Free Speech

Compelled Nondiscrimination in the Subscription Function

|

Another excerpt from the First Amendment section of my Social Media as Common Carriers? article (see also this thread); recall that the key First Amendment arguments are  in this post, which relies on the PruneYardTurner, and Rumsfeld precedents, and in this one, which explains why Miami HeraldHurley, and the various other "common theme" precedents don't apply.

[* * *]

Of course, social media platforms do much more than letting people visit a particular page or view a particular video. Among other things, they let people subscribe to others' materials, so that all or some of those materials appear in the subscribers' feeds. (This is the "follow" feature on Twitter and Instagram, the "subscribe" feature on YouTube, and, more or less, the "friend" feature on Facebook.) Indeed, this subscription function in large measure distinguishes social media from mere user-generated content.

Could the government bar platforms from discriminatorily declining to show subscribers certain materials posted by users to whose feeds they have subscribed? I'm inclined to say that it can, though the argument there is more complex than it is for the pure hosting function.

The key precedent on this in Rumsfeld. We've discussed above how in that case, like in PruneYard and Turner, the Court held that property owners (there, universities) could be required to host speakers they disliked. But say that a university told recruiters: "OK, we have to let you on our property, so you can be in Room 217. But we won't speak to anyone about your being in Room 217: We won't include this in any printed materials where we list all the recruiters, and we won't send out the e-mails with information about you the way we do about other recruiters. Good luck with students finding you!"

Rumsfeld held that this too is constitutionally unprotected: As part of requiring universities not to discriminate against military recruiters in choosing whom to host on its property, the university could also be required not to discriminate in choosing whom to inform students about.

The compelled speech to which the law schools point is plainly incidental to the Solomon Amendment's regulation of conduct, and "it has never been deemed an abridgment of freedom of speech or press to make a course of conduct illegal merely because the conduct was in part initiated, evidenced, or carried out by means of language, either spoken, written, or printed." Giboney v. Empire Storage & Ice Co., 336 U.S. 490, 502 (1949). Congress, for example, can prohibit employers from discriminating in hiring on the basis of race. The fact that this will require an employer to take down a sign reading "White Applicants Only" hardly means that the law should be analyzed as one regulating the employer's speech rather than conduct. See R.A.V. v. St. Paul, 505 U.S. 377, 389 (1992) ("[W]ords can in some circumstances violate laws directed not against speech but against conduct"). Compelling a law school that sends scheduling e-mails for other recruiters to send one for a military recruiter is simply not the same as forcing a student to pledge allegiance, or forcing a Jehovah's Witness to display the motto "Live Free or Die," and it trivializes the freedom protected in Barnette and Wooley to suggest that it is.[212]

The record in the case suggests that this means military recruiters could likely have to be included in "recruiting receptions,"[213] and that "an unwilling institution" would have to "distribute, post, and maintain the military's literature, send emails promoting the military, include the military's listing in printed publications, and make introductions and arrange meetings."[214]

Now as with most discussions of the Giboney "conduct … carried out by means of language" doctrine, this analysis is quite opaque;[215] and the analogy to the "White Applicants Only" signs is not entirely helpful, because it involves speech restrictions rather than speech compulsions. But the underlying principle does indeed arise in a vast range of antidiscrimination rules.

Private schools and private universities, for instance, have broad First Amendment rights to speak, including for instance to "promote the belief that racial segregation is desirable."[216] Nonetheless, the government may require them not to do discriminate based on race in admitting students.[217] And given that this is so, it must be equally permissible for the government to compel the schools and universities to speak to the students in ways necessary to give those students equal access to the education—to hand students schedules of classes, homework assignments, grades, feedback on papers, and the like. Likewise, the schools and universities could be required to speak to third parties about their students without regard to race, for instance by sending copies of transcripts to anyone who asks. (Whether the schools or teachers could be required to provide letters of recommendations, which include opinions about the qualities of the student, is a harder question.[218])

Nor is this limited to bans on discrimination based on race, religion, and the like. Say a phone company tells the Socialist Party, "we understand that we have to allow you to use our property for your evil anti-private-property speech, but we can't be compelled to speak to you, so we won't inform you of the phone number that we have selected for you." Surely common carriage laws can forbid that, and require the phone company to communicate information equally to all customers. Likewise, such laws can require the phone company to equally communicate information about all customers, for instance in its telephone directories.[219]

The same, I think, may apply for platforms' nondiscrimination obligations with regard to their users: If I'm right that platforms can be required to host all users' speech (or at least not to discriminate based on viewpoint), they might also be required to provide that speech to the users' subscribers, on a viewpoint-neutral basis. Under that approach, if someone goes to, say, http://twitter.com/VolokhC, or follows @VolokhC while viewing Tweets in reverse chronological order mode,[220] Twitter would have to show them the posts from my Volokh Conspiracy blog, without blocking any based on viewpoint.

Now the platforms might reasonably want to provide subscribers some subset of all the posts from accounts to which they are subscribed. If a Twitter user is following 500 Twitter accounts, for instance, perhaps Twitter might want to show not every single post from those accounts (which in practice would just mean that the follower would see only the few most recent posts), but some subset, for instance the most retweeted posts.

Yet Rumsfeld suggests that the law could require that any such screening be done in a suitably neutral way (e.g., without discrimination based on viewpoint or based on whether the platforms views certain claims expressed in a post as accurate). By way of analogy, say that a university was sending out a special e-mail about recruiters who are hiring people at starting salaries of over $100,000, or was conducting a job fair for students interested in public interest impact litigation. The statute upheld in Rumsfeld (the Solomon Amendment) wouldn't require the university to include military recruiters there, because they wouldn't fit the neutral criteria for that particular mailing: The Solomon Amendment merely forbids discrimination against military recruiters, rather than compelling the inclusion of recruiters in every item of speech related to recruiting. Likewise, a requirement that platforms not discriminate based on certain criteria in implementing users' subscriptions may be valid even if it leaves platforms free to use other criteria.

To be sure, Rumsfeld isn't a perfect analogy here. (That's why it's just an analogy.) The Solomon Amendment, for instance, prohibited universities from discriminating against military recruiters even in its speech to students as a whole, not just in speech to students who had expressed an interest in military jobs. My analysis here is limited to discrimination in implementing subscriptions, where the recipients of the speech had expressly asked to be shown material from certain users.

Still, I think Rumsfeld is important because it reaffirms that the government may require that platforms (whether universities or social media platforms) provide speech to listeners on a nondiscriminatory basis. Such a requirement, when coupled with a mandate that a platform provide space for speakers on a nondiscriminatory basis, doesn't violate the First Amendment's prohibition on compelled speech.

[212] 547 U.S. 47, 62 (2006) (cleaned up).

[213] Rumsfeld v. FAIR, 390 F.3d 219, 240 (3d Cir. 2004) (discussing such compelled inclusion), rev'd, 547 U.S. 47 (2006).

[214] Reply Brief for Appellants, FAIR v. Rumsfeld, No. 03-4433, 2004 WL 5215238, *5–(3d Cir. Mar. 5, 2004) (citations omitted).

[215] Eugene Volokh, The "Speech Integral to Criminal Conduct" Exception, 101 Cornell L. Rev. 981, 997 (2016).

[216] Runyon v. McCrary, 427 U.S. 160, 176 (1976). The Court stated that "it may be assumed that parents have a First Amendment right to send their children to educational institutions that promote the belief that racial segregation is desirable, and that the children have an equal right to attend such institutions." But while this is cast in light of the rights of parents and children, it presupposes a comparable right on the part of schools, including their rights to engage in "the '[e]ffective advocacy of both public and private points of view, particularly controversial ones' that the First Amendment is designed to foster." Id. at 175.

[217] Id.

[218] See Eugene Volokh, Professor Refuses to Write Letters of Recommendations for Creationists, Volokh Conspiracy (Jan. 30, 2003, 11:‌46 am), https:‌‌//‌perma.cc/‌73K2-TELJ#90255714.

[219] See, e.g., US W. Commc'ns, Inc. v. Hix, 93 F. Supp. 2d 1115, 1131 (D. Colo. 2000) (applying the Telecommunications Act of 1996, 471 U.S.C. § 251(b)(3), which requires that telephone companies provide "nondiscriminatory access to 'directory listings,'" including to customers of competing telephone companies); US W. Commc'ns, Inc. v. Colo. Pub. Utils. Comm'n, 978 P.2d 671, 674 (Colo. 1999) (likewise, with regard to similar state law requirement).

[220] See supra note 101.

NEXT: A Modest Proposal: Livestream SCOTUS Oral Arguments With 7-Second Tape Delay To Account for Disruptions

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of Reason.com or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. EV’s several posts suggest strongly that comon carrier status would stand legal scrutiny in the federal courts.
    WHile I have read many arguments pro and con about the wisdon of this approach, I have read liitle about the ways in which common carrier status has actually damage present common cariers or the public at large.
    As an experimentalist, I’s like to see hard evidence about those negative effects.

    1. It’s vile, but if the courts will refuse to deal with threats to the First Amendment via threats to change section 230, delete it, or break these companies up unless they ban harrassment, oh, look, our political opponents’ posts are harrassing, nay, dangerous! then some retaliatory, fight fire with fire BS like this may be the only option.

      These are politicians and power hungry scum looking out for themselves, Constitution and Freedom itself be damned. But I repeat myself.

      1. I wish there were more reasonable comments on this blog. Oh well, muted.

        1. IP, yet another lawyer subject in utter failure, crushing innovation, and empowering encrusted monopolies.

          1. It’s only success, of course, is your $600/hour fee.

    2. EV’s several posts suggest strongly that comon carrier status would stand legal scrutiny in the federal courts.

      EV’s posts are primarily about Rumsfeld — which is so inapposite that the fact that his argument relies on it shows how weak it is — Turner, and PruneYard, none of which are common carrier cases.

      1. You go to press with the authorities you have, not the authorities you would like to have.

      2. If courts have upheld those restrictions for companies that are not common carriers, doesn’t that make it even more likely that they would support such restrictions on common carriers?

        1. Michael P: Indeed — as I note in my article, PruneYard, Turner, and Rumsfeld are the key First Amendment cases that support the constitutionality of certain hosting compulsions. Such compulsions may apply to common carriers, or, as in those cases, non-common-carriers.

          As to Rumsfeld, I think it’s pretty relevant here, though I think the combination of the three cases does more work than any one of them. I’m not sure, David, why you think it’s so inapposite; if it’s just because it involved military recruiting, I don’t think the First Amendment analysis is any different for regulations of civilians aimed at promoting military recruiting than for other regulations of civilians. And indeed Justices have routinely cited Rumsfeld since then in cases that had nothing to do with the military. See, e.g., Nevada Com’n on Ethics v. Carrigan; Davenport v. Wash. Ed. Ass’n.

          1. Rumsfeld is inapposite because the “compelled hosting” in that case involved permitting the military to set up a booth at a career fair. It didn’t involve compelling a university to accept military recruitment content and to affirmatively distribute it to their own students in whatever way they distributed other recruiting materials.

          2. I don’t deny that Rumsfeld is a valid decision, so the fact that the Supreme Court cites it from time to time doesn’t really respond to my point. I deny that Rumsfeld is more than distantly relevant to this discussion.

            To address your first point, the Rumsfeld decision expressly argued that the military aspect of the law was important:

            The Constitution grants Congress the power to “provide for the common Defence,” “[t]o raise and support Armies,” and “[t]o provide and maintain a Navy.” Art. I, § 8, cls. 1, 12-13. Congress’ power in this area “is broad and sweeping,” O’Brien, 391 U. S., at 377, and there is no dispute in this case that it includes the authority to require campus access for military recruiters. That is, of course, unless Congress exceeds constitutional limitations on its power in enacting such legislation. See Rostker v. Goldberg, 453 U. S. 57, 67 (1981). But the fact that legislation that raises armies is subject to First Amendment constraints does not mean that we ignore the purpose of this legislation when determining its constitutionality; as we recognized in Rostker, “judicial deference . . . is at its apogee” when Congress legislates under its authority to raise and support armies. Id., at 70.

            So I think it a bit misleading to simply handwave that aspect away.

            But the larger point is illustrated by the cases you mention. Rumsfeld (as Carrigan cites it for) held that discriminating against military recruiters wasn’t speech at all, but conduct, with no direct expressive content at all. It then went on to analyze (as Davenport touches on) the Solomon amendment as an expressive association claim, mentioning two considerations; whether the law compels the admission of undesired members, and whether the law makes membership less attractive to others. I don’t think that consideration of either of those factors would actually help the chances of a must-carry law for FB/Twitter, but there’s no need to analyze them at all, because FB/Twitter are raising a compelled speech claim, not a freedom of association claim. A law that said that Twitter must allow everyone to have an account — but not to post! — might pass constitutional muster but would not address the speech issues at play here.

        2. Eugene has failed to make the argument that social media ISPs are or should be considered to be common carriers. Or, at any rate, he has failed to do so in a way that wouldn’t sweep up any other online content provider.

          Facebook is as much a “common carrier” as Reason is, for our comments.

      3. Are you sure you’re a lawyer?

        And, what happened to you? Once, your arguments would show insight or at least be informative in some way. Even if I didn’t find them convincing, they were worth reading. But now, it’s all vacuous, knee-jerk Twitter-y responses. Do you actually find this fun or edifying in any way?

        1. If you are referring to Mr. Nieporent, an exchange a few comments below this one is mighty inconvenient for you.

          Carry on, clingers.

        2. Yes, his comments are vacuous, now that you disagree with his conclusions. Funny how that happens.

  2. I would say that when a social media platform carries messages between posters and subscribers, it is much more clearly resembling a telephone company or mail carrier than when it transmits messages to the entire general public. What is doing is no different from transmitting an email sent to a group of recipients.

    Congress can pass the equivalent of wiretapping and mail tampering laws flatly prohibitting the current business model – not just requiring social media platforms to carry all comers, but prohibiting them or anyone not on the identified recipient list from listening in on transmissions for any purpose, including passing them on to advertisers, without a warrant or similar.

  3. Rumsfeld held that this too is unconstitutional:

    I think you mean that Rumsfeld held that the government could ban this.

    1. Whoops, yes, thanks, meant to say constitutionally unprotected — fixing here, and will fix in the draft.

  4. May I suggest the following?
    Twitter free Tuesdays.
    WeChat free Wednesdays.
    TikTok free Thursdays.
    Facebook free Fridays.
    Snapchat free Saturdays.
    Etc.
    Slowly wean the addicted public off (anti-)social media.

  5. Here’s an article about an incident where some black customers alleged that a bar wouldn’t let them sing Karaoke because of their race.

    I don’t know how it turned out, but the ACLU sided with the customers, not the bar owners. The bar denied that it discriminated, of course.

  6. “And given that this is so, it must be equally permissible for the government to compel the schools and universities to speak to the students in ways necessary to give those students equal access to the education—to hand students schedules of classes, homework assignments, grades, feedback on papers, and the like.”

    Presumably the government can also compel the universities to let all students participate in class discussion, to listen to all students, and to require other students to listen to their views, without discriminating based on race, etc.

    1. Why does it follow that if the government gets to prohibit a phone company from tapping your phone, listening on your conversation, selling its contents to advertisers, and deciding whether your conversation sufficiently follows its line as to be allowed to continue — all of which a phone company has long been prohibited bu well-settled law from doing – then a university can’t control the contents of its courses? The two have always been regarded as quite different. And a company that whose business is drlivering user-created general messages to other users much, much more resembles a phone company than a university.

      Such a company doesn’t do any speaking at all. It merely supplies a utility. Should the electric company get to control your speach merely because it happens to provide a service that makes transmission of your speech possible?

      1. “Why does it follow that if the government gets to prohibit a phone company from tapping your phone, listening on your conversation, selling its contents to advertisers, and deciding whether your conversation sufficiently follows its line as to be allowed to continue — all of which a phone company has long been prohibited by well-settled law from doing – then a university can’t control the contents of its courses?”

        It doesn’t necessarily follow. But the OP says that universities can be required to admit students that they don’t want to admit, and presumably that means providing a platform for their speech.

        And in California, the Leaonard law requires some private universities to provide a platform for speech they don’t want to.

        I’m not sure what the distinction is between compelling private universities to provide a platform for speech they don’t want to provide a platform for, and compelling social media to do the same thing.

        Of course, maybe the answer is that Universities shouldn’t be compelled to provide such a platform either.

  7. kiwifarms, basically a gossip site for internet figures has just been deplatformed from their host due to an unconfirmed twitter rumor that somebody committed suicide because of them. This is the latest a long line of host deplatformings and the services continue to consolidate making finding new hosts more and more difficult. They started with white supremacist groups but have been gradually expanding their verboten list over time. So this stuff extends beyond social media platforms.

  8. You wrote an entire paper about how Google’s editorial judgement about which results to include (among many possible results) is fully protected first amendment speech.

    Yet now, you are arguing that selecting which results to include in a subset of a feed (among many possible results) is not protected at all. It would seem to me that if social media companies have full first amendment protection for their “recommendation function” (as you argue in your latest paper, and as your previous paper implies), the same argument would equally apply to selecting the results to show in a subset of a user’s feed (if the user chooses to enable the subset functionality).

    http://volokh.com/wp-content/uploads/2012/05/SearchEngineFirstAmendment.pdf

    1. (And by a user’s feed, I mean the subset of posts from all posts from people the user has subscribed to.)

  9. I suspect you are wrong here Eugene, in the main.

    The better analogy would be, what if Yale held a “preferred recruiters” reception, which highlighted the recruiters that Yale wished to promote based on shared values? (Assume also that it’s obviously selective, that is, it doesn’t include all non-military recruiters.)

    Such a reception is clearly curated and editorial in nature. I find it hard to believe that SCOTUS would require Yale to invite the military. (I feel like this was almost explicit in the Rumsfeld decision, iirc.)

    Similarly, to the extent a social media site shows users a curated, editorial selection of posts on their aggregated feed, the First Amendment favors honoring those editorial decisions.

    I think you may be right that as a common carrier, social media could be required to show all (legal) posts in some form, such as if a user selected a specific poster’s channel to view. In the context of their own channel, the poster is making the editorial decisions, not the host. After all, the right to speak is pretty meaningless if the host prevents anyone from listening.

    You seem to sort of agree with all that, but then pivot hard on the user action of “subscribing.” I don’t see what relevance that has to the host’s First Amendment editorial rights regarding a curated, aggregated feed. As long as the host provides a way to see the posts, how does subscribing change the constitutional analysis?

    1. You talk about a “curated” feed, but Prof. Volokh was (I think) talking mostly about user-driven selections: subscriptions to updates, or search results. For the same reasons that it wouldn’t be permissible to delete someone’s posts, it would not be permissible to hide them from interested parties through mechanisms that would normally show them.

      1. Yes, you are correct that Eugene’s arguments by analogy proceed largely by mischaracterizing how Facebook and Twitter works.

      2. He describes curated feeds pretty accurately actually:

        “Now the platforms might reasonably want to provide subscribers some subset of all the posts from accounts to which they are subscribed. If a Twitter user is following 500 Twitter accounts, for instance, perhaps Twitter might want to show not every single post from those accounts (which in practice would just mean that the follower would see only the few most recent posts), but some subset, for instance the most retweeted posts.”

  10. For once I’d like to see an example where we FIRST ask what would be best, and then figure out what law and legal traditions will need to be changed to reach that goal. It seems we are forever starting with the horse and not the cart.

    Or maybe that’s a discussion for a different blog.

    1. I think it would be best if a political party wasn’t picking and choosing what was allowed to be spoken of on monopolistic platforms. I’m not sure anything needs to be changed to rule this is government interference in speech.

      1. You figure the Democrats are instructing Facebook, Twitter, Google, and others in this regard?

        This seems to be the point at which being a birther-class, QAnon-curious, stolen election’ kook and roundly bigoted, downscale right-wing loser interferes with sound judgment.

        1. or CNN

          “The chief executives of Facebook, Google and Twitter faced withering criticism from members of Congress on Thursday about their handling of misinformation and online extremism during their first appearance before lawmakers since the Jan. 6 Capitol riots and the rollout of the coronavirus vaccine.”

          1. Oh, no! Withering criticism! Whatever shall they do?!? Besides show up for the made-for-tv event and then forget about it 30 seconds later, I mean?

            To be sure, I think such criticism is improper. And to be sure, if a government official with power to punish them used that power to coerce them to censor someone, that would be a 1A violation. But criticizing them is not coercing them.

            1. “ But criticizing them is not coercing them.”

              Ok. Is it ‘instructing them’?

              1. Only if you’re as paranoid as Brett.

                No, there isn’t a secret coordination between big tech and the US government to suppress the Real Truth.

                1. That’s a nice store you’ve got there, be a shame if something happened to it.

        2. The Democrat Party is owned by the tech billionaires. They said, no. We want Biden, a dotard. They got him. They said, Harris, with zero popularity. The got her. She is from the San Fran area, and represents their interests, may soon replace Biden.

      2. What will you be whining about when you’re proved stupid by Trump’s lawsuits getting laughed out of court? Probably that the judge was afraid of angry black people or court packing. Grow some man-balls BillyG, wipe the tears from your cheeks, turn off Fox News, and do something actually patriotic like getting a grip on reality.

    2. That’s called legal advocacy. Both sides in our adversarial legal system do that and judge’s decide who’s right. The problems arise when judges do it too intentionally or explicitly. Eugene here, to contrast, is basically (in my opinion) offering a tentative roadmap to get where he might maybe want to go someday. Thinking about it. Dipping his toes in the water of social media platforms as common carriers, if you will. Or offering authority that can be cited by a court of someone making legal advocacy in an actual lawsuit.

  11. For once I’d like to see an example where we FIRST ask what would be best, and then figure out what law and legal traditions will need to be changed to reach that goal. It seems we are forever starting with the cart and not the horse.

    Or maybe that’s a discussion for a different blog.

    1. The starting point for “what’s best” (for questions like this in our society) is the market. People choose the products they like, and other people attempt to make products that people will like.

      If that doesn’t work out, then we have to identify what’s not working about it. I think that’s where we are here. But it forces us into a posture of fixing what’s there, not starting from scratch.

      If you have ideas that involve starting from scratch, then yeah, get a job at Facebook, or better yet, create your own Davebook that works the way you think it should.

      1. Randal, EV’s argument is that if Davebook practices viewpoint-based editing, the federal government can prohibit that. It is an argument that government can prohibit Davebook from working the way its publisher, Dave, thinks it should—at least for certain values of what the government permits Dave to think.

        1. That depends, what does Davebook claim to be? Is Davebook a personal blog wherein Dave shares his thoughts and opinions, and allows for people to comment on his posts?

          Is Dave book a public website that claims to allow people to choose to participate in their own content creation and discussions, independent of any input from Dave?

          Again, SL just pretends to make coherent arguments.

          1. Or maybe Davebook is a personal page on a website like WordPress. And Dave posts, but WordPress keeps censoring him, even though his 100 friends want to read what he has to say.

Please to post comments