Candeub and Volokh on Section 230 and Removing "Otherwise Objectionable" Content From Websites

 At Volokh Conspiracy, Eugene Volokh flags a brief article that he coauthored with Adam Candeub that was recently published in the Journal of Free Speech Law. Here's the abstract:

[Communications Decency Act] Section 230(c)(2) immunizes platforms' decisions to block material that they "consider[] to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable." The ejusdem generis interpretive canon suggests that "otherwise objectionable" should be read "to embrace only objects similar in nature to those objects enumerated by the preceding specific words."

In this instance, the similarity is that all those words refer to material that was traditionally viewed as regulable in electronic communications media—and was indeed regulated by the Communications Decency Act of 1996, as part of which § 230 was enacted. And restrictions on speech on "the basis of its political or religious content" were not viewed as generally permissible, even in electronic communications.

Candeub and Volokh argue that "otherwise objectionable" material should not be interpreted as a catch all category, but should instead be construed in light of the other terms in Section 230(c)(2)'s list of materials. As a result, they urge that Section 230(c)(2) be read as immunizing Internet companies and websites from restricting "obscene, lewd, lascivious, filthy, excessively violent, or harassing communications," and that the statute not be read to immunize these actors from the removal of content that is objectionable based on its political content.

I think there are several problems with this argument. With the caveats that these are my initial impressions and that it's been a while since I've done intensive research into Section 230, these concerns are listed below. I've tried to list my objections in order of most technical and specific first, followed by broader points.

First, I think the ejusdem generis interpretive canon has less power in the particular context here because Section 230(c)(2) immunizes platforms' decisions to block "otherwise objectionable" content rather than "other objectionable" content. The latter wording would be a clear example of specific terms followed by a general term, and would be in line with most examples in Candeub's and Volokh's article which involve statutes listing examples, followed by "other" things. But the way the statute is actually worded includes the term, "otherwise," which suggests material that is objectionable in ways that are distinct from the preceding examples. Candeub and Volokh do not address this, and their esjudem generis argument therefore seems stronger than the language of the statute warrants, as it implicitly reads the statute as prohibiting "other objectionable" content rather than "otherwise objectionable" content.

Second, this reading takes an overly narrow view of the purpose of the provision it analyzes and the Stratton Oakmont, Inc. v. Prodigy Services Co. case that played a significant role in prompting the inclusion of the provision. In Stratton Oakmont, the plaintiff sued a website message board for defamation based on several allegedly defamatory posts from message board users. The plaintiff alleged that the message board was liable because it exercised editorial control over the message board, noting that the site set forth detailed content guidelines that discourage the posting of "insulting" notes, as well as notes that "harass other members or are deemed to be in bad taste or grossly repugnant to community standards, or are deemed harmful to maintaining a harmonious online community" stating that such content would be removed when brought to the site's attention. The court cited this choice to engage in this editorial control as opening up the site to greater liability than other sites that did not take such measures. 

Congress was aware of this case when it passed Section 230, which Candeub and Volokh recognize. But they only focus on remarks from legislators who cite the blocking of material that is not "family-friendly," which (sort of) lines up with their restrictive reading of the statute. But this narrow focus ignores the extent of editorial restrictions at play in Stratton Oakmont, which included guidelines for the removal of "insulting," [non]-harmonious," and "bad taste" comments. It also suggests a narrow view of "family friendly," content, which may well exclude truthful, politically significant content such as images and discussions of war or terrorism, litigation and legislation over sexual privacy and abortion, and other such political content. Stratton Oakmont illustrated the danger of any sort of editorial policy giving rise to heightened liability, and Section 230(c)(2) was the response. Candeub and Volokh's narrow reading is inconsistent with this broad concern that prompted the immunity provision.

Third, a reading of the statute that removes immunity concerning the removal of political content while immunizing the removal of other content would run into significant First Amendment problems. Those who run websites or forums may have their own political preferences, and may moderate content based on these preferences. Limiting this ability would infringe on those providers' First Amendment rights--an aspect of online content moderation that is often ignored by Section 230 critics (and misinterpreters) who prefer to focus only on the purported rights of users.

To Candeub's and Volokh's credit, they recognize that the First Amendment may be relevant, but set it aside for the purpose of their article. Because of this, though, the article's analysis is in a misleading vacuum.  What is the point of this discussion about ejusdem generis if the First Amendment will ultimately require a different reading? Indeed, the First Amendment may well be relevant to discussions of interpretive canons: to the extent that the narrow reading of "otherwise objectionable" content leads to First Amendment problems, the canon of constitutional avoidance would weigh in favor of an expansive interpretation that avoids this issue.

These are my initial concerns with the conclusions in the article, but I'll be interested in seeing what others have to say--particularly those who are more involved in Section 230 research and litigation than me.

Labels: