversations. The common carrier change could be accomplished by statute, rather than in the courts, or by FCC rulings.

For example, a law saying that Twitter, FB and maybe others, shall be treated as common carriers.

[David Nieporent]

The common carrier change could be accomplished by statute,

I mean, a legislature could pass a law saying “Twitter is henceforth defined to be a common carrier,” sure. That does not mean such a law would be constitutional.

The government can’t imprison someone without due process. The government can throw stray cats in cages. That does not mean that the government can throw undesirable people in cages by passing a law declaring them to be stray cats.

[Voize of Reazon]

This leads to the pressing constitutional question: Can the government sidestep the First Amendment by coordinating with private companies to shut down “misinformation” and “disinformation”?

The intimation here is pretty shocking – that the government is planning to control private entities to restrict speech in ways the First Amendment would prevent it from doing directly. If a sidestep like that were in the works, I’d expect everyone to be concerned.

But when I read the leaked documents to see how they planned to respond to misinformation, I found this:

Response: Factually countering disinformation through public communications channels to mitigate related threats, increase awareness, and improve public safety.

Apparently what the essay authors see as a “sidestep” is that the government will listen for what it considers misinformation, and counter it with the government’s own speech. These are activities it is entitled to do directly since they don’t involve censoring or “shutting down” anyone else’s speech.

I don’t know whether Profs. Candeub and Hamburger sounded the alarm because they misread the papers or are relying on undisclosed evidence or just let their imaginations get away from them.